Agensi 1
Agensi 1
Agensi 1
OF PUBLIC BUDGETING, ACCOUNTING & FINANCIAL MANAGEMENT, 10(2), 173-191 SUMMER 1998
ABSTRACT. Federal agencies rely heavily on the annual audit report to monitor the
stewardship of public funds. This study examines the impact of audit delay on the monitoring
of low-income housing projects. Specifically, project characteristics correlated with delayed
audit reports are determined. Significant differences between projects filing audit reports on
time and those filing late are also analyzed. Importantly, samples of both Farmers’ Home
Administration (FmHA) and Housing and Urban Development (HUD) projects are
examined, with private and public/municipal housing projects included in each sample. The
results of this study have monitoring applications for users of the audit report and implications
for federal regulators.
INTRODUCTION
The demand for federal agency accountability creates a demand for
monitoring. Monitoring is part of the stewardship function. Governmental
attempts to provide accountability to the taxpayers concerning federal funds
have relied almost entirely on auditing (Dropkin and Kitrosser 1982). These
audits of accounting information are used to assist in allocating federal
resources and monitoring their subsequent use.
___________________
* Janice Lawrence, PhD, CPA, is an Associate Professor of Accountancy at the
University of Nebraska and the Associate Director of the Business, Ethics and Society
Program. Her major research areas of interest are fraud, ethics, and foreign direct
investments. Her teaching interests are in the areas of financial accounting and
business ethics. Barry J. Bryan, Ph.D, is an Assistant Professor in the Auburn
University School of Accountancy where he teaches financial accounting and auditing.
His research areas of interest are auditing, peer review, quality control and
accounting education.
MOTIVATION
Audit Delay
Audit Delay is measured here as the length of time from the date of the
project year-end to the date the auditor submits the annual audit report to the
district office. This audit delay directly affects the timeliness of information
received and, therefore, the decision-making and monitoring capabilities of
the district office. The length of the audit is described by Givoly and Palmon
(1982) as the single most important determinant of the timeliness.
This timeliness is affected by numerous factors, such as the number of
hours of audit time required, the amount of interim audit work performed, the
number of audit personnel assigned to the engagement, and the number of
overtime hours worked. As Ashton, Willingham and Elliott (1987) point out,
the act of subjecting financial statements to an audit introduces an activity
which may conflict with the requirement of timely reporting. "To the extent
that auditing is a time-consuming activity, the release of the ... statements will
be delayed" (Ashton, Willingham and Elliott, 1987: 276). They found longer
AUDIT DELAY IN LOW INCOME HOUSING PROJECTS 175
audit delays for firms with larger asset bases, worse financial conditions and
busy season year-ends. Other reasons proposed for audit delays were
auditor-client disagreements, normal accounting and auditing problems (Dyer
and McHugh, 1975).
Variations in audit delay between projects can also be attributed to a
wide range of project, borrower or auditor characteristics. For example,
discovery of internal control weakness in a project could result in a longer
than average audit delay. Several prior audit delay studies (Bamber, Bamber
and Schoderboek, 1993; Carslaw and Kaplan, 1991; Newton and Ashton,
1989; and Williams and Dirsmith, 1988) focus on the relationship between
audit delay and firm characteristics such as size.
Past literature has also shown timeliness of information to be associated
with "good" or "bad" news. For example, Chambers and Penman (1984)
found an association between "late" announcements and "bad" news. Users
might expect "bad news" for projects with consistently long audit delays or
increases in audit delay. Such "bad news" could cover a wide range of
events. For example, if a project is under financial stress, or if an auditor has
difficulty resolving certain items or issues with management, audit delay can
be expected to increase. Discovery of problems, such as internal control
weaknesses or noncompliances, during the audit could also result in increased
audit delay.
Audit Quality
Because of the reliance on the audit to provide monitoring of the
stewardship function, audit quality becomes a natural focus. A report
prepared for Accountants for the Public Interest (API) disclosed problems
with the performance of audits of federal funds and the manner in which
federal agencies review the work of independent CPAs.
A poor quality or deficient audit may not successfully monitor the
projects. This could have severe consequences, resulting in failure to detect
misallocation of federal funds and wasting of public resources. The API
report found: ineffective audits, deficiencies in staff monitoring of federal
agencies, and a widespread believe of government waste and mismanagement
among taxpayers (Dropkin and Kitrosser, 1982).
