Cash Management

A Toolkit for Microfinance Institutions

 This publication was made in collaboration with:

2
FOREWORD

Providing women with the means to create their own economic security has been at the heart
of WWB’s mission since the organization was established more than 30 years ago. We work
with network members to create innovative approaches to provide credit, savings and insurance
products specifically designed for the unique needs of women and scaling them to other network
members and partners.

This work is built on a commitment to capacity building with our partner institutions. Given the
diversity of institutions within our network, institutional development can take many forms. It
may mean helping to expand marketing capability, training staff or strengthening internal systems.
One of WWB’s broadest forms of capacity building is through sharing best practice and lessons
learned. WWB has played a leading role in building the financial capacity of women-focused
institutions by providing one-on-one trainings and workshops as well as through the development
of self-study toolkits.

Efficient cash management is essential for MFIs not only to ensure proper cash flow forecasting
and effective use of funds, but also for institutions that are beginning to mobilize deposits
Although the principles of cash management remain the same, institutions must now meet more
strict regulatory requirements, including liquidity restrictions, at both the branches and the main
office. Additionally, many institutions are also using alternative delivery channels and accessing
new technology to increase their outreach. Management must consider the implications of new
technology in their cash management practices to ensure that adequate cash is being maintained
at all levels of the institution. This toolkit is a complement to WWB’s Financial Risk Management
toolkit and provides both the theory and the practical tools for safe, efficient, and optimal handling,
managing, and investing of cash, within the broader context of liquidity management.

This toolkit is part of a 5-volume series of performance optimization toolkits on topics including
equity valuation, financial risk management, capital structure optimization, cash management and
portfolio analysis. This series was developed in response to demand from our network members to
have practical tools that they could apply to their institutions. Corresponding trainings are being
delivered to the network to ensure they have the financial tools to support the innovative solutions
they are implementing to expand financial services to women and girls.

I would like to thank the Corporación Andina de Fomento (CAF) and AECID for making this
publication possible. We hope you find it useful.

Mary Ellen Iskenderian,

President and CEO, Women’s World Banking
ACKNOWLEDGEMENTS
Women’s World Banking would like to thank the Frankfurt School for the support that made
this publication possible. Many WWB staff contributed to the paper: Jaclyn Berfond, CJ Juhasz,
Rebecca Ruf, and Julie Slama as well as consultant Laura Cobos.
1. INTRODUCTION........................................................................................................................... 5

2. CASH HANDLING PROCEDURES AND CONTROLS................................................................ 8

2.1 Introduction...............................................................................................................8
2.2 Vault Cash Procedures and Controls.....................................................................10
• Physical Security.......................................................................................10
• Teller Operations......................................................................................12
• Off-Site Transactions in Physical Cash....................................................14
2.3 Petty Cash & Operational Cash Expenses............................................................16
2.4 Anti-Money Laundering and MFI Cash Operations.............................................17
2.5 Book Money Transactions...................................................................................... 19
2.6 Bank Reconciliations..............................................................................................22
2.7 Correspondent Banking Relationships..................................................................24
Key Concepts and Next Steps.....................................................................................28

3. VAULT CASH OPERATIONS...................................................................................................... 29

3.1 Introduction.............................................................................................................29
3.2 Determinants of Optimal Vault Cash.....................................................................30
3.3 Basic Operating Rules for Vault Cash....................................................................33
3.4 The Miller-Orr Control Limit Model........................................................................40
3.5 Combining Trigger Limits with Short-Term Flow Forecasts................................43
3.6 Additional Considerations in Vault Cash Operations...........................................46
• Seasonality Analysis of Vault Flows........................................................46
• Utilizing Buffer Checks.............................................................................48
• Managing Multiple Currencies in the Vault............................................48
3.7 Implementing a Vault Cash Planning System.......................................................49
Key Concepts and Next Steps.....................................................................................52

4. FUNDAMENTALS OF LIQUIDITY MANAGEMENT................................................................ 53

4.1 Introduction.............................................................................................................53
4.2 Measuring Liquidity on the Balance Sheet...........................................................56
• The Rationale for Balance Sheet Measures............................................56
• Maturity Gap Report.................................................................................56
• Balance Sheet Liquidity Ratios................................................................59
4.3 Dynamic Liquidity Management based on Cash Flows......................................69
• Objectives of Cash Flow Planning...........................................................69
• Net Funding Requirements Forecast .....................................................70
4.4 Contingency Planning & Liquidity Stress Testing................................................72
Key Concepts and Next Steps.....................................................................................76

3
 5. INSTRUMENTS OF SHORT-TERM BORROWING AND INVESTING.................................. 77
5.1 Introduction.............................................................................................................77
5.2 Managing Traditional Liquid Assets......................................................................78
5.3 Investing in Short-Term Securities........................................................................79
5.4 Short-Term Borrowing............................................................................................81
Key Concepts and Next Steps.....................................................................................86

6. CASH MANAGEMENT IMPLICATIONS OF INNOVATIVE OUTREACH TECHNOLOGY.... 87

6.1 Introduction.............................................................................................................87
6.2 ATMs and Self Service Points................................................................................88
6.3 Merchant Service Points.........................................................................................89
6.4 Mobile Banking ......................................................................................................91
6.5 Handheld Terminals for Field Staff........................................................................92
Key Concepts and Next Steps.....................................................................................93

APPENDIX 1 – CASH FLOW FORECASTING.............................................................................. 94

ABBREVIATIONS
ALCO Asset Liability Management Committee
AML/CFT Anti Money Laundering / Combating the Financing of Terrorism
ATM Automatic Teller Machine
CD Certificate of Deposit
CIDA Canadian International Development Agency
DID Desjardin International Development
EU European Union
EUR Euro
HO Head Office
MDI Micro-Deposit Taking Institution
MFI Microfinance Institution
PDF Portable Document File format
PEP Politically Exposed Persons
POS Point of Sale (terminal)
SMS Short Message Service (mobile phone text message)
UN United Nations
USAID US Agency for International Development
US$ US Dollar
WWB Women’s World Banking
XML Extensible Mark Up Language

4
INTRODUCTION
This Cash Management Toolkit is part of a suite of practitioner guides and tools developed by
Women’s World Banking for use by microfinance institutions (MFIs). It integrates closely with
the Tool for Developing a Financial Risk Management Policy (WWB, 2005) and expands on the
fundamental guidance regarding liquidity risk management provided there. Cash Management
includes all activities related to the efficient planning, procurement, investment and control of
cash in a financial institution.

Cash by its nature is risky and costly. More than any other asset, cash is exposed to fraud, theft and
accidental loss. At the same time, cash does not earn a financial return and it requires resources
for protecting and controlling it. The analysis of cash must therefore always be two-dimensional
and consider aspects of control and risk mitigation alongside the questions of efficiency and cost
management.

This toolkit is about managing cash: be it in the form of physical bills in the safe or in the care of
field staff, in the form of balances held at other banks, or as liquid investments that can quickly
be converted back to cash. For the purpose of the toolkit we define cash as immediately available
and generally accepted means of payment, thus including both physical bills and coins, so-called
vault cash, as well as deposits with banks. Most banking textbooks give very little room to the
management of vault cash. This is very different in microfinance where clients are often compelled
to transact in cash. MFIs must therefore find ways to safely and efficiently handle a much higher
proportion of their business in vault cash than traditional banks.

In this context, cash management is esssential in the increasingly competitive business of

microfinance. Investing in good cash management is repaid in client goodwill, reduced cost of
holding cash and improved yields on liquid assets.

The target audience for this toolkit are those responsible for branch operations, treasury, risk
management, accounting, finance or internal control functions. The toolkit should also be
helpful to senior managers and executive directors who want a more detailed explanation of cash
management best practice.

5
 Structure of the Tool
The Cash Management Toolkit is meant to be used as a menu of topics rather than a textbook read
cover to cover. This is a collection of practical tools for MFIs at various stages of development,
including both deposit-taking and non-deposit-taking institutions.

The organization of the modules moves from the basic and most tangible topic of cash controls
to more complex and abstract issues of efficient vault cash operations, planning book money cash
flows, organizing the correspondent banking network and managing liquidity risk. Finally, we
also apply the principles of cash management to innovative new outreach technologies, such as
merchant agent models and mobile banking.

Chapter 2 – Cash Handling Procedures and Controls provides an overview of what every
manager should know about controlling vault cash and book money. This includes a synopsis
of fraud patterns related to cash operations. Effective cash handling processes should anticipate
opportunities for theft and close gaps by deploying targeted controls. This section also focuses on
book money transactions and the efficient use of correspondent banking network relationships.

Chapter 3 - Vault Cash Operations represents the core topic of the toolkit. Many institutions
could save the equivalent of the treasurer’s salary by lowering average vault cash holdings in their
branch network. The tool discusses heuristic operating rules for deriving traditional passive trigger
limit approaches to managing vault cash. It then introduces the Miller-Orr Control Limit Model
for specifying optimal lower and upper trigger limits in a passive vault system. We also propose
optional refinements to the vault flow forecasts at branch level, which can detect weekly or monthly
balance patterns as well as annual seasonality.

Chapter 4 - Fundamentals of Liquidity Management establishes the broader context for cash
management and constitutes the second pillar of this toolkit. Although an essential operating
condition and an important source of efficiency gains, vault cash is generally subordinate to book
balances with banks and other liquid assets when it comes to overall liquidity management. The
chapter first takes a detailed look at the “non-linear” nature of liquidity: i.e. always ample in good
times, gone when you really need it. We then measure liquidity from a static perspective on the
balance sheet using liquidity ratios and a maturity gap report. A more detailed and forward-looking
liquidity management approach is discussed through the use of cash flow projections as well as
stress test scenarios.

Chapter 5 Instruments of Short-term Borrowing and Investing discusses common short-

term treasury instruments and action variables of liquidity management. On the asset side, the
scope of available instruments includes savings accounts, sweeping current account balances
into overnight placements, time deposits with correspondent banks and investments in low risk
government debt securities. Potential short-term borrowing instruments consist of overdraft lines
of credit, repos, unsecured money market borrowings, and the placement of negotiable short-term
obligations.

6
Chapter 6 Cash Management Implications of Innovative Outreach Technology wraps up
the toolkit with a review of the implications of outreach technology such as ATMs, merchant
service points, cell phone banking and mobile teller devices. While the use of technology can
mitigate certain vulnerabilities of cash handling, new systems bring new risks that must be
anticipated and properly controlled. The tools and general principles of cash management easily
transfer to new technologies and are equally relevant in paper-based, computerized and mobile-
enabled environments.

The toolkit is accompanied by a complete set of Excel models that are extensively cross-referenced
in the text. These models are designed as functional sample implementations of all the concepts
and methods introduced in the toolkit.The toolkit includes three Excel documents:

1. Bank Reconciliation – provides bank statement, general ledgers, and bank

reconciliation templates and examples as discussed in Chapter 2.
2. Vault Cash Tools – features six models and tools related to the operations of vault cash
discussed in Chapter 3.
3. Liquidity Manager – includes a sample Maturity Gap report, calculation of Liquidity
ratios and a ratio dashboard, and a complete cash forecasting model. These tools are
discussed in Chapter 4 and Appendix 1.

The exercises throughout the text are designed to provide entry points into the spreadsheet models
and practice concepts from the toolkit.

7
 CHAPTER 2
CASH HANDLING PROCEDURES AND CONTROLS

2.1 Introduction
This chapter spells out basic procedures and control mechanisms for handling vault cash and for
managing transaction account balances at partner banks. The section does not provide a detailed
manual for various teller transactions, but instead gives users a compact checklist of accepted
standards and proven practices for safeguarding cash.

We define cash as any immediately available and generally accepted means of payment, such as
physical currency and sight deposits with banks. Vault Cash (also: physical cash) is cash in the
form of physical bills and coins, while Book Money is defined as the portion of cash that is held as
immediately available account balances with central banks or commercial banks.

The following basic principles of cash handling can safeguard the institution against loss,
discourage individuals from acting on impulse and protect innocent staff from suspicion should an
irregularity occur. Some readers may find these rules rather obvious. However, it is still important
to get an occasional refresher on the basics to gain a new perspective on the ways we handle cash
in daily practice.

Box 1: Opportunity Makes a Thief

At one Savings and Credit Cooperative in rural Uganda, a janitor walked off with
the teller’s entire cash drawer contents. The janitor had observed where the teller
kept the key and while cleaning the teller booth, it appears he also cleaned out
the cash drawer into a garbage bin, put some banana peels on top and walked
out to empty the trash, never to be seen again.

We used to think that only large amounts of cash could tempt staff. The truth is,
even modest amounts can be enough of a temptation, if people are under social
pressure to come up with money for essential expenses, if there is no credible
threat of criminal prosecution and if it is easy to simply stop coming to work and
disappear.

8
Figure 2.1 provides an overview of typical vault cash interfaces in microfinance that will be
managed under the guidelines in the following chapters.

Correspondent Banks

Head Teller / Branch

Cash Courier
Manager

MFI Head Office MFI Branch

Suppliers
In House Cash
Petty Cash
Courier

Field Agents / Loan

Tellers
Officers

Clients

Figure 2.1: Vault Cash Interfaces in Microfinance

9
 The following section provides an overview and specific guidelines of the handling of physical
cash as it relates to four main areas: physical security, teller operations, off-site transactions, and
petty cash and operational cash expenses.

2.2 Vault Cash Procedures and Controls

Physical Security
Among the physical cash held in a bank branch or microfinance institution, we typically distinguish
the main vault reserve versus the smaller amounts of working cash that are issued personally to
individual tellers for their daily client transactions. General guidelines for the safety of the vault
reserve include:
3 Outside of the opening hours for client business, all physical cash in a branch
must be kept in a locked vault or safe.
3 The vault should have individual locked compartments for storing the work-
ing cash issued to individual tellers. Alternatively, without individual teller
compartments, tellers must count and transfer their working cash back to
the vault reserve at the close of business and be reissued a new working cash
balance in the morning.
Security of the Vault Reserve
3 The vault or strong room must have solid walls, reinforced by metal cladding,
solid flooring to prevent tunneling in and a sealed-off and reinforced roof.
3 The vault or strong room must be fireproof and should have an automatic fire
extinguisher (sprinkler) or other fire suppression system.
3 If a safe is used instead of a strong-room, the safe should be of comparable
strength, fireproof and protected against removal (e.g. bolted to the floor).
3 Opening of the vault should be regulated by a delayed opening or time-lock
mechanism, which delays the opening even when the proper keys or combi-
nation are used.
3 The main vault reserve should be under joint custody of two individuals who
each hold different keys (set A & set B), both of which are required to open
the safe.
3 A set of duplicate keys should be held under safe custody (in a safety deposit
box) with another financial institution.
3 Key transfers and delegations during absences must be authorized and re-
corded. Special care must be taken such that no individual ever handles both
sets of keys over the course of time.
3 Locks and keys should be changed periodically.
3 Electronic access codes fall under similar precautions.

10
For banks and deposit-taking MFIs, the central bank often provides detailed physical security
requirements and check-lists for on-site inspectors that would be helpful in determining a
reasonable and compliant safety standard for a particular market.1

Movements in the Vault Reserve

The guidelines below apply regardless of whether the MFI uses an electronic or paper-based
accounting.
• The movement of vault reserves is subject to joint custody and record keep-
ing by the key or access code holders.
• The vault record maintained must show the denominations and amounts of
reserve cash. Like any accounting transaction, no cash movement may occur
without receipt and supporting documentation system.
• The reserve cash should be sorted by currency and denomination and bun-
dled into standard lots. Working with bundles makes it easier to perform the
daily (or weekly) bulk check of the reserve cash: i.e. adding up the bundles
and spot checking some bundles for completeness.
• The total cash in the vault reserve must be kept within prescribed lower and
upper levels in order to avoid stock-outs and to stay within insured limits2.
Insurance
Every bank or MFI should have insurance for cash on premises and in transit against theft or
robbery. Typical insurance rates range between 0.5 percent to 3 percent per annum of the insured
vault cash limit, in extreme risk areas they can be as high as 5 percent. The key drivers of insurance
premiums include:
• The loss history of the institution
• The crime exposure of the insured branch locations
• The safety equipment and deterrents available
(cameras, alarms, armed guards)
• The quality of and the adherence to internal cash management policies
• The types of cash transport permitted under the policy

1 See for example: Bank of Uganda, Statutory Instrument 2004 No. 61. The Micro Finance Deposit Taking Institutions (Licensing)
Regulations, 2004.
2 See Chapter 3 on vault cash operations for more detail on how to set limits and target vault levels.

11
 Teller Operations
Teller operations are a key part of any MFI’s operations. The main guidelines for safe and effective
teller operations are:
3 Every teller generally maintains his or her own working cash fund, for which
she is personally responsible. Management should prescribe and enforce up-
per limits for each teller fund.
3 Each teller should be supplied with her own vault compartment for over-
night storage of keys, stamps and the working cash fund, unless the cash is
returned to the vault reserve at the end of each business day.
3 The teller’s workspace should have a locked storage area to individually
guard his/her cash supply during any and all absences. Any working cash
beyond the amount necessary for ordinary transactions should also be kept
in this locked storage, even while the teller is on duty.
3 Tellers must keep their working cash till out of the reach of customers to
prevent grab-and-run situations.
3 Each teller station should be protected by a silent robbery alarm that goes
directly to a police station or a contracted security company. A low-budget
alternative would be to install a hidden switch at each teller station that acti-
vates a flashing visual alarm on the outside of the building.
3 Special protection (burglar bars, safety glass) should be provided for the
head cashier and any large transaction stations.
3 Cash transfers between tellers must be documented by vouchers that are
verified by both tellers.
3 Together with a supervisor or branch manager, each teller must check his/
her cash daily against a control total.
3 Every teller’s working cash must be periodically verified on a surprise basis
by internal control.
3 Tellers must provide receipts to customers for all transactions. In a paper-
system, receipts should be issued on pre-numbered paper forms from a
tightly controlled stock of forms. The receipt must specify the purpose or
type of the transaction (e.g. savings deposit, loan payment, deposit with-
drawal, etc.) and carry clear reference information for reconciliation, such
as client account number, loan ID or passbook number. If a computerized
system is used, the receipts should be designed to meet the same minimum
requirements.
3 Tellers should be asked to periodically turn over currency and coin invento-
ries, so that the same bundles and rolls are not held indefinitely. Otherwise,
there could be a temptation to replace real bills with photocopies in those
rarely touched bundles on the bottom of the drawer.

12
3 Tellers should regularly be rotated to different branches or between differ-
ent positions within the branch and should be required to take at least one
full two-week vacation per year. Branch or position reassignments should be
worked into the overall staff development plan, but should happen at least
every 2 to 3 years. Rotation and regular absences are critical, because fraud
and cover-ups tend to collapse when someone else steps into the fraudulent
individual’s job.
3 Tellers should not have access to accounting records, once processed and
recorded.
3 Teller differences, i.e. cash overages, shortages and counterfeit bills must
be declared and cleared daily. Differences should be recorded and accumu-
lated per teller and must be reviewed by management. The MFI must have
a policy for dealing with accumulated teller differences, which specifically
addresses the circumstances, frequency, amount of loss and form for which
the teller will be personally held liable.

Box 2: Handling Physical Foreign Currency

The cash handling principles described in this chapter naturally also apply
to vault cash holdings and transactions in foreign currency. If the institution
handles foreign cash, the accounting system must be set up for multi-currency
transactions such that the actual currency denomination and amount are stored
alongside a notional valuation in the home and reporting currency. The teller(s)
who are authorized to handle foreign currency receive separate foreign exchange
working cash for their daily operations. The inventory and denominations of
foreign cash are tracked and controlled separately.

13
 Off-Site Transactions in Physical Cash
Field transactions in vault cash are much riskier than in the branch.. The opportunities for fraud,
theft and robbery that arise when cash is handled by staff in the field require additional precautions
and controls.

Patterns of Fraud and Theft in Off-Site Transactions

Below are the typical patterns of fraud and theft associated with the handling of vault cash in the
field to devise the right controls.

After receiving a payment, a staff member may:

• Create a float from delayed receipts: i.e. record the receipts with delay and
substitute the cash received from subsequent clients, thus creating a perma-
nent float of cash for the staff person to use.
• Record the cash receipts, but for lower amounts than actually received.
• Record the cash receipt, but incorrectly add up the total in the cash book.
• Not record the cash receipt and intercept account statements or suppress
other feedback to clients in order to hide the fact that their payments have
not been credited.
• Not record a cash receipt on a seriously delinquent or written-off loan.(no-
body will miss a payment that was not expected and it will be easy to brush
off the delinquent borrower claiming otherwise)
A staff member may also:

• Have a loan written-off (e.g. with a fake death certificate) and continue to
personally collect the installments from the borrower.
• Organize new client groups but not enroll them in the MFIs client register.
Subsequent savings installments collected from the group could then be
pocketed by the field officer.

14
Off-site Transaction Guidelines
As a rule, off-site transactions in cash should be carefully controlled for specific types of
transactions with specifically authorized staff members within a well-defined process. This general
rule should also be clearly communicated to clients, in order to avoid situations where customers
ask employees ad hoc to carry cash to the office.
If loan officers collect installments in the field, it should be an organized and controlled collections
process:3
3 Loan officers should receive a list of installments due for collection before
they go out into the field each morning and must report back and reconcile
receipts and non-payments every evening.
3 Reported no-shows or no-collections should be verified with the client on a
spot check basis.
3 An essential safeguard for accepting cash payments in the field is to equip
staff with personally issued and controlled, pre-numbered carbon copy
receipt books. Clients would be briefed to know that they are to receive such
official receipts for every payment made. Field officers would have to present
and reconcile their receipt books every evening.
3 Loan officers who regularly collect cash installments can be targets for
robberies,and institutions should procure cash-in-transit and personal injury
insurance. In high crime areas, institutions may want to reconsider field col-
lections to ensure the safety of staff members.
In the case of savings:

• Theft of savings deposits and other voluntary payments by clients to field

staff is more difficult to control still, because there is often no immediate
feedback to the client about account activity. One option could be for the
institution to send same-day SMS notifications of all account movements
and ensure clients know to check text messages.

3 Handheld computers and mobile banking applications with receipt printers can facilitate the control of cash receipts in the field,
but they also introduce new opportunities for fraud. Compare our detailed look at the cash management aspects of mobile
banking in Chapter 6.

15
 2.3 Petty Cash & Operational Cash Expenses
In many microfinance institutions, staff meals, supplies, utilities and other small purchases are paid
in vault cash rather than by bank transfer or check.

Larger vault cash payments for operational expenses are processed like a normal teller transaction
on the basis of a journal voucher raised by the branch accountant, which is paid in cash to the
supplier against a receipt signed at the teller window.

For small daily incidental cash expenses, many MFIs maintain a separate petty cash float in the
office, outside of the customer teller stations. The petty cash fund is vulnerable to small-scale
theft and fraud and procedures for handling the petty cash fund must be clearly outlined and
consistently followed.

