Risk-Based Supervision of Pension Funds: Emerging Practices and Challenges

CHAPTER 1

Risk-Based Supervision of Pensions: The Experience of Early Adopters

Gregory Brunner

Richard Hinz

Roberto Rocha

Introduction

This chapter provides a review of the design and experience of risk-based pension fund supervision in several countries that have been leaders in the development of these methods. The use of risk-based methods, which originated primarily in the supervision of banks, has increasingly been extended to other types of financial intermediaries, including pension funds and insurers. The trend toward risk-based supervision (RBS) of pensions is closely associated with movement toward the integration of pension supervision with banking and other financial services into a single national authority. Although similar in concept to the techniques developed in banking, the application to pension funds has required modifications, particularly for defined contribution (DC) funds that transfer investment risk to fund members. The countries examined provide a range of experience that illustrates the diversity of pension systems and approaches to RBS, as well as the commonality of the focus on sound risk management and effective supervisory outcomes. The chapter provides a description of pension supervision in Australia, Denmark, Mexico, and the Netherlands and an initial evaluation of the results achieved in relation to the underlying objectives.

Introduction

Over the past several decades privately managed pensions have evolved from a supplemental form of DC to become an important—and in some cases central—element of social insurance systems. Their supervision has made a similar transition to meet the requirements of this new role. It has evolved from ensuring compliance with tax laws and labor contracts and relatively simple methods to limit investment risk, toward a much more comprehensive approach ensuring proper management of all the risks associated with complex institutions relied on to provide secure sources of retirement income.

The wave of innovation and reforms in Latin America and Central and Eastern Europe beginning in the early 1980s transformed pension funds from primarily employer-sponsored defined benefit (DB) arrangements into more diverse forms, including most significantly the emergence of special purpose financial intermediaries operating on a DC basis. The move toward DC funds largely removed the capacity to rely on employers to guarantee outcomes and placed financial risks squarely on the shoulders of members. This transition shifted the nexus of supervision from controlling agency risks to managing systemic financial and operational risks. Initially the new supervision regimes were based on simple portfolio limits with very proactive compliance enforcement. Limiting downside risk over short periods through investment controls was the primary concern; the risk-return efficiency and effective capital allocation were secondary considerations.

By the beginning of the new millennium several factors combined to accelerate these changes in supervision methods. Private pension funds in a number of countries accumulated asset levels exceeding those of more traditional financial institutions, in some cases more than 100 percent of gross domestic product (GDP), leading to a commensurate increase in attention to their systemic importance. A perfect storm of rapidly declining interest rates coincident with collapsing equity prices exposed the fragility of the loose funding requirements for the remaining DB schemes. Concerns about the capacity of the new DC plans to produce adequate levels of retirement income also focused attention on the efficacy of their design and operation. This attention led a number of countries to begin to adopt supervision systems based on various risk-based approaches that established new standards for the operation of pension funds and guided the conduct of their oversight activities.

This chapter reviews the experience in four of the pioneers in developing RBS for pension funds: Australia, Denmark, Mexico, and the Netherlands. These countries each have large and well-established pension systems that provide a useful initial review because they constitute a representative sample of the characteristics of pension systems worldwide, including occupational and open pension funds as well as both DB and DC arrangements. The other chapters are based on individual country studies that form part of a joint project of the World Bank and the International Organization of Pension Supervisors (IOPS), as well as discussions with pension supervisors and market participants in these countries. The case studies provide a more detailed analysis of individual countries (Berstein and Chumacero 2006a; Hinz and Van Dam 2006; Andersen and Van Dam 2006; and Thompson 2006).

The chapter is structured as follows. The second section reviews the origins of RBS in banking and insurance and the progress achieved in developing further the risk-based approach under the Basel II and Solvency II agreements. The third section provides an overview of the pension systems of the four countries and the factors that have motivated the introduction of RBS. The fourth section—the core of the chapter—provides a more in-depth discussion of the main elements of RBS of pension funds in the four countries. The fifth section is a preliminary assessment of the impact of the new supervisory approach on the pension sectors of each country, as well as some observations on the challenges that supervisors will face in the future. The sixth section draws some preliminary lessons for other countries.