Several variables deserve consideration when the focus is on audit
quality. These variables have been studied and related to audit quality in
prior literature. Among these are audit firm size and auditor changes, as well
176 LAWRENCE & BRYAN
as audit fees, hours, and audit delay. Chow and Wong-Boren (1986)
provided empirical support for the earlier research by DeAngelo (1981), Fama
and Jensen (1983), and Watts and Zimmerman (1983), which contends audit
quality increases with audit firm size.
Research Question
This study begins to address the following research question: does audit
delay impact the effectiveness of the monitoring process for federal agencies?
This question is addressed here by examining project characteristics
correlated with delayed audit reports. Significant differences between the
characteristics of projects filing audit reports on time and those filing late are
also analyzed. Both Farmers' Home Administration (FmHA) and Housing
and Urban Development (HUD) housing projects are considered. The results
of this research as well as extensions could have monitoring applications for
users of the audit report and implications for federal regulation or
enforcement of penalties.
BACKGROUND
Audit Guidelines
The audit of accounting information is used to allocate federal resources
and evaluate the stewardship performance of the project. An agency
relationship exists between the project and the taxpayer, represented by the
government, through its agencies and their district offices.
During recent years, audits of governmental agencies have been the
focus of considerable attention from governmental regulators and agencies,
Congressional Committees (e.g. Wyden Bill), and accounting standard setting
bodies. In response to these governmental and market forces, the audit
guidelines for federal agencies have undergone significant changes in the last
few years. The audit program guidelines were changed in 1989 and audits
are now conducted in accordance Generally Accepted Government Auditing
Standards (GAGAS) with instead of Generally Accepted Auditing Standards
(GAAS).
RESEARCH DESIGN
Data Collection
Data for empirical analysis are collected from archival sources and
interviews. The variable of interest in this study is the length of audit delay.
Other project and audit report variables are determined from information
available to the federal agencies at the time of the audit report review.
Descriptive Analysis
Descriptive analysis is provided for the two samples of projects: FmHA,
Section 515 projects and HUD projects. Univariate summary statistics of
each variable are calculated to provide a descriptive profile for each sample.
Correlations between project characteristics and audit delay measures are also
examined.
T-tests are used to evaluate the differences between the means of the
two classifications for continuous variables. For nominal categorical
variables, the chi-square is used to test whether the distribution of on-time
and late projects is the same with respect to the variables and to examine
significance of the association between the variables and the on-time/late
classification. The Wilcoxon summed-ranks test is used for interval
categorical variables to test whether on-time and late submitting projects
exhibit equivalent population distributions.
entire page of a report was either missing or unreadable. For this reason,
missing data points exist for some of the variables.
Variables
Table 1 provides a description of the variables examined for the FmHA
housing projects. The majority of the variables are self-explanatory, but
several deserve additional discussion here.
AUDITOR. All 40 audit firms contracted by the projects in this study are
small accounting firms. According to Shockley (1982), small firms are more
dependent on an individual client because their audit fee represents a larger
proportion of the audit firms total revenue. DeAngelo (1981) suggests a
lesser degree of independence may decrease the rigor of the audit. This could
possibly impact the length of audit.
TABLE 1
Description of FmHA Sample Variables
___________________________________________________________
Variables Label Variable Measure
-------------------------------------------------------------------------------------------
AUDIT DELAY DELAY Months from year-end to report date
ON-TIME SUBMISSION ONTIME Report on-time: no=0, yes=1
AUDITOR AUDITOR Number of audits assigned to auditor
AUDIT OPINION Audit opinion; clean=0, else=1
SCALED AUDIT FEE AFEE Audit fee/total number of rental units
SCALED CASH FLOW CASHFLOW Cash flow/total number of rental units
SCALED NET INCOME NETINC Net income/total number of rental units
INTERNAL CONTROL WEAKNESS Weaknesses reported: no=0, yes=1
NONCOMPLIANCES NONCOMP Any noncompliances reported: no=0, yes=1
DEVELOPER/ BORROWER Number assigned to project developer
BORROWER
DISTRICT OFFICE DISTRICT Number assigned to the monitoring district
__________________________________________________________
180 LAWRENCE & BRYAN
Sample
For 1990, there are 389 Section 515, rural rental housing projects that are
required to submit an audit report to FmHA. Of these 389 projects, 109
submitted audits during the period of data collection which extended until
April, 1992. The sample includes 100% of the projects submitting audit
reports.