Methods of Petty Cash Theft

Petty cash theft generally starts with the entrusted person “on-lending” funds for personal use.
There are essentially three methods to cover up the shortage in the cash box:

• Increasing the amount on authentic vouchers and receipts

• Creating fictitious disbursements using false vouchers and receipts
• Recycling authentic vouchers and receipts multiple times

Petty Cash Procedures

The established best practice is to maintain petty cash on a so-called imprest basis. At any given
time, the cash and receipts in the cash box shall total the fixed imprest level. The imprest system
allows replenishment only of the amount spent. The replenishment is credited to the petty
cash account and the debits will go to the respective expense accounts. Below are some specific
guidelines for petty cash procedures:

3 Only one designated staff member should handle petty cash. Actual cash
should be spot-checked and verified by the supervisor at least once per
month. The staff person in charge of the fund is personally liable for any
discrepancies.
3 All requests for petty cash expenditures must be signed by an authorized
supervisor on a pre-numbered voucher. A receipt from the merchant must
be kept to document the transaction.
3 The petty cash, vouchers and receipts must be kept in a locked box or safe.

16
2.4 Anti-Money Laundering and MFI Cash Operations
Money laundering is the attempt to legitimize the proceeds of illegal activities. Organized crime
activities in narcotics, arms, human trafficking etc. generate large amounts of physical cash that is
difficult to use directly for purchases of investments, capital goods and luxury life-styles without
attracting the attention of law enforcement. Therefore, money laundering is always about dispersing
large amounts of cash into many small inconspicuous transactions and injecting funds into legal
activities. The objective is that the wealth enjoyed by the criminals may appear to be coming from
ordinary law-abiding businesses and investments.

After the September 11, 2001 terrorist attacks in the United States, the notion of money laundering
was expanded to include the financing of terrorism, which uses many of the same techniques to
obscure the sources of funds and their transmission to terrorist operatives. Anti money-laundering
and combating the financing of terrorism (AML/CFT) is a global effort requiring the collaboration
of law enforcement, tax authorities, central banks and retail financial institutions.4

In recent years, it has become clear that money-laundering is not just a concern for banks in exotic
off-shore jurisdictions but that MFIs in developing countries are in fact quite vulnerable to money
laundering attempts. That is because the small businesses run by microfinance clients operate
mostly on a physical cash basis. Money laundering often includes a dispersion phase, where
many helpers run the funds through superficially legitimate small transactions. Fictitious micro-
businesses are an ideal cover for this type of cash dispersion.

The practical obligations of those who handle cash at microfinance institutions under AML/CFT
rules are quite simple. Essentially, the obligations are about:

1. Documenting and verifying the identity of the client and ultimate beneficiary of a
transaction or account, and

2. Reporting suspicious transactions to management or a compliance officer in-house,

who then determines whether further investigation and/or a report to the national
authorities is required.

4 The best entry point into the world of anti-money laundering is the website of the Financial Action Task Force (FATF) at www.
fatf-gafi.org.

17
 Typically, the national AML/CFT regulations will determine a limit for vault cash deposits by and
for disbursements to non-clients, for which the identity of the walk-in customer must be properly
established and retained by the institution.

More specifically, here is what tellers, loan officers and other client contact staff should look
out for:

• Frequent small cash deposits below the identification threshold by various

non-clients into a client account that have no plausible connection to the
client’s business. Often these deposits are followed by rapid withdrawal or
onward bank transfer by the client.
• Microloans settled early by the client (not by a competing institution)
without an apparent and plausible business background. This could be an
indication that the loans are part of running a fake business front for recy-
cling cash.
• Multiple new clients apparently transacting along the same patterns that are
not obvious by the nature of their business.
If a suspicious situation along the lines of the examples above arises, the MFI staff should:

• Never confront the client with the suspicion. Accept the requested transac-
tion. Show polite interest in the business activity and ask a few questions. If
the suspicion remains, notify the branch manager, the compliance desk or
internal control as per your institution’s policy.
• Management should then assess whether the suspicion is warranted. If it
is, the transaction should be frozen while a suspicious transaction report is
transmitted to the authorities, typically to a Financial Intelligence Unit at the
central bank. The Financial Intelligence Unit should notify the institution
whether to release the transaction or hold the account frozen pending an
official investigation.
At most MFIs, the volume of operations and the financial resources do not justify the investment
in specialized systems for detecting money laundering activities. Yet, the internal control staff
should make it part of their regular duties to review account level transaction data for evidence of
dispersion activities or of excessive and implausible uses of money transfer services. Short of using
specialized software, this can be done by downloading structured transaction data into a standard
spreadsheet or database tool. Embargo and PEP lists can be downloaded from the UN, EU and
other sources for free. With some intelligent filtering or database queries, one can pick up many
suspicious patterns quickly and without excessive manual effort.

18
Every MFI should have a AML/CFT policy. The policy should designate either the branch manager
- or better - a compliance officer or internal control staff member, to serve as first-level recipient of
suspicious transaction reports from front-line MFI staff. This staff member should have received
training on the principles of AML/CFT along the guidelines of the Financial Action Task Force5
as well as on the specific AML laws and regulations of the country where the MFI operates. As a
general rule, financial institutions are not expected to investigate clients. If the suspicion passes a
basic plausibility check, the transaction/account is temporarily frozen and a standard suspicious
transaction report is forwarded to the authorities for further action.

The following section focuses on book money transactions and provides general guidelines for
establishing controls of account balances, ensuring effective bank reconciliations, and maintaining
efficient and strategic correspondent bank relationships.

2.5 Book Money Transactions

General Principles
In addition to physical bills and coins, our definition of cash also includes book money. More
specifically, we define book money as immediately available account balances that the MFI may
hold with commercial banks or the central bank.

Because the consequences of a security breach in bank accounts are so grave, account signature
authorities have always been the most tightly controlled aspect of any business. Signatories should
be limited to a very small number of senior staff. Account authorities should be more restrictive
than ordinary corporate representation and signature rights. Additional guidelines include:

3 Two authorized signatures should be required on any payment instrument

(check, payment order). If payment instructions are given through an on-
line banking interface, two authorized signatories must independently ap-
prove the queued transactions with their personally issued access credentials
and transaction codes.
3 Any payment instrument must first be raised internally with a journal vouch-
er and necessary authorization that ensures that the corresponding internal
accounting entries are made.

5 See www.fatf-gafi.org for additional information.

19
 Box 3: Don’t share bank account credentials!
The above principles seem obvious, but the reality of daily practice often tells a
different story. In one example, three senior managers would regularly come to
an unlocked desk drawer in the shared office to take out a bank-issued smart card.
They used the smart-card in an external reader on their computers to authenticate
payment orders for loan disbursements that had been queued by the system.
This is quite an incredible breach of security! The smartcard was clearly issued to
a particular individual, together with personal passwords and transaction codes.
If the staff shared the smart card, it is clear that they must also have been sharing
the passwords and transaction codes. On top of the mistake of sharing access
credentials, the drawer was unlocked while the card readers and some computers
still running.

Book Money Disbursements and Controls

Book money payments include payroll, supplier invoices, taxes, as well as investment and funding
related transfers. Yet, the most frequent and at-risk book money payments in microfinance are
client loan disbursements.

Loan disbursements may occur:

• Directly in vault cash through a teller at the MFI

• To a client’s current account at the microfinance bank itself or with another
institution including mobile banking accounts,
• Via check for cash disbursement at a correspondent bank, or
• Via payment order for cash pick-up against identification at a partner bank
(similar to a Western Union transfer).
All book money transfers carry specific risks and require special controls. In the context of outgoing
payments by check or bank transfer, fraud might occur if staff were to:

• Create a fictitious transaction or borrower, the payment for which is chan-

neled to the staff member directly or indirectly,
• Redirect the proceeds of a legitimate transaction,
• Inflate a legitimate transaction and siphon off the overpayment
• Pressure clients for kick-backs at time of loan disbursement

20
The fraud can be hidden a variety of ways:

• Creating false invoices from suppliers

• Increasing the amount on a legitimate invoice
• Using supplier invoices twice to support payment
• Creating ghost loans for non-existing clients
• Creating duplicate loan with disbursement going to self or accomplice
• Manipulating summary reports to hide the missing funds or duplicate loans
• Changing disbursement modality. For example, changing from “pay into cli-
ent account” to “issue check”, and intercepting the check and cash while the
client is waiting for a credit to the bank account. Disbursements should fol-
low exactly the instructions on the signed loan application/approval docu-
ments and may not be changed without additional documentation signed by
the client.
Loan Collections and Savings Deposits via Correspondent Banks
If payments made to the MFI come through the network of bank accounts with partner institutions,
it is critical to obtain clear references as to its purpose directly from the correspondent banks.
Reference data should include the name of the depositor and the loan number or savings account
number.

In practice, reliable information on loan installments can be ensured if clients receive pre-printed
deposit slips with full reference details to make the payment. MFIs should only work with partner
banks as cash acceptance points that can provide electronic account statements in machine
readable format. This will help the MFI automate its daily reconciliation of client accounts. This
also applies to savings deposits at third party counters.

21
 2.6 Bank Reconciliations
Accurate and timely bank reconciliation is a key factor in maintaining internal control over cash.
Bank reconciliation requires matching all transactions in the account as shown on the official
statement to those transactions that should have occurred per the MFI’s internal accounting.
Reconciliation thus looks at the MFI-bank relationship from each institution’s perspective and
makes sure that they match. Should differences emerge, they need to be investigated and resolved.
From the perspective of the bank, the statement for a current account is an excerpt from the bank’s
accounting system regarding their liability to the MFI. The MFI in its own accounting system,
maintains a mirroring asset account with the bank as debtor, where it records its view of the same
transactions that concern the bank relationship.
In the MFI’s accounting, the various receipts from clients shown in the bank statement must be
posted as a debit to the internal debtor account for the bank and as a credit to the client loan
account. Conversely, a loan that was paid out by debiting the client loan account and crediting the
internal debtor account for the disbursing bank, must be matched to a debit in the bank statement
as confirmation that the bank has actually paid out the funds to the borrower.
A bank account that is used for frequent client transactions should be reconciled daily, other
bank accounts used less frequently for non-client transactions such as expense payments and
investments can be reconciled less often, but never less than on a monthly basis.
Figure 2.2 below is a simplified paper format that illustrates the basic idea of reconciling a bank
account to the internal accounting.
Bank Reconciliation at 31-March-2010
Bank of Africa, Kigali, Account 789-234589-00

A. General Ledger Data Debit Credit

Balance per general ledger, beginning period: 2.500
Total Receipts: 8.900
Adjustments: 400
Total Disbursements 11.000
Adjustments: 500
Balance per general ledger, end of month: 500

B. Bank Statement Debit Credit

Ending balance per bank statement, current month 550
Outstanding checks for deposit (detailed list) 150
Other pending incoming credits (detailed list) 100
Written checks not yet debited (detailed list) 125
Other pending outgoing debits (detailed list) 175
Reconciled bank balance, end of month 500
Figure 2.2: Simplified Bank Reconciliation Template.

22
Part A of the form summarizes the activity in the internal bank debtor account for the month.
The first line is the opening balance at the beginning of the month, taken from the general ledger.
This should match the previous month’s closing balance. The monthly totals of receipts and
disbursements posted to the particular bank account are listed.

Adjustments should be made for charges or interest received on the account that have not been
posted in the internal debtor account. Those adjustments are also listed in Part A. The same applies
to receipts from clients that have not yet been posted to the client accounts on the general ledger.

Part B begins with the closing month end balance from the bank statement. Reconciling begins
by matching totals and individual items between the two sources. Possible appropriate differences
between the bank balance in Part B and the general ledger balance in Part A could be:

• Check payments received or vault cash shipped to the bank for deposit
included in Part A that have not yet been credited to the bank account. These
are listed as outstanding deposits.
• Checks listed in the disbursement journal and included in the general ledger
total that have not yet cleared the bank. These items are reported as out-
standing checks.
The ending balances of the bank statement should match the ending balance of the general ledger,
see Figure 2.2.

Exercise 2.1: See Excel File BankReconciliation.

The file provides an electronic bank account statement (Sheet: BankStatement) as well as a list of
journal vouchers from the accounting system that concern this bank account (Sheet: GeneralLedger).
Your task is to perform the reconciliation between the two sets of transactions. Please use the sepa-
rate reconciliation workspace that is provided in the sheet Reconciliation. Before you start, delete
the contents of lines 48 and 53 in the sheet Reconciliation, which already contain the solution. Now
perform the reconciliation yourself by trying to identify the source of the differences between the
internal accounting and the bank statement. Capture the identified differences in the workspace “C.
Reconciliation of Differences” and track the residual differences in the blue control section above. It
usually is helpful to first narrow down the source of the discrepancy by looking at the subtotals by
transaction type in the side-by-side summary in sections A) and B). You can then filter the accounting
transactions and the electronic bank statement by transaction codes and other criteria to identify the
missing items.

23
 2.7 Correspondent Banking Relationships
Most MFIs connect to the financial system through a network of correspondent accounts. These
are generally called current accounts, checking accounts, settlement accounts or nostro accounts,
and essentially all serve the same purpose. They facilitate the incoming and outgoing transfer of
book money balances and allow for the redemption of book money against vault cash and vice
versa.

One of the major challenges at many microfinance institutions today is that there are simply too
many correspondent accounts, often with the wrong institutions. This leads to a fragmentation
of account balances that makes it harder to invest cash surpluses profitably and multiplies
administrative efforts.

This section focuses on the efficient use of correspondent bank networks, i.e. how can we reduce
complexity and workload in managing bank accounts. As we reduce complexity, we also facilitate
control and mitigate risk.

Streamlining the Network & Reduction of Unallocated Receipts

There are three main reasons why many MFIs have seen an inflation of bank accounts:
1. The desire to offer convenience to clients who may want to make loan payments or
other deposits at various bank branches, rather than coming to the MFI counters;
2. The need to facilitate vault cash transfers with banks in the proximity;
3. The lack of reference information on client payments received via correspondent
banks, which makes it necessary to compartmentalize the flow of incoming payments
to localized bank accounts.
While the above are valid explanations, they are not insurmountable obstacles in realizing greater
efficiency. This section will first develop a vision of what efficient Treasury operations might look
like and then will discuss ways for getting around the three challenges mentioned above.

Strategically, MFIs should limit their correspondent network to no more than 2-3 primary bank
partners. There may be additional single purpose relationships, say for money market placements
or an international remittance service, but for full-service relationships including the mainstream
client-related transactions, there should be a maximum of three.

Further, account relationships should only be head-office-to-head-office and controlled by the

MFI’s treasurer. Even in remote banking markets, this should not be a problem. A bank that does
not have all of its branches on-line today is most likely not a credible correspondent partner for
the MFI.

24
The benefits of centralized accounts are many:
• No balance fragmentation
• One single overdraft facility (per correspondent)
• Immediate compensation of inflows and outflows in a central cash pool
• Consolidated overnight investing of excess balances
This does not preclude certain decentralized transactions by the MFI branch on that central
account. For example, the branch manager could be an authorized co-signer on the head office
account but the transactions are all booked against the single head office account.

In the interests of efficiency and operational control, the MFI treasurer should strive to consolidate
all outgoing payments into a single settlement account with its preferred bank partner. This reduces
the total settlement balances required for covering payments and improves the negotiating power
with that correspondent to lower per-transaction fees on the consolidated volumes.

The central settlement account will be used for all outgoing check and electronic transfer payments
issued by the MFI, including supplier payables originating at the branch level, see Figure 2.3.

MFI Head Office Primary Bank Correspondent

Finance Dept.
Daily Payment Batch Settlement MFI’s Settlement
Payments
Instructions Debits Balances
Queue

MFI
Clients
Electronic Disbursements Sweep of
Branch X Payment Receipts
Requests Bank
MFI Branch A
Clients MFI
Branch Y Supplier

Bank
MFI MFI Branch B
Clients MFI MFI
Clients Clients Clients
Branch Z
Bank
MFI
Branch C
MFI
Supplier Clients MFI
MFI MFI MFI Clients
Clients Clients Supplier MFI
Clients

Figure 2.3: Transaction Flow in an Efficient Correspondent Network with Central Payments Queue

25
 Instead of branch accountants initiating loan disbursements or supplier payments via decentralized
bank accounts, the branch triggers electronic payment requests to Head Office that are fulfilled
through the central settlement queue. This queue can then be professionally managed with a
view to fully utilizing contractual payment delays on supplier invoices, for example. The central
electronic queue also gives excellent visibility of all upcoming loan disbursements and other
scheduled payments and is helpful in forecasting near-term cash requirements.

Should a particular MFI branch not be located within a reasonable distance of a branch in the
primary correspondent’s network, vault cash requirements can still be settled through one of
the alternate link banks. Yet, this would happen on the basis of liquidity specifically provisioned
to that bank’s head office account. Going back to the three issues that have often led to bank
account overload at MFIs, we can see that a centralized account management vision can in fact be
complementary.

1. Convenience for clients: The MFI may continue to accept loan installments at banks
other than the primary correspondent(s). These receipts, however, will all be posted
directly into centralized accounts. Because these secondary head office accounts
are used only for incoming credits, the MFI does not have to provision them with
settlement liquidity to cover the odd check that might be written by a branch against
this account. Instead, one can have standing orders in place to sweep any credit
balances (exceeding a convenience threshold) into the primary settlement account.

2. Facilitation of vault cash transfers: Ahead of making a withdrawal directly at the bank
branch, the branch manager first notifies the MFI Head Office of the amount required,
which will make sure that the withdrawal is covered by sufficient balances in the
head-office-to-head-office account. If necessary, the MFI treasurer initiates a real-time
transfer of cover to that particular account. The MFI manager then cashes a check at
the bank branch drawn on the account maintained by the MFI Head Office

3. Insufficient reference information and the complexity of reconciling a large single

receptacle with a multitude of Treasury and client transactions: The problem arises,
because tellers at correspondent banks often do not capture with sufficient precision
the essential information required to identify the depositor and the purpose of the
payment (loan account, savings deposit). Unallocated receipts are an accounting
nuisance and absorb much time in pursuing deposit slips from borrowers or
sending inquiries to the bank for investigation. Moreover, if a significant backlog of
un-reconciled loan installments sits in the bank account, the MFI can experience
significant challenges related to arrears management

26
To minimize this third challenge a MFI should follow these guidelines:

3 Clients should receive deposit slip booklets that are pre-printed or pre-
stamped with their name and the particular loan or savings account number
in the reference field.
3 The design of the form should emphasize this reference information and
remind the accepting bank that this information must be captured into the
system.
3 In parallel, the MFI should message the importance of capturing the unique
loan or savings account number along with the client name to its correspon-
dent bank when accepting vault cash deposits at the teller window.
3 Service conditions should be negotiated with the bank to ensure that this
will be enforced with their tellers. If the information is still not captured, the
service level agreement should obligate the bank to produce image copies of
the original deposit slip from the teller files at no cost within a few days.
MFIs should use an electronic banking interface to the main correspondent banks to download
the account transactions in structured, machine readable format. If this data is mapped against the
import interface of the debtor management system, client installment receipts can be posted and
reconciled into the system on a same day basis.

Strategic Considerations of Correspondent Bank Relations

It is important emphasize the strategic opportunities in the correspondent relationship: a
partnership with an innovative and dynamic commercial bank can facilitate cooperation in new
valuable financial services for the MFI’s clientele.

The introduction of a single payment queue does not have to result in the breaking off of business
relationships with other major banks in the local market. One should still retain the head-office-
to-head-office relationships and continue to cultivate wholesale placement and borrowing
opportunities, but not use these accounts for transacting ordinary daily client business and
operating expenses.

27
 Chapter 2 - Key Concepts and Next Steps
• In this chapter we reviewed the basics of physical security for vault cash, the
principles of teller operations, petty cash and off-site cash transactions. We
also highlighted typical fraud attempts related to vault cash transactions.
The next step for a microfinance manager should be to take a critical look
at the daily practice of cash handling in his/her own institution together
with internal audit to determine where the risks are and whether they are
properly contained.
• Money laundering may not have been on the radar screen of many managers
in microfinance, but the vulnerability is real. It is important to sensitize
front-line customer contact staff for the magnitude of the global money
laundering problem and the typical dispersion methods. There is rising
awareness of money laundering vulnerabilities in microfinance among
international investors and sponsors.
• Book money in bank accounts must be safeguarded by strict limitations
on transaction authorities and tight integration with the MFI’s financial
accounting. Prompt reconciliation between internal accounting and bank
statements is essential. One should make sure that Internal Audit regularly
performs an inventory of all bank accounts, signature authorities and
electronic access credentials.
• A vision of efficient correspondent banking was developed that limits the
network of full service bank partners to 2 to 3 institutions. All banking
relationships should be head-office-to-head-office and there should ideally
be only one single outgoing payments queue. When there is an opportunity
to make adjustments due to new products or changes in the branch network,
opt for a reduction of complexity and consolidate transactions in head-
office-to-head-office accounts. In addition, MFIs must improve the reference
information on receipts collected through their correspondent network
and imported electronically into the MFI accounting system. Tracking the
average number of unallocated receipts in management reports can be an
effective measure of increasing awareness of the problem and streamlining a
solution.

28
CHAPTER 3
VAULT CASH OPERATIONS

3.1 Introduction
Chapter 2 focused on vault cash from the perspective of safety and control. This next chapter
adds the third important dimension of efficiency. Chapter 3 provides a set of practical tools and
methodologies for determining the optimal level of cash to hold in a branch network.
Efficient vault cash operations must reconcile three competing objectives:
1. All cash points must have sufficient balances to meet all customer payment requests.
2. The overall balances of vault cash in the branch network should be minimized in order
to reduce the holding cost of cash.
3. The frequency of cash shipments and the associated transport costs should remain
within acceptable limits.

Importance of Vault Cash Operations

Vault cash planning is essential because holding physical cash is a very expensive commodity. Vault
cash does not earn a financial return and therefore has a high opportunity cost, given that it could
be earning a yield through the provision of loans to clients, making short-term investments (and
earning a small return), or paying off existing debt. Also, as indicated in the previous chapter, vault
cash handling has a significant human resource cost.
Most MFIs tend to hold too much vault cash across their branch network. This does not imply
they are overly liquid, but that too much liquidity is held in the most expensive form of physical
cash. Because the costs of vault cash are not always understood, managers may prefer to hold the
maximum levels of insurable cash, instead of actively managing vault movements. This chapter will
be looking at tools that can help you do this rationally and safely.

Approach to Efficient Vault Cash Operations

This toolkit recommends a bottom-up approach for analyzing vault cash movements. Each cash
point (branch, or stand-alone cash acceptance and disbursement point) must first be considered
separately and individually. This is necessary because cash points may have different order costs
and distinct patterns of cash demand, which would lead to different levels of optimal vault holding
and movements. It is therefore important to capture these differences on a per-branch basis first
and then consolidate the individual vault plans into a network-wide view. At the same time, it is
unnecessary to drill down this type of analysis and planning at the individual teller level, as the
time and cost this would entail would likely not yield significant benefit to the results.

29
 It is important to highlight a clear distinction between vault cash analysis and liquidity planning.
Vault cash is only a minor component of the necessary liquid assets that an institution should
hold as a safety buffer against the volatility of loan demand and deposit supply to smooth out gaps
of long-term funding. Chapter 4 further outlines how vault cash planning fits into the broader
context of liquidity management.