Conceptual Origins of Risk-Based Supervision: Basel II and Solvency II

The movement toward RBS approaches can be traced to the development of early warning systems for banks. The earliest of these systems was the CAMEL—a rating system that comprised the components of capital adequacy, asset quality, management factors, earnings, and liquidity—which was adopted by the United States in the 1980s.¹ In 1988 the Basel Committee on Banking Supervision implemented the Capital Adequacy Accord (Basel I), which provided a risk-based framework for assessing the capital adequacy of banks to cover credit risks. The development of this framework was an important step in the path to RBS; the framework sought to ensure an adequate level of capital in the banking system by applying weighting to credit exposures based on broad risk classifications.

During the 1990s a number of supervisors implemented risk assessment and early warning systems. In 1993 the Bank of Italy implemented an offsite monitoring system called PATROL; PATROL comprised the components of capital adequacy (PATrimonio), profitability (Redditività), credit risk (Rischiosità), management (Organizzazione), and liquidity (Liquidità). In 1997 the German Federal Supervisory Office introduced an early warning and monitoring system called the BAKred Information System (BAKIS). In the same year the French Banking Commission introduced an offsite supervisory bank rating system called the Organization and Reinforcement of Preventative Action (ORAP). In 1998 the Financial Service Authority in the United Kingdom introduced its Risk Assessment, Tools of Supervision, and Evaluation model (RATE), a comprehensive bank risk-rating system; the Dutch National Bank (DNB) implemented a comprehensive system called Risk Analysis Support Tool (RAST), which has evolved into the Financial Institutions Risk analysis Method (FIRM) model applied to all financial entities regulated by DNB today.

In 1999 the Basel Committee began the process of replacing the Basel I accord with a more contemporary framework that requires banks to improve risk management and corporate governance in conjunction with improved supervision and transparency. The new framework, known as Basel II, is designed to encourage good risk management by tying regulatory capital requirements to the results of internal systems and processes assessment, thereby creating incentives for improvements in risk management. In addition to making the calculation of regulatory capital more risk sensitive and recognizing the quality of internal risk management systems, the framework added two pillars to the model: the supervisory review process and the market discipline. The three pillars of the new model are shown in table 1.1.

Table 1.1 The Three Pillars of Basel II

Source: Bank for International Settlements 2006; authors.

Note:

a. The interest rate risk on the banking book (loans, deposits) is not included in pillar one of Basel II, but is included in pillar two, as well as in pillar one of Solvency II.

The Basel II framework provides banks with a choice between a standardized approach to calculating credit risk using specified risk factors and an internal ratings-based approach subject to explicit approval by the bank supervisor that would allow banks to use their internal ratings systems for credit risk. The framework has been built through a process of extensive exploration by regulators of emerging industry practices in risk management and considerable testing and calibration.

Pillar one requires implementation of an effective and comprehensive risk management system that includes a proper organizational structure; policies; procedures; and limits for credit, market, and operational risk. Banks are required to have an integrated approach to risk management that covers the risks in particular business segments as well as the bank as a whole.

Pillar two—supervisory review—allows supervisors to evaluate a bank’s assessment of its own risks and assure themselves that the bank’s processes are robust. Supervisors will have the opportunity to assess whether a bank understands its risk profile and is sufficiently capitalized against its risks. This pillar will encourage adoption of risk-focused internal audits, strengthened management information systems, and the development of risk management units.

Pillar three—market discipline—ensures that the market is provided with sufficient information to allow it to undertake its own assessment of a bank’s risks. It is intended to strengthen incentives for improved risk management through greater transparency. This should allow market participants to better understand the risks inherent in each bank and ultimately support banks that are well managed at the expense of those that are poorly managed.

The movement toward greater risk focus is also reflected in the insurance industry. The International Association of Insurance Supervisors (IAIS) is working to develop a common international framework for assessing the solvency of insurers. At a regional level, work is underway in Europe on the Solvency II project that aims to adopt a risk-based approach to capital requirements for insurance companies and introduce qualitative requirements for senior management, risk management, model validation, and internal controls. There will also be recognition by supervisors of internal modeling in collaboration with the actuarial profession. Solvency II will involve a three-pillar approach similar to that of Basel II, introducing a supervisory review process and enhanced transparency.