AUDIT DELAY IN LOW INCOME HOUSING PROJECTS 181
One hundred of these were private projects which are required to submit
their audit report to the district office within 60 days of their fiscal year-end.
The other nine public housing projects fall under the Single Audit Act (and
OMB Circular A-128) and, therefore, are allowed 13 months from fiscal
year-end to submit their annual audit report. The period of data collection
was extended through April, 1992, to provide 16 months for the projects so
affected to submit audits.
Table 2 provides selected descriptive data on the sample of FmHA
housing projects.
Sample Classification
Two audit delay variables are used in this study. First, audit delay is
measured by the number of months between the "as of" date of the financial
statements and the audit report date. This measure is then
TABLE 2
Selected Descriptive Statistics for FmHA Sample (n=109)
___________________________________________________________
Variables Mean Minimum Maximum
------------------------------------------------------------------------------------------
Age (years) 6.3 0.4 16.4
Units 46 26 120
Audit Fee 2,029 0 11,018
Loan Principal 1,014,641 70,756 2,381,798
Total Assets 1,000,363 101,711 2,132,622
Net Income -19,998 -82,591 54,135
Cash Flow 4,580 -56,866 61,127
___________________________________________________________
compared with the date the audit report is required to be submitted and the
projects are classified as submitting the audit report on-time or late. Of the
projects actually submitting an audit report, only 17 percent were submitted
on-time.
Descriptive Statistics
182 LAWRENCE & BRYAN
The audit delay for the municipal/public held housing projects ranged
from 2 to 8 months, with a mean audit delay of 4.6 months, while private
housing projects had a mean audit delay of 2.9 months, with a range of 1 to 8
months. This indicates, that while the public projects submitted later than the
public projects, none of the public projects actually submitting a report
utilized the entire 13 month extension granted because of the single audit act.
Of course not every project submitted a report by the time the study ended
(16 months) and information on these projects is not available.
Table 3 presents descriptive statistics on the variables examined in this
study. While an other-than-clean audit opinion occurs in only 6 percent of
the project audits, 39 percent report internal control weaknesses and 72
percent have noncompliances with their loan agreements. It is these
noncompliances which normally account for the
TABLE 3
Variable Descriptive Statistics for FmHA Sample (n=109)
___________________________________________________________
Variables Mean Minimum Maximum Percentage*
------------------------------------------------------------------------------------------
DELAY 2.9 1 8
ONTIME 0 1 17
OPINION 0 1 6
AFEE 48 0 211
CASHFLOW 114 -948 1,411
NETINC -458 -2,277 1,177
WEAKNESS 0 1 39
NONCOMP 0 1 72
___________________________________________________________
* Percentage of projects for which 0/1 variables are coded as 1. See Table 1
for description of variable coding.
TABLE 4
Selected Correlation Coefficients and Significance Values
for FmHA Sample
___________________________________________________________
Variables ONTIME Prob>|R| under N
Ho: Rho=0
AFEE .2093 .0696 76
NETINC .0995 .3565 88
CASHFLOW -.0439 .6921 84
OPINION .2417 .0293 81
WEAKNESS -.2417 .0267 84
NONCOMP -.1938 .0704 88
AUDITOR .0425 .6890 91
DISTRICT -.2078 .0481 91
BORROWER -.2396 .0222 91
___________________________________________________________
Note: Pearson correlation coefficients are given for variables with
cardinal values and Spearman correlation coefficients are given for
variables with nominal or ordinal values.
184 LAWRENCE & BRYAN
TABLE 5
Comparative Analysis of Time/Late Projects
___________________________________________________________
Test Statistics
Variables t-value* Chi-square Significant Level
------------------------------------------------------------------------------------------
OPINION 2.051 .152
AFEE -1.841 .070
CASHFLOW .398 .025
NETINC -.927 .357
WEAKNESS 3.581 .058
NONCOMP 2.179 .140
_________________________________________________________________
* 1-tailed test
Sample
A total of eighteen HUD projects were selected for review by the
regional office during the years 1989-1991. The sample includes 100 percent
of the projects which were reviewed. All eighteen of the projects were
required to submit their audit report to the regional office within 60 days of
their fiscal year-end.