3.2 Determinants of Optimal Vault Cash

There are three variables that determine the optimal levels of vault cash to be held at a branch:
1. Proportional vault holding costs are the cost of maintaining a certain level of physical
cash at the vaults and can be estimated by adding the following costs:
• The opportunity cost of not earning interest on vault balances
• The physical security of maintaining the vault
• The staff time related to cash controls
• The insurance premium against theft or fraud
Vault holding cost is generally proportional to the amount of cash held: the higher the amount
held, the higher the cost, and is generally expressed as an annual percentage relative to the average
vault balances.
2. Vault cash order cost is the unit cost of triggering, transporting and controlling a vault
cash shipment to or from a vault point, and can be estimated by adding the following
costs:
• Staff time
• External service charges (armored vehicle service, guards etc.)
• Cash-in-transit insurance premium, where applicable.
Vault cash order costs are generally not proportional to the amount transported and can be
considered a fixed cost per shipment.
3. Cash demand functions are the demand patterns within each cash point based on the
market’s trends, and seasonality, which will yield the size and volatility of the demand
for cash.

30
Size of Vault Reserve versus Shipment Frequency
A fundamental relationship of vault cash operations is the trade-off between the size of the vault
reserve and the frequency of cash shipments. The less cash one holds in a branch on average, the
more often one will be compelled to bring in additional cash or ship away excess holdings. The
relative costs of holding cash versus the cost of transporting cash will therefore determine the
optimal target vault holding for a particular branch.
As an illustration, vault cash planning can be compared to managing a retail distribution operation,
such as supplying milk to grocery stores in a chain: one could minimize in-store refrigeration cost,
spoilage and floor space requirements by holding just a few bottles of milk and sending a delivery
van to the distribution center several times a day, which would maximize shipment costs. Or he
could minimize shipment costs by ordering one large shipment of milk per week, which would
require more refrigeration and storage and increase spoilage. In order to determine the optimal
amount that should be held, a third variable is then considered: the cash demand parameters.
Because meeting all cash payment demands is a necessary condition, a cash point with more
volatile flows will require a higher safety level of vault balances and consequently fewer shipments.

In summary, an optimal, cost-efficient vault cash system will be driven by three main input factors,
as indicated in Figure 3.1:

• Proportional holding costs

• Fixed per order costs
• Cash demand functions

Holding Costs -
proportional to the vault Order Costs -
reserve fixed per shipment

Cash Demand Function: volatility, trends, seasonality etc.

Figure 3.1: Determinants of Optimal Vault Cash

Later in this chapter, all three factors will be combined in a model that mathematically derives
optimal vault cash levels with the Miller-Orr Control Limit approach. Before introducing
the formal model, some simple practical alternatives will be described that might also lead to
satisfactory vault cash management.

31
 Exercise 3.1: Estimate Holding and Order Costs.

You would like to apply some of the vault cash planning techniques discussed in the toolkit at your
institution. The first step might be to determine what the vault cash operations actually cost your MFI.
Your accountant has evaluated the first quarter 2011 and reports the figures below. Estimate Vault
Cash Holding and Order Costs from this data:
• Total Brink’s Armored Transport invoices for the first quarter of 2011 are US$ 2,500. This includes 9
trips to Upcountry Branch A and 15 deliveries to Suburb Branch B. The distance and travel time to
Upcountry Branch A is about three times that of Suburb Branch B.
• From your time as a branch manager, you remember that it took you about 15 minutes together
with the head teller to process an arriving shipment into the main vault reserve, or package and
verify an outgoing dispatch. The head teller spends a further 30 minutes raising and transmitting
each cash shipment request.
• The average all-inclusive personnel cost of a branch manager and a head teller are US$ 1,800
and 1,000 per month respectively.
• The weighted average annual yield of liquid assets investments at your institution was 9.5 per-
cent during the first quarter of 2011.
• Your accounting statements show the opening vault balances on 2-Jan-2011 as Branch A US$
25,000 and Branch B US$ 8,000. Closing balances on 31-Mar-2011 are Branch A 45,000 and
Branch B 12,000.
• Vault shortages written-off during the quarter amount to US$ 750 for both branches combined.
• The cash insurance invoice for the first quarter 2011 was US$ 300 for both branches combined.
Sufficient cash-in transit insurance is carried by Brink’s and is not itemized on their invoice to the
MFI.
The solution is provided in the accompanying Excel file [Exercise3.1].

This toolkit outlines two approaches for determining the optimal levels of cash at the branch level:

• Basic Operating Rules for Vault Cash Operations

• The Miller-Orr Control Limit Model

Both of these approaches are based on a passive vault planning approach, which focuses on an
analysis of historical vault transactions of a particular cash point to establish a base level of vault
reserves. This approach does not attempt to predict specific future vault movements, but only
focuses on historical transactions. There are various limitations regarding the passive vault cash
approach, which are discussed later in this section.

A third approach that builds on these two methodologies and combines them with short-term
cash flow forecasts is then described.

32
3.3 Basic Operating Rules for Vault Cash Operations
This approach uses rule-of-thumb operating parameters based on common-sense assumptions
from historical data. As a result, they should be tested against each institution’s unique operating
characteristics and environments.

A note about Vault Cash Reserve Ratios

When considering the amount of vault cash to hold, one may come across various vault cash
reserve ratios such as vault cash as a percentage of total deposits or total assets. A vault reserve
ratio can be a starting point for determining vault holdings, but is not considered adequate for
operational vault planning. Loan disbursements, deposit withdrawals, and other transactions can
vary between individual branches because of their location (e.g. near a central market or train
station), seasonal patterns, and other market demand factors. A ratio provides a static view of the
branch holdings. Customer behavior and vault cash flows must be analyzed more closely to find
optimal levels of cash to hold.

Determining Trigger Limits

When determining the optimal level of cash held, one must first establish a base level of reserves
based on historical vault transactions. These levels are called trigger limits, which are thresholds
of the amount of cash that the institution is planning to hold that when reached will trigger an
inbound or outbound order of additional cash:
• Lower Trigger Limit is the vault cash threshold that will trigger an inbound
vault cash order (ie additional cash is brought in)
• Upper Trigger Limit is the vault cash threshold that will trigger an outbound
vault cash order (ie additional cash is sent out)
If the actual vault level drops below this point it triggers an incoming cash shipment. If the upper
trigger limit is breached, cash is shipped away and credited to the MFI’s bank account.

Steps to Determining Basic Operating Rules for Vault Cash Operations

There are four steps to determine basic operating rules for vault cash operations:
Step 1: Define Lower Trigger Limit
Step 2: Define the Standard Incoming Shipment Size
Step 3: Determine Upper Trigger Limit
Step 4: Define the Standard Outgoing Order Size

33
 Step 1: Define Lower Trigger Limit
We must first determine the lower vault cash threshold that would trigger an inbound order. The
lower trigger limit is the sum of the minimal safety buffer and the approximate amount that would
cover the large daily outflow for the number of days it takes for new cash to come in. The following
formula summarizes the three variables used:

Lower Trigger Limit = Iron Reserve + Days Shipping Cycle *

Large Net Daily Outflow90%

Iron Reserve: a minimal safety buffer of vault cash that should be held by a branch at all times, even
after an unusually large outflow of vault cash and while awaiting the arrival of inbound cash orders
that may have already been triggered.

Days Shipping Cycle: the average number of days it takes for cash to be delivered.

Large Net Daily Outflow: estimate of the net daily outflows of an institution based on historical
figures; the assumption of a large outflow based on percentile of observations.6

Figure 3.2: Screenshot from Accompanying Vault Management Tools

Above is an example of these calculations based on the Vault Cash Tools model in the work sheet
SampleVaultProfile. Columns A and B include a fictitious random daily time series of vault closing
balances, from which we calculate the daily net inflow or outflow as the difference of the closing

6 A percentile is a simple statistic on a number of observations: if one puts all the observed outflows in ascending order, the 90%
percentile marks the point that divides 90% of the observations with lower outflow values and the 10% of cases with higher
observed outflows.

34
balances from one workday to the next. The Large Net Outflow is the 90 percent percentile of all
the observed daily outflows. A year’s worth of daily observations is used in the example as a way
to ensure that annual events (year-end holidays, religious festivals) that might lead to particularly
large vault flows are properly represented in the sample. It is important to use the observed vault
closing balances before the impact of incoming and outgoing cash orders by backing out these
shipments.

The Iron Reserve is based on an assumption of $500. It is difficult to generalize what a rule-
of-thumb ratio for sizing an iron reserve should be. This value will be highly dependent on the
business nature of the institution: if it mobilizes deposits, what type of savings products and how
volatile they are, the location of the branch, and many other factors such as the time required and
cost related to the emergency replenishment of the vault.

Finally, the days shipping cycle is assumed to be 2 based on the number of days it would take cash
to arrive once a shipment is triggered.

The “Key Vault Parameters” in Figure 3.2 are calculated based on the rules of thumb developed in
this chapter.

Once the lower trigger limit is calculated, users could consider adjustments based on business
growth or economic factors. For instance, in a country with high inflation, where the size of cash
transactions and the necessary vault balances grow constantly, it would be reasonable to revise the
lower trigger limit upwards once a year by multiplying it with (1 + balance sheet growth rate) of
the particular MFI.

Again it is important to note that these calculations are based on rule-of-thumb assumptions, and
the user must test the results. For instance, it may turn out that the lower trigger limit calculated
using a 90thpercentile outflow is far too high because the MFI rarely sees two very large withdrawal
days in a row. If the result of the rule-of-thumb appears unreasonable, one should gradually start
reducing the trigger limits and work towards a more practical minimum.

35
 Step 2: Define the Standard Incoming Shipment Size
The next question is: how much cash should the MFI order, once it hits the lower trigger limit?
This may already be determined by external limitations out of the MFI’s control which include
transport modalities and insurable amounts. For instance, the minimum order size may be limited
by the fact that the commercial bank where the MFI orders the cash will only deliver a certain
amount or there may be a maximum amount that a staff member may be insured against for
transporting cash in a briefcase. After estimating the lower limit, the next step is to calculate the
size of the order. The following formula can be used as an illustration of estimating incoming order
size:

Incoming Order Size = min { Daily net outflow90%;(upper trigger limit – lower trigger limit)/2 }

This formula assumes that the incoming order size will be calculated based on either an estimate of
the daily net outflow, which ensures that there will not be more than one delivery per day and the
mid-point between the upper and lower trigger limits. This is to ensure the opposite – that not too
much cash is triggered and therefore needs to be shipped away.

Example of an Order Size Rule

The following table provides an example of an order size rule. One possible rule of thumb
would be to make the standard size of a cash shipment equal to 90th percentile net daily vault
cash withdrawal over the last year (adjusted for business growth and inflation), using $10,000 in
this example. This would allow the branch to passively follow the actual vault cash trend in small
increments. In addition, it would avoid a situation in which several orders were triggered on the
same day. A look at Figure 3.2 makes this clear. An inbound order is initiated, as soon as the vault
balance plus incoming shipments in transit falls below the lower trigger limit.

If the shipment size was only $5,000 on a day that the branch experiences the large withdrawal of
$10,000, we could have an inefficient situation where there would be one shipment request in the
morning and a second order later in the day, when the withdrawals first exceed $5,000.

While a standard order size as large as the 90th percentile of daily net outflows safeguards against
more than one delivery order per day, it might bring the cash balance so close to the upper trigger
limit that we will soon be shipping money away again. One should therefore make sure that the
standard order size is in reasonable proportion to the upper trigger limit, potentially less than half
the distance between the lower and the upper trigger limit.

36
Assumptions for Figure 3.3:

Lower Trigger Limit: 40; Incoming Order Size: 10; 90th Percentile Daily Net Outflow: 10,
Shipping Cycle: 2 days

$’000 End of Day: 0 1 2 3 4 5 6 7

Net vault cash flow -10 -12 -5 -2 +5 0 +5
Effect shipments 0 0 10 10 10 0 0
Actual end of day vault cash 40 30 18 23 31 46 46 51
Note: Shipment Triggered 10 10 10 0 0 0 0

Shipments in transit 10 20 20 10 0 0 0
Lower Trigger Value = Actual end of day vault cash 40 40 38 43 41 46 46 51
+ Shipments in transit

Figure 3.3: Example of a Passive Vault Cash Ordering System

Step 3: Determine Upper Trigger Limit

In Step 3, we need to determine a reasonable rule for the upper trigger limit. Once the upper
trigger limit is reached, the branch would automatically start shipping cash away in certain standard
batches. The upper limit is therefore defined as a function of the lower trigger limit, the large daily
net inflow percentile and the length of the shipment cycle from order to delivery.

Upper Trigger Limit = Lower Trigger Limit + Days in Shipping Cycle * Large Net Daily Inflow90%

Most MFIs will want to avoid a situation in which there are simultaneous pending orders for
incoming and outgoing shipments to the same branch. In other words, it does not make sense
for an armored car taking cash away from the branch to pass another armored car on the highway
that is bringing cash to the branch. If the upper trigger limit is too close to the lower limit, the cash
balance could easily dip below the lower limit in the morning, triggering an incoming order and
then rise above the upper limit in the afternoon, releasing an order for an outgoing shipment.

As indicated earlier, the upper limit is first a function of the allowable amount based on cash
insurance limits and the actual storage capacity in the vault.

37
 Step 4: Define the Standard Outgoing Order Size
The last parameter that we still need to define in this rule-of-thumb vault cash system is the
standard size of the outgoing cash shipment. Again, the objective is to keep the shipment small
enough, so that we do not catapult the cash level below the lower limit just a day or two after
we shipped money off. At the same time, the shipment should be big enough, so that we avoid
triggering more than one shipment per day. The following formula can be used as an illustration of
estimating outgoing order size:

Outgoing Order Size = min {Large Daily Net Cash Inflow90%; (Upper trigger limit – Lower trigger limit) / 2 }

As is the case in the incoming order size, the formula assumes that the order size will be calculated
based on either an estimate of the daily net inflow, which ensures that there will not be more than
one delivery triggered per day and the mid-point between the upper and lower trigger limits so
there is little risk of triggering an incoming shipment shortly after the outgoing dispatch.

As an illustration, the worksheet titled OperatingRulesApplied of the VaultCashTools model

shows the calculations of the simple vault operating rules based on the four steps outlined earlier:

Figure 3.4: Screenshot from VaultCashTools, sheet: OperatingRulesApplied.

38
The formula in cell F23 shows the logic for determining whether a shipment is triggered on
November 17: The vault balance that is tested against the trigger limits consists of the prior
day’s closing balance plus today’s client driven cash flows (C23) plus any shipments in transit on
November 16 (G22). The shipments in transit on November 16 still include the portion that takes
effect on November 17 (D23), which is why we do not need to add D23 to the test balance. The
formulas must be built in this iterative and somewhat indirect way in order to avoid circular cell
references in Excel. The following exercise will make this clearer.

Exercise 3.2:

See Excel File VaultCashTools: Sheet “SampleVaultProfile” & Sheet “OperatingRulesApplied.”. There
you will find all of the above parameters under a rule of thumb system applied to a numbered ex-
ample. Examine and try to follow the formulas provided in the worksheets. Interpret the results.
By pressing F9 you will generate a new random series of vault closing balances and a new set of
vault parameters. You can also overwrite the time series of vault balances with real data from your
institution or with another fictitious sample to see the effects on the vault parameters.
When using raw institutional data and you find that there are either too many shipments, or actual
stock-outs occur frequently, you should adjust the parameters by varying the iron vault reserve and/
or the percentile threshold for determining “large” daily inflows and outflows.

39
 3.4 The Miller-Orr Control Limit Model
The second approach of passive vault management is the Miller-Orr model. This model derives
an optimal point, instead of formulating lower and upper trigger limits and order sizes based on
rules of thumb. The Miller-Orr model assumes that the daily movements of the vault cash balance
are the consequence of many individual actions that appear random to the financial institution.
On any given day, the cash balance either goes up or down, independent of the prior day’s change.
In probability theory, this situation is known as a random walk. The model uses the statistical
properties of the random walk and combines these with the holding and order cost functions. On
the basis of these assumptions, one can calculate an upper and a lower trigger limit as well as an
optimal return point.

Derivation of the Optimal Cash Parameters

The optimal values in the Miller-Orr model are calculated based on an optimal return cash level
that is given by:

The upper trigger limit is three times the optimal return cash level:

Upper Trigger Limit = 3 * Optimal Return Level

We will skip the mathematical justification for the optimal results according to the Miller-Orr
model.7 It is easy to see, however, that the inputs into the formula are consistent with our intuitive
argument about the important factors influencing cash levels in the passive vault approach. The
formula explicitly takes account of the holding costs of vault cash and the fixed costs per order. The
cash demand is represented by the variance of daily changes in the cash position8. The variance is a
measure of the volatility of vault cash demand. A larger variance leads to a higher return cash level,
a higher upper trigger limit and thus to larger cash shipments.

7 See M.H. Miller and D. Orr: An Application of Control Limit Models to the Management of Corporate Cash Balances,”
Proceedings of the Conference on Financial Research and Its Implications for Management, ed. A.A, Robichek, New York, John
Wiley and Sons. 1967.
8 The variance is calculated as the mean of the quadratic deviations from the average:

where n is equal to the number of days sampled, xi is the vault cash balance of day i, and x is equal to the average daily
balance over all periods under consideration.

40
Using the Miller-Orr Model
Figure 3.5 illustrates the use of the Miller-Orr model. The cash manager only takes action when
the vault balance hits either the upper or the lower trigger limit. When this happens, the difference
between the actual cash balance and the optimal return level is the amount of cash either transferred
to or transferred from the branch.
$

Upper
limit

Cash shipped
out
Optimal
return
point Cash shipped
in
Lower
limit

0 Time
Figure 3.5: Miller-Orr Control Limit Model

Below is a simple example of an application of the Miller-Orr model. The vault manager estimates
the fixed ordering costs at $100 per transaction, the total annual holding cost of cash is 18 percent
and the variance of daily cash balances is $35,000. For now, we hold the minimum cash reserve at
zero and then calculate the optimal return cash level and the upper trigger limit as follows:

With a lower trigger limit of $1,000 derived as per the Simple Operating Rules Approach, the
revised vault cash parameters become:

Lower trigger limit = $1,000

Optimal return level = $2,746
Upper trigger limit = $6,238.

41
 Exercise 3.3:

See Excel file VaultCashTools, sheet MillerOrr. You should experiment with the model by examining
the formulas and varying the holding cost and per shipment cost assumptions in order to see how
the optimal vault parameters respond. Now calculate the optimal Miller-Orr vault parameters for the
Branch A from Exercise 3.1. The correct results from Exercise 3.1 were: Branch A order cost of $186.07
with an annual holding cost of 18.83 percent.

You will need additional information on the variance of the daily closing balances and also need to
make a reasonable assumption about the value of the lower trigger limit. Use the following daily
closing balances for the month of March 2011:

Day Balance Day Balance Day Balance

1 26.700 8 28.100 15 28.900

2 30.000 9 21.700 16 29.500

3 30.700 10 27.500 17 24.200

4 29.300 11 30.500 18 26.300

5 20.700 12 23.200 19 24.700

6 18.900 13 21.700 20 27.200

7 27.100 14 28.900 21 18.600

42
3.5 Combining Trigger Limits with Short-Term Flow Forecasts
Limitations of the Passive Vault Cash Approach
The Passive Vault Cash Approach has various limitations. As mentioned this approach is based on
historical analysis and is not forward looking. Consequently, it does not take into consideration
seasonal patterns of loan disbursements and deposit movements which may impact cash flows
during certain times of the year. For instance, if management knows from past history that the
week before the year-end holidays will likely see deposit outflows three times as large as during a
normal week, it could consider increasing the size of cash shipments for that week.

Additionally, the passive vault cash approach considers loan disbursements as random events, when
in fact, they can be predicted when, for example, the status of loan files moves from “application
received”, to “approval” and “pending disbursement” in the MIS.

Recommended Approach: Hybrid Cash Planning

The Hybrid Cash Planning Approach combines the upper and lower trigger limits utilized in the
passive approach with short-term forecasts of upcoming vault flows. This is the recommended
approach as it is the most comprehensive and it involves more forward-looking methodology.

The Hybrid Cash Planning approach is based on four steps:

1. Estimate Future Daily Vault Cash Flows

Each Branch Managers is asked to forecast daily cash inflows and outflows for the upcom-
ing week, including expected loan disbursements and collections, deposits flows, other
inflows such as remittances, and any expected cash expenses. Based on this estimates, a
daily net vault cash flow can then be estimated.
2. Estimate Upper and Lower Trigger Limits
Upper and Lower Trigger Limits are calculated based on either the Simple Operating Rules
or Miller-Or Model discussed in the Passive Vault Cash Approach discussed above.
3. Determine Minimum Target Vault Levels
The upper and lower trigger limits of that day are compared to the expected daily net vault
cash flow of the following day. If a net inflow is expected, then the lower limit in the vault
should suffice. If an outflow is forecasted, then the forecasted outflow is added to the lower
limit. This amount is adjusted to ensure that if there is any outflow, it reflects the minimum
shipment size desired by the institution.
4. Determine Cash movements
The next step is to determine whether cash movements should be triggered. These
amounts are estimated using the Simple Operating “rules of thumb”. If the forecasted clos-
ing balance is lower than the target vault level, an incoming order must be triggered. If the
forecasted closing balance is higher than both the Upper Trigger Limit and the Minimum
Target Vault Level, then an outgoing order should be triggered.

43
 The following example is based on the Vault Cash Tool, see worksheet “Weekly Cash Planner.”
The first step is to estimate Future Daily Vault Cash Flows. Each Branch Manager fills out a
planning form (see template below) on Friday for the upcoming week and estimates the day-by-
day inflows and outflows that will come through the branch vault. The Daily Net Vault Cash Flow
is the sum of forecasted inflows and outflows for the day. The Forecasted Closing Balance is the
opening balance plus the daily net flow.
In step two, the Upper and Lower Trigger Limits are then estimated based on one of the approaches
discussed earlier, as can be observed in Table 3.6 below.

Figure 3.6: Hybrid Weekly Cash Planning Template

In step 3, Minimum Target Vault levels are estimated. In Figure 3.7 below we can see that, cell
E40 shows the Minimum Target Vault Level at Tuesday close of business. The formula looks
ahead at the vault cash flow forecasted for the next day in cell F34. If that will be a net inflow, the
lower limit in the vault should be sufficient for the next day. Hence, the target closing balance on
Tuesday would be the Lower Trigger Limit as per cell E39. If a net cash outflow is forecasted for
Wednesday, i.e. F34<0, then the target balance for Tuesday closing is the lower trigger limit plus
the forecasted outflow, i.e. E39-F34.

44
The Minimum Target Vault Level is adjusted to ensure a reasonable minimum shipment size. In this
case the minimum shipment increment was calculated as a quarter of the distance between the lower
and the upper trigger limits. So, if there is an outflow on the next day, the target balance is the minimum
limit plus either the forecasted outflow or the minimum shipment whichever is larger.