The current solvency framework in Europe dating from the early 1970s defines capital requirements for insurers in terms of solvency margins typically based on simple rules applied to technical provisions or premiums. Under Solvency II the first pillar will define the resources that a company needs to be considered solvent. It will define two thresholds for capital: the solvency capital requirement will set a threshold for supervisory action, and the minimum capital requirement will provide a basis for stronger action or even withdrawal of the company’s license to write new business. As with Basel II, the capital requirement can be calculated using either a simple standardized model or an internal model approved by the supervisor. Pillar two will take into account qualitative measures of risk control focusing on risk management processes, individual risk capital assessment, and aspects of operational risk, including stress testing. Pillar three will address disclosure requirements incorporating more consistent international accounting standards. In many European countries that operate DB pension schemes or guarantee arrangements involving technical reserving, the rules applying to insurance companies may also apply to pension entities.

Across the globe the trend is inexorably moving toward improved risk management based on the three key elements outlined in figure 1.1. First, institutions themselves are focusing on improving their own risk management. They are developing risk management strategies, and they are measuring and assessing risk in a more comprehensive manner. In many institutions this process involves the creation of dedicated risk management units. These units are implementing controls to ensure that risk management polices are followed and that information is presented to management and board in a meaningful fashion.

Figure 1.1 The Basic Risk Management Architecture

Source: Authors.

Supervisors are responding by building up their ability to assess risk. The basic tools of onsite and offsite supervision are taking on a risk focus, and specialist risk units are being created with expertise to tackle complex issues. Many regulators are facilitating improved risk management by implementing regulatory standards and providing guidance. Finally, more external parties are encouraged to take a role in the risk assessment process, either through broadening the role of some traditional players like auditors and actuaries, or through encouragement of greater scrutiny by outside parties by means of greater transparency of reporting.

Introduction of Risk-Based Supervision for Pensions

Overview of the Four Pension Systems

An overview of the private pension systems of the four countries provides an understanding of the factors that motivated the introduction of RBS. Further background information on the pension systems of these countries is provided in the individual country papers (Andersen and Van Dam 2006; Berstein and Chumacero 2006a; Hinz and Van Dam 2006; and Thompson 2006).²

As shown in table 1.2, all of the countries have mandatory or quasi-mandatory private pension systems. In Australia and Mexico, contributions to private pension plans are imposed by legislation. In Denmark and the Netherlands, contributions take place in the context of collective labor agreements that are classified as quasimandatory because most workers are covered by these agreements. The mandatory or quasimandatory nature of contributions results in high coverage rates except in Mexico. The lower coverage ratio in Mexico, despite the legal obligation to contribute, is explained by the large share of the labor force in the informal sector and the lower number of active contributors relative to the total universe of pension fund members.³

Table 1.2 Main Characteristics of the Four Private Pension Systems, December 2005

Source: Hinz and Van Dam 2006; Andersen and Van Dam 2006; Thompson 2006; Berstein and Chumacero 2006a; and Rofman and Luchetti 2006.

Note:

a. Denmark: 44 corporate funds, 30 industry-wide funds, 37 life insurance companies,

b. Australia: 681 corporate funds, 86 industry-wide funds, 194 retail funds, 43 public sector funds.The figures do not include small funds. VaR = value at risk.

The pension systems in these countries are very large, with assets exceeding 100 percent of GDP in all cases except Mexico. The relatively small size of assets relative to GDP in the Mexican case is due to the lower coverage ratio and the fact that the Mexican system is much younger, having started operations only in 1998. However, the mandatory nature of contributions to individual accounts implies that the private pension system will continue growing quickly and increase its share in the financial sector.

Three countries have a large number of funds, ranging from 111 in Denmark to 1,000 in Australia; these funds may operate more than one pension plan. Many of these are occupational funds structured as nonprofit trusts or foundations originally created on a voluntary basis and operating for several decades. They include single funds and larger multi-employer or industry-wide funds. Australia and Denmark also have several for-profit commercial institutions managing pension funds—including life insurance companies in the Danish case.

Mexico has only 18 funds currently licensed. The difference in the number of funds is a result of the different origins and characteristics of the Mexican system. The Australian, Danish, and Dutch systems have their roots in voluntary arrangements with employers. Most funds were initially established with liberal licensing and authorization rules designed to encourage participation and coverage. By contrast, the Mexican system was established as a mandatory system of open funds subject to a strict regulatory framework, including much stricter licensing rules.⁴

Dutch pension funds manage primarily DB plans⁵; the Netherlands has been one the few countries that has successfully resisted the move toward DC plans. The Danish system is a DC system that offers benefit guarantees and operates on a risk-sharing or profit-sharing basis. The guarantees introduce a core liability and the risk of insolvency of the provider. Therefore, the Danish system exhibits some of the characteristics of a DB system, although it operates with more flexible rules than pure DB systems and seems to be moving in the direction of DC plans with fewer guarantees.