TABLE 6
Description of Variables for HUD Sample
___________________________________________________________
Variables Label Variable Measure
------------------------------------------------------------------------------------------
Audit Delay Days Number of days from year-end to report date
Quality Quality "Acceptable" Quality Rating: no=0, yes=1
Size of Auditee Assets Total assets of the project audited
Size of Audit Firm Staff No. of staff employed by auditor
Independence Fees Auditee still owed prior year audit fee: no=0, yes=1
Awareness of Year Quality review performed in year following
Review initial year of the program: no=0, yes=1
History of Sub- Problems Auditee previously received unacceptable quality
Standard Quality review: no=0, yes=1
_________________________________________________________________
Descriptive Statistics
The audit delay for the HUD projects ranges from 18 to 356 days, with
a mean audit delay of 90.11 days. Table 7 presents descriptive statistics on
the variables examined in this study.
Correlation analysis, the results of which are summarized in TABLE 8,
finds negative correlation between quality and audit delay, although not at a
significant level. A negative correlation is also noted for assets, although not
significant. Another interesting finding is the independent variable, Fees.
Although not significant, a negative correlation was shown, implying that
when companies owe previous year audit fees, delay is longer, which is not
186 LAWRENCE & BRYAN
surprising. The only significant variable was Staff, but surprisingly the
coefficient was positive indicating longer delay for larger firms. Perhaps this
can be explained by the fact that larger firms may have more client
responsibilities during the traditional busy season.
TABLE 7
Variable Descriptive Statistics for HUD Sample (n=18)
___________________________________________________________
Variables Mean Minimum Maximum Percenta
--------------------------------------------------------------------------------------------
Quality 0 1 61
Assets 2,246,489 33,568 8,054,046
Days 90.11 0 211
Staff 10.94 -948 1,411
Fees 0 1 33
Year 0 1 50
Problems 0 1 6
___________________________________________________________
* Percentage of projects for which 0/1 variables are coded as 1. See Table 1
for description of the variable coding.
TABLE 8
Selected Correlation Coefficients and Significance Values
for HUD Sample (n=18)
___________________________________________________________
Variables Days Prob>|R| under
Ho: Rho=0
------------------------------------------------------------------------------------------
Quality -.02587 .9188
Assets .05909 .8158
Staff .84619 .0001
Fees -.09622 .7041
Year .11211 .6578
Problems .02383 .9252
___________________________________________________________
AUDIT DELAY IN LOW INCOME HOUSING PROJECTS 187
STUDY LIMITATIONS
Not every project submitted a report by the time the study ended (16
months) and, therefore, information on these is not available. These
extremely delinquent projects may well be the most interesting.
There is a threat to the external validity of this study resulting from
limiting study to Texas, for the FmHa projects, and Arkansas, for the HUD
project audits. As a result, it may not be possible to generalize the results and
draw conclusions from this study that will pertain to audits of low income
housing projects across the United States, audits of other federal agencies in
Texas and Arkansas, or audits of governmental agencies as a whole.
A threat to internal validity could arise from the limitations in the field
study methodology. Gathering data manually from individual paper
documents in the state offices allows errors of omission to occur as a result of
lost or misfiled documents.
ACKOWLEDGEMENTS
The first author wishes to thank Wanda Wallace, Gary Giroux, and
workshop participants at Texas A&M University for helpful comments on
earlier versions of this manuscript.
REFERENCES
Ashton, R., Willingham, J. and Elliott, R. (1987), "An Empirical Analysis of
Audit Delay," Journal of Accounting Research, 25(2): 275-292.
Bamber, E., Bamber, L. and Schoderboek, M. (1993), "Audit Structure and
Other Determinants of Audit Report Lag: An Empirical Analysis," Paper
presented at the annual meeting of the American Accounting
Association, Washington, D.C.
Chambers, A. and Penman, S. (1984), "Timeliness of Reporting and the
Stock Prince Reaction to Earnings Announcements," The Journal of
Accounting Research, 22 (1): 21-47.
AUDIT DELAY IN LOW INCOME HOUSING PROJECTS 189