Figure 3.7: Decision logic for deriving the Target Vault Level, line 40.
Finally, in step four cash movements are determined. In cell E42, we now decide if a cash movement
should be triggered on Tuesday, see Figure 3.8. An incoming shipment should be ordered,
if the closing balance is less than the target vault level, i.e. IF(E40>E36) is true. The incoming
shipment is estimated as the larger of the target balance minus Forecasted Closing Balance
and the efficient minimum shipment size: MAX(E40-E36,(E38-E39)/4),0). If the Forecasted
Closing Balance is greater than both the Upper Trigger Limit and Minimum Target Vault Level,
i.e. IF(E36>MAX(E38,E40) is true, then we should order an outgoing shipment. This outgoing
shipment is estimated as the absolute larger of the difference between Forecasted Closing Balance
and Upper Trigger Limit and the minimum shipment of ¼ of the distance between the two trigger
limits: MIN(E38-E36,(E39-E38)/4). Just like for the incoming orders, a rule of thumb is used in
rounding up the outgoing shipment size to at least a quarter of the distance between the trigger
limits, in order to avoid very small cash transports.

Figure 3.8: Decision logic for determining cash shipments, line 42.

It is important to note that the above cash planner assumes that the branch has a cash transfer cycle
time of 1 day (i.e. it is within reach of a daily armored transport route, or a MFI staff member can make a
same-day trip to the bank for smaller shipments, such that an order triggered on Tuesday arrives before
close of business on Wednesday). If the cycle time is longer, for example two days, the formulas should
look ahead at the cumulative flows over the next two days.
Column I sums up the transactions for the entire week and gives a net view of the vault forecast for
a weekly planning increment. 45
 Exercise 3.4:

Open the accompanying Excel file VaultCashTools and go to the sheet WeeklyCashPlanner. Try to
capture a vault cash forecast for next week for the specific branch where you work or for your entire
MFI. Adjust the Lower and Upper Trigger Limits to fit with the overall dimensions of your vault cash
operations. Derive possible Lower and Upper Trigger Limits for this example using the Miller Orr
Control Limit Model as prepared in the sheet MillerOrr. Interpret the results in the Daily Vault Cash
Plan and make adjustments to the parameters and the decision logic as necessary to obtain plausible
and practical results.

3.6 Additional Considerations in Vault Cash Operations

Seasonality Analysis of Vault Flows
A Seasonality Analysis examines the vault cash flows and the resulting daily closing balances for
a particular branch for certain recurring patterns within each month. For example, a branch that
serves many formally employed workers will likely see large vault cash outflows within the first few
days after payday. Small merchants and informal traders may make larger than average deposits
during the first two weeks of the month when the workers spend their wages.

Although it is impossible to determine in advance the specific and repetitive patterns that can
emerge from these multiple factors, recurring intra-month patterns in past vault data can be
measured through the use of trend indeces based on historical data.

A percentage index can be estimated relative to the underlying trend for a particular day using
historical values. For example, assume there is an overall long-run trend that vault balances grow
over time. Instead of simply making the continuation of this trend line, the forecast of balances
can be instead multiplied by the trend forecast with a pattern index for the particular day of the
month, say 91% for the 3rd of the month. This is based on an analysis of past observations that
show that on average, the 3rd working day of the month closes with only 91% of the underlying
trend balance.

The sheet “VaultSeasonality” in the accompanying Excel file “VaultCashTools” calculates an

example of a daily pattern index extracted from 12 months of closing balance observations. The
example uses a numerical day-of-the-month index from 1 to 31 to analyze whether all the 21st
days of the month have higher or lower balances relative to trend.

Figure 3.9 provides a snapshot of these average index values relative to trend for each day of
the month calculated from one particular set of randomly generated vault closing balances over
the course of a year. In this example, heavy vault inflows can be observed on the 20th and 21st,
followed by big outflows on the 22nd through 25th of the month. On average over the course of
the last year, the 21st, for example, closed at 137 percent of the underlying trend balance.

46
 1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Figure 3.9: Average Daily Index of Vault Closing Balances Relative to Trend

The daily pattern index helps to analyze client behavior patterns that can be then incorporated into
the vault cash forecast. Based on this pattern analysis, the branch manager should simply make an
adjustment to the plan by multiplying the projection by the appropriate index.

In addition to monthly cycles, intra-week seasonal cycles can also be analyzed to determine
whether certain days of the week have commonly lower or higher balances relative to the global
trend. The calculation would be identical to the example in the sheet VaultSeasonality except that
the function WEEKDAY() is used instead of DAY() to obtain a day index ranging from 1-7 for
which average indeces can then be calculated relative to trend.

Exercise 3.5:

See Excel file VaultCashTools, sheet VaultSeasonality. Try to follow the logic of the calculations. Re-
place the historical vault balance figures with your own assumptions, with other random data or with
observations from your own institution. Interpret the results and decide whether it would be useful to
apply the pattern index in forecasting the next month’s vault balances under these circumstances.

Make a copy of your VaultSeasonality worksheet and analyze weekly patterns rather than a full
day-of-the-month cycle. Instead of the DAY() function, you could use the WEEKDAY() function to
determine the positioning of a particular date in the week.

47
 Utilizing Buffer Checks
Buffer checks are checks drawn on a correspondent bank near a particular branch issued for a
specific amount should an emergency need arise. The check already bears the control signature of
the treasurer and the branch manager is authorized as co-signatory on the account so may redeem
the check for vault cash at the commercial bank should there be a need. The advantage of a buffer
check is that the funds are only debited to the MFI’s account as and when it is actually cashed, thus
reducing the opportunity cost compared to actually holding the amount as an iron vault reserve.

Buffer checks are not a shortcut to the vault cash planning problem, as they are not equivalent
to actual cash in the vault. They are considered a practitioner’s instrument to realize a small
emergency cash order within a short cycle time. There is still action required to: (1) discover that
an emergency order will be necessary, (2) to mobilize the authorized check signatory and (3)
to make the trip to the bank branch for cashing the check. One should also consider that bank
teller opening times might not always match with MFI client service requirements and that the
bank will also often require at least a day’s notice, before they can guarantee to have funds for
cashing a large buffer check. Nevertheless, buffer checks are useful tools that branch managers
and Treasurers can consider if available in their particular context.

Managing Multiple Currencies in the Vault

The realities of a weak local currency and a creeping dollarization of the economy often make
it necessary to carry out transactions in a foreign reference currency in addition to the national
currency, should it be allowed to do so. If the MFI denominates some of the loan portfolio and /
or its deposits in the dominant parallel foreign currency, then generally it must also be prepared to
accept and stock physical foreign cash in its vault. These foreign exchange cash stocks are planned
like a set of separate branch vaults. Each branch essentially has two vaults, one in local and a
second in foreign currency with separately tracked balances and flow forecasts etc. This is different
from occasional foreign exchange bureau transactions where the MFI might buy foreign cash, but
would not be expected to always hold enough foreign cash stock to be ready to sell to clients at
any time.

Also, one need not be concerned that the multiple currency vault planning idea would lead to
an excessive multiplication of distinct currency sub-vaults that need to be planned per branch.
In most countries where there is a weak local currency, there is generally one dominant foreign
reference currency that is used as an anchor for valuing big ticket items, salaries, rents and prices
of daily commodities. It is fully sufficient to maintain planned stocks in that dominant parallel
currency, typically US$ or EUR. Transactions in any other foreign currencies would be treated as
exotic requests and the MFI would be willing to act as a foreign exchange bureau in buying and
selling the currency as a courtesy, but the MFI would not systematically hold inventory in third
currencies.

48
 3.7 Implementing a Vault Cash Planning System in a Branch
Organization
Network-Wide View of Vault Cash
So far, the cash planning approaches have only considered a single branch or stand-alone cash
acceptance point as the unit of analysis. The next step is to look at the MFI’s entire network of
branches. Vault cash in a branch network really is a collection of independent vault service points.
The MFI must be able to meet all vault cash payment demands in their specific location, so there
cannot be any inter-branch averaging, where a cash shortage in one branch is arithmetically
compensated by an excess balance in another.

Network-wide vault management must therefore be based on a side-by-side analysis of the vault
parameters of all branches. The following vault cash overview in Figure 3.10 is adapted from the
Treasury tools of a Ugandan MDI.

Figure 3.10 shows a vault cash snapshot of all 22 branches and the Head Office (HO), which
no longer has a separate vault. The light yellow rectangles give the allowable vault ranges, that is
the space between the trigger limits for each individual branch. The diamonds mark the last daily
closing balance and the red line indicates the expanded potential balance as the sum of the last
closing balance plus any buffer checks that are already issued and available at that branch.
140
Millions

120

100

80

60

40

20

0
ADJ BMB BKY BSA BSK BWR CORP HO KJS KST KYG KBG KSZ KSR KGR LWR LTD MTN NKW RSR TRY TRR ZBW

Min - Max Current Cash Current Cash and Buffer Checks

Figure 3.10: Daily Vault Cash Overview for Multiple Branches, Uganda MDI.

A vault cash overview as in Figure 3.10 should be integrated into the weekly Vault Cash Planner.
Each branch submits vault cash forecasts for the upcoming week to Head Office Treasury on
Friday afternoon, where the individual branch plans are consolidated into a current actual and
daily forecasted vault cash overview. Head Office Treasury should review the branch forecasts for
accuracy and might fine-tune the planned vault movements in order to combine shipments where
possible or optimize the armored car routes.

49
 A Vault Cash Dashboard
The Final step of the Vault Cash Planning System is for the development of a comprehensive vault
cash dashboard that gives a concise summary of the consolidated vault cash position of a branch
operation. The dashboard should be reviewed by Treasury on a daily basis. A comprehensive
dashboard should include the following elements:

• Current actual opening balances in all branch vaults

• Cumulated net flow forecast one week forward by branch and across the
network
• Lower – upper trigger limit ranges per branch
• Available buffer checks by branch and by correspondent bank
• pending inbound/outbound shipments by branch and for total network
• Forecasted end of week Closing Balances by branch and for entire network
Worksheet Vault Dashboard of the Vault Cash Tools model includes a template for a vault cash
dashboard that condenses all of the data elements outlined above into a compact graphical view of
a decentralized vault cash system.

Vault Cash Dashboard

Vault Parameter Summary for All Branches

From Weekly Vault Plan: Branch A Branch B Branch C Branch D Branch E Branch F Branch G Branch H Totals
Opening Balance 9,500 15,000 45,000 7,000 25,500 5,000 9,000 6,500 122,500
Net Vault Cash Flow -2,000 3,000 -13,000 -3,000 -9,000 -2,500 2,000 -3,000 -27,500
Forecasted Closing Balance 7,500 18,000 32,000 4,000 16,500 2,500 11,000 3,500 95,000
Upper Trigger Limit 11,000 12,000 40,000 10,000 30,000 8,000 10,000 8,000 129,000
Lower Trigger Limit 3,000 5,000 15,000 2,000 5,000 2,000 2,000 2,500 36,500
Minimum Target Vault Level 5,000 5,000 28,000 5,000 14,000 4,500 2,000 5,500 69,000
Cash Delivery / Deposit Triggered 0 -6,000 0 2,000 0 2,000 -2,000 2,000 -2,000
Closing Balance after Shipments 7,500 12,000 32,000 6,000 16,500 4,500 9,000 5,500 93,000

Buffer checks Bank A 1,500 1,000 500 500 3,500

Buffer checks Bank B 500 1,000 1,000 2,500
Buffer checks Bank C 3,000 500 3,500
Total Buffer Checks by Branch 1,500 1,500 3,000 500 1,000 500 500 1,000 9,500

Shipments Total Closing Cash Sum Upper Sum

Total Opening Cash in Network: 122,500 Net Vault Flow for the Week: -27,500 -2,000 93,000 129,000 36,500
Triggered: after Shipments: Limits: LowerLimits:

50,000

45,000
Opening Balance

40,000 Trigger Limit Range

35,000
Closing Balance Forecast

30,000 Closing Balance after Shipments

25,000
Available Buffer Checks

20,000

15,000

10,000

5,000

0
A B C D E F G H

Figure 3.11: Vault Cash Dashboard. Screenshot from VaultCashTools.

50
Organizational Aspects of Implementing a Vault Cash Planning System
One of the biggest challenges in improving vault efficiency across the branch network of a MFI
is that most branch managers are not fully aware of the holding cost of vault cash. Most MFIs do
not have fully-fledged transfer pricing and cost accounting systems that allocate vault holding and
transport costs down to branch-level profit centers.

The starting point for more efficient vault cash operations therefore should be that Head Office
Treasury sensitizes branch managers on the full cost of vault holdings and the importance of
collaborating in a weekly vault cash planning cycle. The next practical steps towards implementing
a weekly branch cash planning system are as follows:

1. Head Office Treasury and branch managers jointly perform an analysis of vault
shipment unit costs, cycle time, typical vault flows per branch and reconstitute a daily
closing balance time series (excluding the effect of shipments) over multiple months,
ideally a full year.
2. With the data inputs from step 1, Treasury calculates lower and upper trigger limits
and order sizes using the Miller-Orr model for each branch and checks these for
plausibility and feasibility against the branch manager’s experience, physical vault
capacity and insurance limits, etc.
3. Branches are asked to participate in a weekly cash planning routine using a format
similar to the template introduced in Figure 3.6.
4. As the branches adopt the vault cash planning framework, they will track the accuracy
of their forecasts against the realized actual flows on a weekly basis.
5. Head Office Treasury will review the performance of the forecasts at least quarterly
and establish target ranges for the forecast accuracy.
6. In the absence of a full profit center cost accounting system, the vault forecast accuracy
could become part of the incentive formula for branch managers. This would keep the
pressure up to plan thoroughly and continuously refine the understanding of customer
behavior and seasonal patterns.

51
 Chapter 3 - Key Concepts and Next Steps
• The objectives of efficient vault cash operations are to (1) provide cash
points to meet all normal customer payment requirements, (2) minimize
the cost of holding idle vault balances, and (3) control the frequency of cash
shipments and the associated transport costs.
• The basic unit of analysis in vault cash operations is a geographically distinct
branch or cash point, where the MFI transacts with clients in vault cash.
A bottom-up branch-by-branch analysis is necessary because various cash
points may have different order costs and a distinct profile of cash demand
and thus will require different holding levels and shipment patterns in order
to avoid stock-outs while minimizing cost.
• A passive vault cash system does not anticipate future inflows and outflows
but simply establishes lower and upper trigger limits on vault balances that
release an order for transferring vault cash when reached.
• A hybrid vault cash system combines passive trigger limits with a very short-
term forecast of vault cash flows. A hybrid system thus makes use of the
forward visibility of vault cash flows through the IT system and leverages the
experience of front-line staff in predicting the cash transaction patterns of
their clients.
• The toolkit recommends to move towards a hybrid vault cash approach.
To get started, it would be useful to perform an analysis of the holding and
order costs, cycle times and recent actual vault flows as a joint initiative
between treasury and branch managers. Even if one does not progress right
away to the perfect hybrid planning system, the act alone of analyzing vault
more closely will already improve awareness among branch staff of the costs
at involved.

52
CHAPTER 4
FUNDAMENTALS OF LIQUIDITY MANAGEMENT

4.1 Introduction
After delving into the finer issues of vault cash efficiency, we should take a step back and remember
that planning physical vault balances is a small piece in the much broader topic of liquidity
management. Vault cash is a short-term sub-cycle to the management of book balances with banks
and other liquid assets. Poor vault cash management rarely rises to the level of a financial risk that
could threaten the viability of the MFI. The following section focuses on the broader topic of
Liquidity Management as an essential aspect of Cash Management. The section first highlights the
objectives and importance of liquidity management, and then focuses on some of the challenges.
Then, we focus on measuring liquidity management through two methods: maturity gap reports
and liquidity ratios. Finally, we discuss the importance of cash flow forecasting as a liquidity
planning and monitoring tool.
For a more in-depth explanation of liquidity risk in the context of Financial Risk Management,
please reference WWB’s Toolkit for Developing a Financial Risk Management Policy9, which
provides an organizational framework for managing financial risk management and addresses
Asset Liability Management in detail, including liquidity risk.

Importance of Liquidity
Liquidity is the ability of a financial institution to honor all commitments of cash payments as they
fall due. These commitments can be met either by drawing from a stock of cash holdings, by using
current cash inflows, by borrowing cash or by converting liquid assets into cash.
Liquidity is a vital condition for any business and an even more crucial for financial institutions
because they are particularly vulnerable to unexpected and immediate payment demands. To stay
in business, a MFI must be able to pay out legitimate deposit withdrawals and credit requests
instantly. Liquidity Risk is therefore the possibility of negative effects on the interests of owners,
customers and other stakeholders of the financial institution resulting from the inability to meet
current payment obligations in a timely and cost-efficient manner.
In the vast majority of daily transactions, a financial institution does not act on its own behalf, e.g.
paying rent for bank offices or buying photocopy paper, but rather functions as an intermediary
between savers and borrowers or as a payment agent for transfers between businesses or
individuals. For this reason, the failure of a financial institution can have far-reaching economic
effects on the entire national financial system. A systemic effect can arise even from a liquidity
crisis at a relatively small MFI or cooperative bank, because the default of one such institution can
compromise the public trust in the entire microfinance industry in that country. For instance, in
2002/2003, a string of small cooperative banks failed in the Indian state of Gujarat after losses on
speculative margin loans to stock brokers and other irregularities.10 These events discredited the
entire cooperative banking movement in the state and led to “contagion” and deposit withdrawals
at many other well managed institutions.

9 Tool for Developing a Financial Risk Management Policy. Louise Schneider-Moretto, WWB 2005.
10 Madhavpura Mercantile Cooperative Bank, Charotar Cooperative Bank,Vishnagar Cooperative Bank, Laxmi Cooperative Bank, 53
Diamond Jubilee Cooperative Bank, Suryapur Cooperative Bank, General Cooperative Bank and Baroda People’s Cooperative
Bank etc.
 As more MFIs offer a full range of banking services including savings, checking, transfers and
even insurance, liquidity planning becomes more important. Dealing not only with the fluctuating
demand for loans but also with volatile deposits and unpredictable transfers makes the task of
liquidity management quite complex and requires planning. At the same time, the growing size
of the microfinance industry means that MFIs’ liquidity has become an issue of macroeconomic
importance, at least at a local or regional level. Liquidity therefore is an essential survival skill for
every microfinance institution.

Objectives of Liquidity Management

The general goals of liquidity management are to:

• Honor all cash outflow commitments on a daily and ongoing basis

• Minimize the cost of foregone earnings on idle liquidity
• Satisfy minimum reserve requirements and other prudential liquidity stan-
dards
• Avoid the additional cost of emergency borrowing and forced liquidation of
assets
The time horizon for liquidity management is short. It involves detailed estimations of the size and
timing of cash inflows and outflows over the next few days and weeks. Often, liquidity projections
are extended up to a year with diminishing detail on the far end of the time line.

As an illustration, some practical questions that a liquidity manager needs to address in this context
can include:

• How can a MFI in an agricultural community predict seasonal deposit varia-

tions resulting from crop production cycles?
• How does loan demand vary with macroeconomic cycles of boom and reces-
sion?
• What is the expected day-by-day volatility of the aggregate deposit supply
raised from many small savers?
• How will the depositor and borrower behavior change during a macroeco-
nomic crisis scenario?

54
Challenges in Liquidity Management
Liquidity management operates in an environment of uncertainty. There is uncertainty about
future customer behavior, about general macroeconomic conditions, about weather patterns that
influence agricultural production etc. Liquidity management therefore is not about determining a
single optimal level of cash to hold. It is about charting a reasonable compromise between the risk
of a liquidity shortage and the loss of income from not investing idle resources in interest earning
assets.

What makes liquidity management even more complex is that most of the factors determining
liquidity are interdependent. Loan demand, for example, can be closely linked with deposit flows.
MFIs may serve the same type of clientele on the asset and liability side, for example a community
composed largely of farmers of the same crop. Most customers probably will want to withdraw
a large share of their deposits during the planting season to buy fertilizers and supplies. This is
precisely the time when demand for loans is highest for the same reason. At harvest time, however,
customers will tend to replenish their savings while others repay their seasonal loans. Without
proper diversification of its customer base, a MFI may well face seasonal swings between periods
of unproductive excess liquidity and lost lending opportunities because of liquidity shortages.

There are many more explicit and implicit inter-relations between liquidity factors to consider.
Tight liquidity, for example, which may require a MFI to refuse a loan to a qualifying customer, is
not just a missed business opportunity. Turning a legitimate loan demand away can have severe
implications for customer confidence. If word spreads that loans are refused, depositors might
conclude that the MFI is in financial trouble and will rush to withdraw their funds. Few banks can
expect to survive an outright run on their deposits.

Liquidity management resembles a delicate balancing act. In retrospect, institutions that survive
always appear to have had excessive liquidity, while banks or MFIs that fail were closed because
they could not meet payment demands. The trick is to have enough liquidity so that the institution
will never be challenged to use it. Conversely, an institution that is overly aggressive in minimizing
liquidity in order to enhance profits may find that its correspondent banks and depositors will
decide to test its liquidity by canceling credit lines and withdrawing deposits precisely when
liquidity is already tight.

Liquidity essentially comes down to market confidence. As long as depositors, funders and
regulators have confidence in the long-term viability of the institution, temporary shortages of
cash can be mitigated by borrowing or by selling high quality assets. But when that confidence
is compromised and depositors and creditors race to get their money back, the collapse of the
institution is almost certain.
A liquidity crisis is typically the final manifestation of other deeper problems, i.e. the symptom
rather than the disease. The underlying problem could be loan losses, excessive operating expenses,
fraud and other operational risk events or losses on foreign exchange positions.

55
 And finally, a word on the difference between solvency and liquidity is important because these
terms can easily be confused: Solvency is having positive equity commensurate with the risks of
the institution, i.e. a surplus of economic net asset value above the economic value of net liabilities.
Therefore, a solvent MFI can be liquid on average, but in reality, liquidity must still be assured on
a daily basis. Conversely, an insolvent institution may still be liquid, i.e. able to pay, but because
the value of equity is already negative, before long, could be liquidated or its existing shareholders
wiped out.

4.2 Measuring Liquidity on the Balance Sheet

The Rationale for Balance Sheet Measures
The conventional place to start in measuring liquidity is on the balance sheet, by looking at certain
ratios between liquid or illiquid assets and liabilities and at the contractual maturity structure of
the balance sheet in the form of a maturity gap report.
Such static balance sheet measures of liquidity generally are only an entry point for a first quick
analysis. Sometimes they are also used as an index to describe the desired positioning outcome,
which has been derived by other more sophisticated planning methods. Never can successful
liquidity management be as simple as saying: “keep this ratio to x% and everything will be fine”.
Nonetheless, ratios and other static balance sheet measures are a useful point of departure. Many
MFIs use ratios in addition to detailed cash flow projections as a tool for high level planning and
for formulating simple operating rules or positioning limits. External analysts, regulatory agencies
and investors find ratios practical, because they can give a quick indication of the overall liquidity
position based on just a published balance sheet. Ratios make it easy to compare liquidity between
different institutions or to calculate average liquidity for an entire sector of the financial industry.
By tracking the evolution of a ratio over time, rather than just taking a single snapshot, one can also
detect interesting trends in the liquidity position. While static balance sheet measures can seldom
provide answers about changes in the MFI’s liquidity or operational efficiency, marked trends or
reversals in their development can point to questions that merit further investigation.
Proper liquidity management always involves a detailed plan of future cash flows that will ensure
that the MFI can systematically meet its payment obligations, not just on average, but every day.
The larger part of this chapter will therefore be dedicated to deriving a dynamic plan of future net
funding requirements, the most important baseline input for active liquidity management.