Australian pension funds manage primarily traditional DC plans with no formal guarantees. There are still some DB plans, but these are mostly restricted to public sector funds and account for a small share of total assets. Australia best represents a pure DC system.

Mexican funds, by contrast, manage their DC plans under a new regulatory framework that includes a limit on downside risk defined by a ceiling on the daily absolute value at risk (VaR). This is a significant departure from the setup introduced in Chile and other countries in Latin America and Central Europe that relied on quantitative portfolio restrictions to manage risks. Most of these countries have introduced minimum relative return guarantees that intensify herding behavior and lead pension funds to base their investment strategies on tracking errors or relative VaRs vis-à-vis the benchmark portfolio. Pension fund managers in these countries are more concerned with relative risk (the risk of deviating from the benchmark and facing a capital call to honor the relative return guarantee) than absolute risk. Other countries are following the Mexican experiment with interest.

Factors Motivating the Adoption of Risk-Based Supervision

Some of the factors that have motivated the introduction of RBS of pension funds are common to all the four countries, while others seem to be country-specific. Table 1.3 summarizes the motivating factors identified in the individual country studies.

Table 1.3 Factors Motivating the Adoption of Risk-Based Supervision

Source: Authors.

Preventing underfunding of DB plans was a strong factor motivating the adoption of RBS in the Netherlands. Dutch funds enjoyed the equity boom in the 1990s and started taking contribution holidays when funding ratios reached levels considered as high. However, these funding ratios proved insufficient to absorb the adverse price movements in the early 2000s—the crash of the equity market combined with the drop in interest rates led several funds to become underfunded or only marginally funded. Regulators interpreted the outcome as indicating a weakness in the supervisory approach that was perceived as lacking sufficient foresight and concern for the risks facing the institutions.

The introduction of a more risk-based approach to supervision in Denmark was also motivated by a concern with the solvency of pension providers, but the surrounding conditions were different from those in the Netherlands.

• First, the new Danish traffic-light system preceded the equity crash in the early 2000s. By the time equity prices collapsed and interest rates declined, the new system was already in place.

• Second, the new system was introduced as a quid pro quo for a more liberal investment regime in which the ceiling on equity investments was raised to 70 percent. Danish funds were allowed to make riskier investments provided that they held sufficient capital to absorb the risk.

• Third, the Danish system operates on a risk-sharing basis, which means that the system has buffers than can absorb at least part of the adverse price movements.

These differences imply that the desire to prevent underfunding was more important in the Netherlands than in Denmark. However, there was still concern with provider solvency in Denmark, justifying the inclusion of this factor.

Concern with adverse price movements was also one of the motivating factors in Mexico, although the Mexican system is a DC system where the investment risk is shifted to the individual and there is little risk of provider insolvency. The policy concern in Mexico was not the risk of provider insolvency, but the exposure of retiring workers to extreme downside losses and the extreme volatility of benefits across cohorts.⁶ It is interesting to note that, as in the Danish case, the adoption of a VaR ceiling in Mexico and the introduction of strict risk management rules were a quid pro quo for the introduction of a more liberal investment regime that allowed pension fund managers to make riskier investments and use derivatives.

The search for efficiency gains was also one of the main motivating factors in Denmark and Mexico. In both cases, the investment regime was liberalized and pension funds were allowed to invest more in equity and other assets perceived as risky. In Mexico, pension funds were allowed to use derivatives, subject to certification by the supervisor. The relaxation of the investment regime was motivated by perceptions that pension funds were constrained below the efficient investment frontier and that there was scope for longer-term improvements in the risk-return tradeoff. The relaxation of investment rules was accompanied by other rules designed to strengthen risk management and constrain excessive risk-taking.

The need to establish rules that enabled pension funds to take advantage of the increasing sophistication and complexity of financial instruments and markets was a motivating factor in all four countries. The creation of these rules reflects a more general recognition by financial supervisors worldwide that it is no longer feasible to monitor all of the operations of financial institutions; a more effective approach entails ensuring that these institutions have sound risk management practices and internal controls.

In Australia, Denmark, and the