Maturity Gap Report

Figure 4.1 shows the typical format of a contractual maturity gap report for a MFI. The gap report
is simply a stratification of the entire balance sheet into time bands by contractual maturity. This
means that the carrying values of individual asset or liability items are classified by when they are
available or payable contractually. The time intervals are counted forward from the balance sheet
date.

56
Take the line retail term deposits on the liability side in Figure 4.1: On the balance sheet day of
this report, the bank had 24,000 in term deposits becoming payable within the next 7 days, 35,000
coming due within 8 to 30 days and so forth. “Contractually due” means that the client gave
specific instructions that the deposit be placed for a certain time.

The Maturity Gaps at the bottom of Figure 4.1 are simply the sum of all the assets minus the
liabilities becoming due during the particular time interval. By adding these period gaps up from
left to right, we find the Cumulative Maturity Gap. By definition, the final Cumulative Gap on the
right must add up to zero, because total assets should be equal to total liabilities and equity.
Contractually Available Within: 1-7 days 8-30 days 1-3 mths 3-6 mths 6-9 mths 9-12 mths 1-2 yrs 2-5 yrs > 5 yrs TOTAL
ASSETS
Vault Cash 12,000 - - - - - - - - 12,000
Current Accounts at Central Bank 13,000 - - - - - - - - 13,000
Mandatory Reserve at Central Bank 11,700 - - - - - - - - 11,700
Deposit Accounts at Central Bank 5,000 - - - - - - - - 5,000
Repo-eligible Treasury Bills and Bonds - - 5,000 10,000 10,000 5,000 5,000 7,950 42,950
Reverse Repo Placements 2,500 2,500 - - - - - - - 5,000
Other Financial Investments - - - - - - 2,000 - - 2,000
Current Accounts at Commercial Banks 4,000 - - - - - - - - 4,000
Deposits at Banks and Financial Institutions 5,100 15,000 7,500 - - - - - - 27,600
Loans to Financial Institutions - - - - - - - - - -
Guarantee Deposits at Banks - - - - - - 2,000 - - 2,000
Gross Customer Loan Product 1 3,000 7,000 25,000 25,000 30,000 15,000 15,000 - - 120,000
Gross Customer Loan Product 2 2,000 6,000 16,000 18,500 25,000 7,500 15,000 - - 90,000
Gross Customer Loan Product 3 1,500 5,500 18,000 26,000 23,000 16,000 20,700 - - 110,700
Loan Loss Reserve 4,200 - - - - - - - - 4,200
Net Customer Loans 2,300 18,500 59,000 69,500 78,000 38,500 50,700 - - 316,500
Other Receivables - - 3,500 - - - - - - 3,500
Fixed Assets - - - - - - - - 20,172 20,172
Other Assets - - - - - - - - 3,000 3,000
Total Assets Available: 55,600 36,000 75,000 79,500 88,000 43,500 59,700 7,950 23,172 468,422

Contractually Maturing Within: 1-7 days 8-30 days 1-3 mths 3-6 mths 6-9 mths 9-12 mths 1-2 yrs 2-5 yrs > 5 yrs TOTAL
LIABILITIES
Current Accounts of Financial Institutions - - - - - - - - - -
Bank Overdraft Borrowings 3,000 - - - - - - - - 3,000
Repo Borrowings 2,500 7,500 5,000 - - - - - - 15,000
Financial Institution Term Deposits 2,000 10,000 5,000 - - - - - - 17,000
Client Current and Transaction Accounts 29,600 - - - - - - - - 29,600
Savings Deposits at Sight 26,000 - - - - - - - - 26,000
Retail Term Deposits 24,000 35,000 15,000 20,000 10,000 9,000 16,000 - - 129,000
Pledged Collateral Deposits - 1,052 2,500 2,500 2,500 2,500 10,000 - - 21,052
Senior Medium & Long Term Borrowing 3,800 4,000 5,000 4,000 5,500 3,500 58,000 60,000 20,000 163,800
Subordinated Borrowings - - - - - - - - 22,720 22,720
Other Liabilities 150 2,000 4,000 - - - - - - 6,150
Accruals & Reserves - - 600 - - - - - - 600
EQUITY -
Paid-In Common Shares - - - - - - - - 18,500 18,500
Share Premium & Other Capital Reserves - - - - - - - - 10,500 10,500
Retained Income - - - - - - - - 5,500 5,500
Liabilities Payable & Equity 91,050 59,552 37,100 26,500 18,000 15,000 84,000 60,000 77,220 468,422

Maturity Gaps (35,450) (23,552) 37,900 53,000 70,000 28,500 (24,300) (52,050) (54,048)

Cumulative Maturity Gap (35,450) (59,002) (21,102) 31,898 101,898 130,398 106,098 54,048 -

Figure 4.1: Sample Contractual Maturity Gap Report

Taken as a whole, Figure 4.1 gives us a snapshot of the maturity structure of the balance sheet on
a certain date. The 1-7 day gap of -35,450 tells us that the institution has 35,450 more payable
liabilities than available assets over the course of the next week.

57
 At first reading this sounds alarming. In reality, this is a typical situation for deposit-taking
institutions as they are funded to a large degree by client savings, transaction accounts and retail
term deposits.

Figure 4.1 shows a preponderance of negative gaps because a large portion of funding sources
are made up of sight deposits and short-term deposits from retail clients. As mentioned, this
situation is not alarming because of the difference between real and contractual behavior of assets
and liabilities. Although savings account holders may contractually be allowed to withdraw
their entire deposits tomorrow. Statistically, only a small fraction of depositors will, and even that
small withdrawal is often compensated by inflows from new clients and additional deposits from
existing customers. Even customer term deposits that have a clearly defined contractual maturity
will frequently be extended or rolled over for a long time.

As a result, deposit-taking institutions must closely study the “real” behavior of depositors in a
normal business environment and also - very importantly - under stress conditions.
Similar differences between contractual and real behavior are also at work on certain types of
client assets. Overdraft utilizations by businesses and households are theoretically due tomorrow
but tend to be rolled forward for a long time. The behavior of a portfolio of short-term working
capital loans to micro-entrepreneurs in aggregates appears to be never really collected. Instead,
most loans are quickly renewed.
Although one can observe several differences in contractual behavior in maturity gap reports, they
are still a useful data input for further study. Many institutions apply the results of studies of
real behavior to the gap report and replace contractual maturities with experiential schedules of
deposit “mortality,” both under normal and unfavorable business conditions. In addition, external
analysts and regulators often collate the gap reports from multiple institutions to get a consolidated
maturity profile of the entire sector, to which they might apply their own industry-wide stress test
assumptions.

Model:

Compare: Excel file LiquidityManager, sheet MaturityGapReport.

58
Balance Sheet Liquidity Ratios
Liquidity ratios are condensed balance sheet measures of liquidity that are calculated to manage
the liquidity position of an institution at a certain point in time. Adequate liquidity measures
must be developed and designed for each institution to limit short-term liquidity exposure and to
monitor growth. Certain regulatory liquidity ratios are also developed as a miniature stress test as
they can include standard run-off assumptions for various types of deposits, for example.
Some common liquidity ratios that are suitable for MFIs are highlighted in the graph below and
explained in detail in this section.

Ratio Name Definition

Cash Position Indicator = (Cash and Due from Banks) / Total Assets
Free Cash Position Indicator = (Cash and Due from Banks less Mandatory
Reserve Holdings) / Total Assets
Liquid Assets Ratio = Liquid Assets / Total Assets
Capacity Ratio = Not Loans / Total Assets
Purchased Funds Ratio = Short-Term Borrowing / Total Assets
Total Deposit Ratio = Retail Customer Deposits / Total Assets
Loan-to-Deposit Ratio = Net Loans / Total Customer Deposits
Reserve Ratio = Cash Assets / Total Customer Deposits
Liquidity Coverage Ratio 30 days = Liquid Assets 30 days / Liability Run-Off 30 days
Operating Liquidity Ratio = Liquid Assets / (Monthly Operating Expenses +
Recent Monthly Net Loan Portfolio Growth)
Figure 4.2: Overview of Liquidity Ratio Definitions

Cash Position Indicator

The cash position indicator compares vault cash and demand deposits at banks, including the
central bank, to the total asset base of the institution at a particular moment in time:

Cash Position Indicator = (Cash and Due from Banks) / Total Assets

Cash Assets (or simply cash) are generally understood to include (1) vault cash, (2) demand and
transaction deposits held with commercial banks or at the central bank, and (3) cash items in the
process of collection resulting from check remittances and payment transfers.

A larger Cash Position Indicator means that the institution holds a larger proportion of its balance
sheet in cash assets and therefore is in a stronger position to meet any immediate settlement
requirements.

59
 If the MFI is subject to mandatory reserve requirements on its deposits, some of the numerator
in the ratio would actually consist of mandatory central bank deposits that are not really available
for meeting payment demands. Mandatory reserves must remain with the central bank at least on
average in order to comply with the reserve maintenance requirements. One could exclude the
required reserve from the numerator to give an indication of the actually available cash position:

Free Cash Position Indicator =

(Cash and Due from Banks less Mandatory Reserve Holdings) / Total Assets

Liquid Assets Ratio

This is probably the simplest and most relevant liquidity ratio of all. The numerator includes the
cash items of the Cash Position Indicator plus any other highly liquid assets that can rapidly be
sold for cash with minimal price risk.

Liquid Assets Ratio = Liquid Assets / Total Assets

Liquid Assets can be defined as cash assets plus other liquid assets that can readily be sold for cash
with negligible price depreciation. What makes these assets liquid is the fact that they can be sold
in an active secondary market at a moment’s notice, so that the cash value is often available for use
the same day.

Despite its simplicity, the liquid assets ratio already incorporates the lessons from the 2007–2009
global financial crisis. The crisis demonstrated that even for the world’s largest banks, asset liquidity
is critical in times of market turbulence. The ability to borrow liquidity instead in the interbank
market is very unreliable and tends to evaporate just as banks would need it most. Hence, the
insistence by regulators that banks should hold large reserves of unencumbered, highly liquid
assets. Such top quality assets include bills and bonds issued by central governments and by other
highly rated institutions that are eligible as collateral for central bank borrowing and have a deep
secondary market. In practical terms, this means that a MFI should hold a substantial reserve in
term deposits and/or Treasury bills and bonds beyond its immediate payment needs.

Narrowly interpreted in respect to liquidity risk, a higher liquid assets ratio is always better.
Depending on the yield of the liquid investments, however, the opportunity loss of not deploying
these resources in more profitable client assets begins to weigh heavily, if we push this ratio even
higher.

In steady-state operations, where we do not plan a major expansion of the loan portfolio and with
a reasonably stable and long-structured funding base typical for microfinance, a liquid assets ratio
of 15% to 20% would be an appropriate target.

60
Capacity Ratio
Compared to the liquid assets ratio, the capacity ratio looks at liquidity from the opposite end of
the asset structure. It measures the largest illiquid asset class relative to total assets. Other than
some fixed investment in property and equipment, the bulk of assets in microfinance are either in
liquid assets or in the client loan portfolio. The capacity ratio therefore tells us the extent to which
the capacity for disbursing client loans with the existing funding base has already been utilized:

Capacity Ratio = Net Loans / Total Assets

Net loans are defined as total gross loans minus the accumulated loan loss reserve.

Because the capacity ratio indicates the extent that an institution’s assets is made up of the loan
portfolio, a higher capacity ratio points to lower liquid assets and higher liquidity risk.

The theoretical range of the capacity ratio is between 0 and 1. Taken together, the liquid assets
ratio and the capacity ratio will typically account for 80-95% of the total balance sheet, with the
remainder constituted by fixed investments and sundry other receivables and intangible assets.

Purchased Funds Ratio

We now turn to the liquidity risks emanating from the funding structure of the MFI. Relying to
a large extent on short-term and potentially volatile sources of funding naturally creates liquidity
risk. We measure this exposure with the purchased funds ratio:

Purchased Funds Ratio = Short-Term Borrowing / Total Assets

Purchased funds are short-term wholesale borrowing instruments such as overdrafts, interbank
borrowings, bills of exchange or large institutional time deposits. The common denominator
is that the MFI actively pursues these deposits or borrowings from professional counterparts
based on competitive rates paid for the credit risk it represents to the investor. Hence, the term
purchased funds. This is interest rate and credit risk sensitive, volatile funding. Purchased funds
are hot money that must be used sparingly and should generally not fund long-term illiquid assets
such as the microfinance loan portfolio.

Thus, the larger the purchased funds ratio, the higher the liquidity risk. In a typical MFI, recourse
to purchased funds should not exceed 15% of total assets, such that volatile sources of funding
remain easily offset by the liquid asset reserve.

61
 Total Deposit Ratio
This ratio measures the available base of customer deposits relative to the overall size of the balance
sheet. All else equal, a high deposit ratio signifies low liquidity risk:

Total Deposit Ratio = Retail Customer Deposits / Total Assets

When adding up retail customer deposits for this ratio, it is important to exclude any inter-bank
deposits or other short-term money market borrowings, which should be classified as purchased
funds. Sometimes this distinction can become blurred as in the case of institutional deposits or
brokered certificates of deposit (CDs) raised from insurance companies or other corporations
through an agent. When in doubt, the litmus test should be the sensitivity of the interest rate. If the
MFI acquired a large deposit by bidding in an interest-rate competitive market, the deposit should
be classified as purchased funds and be excluded from the deposit base.

The Total Deposit Ratio is best interpreted together with the Purchased Funds Ratio: if the total
deposit ratio is low and the purchased funds ratio is high, the MFI relies too much on short-term
funding to finance long-term lending, which is classic liquidity risk. If both the total deposit
base and the reliance on purchased funds are small, then the MFI is in the position to finance
its operations with mainly equity and (concessionary) long-term funding. This would indicate
relatively low liquidity risk.

Loan-to-Deposit Ratio
This ratio is a traditional prudential liquidity ratio frequently found in banking regulations. It is
defined as the name implies:

Loan-to-Deposit Ratio = Net Loans / Total Customer Deposits

The idea of a loan-to-deposit ratio is that loans are illiquid assets, while deposits are understood as
the primary source of stable long-term funding. A high ratio, particularly one that exceeds 100%
indicates liquidity risk, because in this case the illiquid investment in the loan portfolio would
exceed the stable funding base. Implicitly it is assumed that if the loan portfolio is not financed by
stable retail deposits, the necessary funding is probably raised via volatile purchased funds.

Typical prudential limits on the loan-to-deposit ratio imposed on banks are in the 75% – 85%
range.

We should point out that the loan-to-deposit ratio really only allows for a meaningful interpretation
when customer deposits are the dominant form of financing, i.e. in situations where the total
deposit ratio approaches or exceeds 80%.

The loan-to-deposit ratio is meaningless, if banks or MFIs have higher levels of equity or other
significant sources of long-term debt finance from bond issuance or development sponsors.

62
Reserve Ratio
A reserve ratio measures the proportion of cash or liquid assets versus customer deposits and
sometimes also against other short-term liabilities. The most basic definition of a reserve ratio is:

Reserve Ratio = Cash Assets / Total Customer Deposits

Because reserve ratios are frequently mentioned in the context of liquidity management and
are frequently found in prudential regulations, a brief summary is included here. However, it is
important to note that a reserve ratio is not a particularly useful indicator of liquidity risk and
only has limited value as a prudential constraint on the balance sheet positioning of a MFI.11 In
addition, reserve ratios measure liquid reserve assets not against risky funding but against deposits,
which as mentioned earlier are considered more of a long-term source of funding because of their
behavioral characteristics. The traditional minimum reserve requirements that regulated financial
institutions must hold at the central bank use a similar reserve ratio to determine the amount of the
reserve. Such a mandatory reserve is a valid instrument of monetary policy and still used widely
in emerging and developing markets. See more on required central bank reserves in Box 5 below.

11 Reserve ratios are not very relevant in microfinance liquidity management because they usually measure cash assets as a
whole. If the intention is to prescribe minimum levels of freely usable liquidity, one must measure free reserves in excess of the
required reserve with the central bank (i.e. cash assets in excess of required reserved as a percent of total customer deposits).
One could also look more broadly at liquid assets, rather than just free cash assets. In the case of microfinance, however, this is
usually not a binding constraint as for many institutions deposit funding does not exceed 50% of total funding.

63
 Box 4: Mandatory or Required Minimum Reserve
In almost every jurisdiction, commercial banks are obligated by law to maintain
a mandatory minimum reserve with the central bank. A required reserve is an
instrument of monetary policy and a way to regulate the expansion of credit in
the financial system. The higher the reserve rate that banks must maintain at the
central bank, the less funding is available out of every dollar in deposits raised
and the more expensive and harder it will become for clients to obtain bank
credit.

For example, in the euro-system, the reserve rate is 2% on all customer deposits
and short-term inter-bank borrowing. Reserve rates can be drastically higher,
though. In many emerging and developing markets, mandatory reserve rates are
as high as 10%. In Serbia, for example, reserve rates have reached punitive levels
as high as 40% on foreign currency liabilities when the central bank was fighting
to contain a bubble in consumer credit.

From an institutional perspective, the required reserve balances are not usable
liquidity in the normal sense. Because the reserves are mandatory given the level
of existing deposit liabilities, these funds are not available for meeting other
payment obligations or for disbursing additional loans. Counter-intuitively, high
required reserve rates therefore mean low available liquidity and low reserve
rates increase available liquidity. That said, reserve requirements are typically
calculated on a rolling monthly average of deposits and must only be maintained
during a separate forward period or during a partially overlapping maintenance
period. A bank can therefore briefly draw down its reserve balance to meet
settlement obligations, as long as the reserve account average is brought back
up in line with the reserve ratio by holding additional reserves later. However, the
main message remains: required reserves are not freely available liquidity.

Liquidity (Coverage) Ratio

In many jurisdictions, the central bank or regulatory agency requires that banks regularly calculate
a stress outflow coverage ratio. Such a liquidity coverage ratio is sometimes simply called “the
liquidity ratio,” which makes it a bit confusing because the other ratios described above are, of
course, also liquidity ratios in a general sense. This special liquidity coverage ratio is essentially a
simplified liquidity stress test. It compares available and maturating liquid assets over a 30 or 90 day
time horizon with payment obligations resulting from a liability run-off under stress conditions:

Liquidity Coverage Ratio 30 days = Liquid Assets 30 days / Liability Run-Off 30 days

The liquidity coverage ratio should obviously always be greater than 1, such that the institution
would survive a liability run-off as per the stress scenario.

64
A liquidity coverage ratio can only be calculated based on specific instructions regarding the
eligible liquid assets and the run-off proportions that are to be assumed for different types of
liabilities. Here is a sample of such assumptions:

Liquid Assets 30 days = Vault cash + sight balances at central bank + due from banks
+ interbank assets maturing within 30 days, repo-eligible Treasury bills and bonds
+ other Treasury assets maturing within 30 days.

Liability Run-Off 30 days = 10% of retail sight deposits, 20% of retail term deposits,
30% of interbank borrowings and institutional deposits

A maturity gap report as in Figure 4.1 would be an excellent starting point for finding the necessary
data that goes into the liquidity coverage ratio calculation.

Operating Liquidity Ratio

An operating liquidity ratio takes us outside of the realm of standard balance sheet ratios by
benchmarking asset liquidity against income statement and cash flow figures. Operating liquidity
ratios measure the survival horizon not against liability withdrawals (liquidity coverage ratio), but
against the cash absorbed by operating expenses and loan portfolio growth:

Operating Liquidity Ratio = Liquid Assets / (Monthly Operating Expenses + Recent Monthly Net
Loan Portfolio Growth)

If it takes three months to negotiate and receive disbursement under a new credit facility with a
microfinance investor, then it would be wise to have liquid assets to cover at least three months of
business as usual, paying operating expenses and continuing loan disbursements. It is not unusual
to impose liquidity covenants and internal position limits on MFIs that require an operating
liquidity buffer of 6 months, i.e. an operating liquidity ratio > 6.

65
 Liquidity Ratio Values – Industry Cross Section
In Figure 4.3 we assembled a random selection of banks and MFIs from various markets and at
different balance sheet dates for which we calculated some of the liquidity ratios described above.
We focused on those ratios that can quickly be calculated off the balance sheet without any further
inside knowledge of the institution.
In the table, we chose to split the liquid assets ratio into the Cash Position Indicator and the
column Other Liquid Assets / Total Assets. The sum of these two total the Liquid Assets to Total
Assets ratio mentioned above.
Organization Date Cash Other Capacity Total Purchased Loan-to- Reserve
Position Liquid Ratio Deposit Funds Deposit Ratio
Indicator Ass. / Total Ratio Ratio Ratio
Assets
Microfinance Institutions & Microfinance Banks
Anelik Bank, Armenia Dec-07 33.95% 0.10% 50.26% 39.47% 2.59% 127.34% 86.01%
Banco Sol, Bolivia Dec-08 8.67% 13.36% 74.33% 68.87% 8.20% 107.93% 12.59%
Banka Credins, Albania Dec-09 20.85% 6.98% 61.00% 87.21% 2.03% 69.95% 23.90%
BRAC Uganda Dec-08 45.41% 0.00% 54.04% 10.82% 0.00% 499.56% 419.83%
Microfinance
Capitec Bank, South Africa Feb-10 9.29% 8.74% 69.75% 43.28% 10.43% 161.18% 21.46%
Duterimbere MFI, Rwanda Dec-09 6.53% 0.00% 82.44% 31.47% 2.17% 261.96% 20.74%
FINCA Uganda Dec-09 10.05% 1.48% 71.28% 26.71% 0.00% 266.91% 37.65%
First Microfinance Bank, Dec-09 19.09% 29.21% 44.92% 86.01% 0.00% 52.22% 22.20%
Pakistan
Grameen Bank, Dec-09 1.26% 36.65% 54.71% 84.73% 1.62% 64.57% 1.49%
Bangladesh
MI-BOSPO, BiH Apr-10 1.47% 2.56% 72.56% 0.00% 0.00% n/a n/a
ProCredit Kosovo Jun-10 12.65% 17.07% 66.33% 83.99% 3.44% 78.97% 15.06%
XacBank Mongolia Mar-11 7.63% 6.85% 75.79% 49.92% 4.63% 151.82% 15.29%
Commercial Banks
China Merchants Bank Dec-08 13.92% 23.22% 54.16% 78.64% 11.01% 68.87% 17.70%
Deutsche Bank AG, Mar-11 0.78% 66.10% 21.41% 28.94% 46.63% 73.97% 2.69%
Germany
IKB Industriekreditbank, Jun-07 0.05% 2.29% 56.84% 8.61% 28.51% 660.08% 0.56%
Germany
Northern Rock, UK Jun-07 0.68% 14.31% 85.16% 26.54% 46.04% 320.88% 2.55%
Standard Bank, South Dec-10 2.15% 27.72% 53.19% 59.08% 16.66% 90.02% 3.63%
Africa
WellsFargo, USA Mar-11 1.40% 22.90% 60.76% 67.75% 4.41% 89.68% 2.06%

Figure 4.3: Liquidity Ratio Profile of Selected Financial Institutions

At first sight, one might be disappointed that there are no immediately obvious global benchmarks
towards which the various observations converge. The ratio values cover a very wide range and
appear somewhat arbitrary. As we mentioned before, there is no specific target ratio that will
work for all institutions all the time. We should keep in mind that a single snapshot of a ratio is
much less informative than a time series of observations that would allow us to detect trends and
eliminate outliers on a particular date. Also, a particular snapshot does not necessarily represent
that institution’s desired or optimal positioning. It could well be that the cash position at the end of
the quarter was distorted by a large transaction that so happened to settle on that date and would
have been invested or otherwise covered just a few days later.

66
However, certain trends and patterns can still be observed. When observing the Cash Position
Indicator, we notice that some banks and microfinance institutions survive with as little as 1%
or less of total assets in immediately available cash. A low cash position is safe, as long as there
are other liquid assets that can rapidly be converted into cash in efficient secondary markets.
In developing markets, where the sale of short-term investments may not be as immediate, an
institution will have an incentive to hold more cash versus other liquid assets. Another important
driver for the size of the cash position are regulatory minimum reserve requirements, which tend
to be much higher in emerging and developing countries than in Europe or the US (see Box 5).

For large universal banks like Deutsche Bank that run investment banking and commercial banking
operations on the same balance sheet, the Other liquid assets position can be quite inflated by
trading inventory in bonds and other capital markets instruments. At more traditional retail banks
like Wells Fargo, Standard Bank or Capitec, we find more typical liquid asset ratios in the 15% -
30% range.

The table also includes two institutions that did not survive the 2007-2009 liquidity crunch
that was part of the recent global financial crisis: British Northern Rock and the German IKB.
IKB certainly is an example of low asset liquidity with too much exposure to volatile short-term
purchased funds. What made IKB collapse in an already tight liquidity situation was a call on a
back-up credit line extended to an asset-backed commercial paper conduit that was held off balance
sheet. There was not enough liquidity reserve at IKB to honor this disbursement obligation and
weaker, less liquid institutions like IKB were already locked out of the interbank money market in
summer 2007.

Northern Rock appeared to have slightly more asset liquidity before its collapse in September
2007, but they made a beginner’s mistake in liquidity management: i.e. funding 20-30 year
mortgages with rolling 3-months money market borrowings. A purchased funds ratio of 46% is
clearly an excessive liquidity risk for a mortgage bank.

China Merchants Bank might serve as a typical example of a bank in emerging markets that only
has limited access to bond markets and does not raise significant funding from development
finance sources. The liability side is simply deposits, equity and a small amount of short-term
interbank borrowing. In this traditional balance sheet structure, a loan-to-deposit ratio would also
be a reasonable indicator of liquidity risk.

Duterimbere MFI in Rwanda continues to operate normally as of writing in 2011, but its low
asset liquidity in 2009 was a cause for concern. To Duterimbere’s credit, the dependence on
volatile short-term funding was low and the rest of the liability side was comprised of equity, some
concessionary borrowing and micro-deposits. But even with a low-risk funding structure, the
minimal asset liquidity in 2009 left no room for portfolio growth and only a negligible buffer to
deal with unexpected cash requirements. By year-end 2010, Duterimbere’s liquid assets stood at a
more prudent level of 11.8% of total balance sheet.

67
 Model:

All of the above liquidity ratio calculations are already implemented in the accompanying Excel
tool LiquidityManager.xls. Once you have captured a balance sheet history and some income
statement data for a particular institution, the ratios are derived automatically. See BalanceSheet and
RatioDashboard sheets.

Exercise 4.1:

Exercise 4.1: Use Excel file LiquidityManager: BalanceSheet, IncomeStatement and RatioDashboard.
Make a copy of the LiquidityManager workbook and enter your own institution’s balance sheet and
income history into the worksheets. Alternatively, capture the financial statements of any other MFI
that might be of interest. Review the ratio results and interpret the graphs in the RatioDashboard.

Once liquidity ratios are selected and calculated for a specific time period, a Liquidity Dashboard
can be developed to provide a pictorial view of the institution’s liquidity position at different
points in time. The following Liquidity Dashboard can also be found in the excel worksheet
RatioDashboard of the LiquidityManager.xls file.

Liquidity Ratio Dashboard: Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Mar-10 Jun-10 Sep-10 Dec-10 Mar-11 Jun-11

Vault Cash 2.07% 1.92% 1.63% 2.22% 2.03% 2.10% 2.14% 2.15% 2.56% 1.80% 1.88%
Cash Position 4.42% 5.07% 4.41% 5.04% 4.85% 6.51% 4.72% 4.57% 7.19% 5.45% 6.02%
Liquid Assets 14.18% 18.31% 20.02% 22.50% 26.38% 23.27% 13.46% 14.73% 18.69% 12.37% 12.37%
Capacity Ratio 71.97% 66.89% 66.73% 64.10% 61.03% 62.74% 72.99% 72.61% 67.57% 75.81% 74.43%
Purchased Funds 3.27% 4.04% 4.98% 6.17% 6.91% 5.27% 4.12% 4.68% 7.47% 7.56% 9.76%
Customer Deposits 37.02% 37.47% 39.84% 41.08% 39.61% 37.00% 34.51% 36.01% 43.90% 42.27% 41.25%
Stress Coverage 289.02% 353.50% 331.23% 342.31% 386.72% 412.69% 281.54% 282.23% 280.34% 196.94% 208.39%
Operational Liquidity 0.00% 1343.95% 1027.67% 2022.73% 1737.39% 1173.67% 443.86% 658.15% 1070.48% 389.35% 909.47%

35.00% 20.0
32.50% Stress Coverage
17.5
30.00% Operational Liquidity
15.0
27.50%
25.00% 12.5
22.50% Liquid Assets 10.0
20.00% 7.5
Cash Position
17.50%
Vault Cash 5.0
15.00%
12.50% 2.5
10.00% 0.0
7.50%
5.00%
2.50%
0.00%
Apr/10
Apr/09

Apr/11
Feb/09

Feb/10

Feb/11
Jun/09

Jun/10

Jun/11
Dec/08

Dec/09

Dec/10
Aug/09
Oct/09

Aug/10
Oct/10

Figure 4.4: Screenshot from LiquityManager, sheet RatioDashboard

68
4.3 Dynamic Liquidity Management based on Cash Flows
Objectives of Cash Flow Planning
Cash flow planning is the essence of modern liquidity management. Balance sheet ratios can give
a quick sense of the liquidity risk of an institution, but assurance that a MFI will always be able to
meet its payment obligations can only come from a detailed forecast of cash inflows and outflows.
That forward cash flow profile should then be combined with a safety buffer of liquid assets that
are expressed as a liquidity ratio. For example, after covering the forecasted net cash flows, we plan
to maintain a 20% liquid assets ratio as a buffer for unexpected flows, sudden stress events and
business growth.

Cash flow forecasts should cover at least six months forward, if not twelve, and should be very
granular for the immediate future. Typically, one would plan in daily increments for the next
month and update the plan based on actual materialized flows at least weekly, preferably daily.

The following table conveys the basic concept of a dynamic, forward-looking cash-flow chart. The
individual cash flows are captured with their direction, their estimated size and the time interval in
which they are expected to occur.

Time Jan. Jan. Jan. Jan. ...

Cash Flow 1 2 3 4

Savings Deposit Customer A +10

Disbursement Loan B -30
Repayment Loan C +40
Interest Payment Loan D +15
Savings Withdrawal -5 -5
Customer E
Branch Payroll Disbursement -60
Interest Payment Loan F +10
......

Daily Cash Change +5 +25 +5 -60

Cumulative Cash Balance 5 30 35 -25
!
Figure 4.5: Simple Cash Flow Forecast

Expected deposits from various customers, for example, are registered individually with a positive
sign as a cash inflow. In contrast, expected loan disbursements to individual borrowers represent
a cash outflow and are therefore shown as a negative figure.

Cash flow planning must follow a two-step process: we must first forecast the Net Funding
Requirement to understand how cash will be generated or absorbed by client-driven transactions
and other operating expenses and revenues. Then, we must apply appropriate liquidity management
actions (i.e. either we invest excess cash or procure additional liquidity to cover shortfalls) based
on or triggered by this forecast.

69
 Net Funding Requirements Forecast
Because there are so many individual transactions on a daily basis, it is difficult to predict their
exact timing and size. One can therefore use statistical measures to approximate the flow of small
transactions, complemented with information about anticipated, large, individual transactions.
The resulting timeline of cash inflows and outflows allows us to calculate estimates of cumulative
surpluses or deficits at specific points in time which we call the Net Funding Requirement.

When forecasting the net funding requirements, one must first develop detailed projections of
loan portfolio balances and deposit supply as these are the most important portion of a MFI’s
operations. Trend and seasonality elements can be incorporated into these forecasts to ensure the
incorporation of market-based patterns. However, institution-specific strategic changes should
be prioritized when developing these kinds of forecasts. Once loan and deposit estimates are
made, other flows such as interest payments, interest on short-term investments, and other cash
operating expenses and revenues can be derived.

It is important to note that the focus of this forecast must be on the net liquidity effects of the
transactions. Projections of the outstanding loan portfolio should be aggregated after having
totaled expected new disbursements against planned repayments on old loans. A net increase of
the outstanding loan portfolio represents a cash outflow, while a reduction of the outstanding
amount leads to a net cash inflow. The same principles apply to the deposit base.

A detailed description of forecasting loan demand and deposit supply, including utilizing seasonal
patterns, as well as forecasting other cash outflows and inflows can be found in Appendix 1. In
addition, the Excel model titled LiquidityManager.xls includes a comprehensive forecasting model
that can be used for this purpose and is linked to the Liquidity Dashboard and other Liquidity
tools.

Once all the individual cash flows are estimated, they are then be consolidated and the net periodic
change in cash and the cumulative cash position, i.e. the Net Funding Requirements are calculated.
Figure 4.6 shows an illustration of the aggregated net funding requirements after a detailed cash
forecast was undertaken. This table shows the MFI’s projected cash position over the next year.
This is what is expected to happen without any further intervention, i.e. in the absence of active
liquidity management.

70
 !
Figure 4.6: Net Funding Requirements Forecast. LiquidityManager, sheet TotalFlows.

After forecasting the net funding requirements, the MFI must review the results in detail and
analyze whether any liquidity management actions are necessary. In the case outlined above,
the treasurer must procure additional cash to prevent the predicted negative cumulative cash
beginning in September 2011. The positive but very low cash position in August will also
be unacceptable for the institution. If the liquidity risk policy requires a Free Cash Position
Indicator of 2% and a liquid assets ratio of 20%, for example, then the treasurer will have to draw
on short-term borrowing lines from other banks beginning in February. The projected cash flows
from these activities will in turn be added to the Net Funding Requirements. It is only after this
final step that the cumulative cash balances must strictly be positive and, moreover, must meet
the minimum requirements set by management and by the regulatory authorities.
Finally, it is important to note that the objective of cash flow planning method is to derive a
reliable forecast of net funding requirements under normal business conditions, when loan
demand and deposit supply largely follow the institution’s business plan and adhere to previously
observed patterns of client behavior. Such a business-as-usual liquidity plan must always be
paired with stress testing and contingency planning, which is the subject of the following section.

71
 4.4 Contingency Planning & Liquidity Stress Testing
The Nature of Liquidity Stress Events

The Net Funding Requirements describe a scenario where everything goes according to plan.
Every MFI must also think through and prepare for a situation when normal assumptions about
client behavior, asset prices and the availability of borrowing opportunities no longer apply. Such
a crisis can either be of systemic nature, concerning most financial institutions or at least most
MFIs in the national or regional market, or may be specific to a particular MFI.
Systemic crises are typically triggered by macro-economic shocks, such as the collapse of an
asset price bubble, drastic exchange rate adjustments, a sharp recession or combinations of such
factors that undermine the depositor and investor confidence in the financial system.
Institution-specific liquidity crises tend to originate from real or perceived losses and
reputational events such as poor loan recovery, large publicized fraud or theft, speculative
investments gone bad etc. These losses raise questions about the financial viability of the
particular institution, thus creating an incentive to wholesale funders and depositors to reduce
their exposure while they still can, triggering a run on the liabilities of the MFI.
From international experience, liquidity crises rarely occur in pure systemic or institution-
specific form, generally one finds contagion effects that can lead to a systemic impact following
an institution-specific situation, while a system-wide liquidity disruption tends to always affect
certain vulnerable institutions more than others.
Thus, every MFI should have contingency plans in place that will guide the reaction to an
accelerated run-off on its retail and wholesale liabilities that will likely occur at a time when
market conditions in general are also tight, making it difficult to borrow in the money market and
to sell otherwise liquid securities for cash. These should be simulated and regularly evaluated in
a comprehensive stress test to prove the ability of the MFI to withstand such a liquidity stress
situation.

72
Contingency Actions
Below are some principles to guide a MFI’s response to an emerging liquidity crisis. The specific
contingency actions decided by senior management and the Board should be documented in the
MFI’s liquidity policy or in the overall risk management procedures:

Crisis Preparedness:

• A MFI should strive to establish confirmed overdraft or money market bor-

rowing lines with several domestic and regional institutions. Where possible,
this should also include a stand-by liquidity facility available with or guar-
anteed by the international microfinance network that the MFI is affiliated
with.
• These lines should be regularly tested for continued availability by short-
term partial draws during uncritical normal operating conditions, even if no
immediate liquidity requirement exists.
• The MFI should engage with potential counterparties for sales of less liquid
assets in order to develop model contracts and establish reference transac-
tions and prices. Such transactions could include sales of delinquent loan
portfolios, or of repossessed and foreclosed assets, as well as collateralized
borrowing, e.g. against real estate owned by the MFI. Working through mod-
el transactions during normal times should facilitate the rapid implementa-
tion of additional sales at minimal fire-sale losses in time of liquidity stress.
Monitoring and Crisis Detection:

• The Risk Management Committee should regularly review the economic

environment as well as signals from wholesale funders and clients for early
signs of a deterioration of confidence that could lead to liquidity stress.
• Particular attention should be paid to reputational events such as publicized
loan portfolio problems, major fraud, a law suit or other negative press that
could escalate into a public concern about the creditworthiness of the MFI.
• The Risk Management Committee should be the responsible party for de-
claring a crisis or watch event, informing the Supervisory Board and trigger-
ing the implementation of the liquidity contingency plan.
Communication with the Public:

• The Chief Executive Officer and the Executive Management team should
assume exclusive responsibility for communicating with the central bank,
government and the media about the emerging situation at the MFI.
• The issue at hand (loss, fraud, legal problem) should not be denied but fully
disclosed and put into perspective with the overall size of the operation and
the financial resources of the institution.
• Senior management should communicate proactively about the situation
and seek the explicit backing of the regulator and major financial institutions
to reassure clients and stakeholders about the solvency of the MFI.
73
 Operating Conditions during a Crisis:

• Even under liquidity stress, the MFI must continue to meet all retail deposit
redemptions and any wholesale settlement obligations without delay.
• The MFI should not resort to temporary branch closures or limitations of
customer withdrawals. The MFI should quickly draw 100% of any available
standby facility from the affiliate network and 90% of all other available
money market and overdraft lines of credit, before potential limit reduc-
tions and cancellations might take effect. Management should implement
an immediate capital spending freeze, delay accounts payable settlement
for supplier invoices by three weeks on average and consider paying only a
partial payroll to employees with executives accepting the largest deferral of
salary. Using the previously established counterparty network, model agree-
ments and reference prices, the MFI should accelerate sales of distressed
receivables, foreclosed assets and non-essential real estate holdings, in order
to raise additional cash proceeds. In order to conserve cash, the MFI should
consider making loan disbursements only to strategically important clients
and customer segments according to a previously established client clas-
sification. Note that this customer classification does not necessarily have to
be driven by the size of the business and profitability of the relationship. It
can also be built around the social mission of the MFI. This could mean that
small working capital loans to micro-entrepreneurs would take precedence
over consumer loans to salaried individuals. However, utilizations under
confirmed lines of credit and previously communicated customer limits
must always be honored.
Stress tests are a set of scenario assumptions that describe the unfavorable market situation,
the behavior changes among clients and funders and management’s contingency actions. We
recommend that the standard liquidity stress test should focus on an institution-specific crisis
against a backdrop of an overall unfavorable financial market situation. The institution-specific
crisis might be triggered by a loss event or reputational situation that comes at a time when
the microfinance industry is already viewed with concern. An illustration of such a scenario is
described below:

1. Reputational stress event starts today and lasts for 3 months.

2. Unused overdraft lines of credit are canceled by correspondents within first
month.
3. Unsecured overdraft utilizations are paid back within 3 months.
4. Large time deposits run off by 30% over 3 months.
5. Retail depositors run off 10% over 3 months.
6. No new long-term borrowing: Scheduled draws on existing funding lines are
being delayed by the lenders and do not become available during the next 3
months.
7. Monthly loan collection ratio (collected / contractually due) declines to
80% by the second month.

74
 8. Maximum stand-by facility from the international affiliate network is drawn
within first month.
9. The MFI pays 2% higher interest on all retail deposits in order to contain
outflows to the 10% run off.
10. Loan disbursements are limited to prolongations for core clients only.
11. Any capital spending is deferred.

Once the assumptions for the stress test are made, they must be incorporated into the financial
model to test for sufficient liquid asset coverage of outflows and continued compliance with
regulatory and internal minimum liquidity requirements. The objective of the stress test is to
prove that the MFI can survive a moderate liquidity stress event. For that, we need to calibrate
the assumptions such that they actually describe a rare but plausible and potentially survivable
situation.

Once we have settled on a reasonable stress scenario, we should hold it steady and regularly run
the current balance sheet positioning of the MFI against this fixed scenario. If the institution then
fails the stress test, we should not go back and change the assumptions, but actually change the
positioning by adjusting the liquidity risk of the business plan.

Model:

See the sheet StressTest in LiquidityManager. The stress scenario parameters set out at the top of the
sheet are applied to the most recent balance sheet positioning and the base line cash flow forecast
developed previously.

75
 Chapter 4 - Key Concepts and Next Steps
• Liquidity is the ability of a financial institution to honor all cash payment
commitments as they fall due.
• The objectives of liquidity management are to (1) honor all cash
outflow commitments on a daily and ongoing basis, (2) minimize the
cost of foregone earnings on idle liquidity, (3) satisfy minimum reserve
requirements and other prudential liquidity standards, (4) avoid the
additional cost of emergency borrowing and forced liquidation of assets.
• From an external perspective, liquidity can be analyzed using ratios that
measure the proportions of liquid and non-liquid assets relative to other
balance sheet items.
• Internal liquidity planning involves detailed estimations of the size and
timing of cash inflows and outflows over the next few days and weeks,
complemented by less granular forecasts for up to a year.
• Liquidity management is a two-step process: first, one determines future
cumulative surpluses or deficits of cash resulting from client-driven
transactions and the operational business of the MFI, i.e. the Net Funding
Requirement. In a second step, MFI Treasury deploys short-term investing
and borrowing transactions, with a view to managing the cash position into
alignment with the Net Funding Requirements, while observing operational
and prudential constraints on cash and liquid assets.
• In a multi-branch MFI, liquidity management is best organized as a
centralized function at head office. Branch staff should be involved in
planning client-related transactions, particularly those occurring in vault
cash, but the resultant cash flow forecasts must be consolidated into a
centralized view of liquidity at Head Office Treasury, where liquid assets are
managed for the entire institution.
• The single most important take home message from the extensive discussion
of liquidity management is to appreciate the “non-linear” nature of
liquidity: i.e. the fragile balance between having too much liquidity and
being challenged to use it for meeting accelerated outflows. This is why
5% liquid assets to total assets is not enough, even in a non-deposit taking,
long-funded MFI. It may be enough to meet normal planned payment
obligations, but the slightest internal or external trigger event can put this
MFI over the edge. If 5% liquid assets is enough to meet normal forecasted
requirements, then you should make sure that the MFI still has 5% liquid
assets after facing a stress outflow.
• One way to improve liquidity management practice at your MFI is to do
a stress test. Simply enter your financial statements in LiquidityManager,
review the scenario assumptions (sheet StressTest) and you can see instantly
whether your institution can survive with its current level of liquidity.

76
CHAPTER 5
SHORT-TERM BORROWING AND INVESTING
INSTRUMENTS

5.1 Introduction
Once liquidity ratios and reports have been measured, cash flow planning has been finalized, and
the appropriate liquidity management actions undertaken, we must then focus on the investment
decisions to be made related to the management of the MFI’s liquid assets.

The Excel model LiquidityManager captures liquidity management transactions by assuming that
one borrows enough cash or invests excess liquidity in order to accommodate the Net Funding
Requirements derived in Chapter 4. The following table in the LiquidityManager Excel file
titled Liquidity Workbench highlights the net funding requirements before any active liquidity
management actions are undertaken, and then gives the user the opportunity to make the
corrective actions to ensure that the institution is meeting the adequate limits and complying with
appropriate requirements.
Before Active Liquidity Management Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11 Nov-11 Dec-11
Cumulative Net Funding Requirement 22,212 13,976 23,019 11,778 13,195 33,094 24,966 444 -3,638 -9,979 -11,452 -14,889
Free Reserve Ratio - before 6.94% 1.86% 6.96% 0.39% 0.98% 11.59% 6.62% -6.09% -8.01% -10.87% -11.19% -12.68%
Free Cash Position Indicator - before 2.56% 0.69% 2.51% 0.14% 0.36% 4.16% 2.43% -2.30% -3.07% -4.21% -4.48% -5.09%

Liquidity Management Actions Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11 Nov-11 Dec-11

Cash from Short-Term Borrowing

Bank Overdraft Borrowings 5,000
Repo Borrowings
Financial Institution Term Deposits 10,000
Credit Line / Source 4
Cash from Short-Term Investing
Repo-eligible Treasury Bills and Bonds 18,000
Reverse Repo Placements 5,000
Deposits at Banks and Financial Institutions 5,000
Investment Type / Counterparty 4
Total Cash from Liquidity Actions 0 0 0 0 0 0 0 18000 10000 10000 0 5000

After Active Liquidity Management Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11 Nov-11 Dec-11

Cumulative Cash 22,212 13,976 23,019 11,778 13,195 33,094 24,966 18,444 24,362 28,021 26,548 28,111
Free Reserve Ratio - after 6.94% 1.86% 6.96% 0.39% 0.98% 11.59% 6.62% 2.77% 5.22% 6.59% 5.14% 5.84%
Free Cash Position Indicator - after 2.56% 0.69% 2.51% 0.14% 0.36% 4.16% 2.43% 1.04% 2.00% 2.55% 2.06% 2.34%

Balance Control Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11 Nov-11 Dec-11
Beginning
Bank Overdraft Borrowings 3,000 3,000 3,000 3,000 3,000 3,000 3,000 3,000 3,000 8,000 8,000 8,000 8,000
Repo Borrowings 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000
Financial Institution Term Deposits 17,000 17,000 17,000 17,000 17,000 17,000 17,000 17,000 17,000 17,000 27,000 27,000 27,000
Credit Line / Source 4 0 0 0 0 0 0 0 0 0 0 0 0

Repo-eligible Treasury Bills and Bonds 42,950 42,950 42,950 42,950 42,950 42,950 42,950 42,950 24,950 24,950 24,950 24,950 24,950
Reverse Repo Placements 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 0
Deposits at Banks and Financial Institutions 27,600 27,600 27,600 27,600 27,600 27,600 27,600 27,600 27,600 22,600 22,600 22,600 22,600
Investment Type / Counterparty 4 0 0 0 0 0 0 0 0 0 0 0 0

Figure 5.1: Screenshot from LiquidityManager, sheet LiquidityWorkbench.

If one keeps the financial statement and cash flow forecasts in the tool updated, the liquidity
workbench can be used as a regular planning tool for the treasurer. The workbench can help us
make daily treasury decisions about investing, borrowing and moving cash around between
various accounts, while immediately simulating the effects of these decisions over the coming
months.

77
 The next decision is which investment and borrowing instruments are available and appropriate.
This chapter will therefore focus on the various instruments that might already be available to the
MFI or that one should pursue in the future. The section first starts with instruments from the
asset-side of the balance sheet, following by short-term borrowing opportunities.

Before we go into the details, we want to caution not to overcomplicate the question of liquidity
management instruments. The success factors in liquidity management instruments are the
minimization of credit risk, market risk and transaction costs, while improving speed and
flexibility. A MFI treasury can be managed very effectively with just two or three current accounts
with correspondents, some bank time deposits and a few overdraft lines of credit.

5.2 Managing Traditional Liquid Assets

The asset side of the balance sheet is where the MFI stores its liquidity. This liquidity stock can be
held in any of the following forms: vault cash, demand balances or short-duration time deposits
with other banks, as well as highly liquid investments. Vault cash is the most expensive form of
liquidity to hold, while short-term investments offer the smallest loss of income when compared
to typical earning assets, such as micro-loans.

The key to minimizing the cost of liquidity thus is to cultivate safe, highly liquid, short-term
investment opportunities with reasonable yield. As a first step, a small MFI can try to negotiate
interest to be paid on the balances that it holds at formal commercial banks. If there is resistance
to paying interest on ordinary demand deposits, the MFI can try to keep only a small base of
transaction balances while frequently “sweeping” any excess into an interest earning savings
account or an overnight money market placement. These balances could be transferred back into
the transaction account as needed on a daily basis.

Such sweep arrangements are particularly useful if a MFI maintains multiple transaction
accounts with a correspondent bank for the convenience of its customers who may pay their loan
installments at these branches rather than coming to the MFI counters. With a daily sweep, it can
avoid scattering idle transaction balances across many accounts and can at least earn some interest
by investing the consolidated balance overnight.

Time Deposits
The next step can be to invest demand balances that are not immediately needed in higher-earning
time deposits at the MFI’s main correspondent bank. Since this money represents liquidity that
should be available to cover transaction needs at any time, it would be important to arrange for an
overdraft borrowing facility with the time deposit as collateral. Borrowing your own money back
can make sense, if one needs to cover a peak in cash demand for just a few days. Instead of breaking
the time deposit, forfeiting any accrued interest and possibly even paying a withdrawal penalty, the
MFI would briefly carry a deposit and a short-term debt in parallel. Over time, the MFI should be
able to negotiate down the full cash collateral requirement. With only partial cover, one would be
able to borrow a multiple of the time deposit pledged as collateral.

78
Money Market Deposits
The next step up for a MFI with sizable cash reserves is to place large overnight or time deposits
in the interbank money market. These deposits are usually handled by money market traders in
standardized lots with rather large minimum amounts (say US$ 500,000). The most common
interest calculation convention for time deposits and larger money market placements is
Actual/360. This means that the daily interest multiplier is obtained by dividing the quoted annual
interest by 360 days and then applying this rate for the exact number of days placed. Interest under
money market conventions is always paid at maturity together with the return of the principal.

When considering an interbank money market the first step is to place money market deposits with
a MFI’s main correspondent, who knows the institution’s business and will be more flexible when
it comes to minimum lots and the possibility of early withdrawal in the case of an acute liquidity
shortfall. As soon as placement volumes allow, a MFI would be well advised to open correspondent
accounts and begin money market relationships with one or two additional commercial banks.
Having more than one established outlet for placing money market deposits will allow the MFI
to shop around for the highest interest rates and avoid a concentration of counterparty credit risk.

Regular dealings with the traders at the correspondent banks will also provide an opportunity to
establish a name for the MFI in the money market. The reputation as a reliable business partner
and the personal relationships that the MFI builds in the process are essential, if the institution
desires not only to place deposits in the market but also wants to take in funds from commercial
banks at a later date.

5.3 Investing in Short-Term Securities

Most commercial banks store a significant portion of their liquidity in readily saleable, high quality
securities. Again, the motivation is to earn as much interest on the liquidity reserve as possible,
while still having access to the funds at a moment’s notice. MFIs can also have a similar strategy
even though their placement volumes may be smaller, as long as the development of their local
financial markets permits.
Below is a checklist of what characterizes suitable liquid asset investments in marketable securities.
The short-term investment instrument should:
• Represent a Debt Instrument Only: Short-term investments for liquidity
purposes should not be speculative and expose the MFI to significant price
risk. This rules out any equity security, such as common stock, the value of
which changes daily as a function of the profit outlook of the company.
• Provide Full Repayment of the Principal Amount Plus Interest: Liquid
short-term investments should only be in debt obligations that require the
debtor to pay a fixed amount of interest in addition to the full principal. Such
obligations are also often referred to as fixed-income securities.
• Have Negligible Credit Risk: Even though a fixed-income security offers
set payments at precisely defined times, there is still the default risk of
the borrower to consider. Unless the fixed-income security comes with
additional guarantees or collateral, the debt might not be paid back, if the

79
 borrower goes bankrupt or experiences financial difficulty. This credit risk
should be minimized for liquidity investments. In many developing markets,
the only reasonably liquid securities are short term Treasury bills. The only
potentially eligible other issuers would be utility companies or the national
telecom provider, for example. Any investment in private issuer debt should
be within strict risk diversification limits.
• Have a Short Maturity and an Active Secondary Market: Debt securities
that are eligible for the liquidity reserve should not have maturities longer
than a year, typically even less than six months. Intuitively, one might think
that the reason for the short maturity is the desire to get your investment
back quickly when you need the liquidity. This is not the main motivation,
however, because six months is still a long time to wait for the borrower to
pay, if you are running out of cash. Rather, you would sell the security right
away. Hence, the requirement for a secondary market. The real reason for
the short maturity is that the price of a debt security changes depending
on the current interest rates in the financial markets. This is important for
the liquidity manger to consider, because the MFI may be forced to sell the
security before its maturity in order to raise cash. If market rates rise, the
MFI may take a loss on the sale of the bond. This price risk becomes larger
for securities with longer maturities.
• Consider Repo Eligibility: Longer dating government bonds would only
be a suitable liquid investment if the MFI has access to the collateralized
money market or to direct borrowing from the central bank. These collat-
eralized transactions are generally structured as a repurchase agreement, or
repo for short. In a repo, the potential loss in market value of the bond is not
realized. Instead, the MFI effectively only borrows against the bond hold-
ings as collateral. However, one should note that the collateral valuation is
reduced by a safety amount, which also grows with the maturity of the bond.
This means that one cannot borrow as much against a long-maturity bond as
one could in a repo against shorter dating paper.
• Offer Low Transaction Costs: For securities operations to be worth the
effort, transaction costs should be reasonable compared to investment vol-
umes. One has to carefully weigh broker commissions, possible stamp duty
and taxes against the interest earned on the investments. In addition to the
cost, the MFI has to be concerned about the speed and the ease of executing
a transaction. If it takes several days to place an order over the telephone and
receive the proceeds in the MFI’s transaction account, then the additional
interest earned on the investment might not justify the liquidity risk from
delayed access to the funds.
• Not Add Exposure to Currency Risk: It is important to consider poten-
tial foreign exchange risks that might arise from short-term investments in

80
 securities. A MFI might be faced with a situation where suitable short-term
securities are only available in US dollars or euros, but the liquidity require-
ment from the MFI’s operation is in local currency. This would introduce an
additional dimension of foreign exchange risk into the liquidity plan. Gener-
ally, MFIs should avoid mismatching the currency of their liquidity reserve
and the currency in which the use of liquidity occurs. However, under cer-
tain conditions it might be reasonable for a MFI that operates in a weak local
currency to hedge its capital adequacy ratio or even the full foreign currency
counter-value of its equity by way of a structural open long foreign exchange
positions.12 Foreign currency dominated government securities or time
deposit placements with prime international banks in foreign currency could
be appropriate liquid assets for holding such an open position.

5.4 Short-Term Borrowing

Overview
As we have mentioned, and as the 2007-2009 global financial crisis has demonstrated, liquidity
stored on the asset side of the balance sheet is always safer than relying on borrowing opportunities
to meet unexpected cash needs. Nonetheless, a fully featured Treasury operation should also have
access to some short-term borrowing instruments. Typical instruments include overdraft lines
of credit, large institutional deposits, money market borrowings from banks, short-term notes,
certificates of deposit (CDs) or bills of exchange placed with insurance companies, corporate, or
other institutional investors. These instruments are identical to the scope of what we previously
have called purchased funds.

Strategy for Managing Purchased Funds

Purchased funds are different from retail sight or term deposits insofar as the institution actively
goes out and contracts them in large individual transactions at interest rates that are imposed by the
financial market. These funds can be obtained quickly over the telephone and are much cheaper
to handle than small retail deposits. In contrast to retail deposits, however, they carry a substantial
financial cost. The financial cost consists of the interbank market rate plus a risk premium that
depends on the standing of the individual MFI. Purchased funds are also extremely sensitive to
credit risk and interest rate changes. They are placed by professional money managers, who will
move the deposit for a slightly higher interest rate and will withdraw the funds abruptly at the
smallest sign of financial difficulty on the part of the MFI.

Despite these cautions, short-term commercial financing is a part of the funding mix that an
ambitious, growing MFI should not neglect entirely. Wholesale deposits and commercial short-
term loans are essential instruments of liability-side liquidity management.

12 It is beyond the scope of this toolkit to discuss the detailed pros and cons of structural foreign exchange positions. For an
entry point into the discussion on structural positions, the reader is referred to the Basel II documentation, Art. 718(xxxvii) &
718(xxxviii), BSCBS International Convergence of Capital Measurement and Capital Standards - A Revised Framework, June
2006.

81
 Working with the local commercial banking sector is a successful strategy to diversify funding
sources and reduce reliance on developing finance sources. Other benefits include access to
additional services such as international banking, overdraft facilities, equipment finance, business
advice etc.

Regarding deposits at commercial banks, the first step is to work with the correspondent banks
that hold the MFI’s own deposits and transaction balances. We already mentioned the importance
of building strong business relationships with the correspondents in the formal banking sector
when discussing short-term investment opportunities. It is in the area of short-term borrowing
that this relationship-building really pays off. Cultivating these borrowing windows is important,
even if initial volumes are small or if there is no current need for the funds. Available borrowing
capacity is an important element in the liquidity management toolbox.

In order to negotiate overdraft facilities or a money market borrowing line, the MFI needs to
make its case with the commercial bank by providing financial statements and other background
about the size and the quality of its operations. The following details should be negotiated and
documented, typically in a framework agreement, well before the need for drawing funds arises:

• Definition of terms and interest rate conventions

• Credit limit
• Minimum and maximum transaction sizes, available tenors
• Collateral requirements (e.g. partial cash collateral, pledge of securities etc.)
• Interest rates (defined as spread relative to a market index)
• Communication channels (who is authorized to give instructions over the
phone, for example)
We should stress again that borrowing capacity is risk-sensitive and tends to evaporate just when
you need it most, i.e. at a time of liquidity strain. This is why it is important to regularly negotiate
money market trading lines and then test and use them even when there is no need for the
liquidity. This is simply an assurance that a diversified borrowing network is in place for the day
that a liquidity need arises. Moreover, if you only use the borrowing facility when you really need
it, the counterpart might sense that there is an issue and could well refuse the loan.

Purchased Funds Dashboard

The most appropriate use for these volatile and high-cost funds is for short-term, seasonal liquidity
needs. MFIs should establish safety rules that do not allow purchased funds to exceed a certain
percentage of the balance sheet total and that limit the utilization of the various facilities under
normal operating conditions. For example, the liquidity management policy could stipulate a
maximum utilization of 50% of all available overdrafts and money market borrowing lines across
all bank counterparts during normal operations. At the same time, any individual facility should
not be drawn to more than 75% of the available limit. The point is to not signal liquidity tightness
to the bank partners and to keep a significant amount of short-term borrowing capacity available
for immediate and unexpected cash requirements.

82
The purchased funds dashboard in Figure 5.2 gives a concise overview of all overdrafts and money
market borrowing lines available to the MFI. This dashboard is adapted from the treasury tools
used by the same Ugandan MDI as in the vault cash example in Figure 3.10.

A purchased funds overview like in Figure 5.2 makes it easy to track the current utilization and
available limits under various short-term borrowing facilities. The dashboard also visualizes the
limit encumbrances from outstanding vault cash buffer checks that could be presented for payment
against the particular account by a branch at any time. These checks need to be covered by positive
balances or available limits with the bank on which the check is drawn. By showing the effective
borrowing cost alongside the available limits and utilization, the treasurer can use the dashboard
to steer utilizations towards limits with the lowest cost and move available cash balances towards
paying down the most expensive overdrafts first.
0 0 0 0
0
-182 -134
-241

-542 -485 -98 -216

(320)
-717 (350)
-222
-1041
-600 -19 (645)
-626 -173
(800)
-627
-1015

-159
-1200 (1,200) -1033

(1,500)

(1,750)
-1800
Bank
Stanbic, A x%
14.5% Bank
Nile B15%
Bank 1, y% Bank
Nile BankC
2, x%
15% Bank
Allied 2,D x%
15% Bank
Allied,E y%
18% Bank
CitiBank,F19%
x% Bank
DFCU,G x%
20%

Used OD
Current Utilization
Unbanked Buffer Checks
Available Overdraft
Available OD Current OD OD Limit

Outstanding Buffer Checks Overdraft Limit

Figure 5.2: Purchased Funds Dashboard adapted from a Ugandan MDI.

Short-Term Notes, Bills of Exchange, Commercial Paper

Instead of purchasing short-term funds from a commercial bank, larger MFIs might also consider
issuing negotiable debt instruments to a broader range of investors. Such negotiable notes,
certificates of deposit (CDs), bills of exchange13 or commercial paper14 are all legally abstract and
irrevocable promises to pay. These titles are typically part of a “program” arranged with the help of
a bank that allows the regular placement with investors under standardized terms and conditions.
Typical buyers of short-term paper are institutional investors including banks, insurance
companies, social security institutions, government entities, and money market investment funds.
These issuance programs are generally arranged as private placements and are typically held to
maturity by the buyers. What type of instrument to use depends on the legal framework and the
conventions in the particular market.

13 A Bill of Exchange is an unconditional written order, calling on the person to whom it is addressed to pay, on demand or at a
fixed future time, a sum of money to the order of a specified person or to bearer. A promissory note is particular form of a bill
of exchange that is initiated by the entity obligated to pay rather than the beneficiary of the payment as in standard commercial
bills of exchange.
14 Commercial Paper is a form of unsecured, short-term debt obligation evidenced by negotiable promissory notes issued by
large corporations and financial companies.
83
 Accessing the capital markets as a means to broaden sources of short-term funding is an important
strategy to consider for growing and profitable MFIs that are ready for the additional complexities
and reporting requirements. However, this is not necessarily the suitable strategy for many MFIs
and proper caution and analysis must be undertaken before doing so.

Liquidity Pools and Apex Institutions

Another frequently cited option for short-term funding is a liquidity pool created by a group of
MFIs that operate in the same region or national market. Such liquidity “self-help groups” are also
often discussed in the framework of apex institutions. An apex is a legally registered wholesale
institution that provides financial management and other services to retail MFIs.
Apex institutions come in two basic variations: top-down or bottom-up. Top-down apex
organizations are essentially designed as efficient vehicles for the delivery of development funding
and technical assistance to a number of retail MFIs in the area. Since many MFIs lack access to a
lender of last resort, such as a central bank, a borrowing window offered by the apex institution is
particularly important for providing emergency liquidity. Some microfinance networks also play
a similar role for their member institutions and may either fund a stand-by liquidity facility or
guarantee such a facility through a local commercial bank.
A bottom-up apex model is built around funds that are mobilized by member MFIs for on-lending
to other members in need of liquidity. Such networks between peer organizations may reduce
the reliance on continuous donor or government support. By cooperating, these MFIs should be
able to achieve economies of scale that render their financial intermediation more efficient. Often
enough, such bottom-up apex organizations are the only wholesale source of additional liquidity,
because the commercial banking sector may not be willing or able to work with MFIs. Examples of
bottom-up arrangements are frequently found in the mutual and cooperative banking movements,
where they are called unions of cooperatives or cooperative central banks.
Bottom-up apex models can also be risky as they can easily deteriorate into a cross subsidy
mechanism, where stronger members finance losses at weaker member institutions. Generally,
cooperative apex institutions can work succesfully if the apex has the power to enforce financial
discipline and takes on some form of self-regulatory role within the network.

84
Box 5: Bank Andara - A promising Top-Down Apex Institution in
Indonesia
Bank Andara was created in 2009 as a dedicated wholesale bank serving
microfinance institutions in Indonesia. Shareholders in Bank Andara are Mercy
Corps, IFC, Hivos-Triodos Fonds, KfW, Cordaid and the founder of the regional
Bank Sri Partha, which was acquired by the international shareholders and
transformed into the legal platform for Bank Andara.

The vision of Bank Andara is to serve as commercially viable apex bank for MFIs
throughout Indonesia. It offers treasury and investment products, a range of
funding solutions including long-dating senior refinancing, as well as technology
services and innovative product partnerships. Bank Andara is committed to
remaining competitively neutral and will not provide retail financial services to
its partner MFI’s clientele. The fundamental challenge in Bank Andara’s strategy
is to achieve profitable scale in this narrowly defined market against a backdrop
of competition from the leading national banks, such as Bank Mandiri and Bank
Danamon. Areas where Bank Andara truly can add value are shared technical
services, the design of innovative products and the smart use of new technology
for reaching out to underserved rural populations.

85
 Chapter 5 - Key Concepts and Next Steps
• The action variables of liquidity management are short-term investing and
borrowing. From an operational perspective, one must plan when and
what amounts to draw down or pay into, and which short-term investment
and borrowing products to use and when. The accompanying Excel model
LiquidityManager provides a workspace for planning these specific liquidity
actions on a daily basis while immediately simulating the effects of the deci-
sions over the coming 12 months.
• The main criteria in selecting appropriate liquidity management instruments
are the minimization of credit and market risk, speed and flexibility, as well
as the reduction of transaction costs.
• On the asset side, the scope of available instruments includes interest-bear-
ing savings accounts, sweeping of current account balances into overnight
placements, time deposits with correspondent banks and investments in
low-risk government debt securities.
• Potential short-term borrowing instruments include overdraft lines of credit,
repos, unsecured money market borrowings, and the placement of negotia-
ble short-term obligations (certificates of deposit, commercial paper or bills
of exchange).
• The most effective approach to developing access to appropriate instruments
for short-term investing and borrowing at better terms and conditions is to
work with your network of correspondent banks. This is where a growing
MFI can cultivate strategic relationships and leverage its bargaining power
based on transaction volumes and investable amounts. Close relationships
with international donors are a nice bonus, but a successful commercial MFI
must seek closer integration into the national or regional wholesale financial
sector.

86
CHAPTER 6
CASH MANAGEMENT IMPLICATIONS OF INNOVATIVE
OUTREACH TECHNOLOGY

6.1 Introduction
This chapter explores the cash management consequences of innovative initiatives in microfinance.
These include setting up automatic teller machines (ATMs) or self-service branches, cell phone
banking, electronic bill payment and micro-remittances, handheld mobile terminals for MFI
officers etc.

For the most part, nothing really changes about the principles of cash handling and liquidity
management. However, with new technologies come new opportunities for theft and fraud that
one must anticipate and discourage with controls. With new partnerships and agency arrangements
also come new settlement accounts and vault cash points that must be provisioned and transactions
that must be reconciled into the in-house accounting system. All of the cash management skills
discussed previously easily transfer to any new product and technology environment.

The most important reminder is to not let internal control and cash management be an afterthought
in designing new products. As a MFI conceives a new service or delivery channel, the security and
liquidity management aspects should be built into the design and implementation from the very
beginning.

87
 6.2 ATMs and Self Service Points
Many deposit-taking MFIs strive to provide bank-like convenience and access to savings for their
clients. In developing markets, ATMs still carry a cachet of progress and having an ATM card is
not only a convenience but also confers a certain prestige. Issuing ATM cards on current or savings
accounts therefore may make good marketing sense for many MFIs. However, owning a whole
network of ATMs and self-service banking stations may not be economically reasonable for a MFI.

In many cases and depending on the size of the institution, the most practical ATM strategy for
MFIs is to arrange access to an existing ATM network through a partnership with a commercial
bank or a payment systems provider. In such a partnership, where a MFI can bring a substantial
number of additional users into an ATM network, there is room to negotiate attractive fees for the
MFI clients. The MFI can consider subsidizing access charges for its clients as a way to encourage
smaller and more frequent withdrawals, thus stabilizing aggregate deposit balances.

The strategy of accessing an existing ATM network also minimizes the cash management
requirements for the MFI. The ATMs are not owned or operated by the MFI, so stocking them
with vault cash is not the MFI’s concern. Their clients’ transactions at the banks’ ATMs will be
settled via correspondent accounts with the bank that owns the ATMs, which adds one more
correspondent account to manage. The MFI’s IT system must be capable of responding to real-
time balance queries from the ATM network for authorizing disbursements to avoid uncovered
withdrawals by MFI clients.

Nevertheless, some MFI’s can successfully implement their own network of ATMs if there
is sufficient scale to support the additional costs associated with owning and operating such a
network. After undergoing a full feasibility and cost-benefit analysis, management can make the
decision whether taking such a strategy would make sense. In some cases, ATMs can also become
independent business units of the institution by generating additional revenues, and targeting a
different clientele. In addition, having their own ATM network can strengthen a MFI’s image and
branding.

88
6.3 Merchant Service Points
Another strategy to maximize outreach and client convenience are microfinance service points in
supermarkets or other retail store chains. Such merchant service points could be implemented in
three different formats:

1. A full-service agency agreement with access to real time balances and a full range of
transactions including deposits and withdrawals, submission of loan applications and
other documentation via image scanner, loan disbursement and installment collections
etc. Such a full service agency agreement would only be an option for a large MFI
with a national chain of stores, because it would require some specific equipment and
agent training at the merchant. This would only be economical if sufficient client traffic
can be generated throughout the merchant branches to recover the investment and
maintain the staff awareness of the MFI services.

2. A limited agreement with an existing network such as the ATM card network
discussed above. MFI clients would use their ATM cards on existing standard point-
of-sale (POS) equipment at the merchant. In addition to paying for purchases, the
POS terminal could also handle balance inquiries, accept a cash deposit to the client
account and allow limited withdrawals using cash back functionality.

3. Leverage mobile payment transfer services like M-Pesa, SplashCash, MTN Mobile
Money, or WIZZIT to turn retail merchants into cash points for MFI clients. This
would require that MFI clients, merchants and the MFI itself are registered users
on the particular mobile payment service. The merchant then simply becomes a
cash acceptance / disbursement agent with settlement across the mobile payment
accounts. A client with available funds in his mobile money account can obtain
physical cash from a merchant by transferring the equivalent plus a possible small fee
to the merchant’s account in real-time. The merchant receives an SMS confirming the
receipt and disburses the cash. A client with cash in hand can deposit the cash with a
merchant and receive a transfer of the same value into her mobile account. From there,
the client transfers the funds to the MFI to pay a loan installment, for example.

In terms of cash management for the MFI, merchant service points are quite simple to handle. The
vault cash procurement and control is entirely outsourced to the merchant. The MFI settles with
the merchant via a dedicated correspondent account at a mutually suitable bank. That account
needs to be reconciled daily and covered with settlement balances as per the standard processes
and controls applied to all correspondent accounts.

Merchants are natural cash point partners for MFIs, because they are otherwise vault cash positive
from their retail sales and can cover disbursements without incurring additional cash procurement
and management costs. Merchants can also potentially benefit from the additional foot-traffic that
this service brings into their store.

89
 Box 6: Go Banking, South Africa
Pick ‘n Pay Stores Ltd is one of South Africa’s leading supermarket chains. In 2002, it
launched Pick ‘n Pay Go Banking as a joint venture with Nedbank Ltd, one of South
Africa’s Big 4 retail banks.

Go Banking was designed as a mass market offering that specifically targeted the
previously un-banked working class. The value proposition is a combination of
convenience, simplicity and low cost. Clients have access to their accounts at all
Nedbank branches, Nedbank ATMs, and self-service terminals as well as at over
5,000 Pick ‘n Pay tills across South Africa.

The basic platform is a low-cost current account called the Go Account and the Go
Credit Card that can be accessed through the Pick ‘n Pay teller stations and offers on-
the-spot discounts and promotions at Pick ‘n Pay stores. The Go Account also comes
with on-line and cell-phone access options.

In a relatively crowded retail banking market, Go Banking had reasonable success

and has grown to about 90,000 accounts. This is enough to prove the validity of
the merchant service point model, but not quite enough to reach profitable scale
as a stand-alone subsidiary. In 2008, Nedank decided to integrate the Go Banking
operations into its mainstream brand and traditional distribution channel.

The service point agreement with the Pick ‘n Pay stores continues, however, and
is even expanded by additional Nedbank branch outlets inside Pick ‘n Pay Ltd
supermarkets including the Boxer and Score branded stores.

90
6.4 Mobile Banking
Mobile banking has increased in popularity in recent years and many MFIs are incorporating or
considering its usage as a way to improve customer service. Mobile banking can have many different
applications. At a very basic level, this could simply be about communicating with clients via SMS.
Many MFIs routinely send SMSs with installment due date reminders and acknowledgements of
receipt directly from their loan management software.

The next step would be to allow certain self-service transactions, such as balance inquiries and
mini-statements via an exchange of SMS messages between the client and an automated interface
of the MFI’s back end software.

In a fully-featured mobile banking application, the cell phone becomes an alternative transaction
interface onto a standard bank account. In addition to the above, this interface would provide
further information services and allow bill payment and person-to-person transfers. For the bill
payment and transfers to be worthwhile for the client, it would be necessary to have a critical
number of users who can be reached via the service.

The cash management consequences of cell phone banking remain the same: We estimate the
percentage of client transactions going through this mobile channel and provision the interface
accounts with settlement cash. Excess balances should be consolidated back to the primary
clearing account with standing overnight sweeping orders. The transactions in the settlement
accounts must be reconciled automatically to internal client accounts via reference data that is
imported from structured, electronic statements.

91
 6.5 Handheld Terminals for Field Staff
Finally, a common mobile outreach technology are handheld terminals issued to MFI field
staff. The functionality of handhelds may vary between that of an electronic receipt block with
printer and a fully functional teller workstation. Some handhelds operate off-line on a local copy
of the accounts database, i.e. with static client data, recent transactions and balances. With that,
client disbursements and payments for savings or loan accounts can be authorized, recorded and
receipted. The transaction data accumulated over the course of the off-line day must promptly be
uploaded and reconciled into the main MFI back end systems.

Box 7: AMIO Teller – Pocket PCs used by Small Credit Cooperatives

in Mali
AMIO teller is software solution on a standard
Pocket PC or smart phone that provides full teller
functionality in the field. It has been piloted with
good success within the Nyèsigiso Credit Union
Network in Mali. The software development and
systems integration was carried out by Desjardin
Développement International with funding from
USAID-Mali and CIDA.

The off-line database on the handheld is synchronized

at the beginning and close of every business day with
the central SAF banking software. SAF handles all
operational and financial transactions for the small
savings and credit cooperatives in the Nyèsigiso network. A detailed account of
the AMIO project is available for download at www.did.qc.ca/en/ pdf/services/DID-
AMIO-Eng.pdf.

Where economically and technically feasible, an on-line handheld terminal can be a better
option. On-line connectivity ensures database consistency in the back end and greatly reduces the
opportunity for fraud. A possible fraud pattern by clients could consist of multiple withdrawals
against the same deposit balance with different field agents on the same day. This would be difficult
to catch, unless the back end database is updated in real time.

Thinking through the fraud opportunities offered by handhelds from the field officer’s
perspective, there are typical patterns such as the short-recording of deposits and over-receipting
or underpayment of disbursements. In this respect, a handheld is no different from a numbered
receipt block. An example of the new opportunities created by the technology could be to book
some payments obtained from clients against a duplicate database on an additional memory chip
that is not turned over in the evening for synchronization with the back end. The client receives an
official-looking receipt printed from the handheld terminal, except that these transactions never
make it to the back end and the cash overage is stolen by the field officer. Most of these schemes
are much harder to carry out in an on-line environment versus an off-line system.

92
Chapter 6 - Key Concepts and Next Steps
• Cash management implications of technology initiatives in the microfinance
sector were discussed, including: access to ATM networks, handheld mobile
terminals, merchant cash points and mobile banking.
• Providing access for MFI clients through existing ATM networks operated
by banks or payment service providers has the cash management advantage
of outsourcing most operational risks including vault cash handling and
equipment maintenance. Vault cash transactions are thus transformed into
simple book money settlements in a correspondent account.
• Merchant cash points are another promising avenue for outsourcing a
portion of vault cash transactions to partners who typically are vault cash
positive and interested in attracting MFI clients into their premises for cross
selling.
• The same principles of internal control, cash handling and efficient liquid-
ity management developed earlier in this toolkit easily transfer to any new
product and technology environment.
• Internal control and cash management should be integrated early, including
changes to the vault cash and book money flows, where the fraud vulner-
abilities are, and how they can be mitigated.

93
 APPENDIX 1- CASH FLOW FORECASTING
This Appendix provides a detailed description of Cash Flow forecasting in the context of liquidity
management utilizing the Excel model LiquidityManager.

1. Forecasting Loan Demand and Deposit Supply

Loan disbursements and repayments as well as savings deposit transactions are the most important
categories of frequent, small, and in many cases uncertain cash flows in a MFI’s operation.
Therefore, cash-flows that are directly linked to these transactions are the first and most important
input to the cash flow forecast.

1.1 Properties of the Loan Portfolio in MFIs

In order to predict future cash flows from loans and deposits, we need to first develop a clear
understanding of the nature of the loan and deposit portfolios in microfinance. Most MFIs make
small, uncollateralized working capital loans for maturities seldom exceeding one year. These
loans are typically structured as installment loans requiring regular weekly or monthly payments
of principle and interest. Many MFIs employ a graduation principle, where borrowers have to
demonstrate their reliability by starting with very small loan amounts that are gradually increased
with each successfully repaid credit. These loan cycles already engender an almost automatic
growth trend in the loan portfolio.
In addition to such a possible built-in loan-expansion factor, MFI management must strive to
understand all other major determinants of its loan demand. How do the borrowers actually use
the loans? Are the loans employed for working capital in a micro-business, as the loan application
states, or are the funds used to finance seasonal or consumer spending instead? Sound data about
the actual uses of loans will make it much easier to explain the seasonal and cyclical variations in
loan demand.

1.2 Behavior of Savings Deposits

On the deposit side, it is now well documented that MFI clients want to save and can accumulate
funds that, if collected efficiently, will provide an attractive capital base for a MFI. In addition to
the safety of savings, which is always the primary concern of a saver, the second most important
condition that micro-savers attach to their deposits is the immediate availability for withdrawal.
Contrary to the apprehensions of many MFI managers, this does not mean that savers will indeed
withdraw their funds frequently. Those who save to prepare for emergencies or to cover a specific,
future expense will typically be very reluctant to take out their money for lesser causes. Even after
an emergency, the poor tend to replenish their savings accounts quickly.

MFI managers often wonder how they can improve the stability of their deposit base while still
allowing their many small savers immediate daily access to their funds. This can be done because
of the great likelihood that many of the small individual transactions will offset each other. While
one saver may experience an emergency and may withdraw his funds, two others might have just
sold their products at the local market and make their weekly savings deposit. Conversely, the
MFI would face a severe liquidity crisis if most of its depositors were to be hit by an emergency at
the same time, such as in a natural disaster. A natural disaster or a harvest failure could not only

94
run down deposits but would also lead to an increased loan demand. Such correlation in client
behavior can only be controlled by appropriate diversification.

1.3 Projecting the Net Cash Effects

Before we start to project the cash flows of our lending and savings transactions, it should be
pointed out that we will focus on the net liquidity effects of these transactions. In other words, the
dynamic liquidity planning model will include projections of the outstanding loan portfolio on
an aggregated basis, after having totaled expected new disbursements against planned repayments
on old loans. A net increase of the outstanding loan portfolio represents a cash outflow, while a
reduction of the outstanding amount leads to a net cash inflow. The same principles apply to the
deposit base.

1.4 Trends, Seasonal Patterns and Macroeconomic Cycle Effects

A widely used practical approach to estimating future loan and deposit development is to
distinguish two components of total deposit and loan variability: a trend element and a seasonal
component.

The trend component represents the long-term growth rate of the loan portfolio or the deposit
base. One could estimate these growth rates by fitting a trend line over the monthly deposit or loan
totals by main product category for a long period of time, say 5 years. The seasonal component
measures how deposits or loans behave in any given interval (month or week) due to seasonal
factors as compared to the underlying long-run trend.

Total outstanding loans

Trend

Seasonal deviation

Total expected loans = Trend loans + seasonal deviation

Jan Feb Mar ... ... Nov Dec

Figure A1.1: Decomposition of Loan Demand in Trend and Seasonal Elements

Figure A1.1 above shows an example of a loan portfolio that follows a long-term growth trend
(depicted by the straight line) but is subject to strong seasonal fluctuations. During the early
months of each year, loan demand tends to be higher than the average trend, while during the
summer months it falls below the expected trend level.

95
 In addition to long-term trends and seasonal effects, one may also want to consider an additional
impact on loan demand and deposit supply from the macroeconomic growth cycle. In times of
strong economic growth there tends to be more micro-business opportunity and consequently
more loan demand as well as a greater willingness to lend by banks and MFIs. In a recession,
business and consumer loan demand tends to decline along with the eagerness of institutions to
disburse.

For the deposit business, the direction of the impact of boom and recession is less clear. In
developed markets, one observes lower savings rates during economic expansion and higher
savings during recession, because people are concerned about losing their job and try to accumulate
a safety buffer. In developing countries and in microfinance, the macro-cycle could actually work
the other way around: financially vulnerable MFI clients have a generally high propensity to save
for adversity and their slightly higher income during a boom might just give them the cash flow
to finally put some money away. In a recession, however, when revenues of micro-entrepreneurs
decline, we could expect clients to draw down their deposits to cover essential expenses. In the end,
the most plausible way to incorporate the macroeconomic cycle effects into the forecast of deposit
supply and loan demand will be based on the actual experience of the MFI in its particular market.
Practically, a macro-cycle adjustment to the forecast would simply be a percentage scaling factor
that inflates or discounts the seasonally adjusted forecast as per the macroeconomic expectations
of the MFI managers.

A MFI that is just beginning to mobilize deposits can estimate trend and seasonal savings without
historical data by studying peer institutions that work with a similar clientele. One could also
obtain data on the aggregate deposit trends in the banking sector from central bank statistics.
However, MFIs should be cautious in relying on official macro-statistics as they might not capture
the economic behavior of the MFI’s target population. If no external data are available, the deposit
forecast should reflect the MFI’s best estimate in terms of the business plan and the expected
marketing effort. The MFI should also attempt to make some plausible assumptions about how
customer behavior might vary according to season. Some original market research on savings
supply and loan demand patterns could help validate these working assumptions.

96
1.5 Forecast Calculations
Calculating a monthly forecast of deposit supply and loan demand by main product group is
actually quite simple when planning with a standard spreadsheet software. Trends, seasonal
indexes and manual adjustments for macro-cycles and institution-specific are already built into
the accompanying liquidity management tool, see Figure A1.2.

Figure A1.2: Deriving Forecasted Monthly Loan Balances by Product. Screenshot from LiquidityMan-
ager, sheet LoanForecast.

One can manage the forecast calculations depending on the availability and relevance of the
historical loan and deposit data with the parameters highlighted in the top section of the forecast
sheet, i.e. cells B15:F19 in Figure A1.2.

You will notice that we often use an exponential trend function instead of a linear trend assumption.
In Excel, an exponential trend can be calculated using the =GROWTH() function, while linear
trends are given by the functions =TREND() or =LINEST(). Applied to a time series of economic
variables, an exponential function frequently gives a better fit compared to a simple linear trend.
Exponential trends are often more suitable, because economic variables are generally subject
to compounding growth effects from growth rate upon growth rate, interest upon interest, and
inflation upon inflation. Such compounding growth leads to an exponential trajectory rather than
a straight line development.

In the Excel model Liquidty Manager, we incorporate possible seasonal patterns in loan demand
and deposit supply by calculating a seasonal percentage index for every monthly observation
point. The seasonal index value is simply the actual observation divided by the underlying trend
for that month. These seasonal index values are averaged across all observations for a particular
calendar month. For example, assume the actual observations divided by trend value gave us 1.25
in June 2008, 1.15 in June 2009 and 1.20 in June 2010. We calculate an average seasonal index of
1.20 for the month of June. If we now want to forecast the June 2011 value, we simply multiply the
trend forecast with the June index of 1.2 to obtain a seasonally adjusted forecast.

97
 Figure A1.3 shows actual seasonal index values that were calculated on the main savings product
groups at Allied Bank in Pakistan with 1996 - 2006 data.15

1.200
Savings Accounts
Current Accounts
1.150 Term Deposits

1.100

1.050

1.000

0.950

0.900
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
!

Figure A1.3: Seasonal Index Values for Savings Products Allied Bank, Pakistan

The seasonal index values for Allied Bank in the chart above show a clear seasonal spike in June
for all three main product groups and a strong drop just for ordinary savings towards the year-end.
Both seasonal effects have a plausible explanation. Traditionally, bonuses were determined based
on June balance sheets, which explains why branch staff made a special effort to bring in deposits
around June and then let the balances taper off again. The decline in ordinary savings accounts
towards year end was explained with the Zakat, a charitable donation that is made during the
month of Ramadan, which fell towards the end of the year during 1996 to 2006. In Pakistan, a
Zakat levy is by law deducted from all savings accounts, but not from current accounts. Hence,
many savers temporarily withdraw their savings or move balances into current accounts to avoid
the Zakat. It would be important to incorporate this very clear pattern “June high – Ramadan low”
into the cash flow forecast by seasonally adjusting the expected deposit balances relative to the
year-on-year growth from the business plan.

1.6 Own Estimate Override

Exponential trends, seasonal indices and the macro-cycle overlay estimates are great for filling the
blanks if we have no other information about the future. However, any specific knowledge about
upcoming activities and our clients’ intentions should take precedent over the default trend. For
example, if the MFI is opening new branches and launching new loan products with a special
marketing campaign, we should utilize the strategic plan forecasts rather than following the trend
from prior years. For this purpose there is a manual adjustment line in the planning tool that
allows the user to replace the trend assumptions with more specific forecast based on additional
information.

15 Stability of Small Balance Deposits in Developing and Emerging Markets. CGAP Technical Note by Joachim Bald, June 2009.

98
2 Cash Flows from Interest Payments
Once a reliable forecast of deposit supply and loan portfolio balances have been estimated, related
interest flows can be calculated by simply multiplying the forecasted balances with the weighted
average interest rate and fee schedules for the particular loan and deposit products. See the sheets
LoanForecast and DepositForecast in the accompanying Excel LiquidityManager

Cash interest payments on long-term debt owed by the MFI will be planned in the same way, i.e.
either by multiplying the total outstanding borrowings with the weighted average interest rates or
by drilling down to individual facilities and their outstanding balances and applicable contractual
rates.

Model:

As a default, the LiquidityManager tool provides an extrapolation of long-term borrowing balances

and related interest expense from the corresponding line items in the prior balance sheet and income
statement. As an override, there are also manual input lines for capturing individually planned long-
term funding transactions and interest payments, see sheet OtherFinancial Flows.

Figure A1.4: Screenshot Liquidity Manager, sheet OtherFinancialFlows.

3 Interest on Short-Term Investments and Borrowing

The remaining components of net interest income, i.e. interest received on short-term investments
and interest paid on short-term borrowings will be accounted for at a later stage of the cash flow
plan. The short-term interest payments depend directly on the pattern of short-term investments
and borrowings. Since short-term investments and borrowings are the action variables of liquidity
management, it would not make sense to average these flows on the basis of historic data. Rather,
we will plan for the short-term interest flows along with the explicit movements in short-term
investments and borrowings in the final stage of the cash budgeting process, see LiquidityManager
worksheet LiquidityWorkbench.

99
 4 Adjusting the Cash Flow for Loan Delinquency
Before combining all of the previous forecasts to derive the total estimated change in cash from
financial operations, we have to make one final adjustment for the cash effect of loan losses. So
far, the cash flow forecast from lending operations is driven by the incremental principal disbursed
as the difference between gross loans outstanding from one month to the next. Obviously, not
every installment of loan principal and interest due will actually be received by the MFI, because
some borrowers will default on their obligation.

We can now adjust the monthly interest flows for loan delinquency by multiplying the contractual
flows with the average collection rate: if the collection rate is 98% and the interest and fee income
per month on the portfolio is $1,000, then we can expect to actually receive $980 in cash from
client installment payments.

The same idea applies to the principal repayments on the existing portfolio, here we will also see
only 98% of the principal flow back, if the average collection ratio is 98%. However, we need to
take into consideration the average maturity and the resulting turnover of the loan portfolio. It
makes a big difference to lose 2% of the principal on every loan within a short-term portfolio that
turns over 3 times a year, versus a portfolio with average maturity of one year. On the short-term
portfolio, we lose 2% of the portfolio capital three times in a row, instead of booking just one 2%
principal write-off on the 1-year loans.

The turnover rate is simply the inverse of the average loan period. If the MFI grants loans with an
average of six months term, then the turnover rate is twice per year or one sixth per month. The
cash loan loss then is calculated as follows:

Monthly Cash Loan Loss = Gross Loan Balance * Monthly Turnover Rate * (1 – Collection
Rate)

For example, the monthly cash loan loss on a $1,500,000 loan portfolio with 9 months average
maturity and a 96% collection rate would be:

Monthly Cash Loan Loss = $1,500,000 * 1/9 * 0.04 = $6,666.67

Model:

The LiquidityManager tool implements the cash loan loss logic as above directly alongside the loan
balance forecasts in the sheet LoanForecast, see line 30 in Figure A1.2.

100
5 Cash Operating Expenses and Revenues
In order to get a complete picture of the future cash flows and net funding requirements, we still
have to incorporate the cash effect of non-interest operating expenses and revenues. Few MFIs
have significant additional revenues apart from interest and similar fees and service charges.
However, if a MFI has important income from transfer fees, consulting services, or collects rent on
leased premises, for example, then these cash flows should obviously be included in the liquidity
plan. Most operating expenses can be extrapolated from the income statement. We just have to be
sure to avoid the typical non-cash accounting items such as depreciation, amortization and loan
loss reserve expenses.

Seasonal Expense Patterns: When planning in monthly intervals, it is a reasonable simplification

to break out the annual operating expenses into a steady flow of 12 installments over the year.
In the accompanying liquidity planning tool, we expand the projections for the immediate next
month and look at daily cash flows. Here, it would be useful to specify exactly on which day certain
large expenses such as the monthly payroll are going to hit the books.

Model:

See sheet OperatingCosts in the LiquidityManager tool. For the detailed daily cash flows during the
first month of forecast period, see the sheet DailyLiquidity.

6 Long-Term Investing and Financing and Other Large Individual Cash Flows
So far, we have accounted for the cash flow effects of the ongoing financial activities and the
ordinary operating expenses of the MFI. The only liquidity changes that still need to be included
in the plan are the cash effects from long-term investing and financing and other large singular
cash flows. These cash flows are relatively easy to predict because they are largely under the control
of MFI management and are generally known well in advance. and can be captured by expected
amount and the time that it will occur. In the Liquidity Manager tool, we combined these cash flow
forecasts with the long-term borrowing transactions and associated interest payments as shown in
Figure 4.9. Examples of cash flows in this category include the following:

• cash proceeds from a capital increase or a long-term loan facility,

• dividend payouts,
• tax payments,
• cash outlays for building a branch office or for purchasing computer equip-
ment.

101
 7 Other Cash Flows: Remittances and Mobile Money Transfers
Many MFIs not only take deposits and make loans but also offer remittances through Western
Union, MoneyGram or other wire transfer companies, including acting as cash points for some of
the newer mobile money transfer services. There is no need to include remittance transactions in
the Net Funding Requirements forecast, because these transfers are liquidity neutral. An incoming
transfer leads to a credit in one of the MFI’s bank accounts and an equivalent disbursement in vault
cash at its counters very shortly thereafter. An outgoing remittance typically represents a vault
cash inflow and a debit in the correspondent account with the remittance service provider. Thus,
transfer services simply lead to shifts between vault cash and book money balances. . We just need
to be sure that the transitions between book money and vault cash are properly factored into the
short-term vault cash plan as we discussed in Chapter 3. Of course, there might be fees that the
MFI is allowed to deduct from the value of the transfers. If the volume of the fees is significant, we
can capture the cash effect of these fees in the operating expense and revenue section of our cash
flow plan, i.e. non-interest revenue - line 13 - in the sheet OperatingCosts.

8 Net Funding Requirements Forecast

Once all the individual cash flows are estimated, they are then consolidated and the net periodic
change in cash and the cumulative cash position, i.e. the Net Funding Requirements are calculated.
The screenshot in Figure A1.5 gives the treasurer a very concise view of the MFI’s projected cash
position over the next year. This is what is expected to happen without any further intervention,
i.e. in the absence of active liquidity management.

Figure A1.5: Net Funding Requirements Forecast. LiquidityManager, sheet TotalFlows.

102