Financial Supervisory Implications of Financial Conglomeration

Regulatory structure refers to the way in which a country organizes the various agencies in charge of financial sector regulation. In principle, there are two fundamentally different models of regulatory structure—one based on sectoral groups (i.e., banking, insurance and securities), and the other based on regulatory functions. Regulatory functions refer to the underlying functions of regulation, namely, addressing the various sources of market failure. Carmichael (2002) identifies four main sources of market failure in the financial system, and how to address them—anti-competitive behavior (competition law), market misconduct (conduct of business regulation/consumer protection), asymmetric information (prudential regulation) and systemic instability (macroeconomic surveillance and stabilization).

In the purest form of the sectoral model, a single regulator responsible for correcting all four sources of market failure is assigned to each sectoral group. In the purest form of the functional model, correcting each of the four sources of market failure is assigned to a single regulator that will be responsible for all institutions that are subject to that particular failure. In practice, regulatory structures around the world typically involve a mixture of functional and sectoral divisions. The most extreme case of regulatory approach would be the single regulator supervisory model, wherein there is only one control authority, separated from the central bank, with responsibility over all financial markets and intermediaries, and concerned with all the objectives of regulation.

It is now generally accepted by banking regulators that banking groups or financial conglomerates need to be supervised on both a solo and consolidated basis to take into account supervisory concerns that may be overlooked at the entity level (Palmer, 2002). In fact, this forms part of the core principles identified by the Basle Committee on Banking Supervision. Prudential and market conduct concerns that result from financial services integration include transparency, contagion, regulatory arbitrage, conflicts of interest, double and multiple gearing; fit and proper requirements; and unregulated group entities, which are interrelated (Skipper, 2000).

Recently, significant attention has been focused on the structural aspects of financial regulation, particularly the desirability of unified regulatory agencies—that is, agencies that supervise two or more of the traditional financial services sectors (i.e., banking, insurance and securities). The primary reason has been the trend towards financial convergence or blurring of product lines, and the rise of financial conglomerates, hence the need for more effective modes to supervise them. Smaller countries are also seeking ways to achieve economies of scale in regulation through better management of regulatory resources (particularly personnel) and infrastructure support (Mwenda and Fleming, 2001).

The debate on financial supervisory structure focuses on two issues: sectoral approach versus functional approach, and the single financial supervisor model (FSA) versus multi- financial authorities model (FMA). However, in the context of increasing financial convergence or conglomeration, the first issue seems to have become moot because the relevance of the sectoral approach rests precisely on the separability of the banking, securities and insurance markets. This is confirmed by the adoption or consideration of various models ("pure" or "mixed") of the functional supervisory approach in an increasing number of countries (Masciandaro, 2003). Table 9 [ PDF 117.2KB | 1 pages ] presents a summary of the pros and cons of integrating financial sector supervision, which Èihák and Podpiera (2006) discuss in greater detail.

The global trend towards integrating financial regulation can be viewed as a trend towards restructuring regulatory agencies along functional lines, particularly with regard to prudential regulation (Carmichael, 2002). At least 46 countries had adopted unified or integrated supervision by 2002, with around half of them creating a single regulator for the entire financial sector and the other half merging two of the main supervisory authorities (Table 10 [ PDF 131.4KB | 1 pages ]).⁷

Recent surveys of the experience of countries with integrated regulators indicate a high degree of consensus in terms of the motivation for establishing the integrated agency, namely: (i) convergence in financial markets and the need for a more consistent approach to regulating financial conglomerates; (ii) the need for greater consistency in the application of policy across different industries; and (iii) the ability to make more efficient use of scarce regulatory resources (Carmichael, 2002; de Luna Martínez and Rose, 2003).

With respect to its applicability to developing countries, the literature cites two key lessons that can be learned from the experience of developed country practitioners (e.g., Abrams and Taylor, 2000; Bain and Harper, 1999; Briault, 2001; Carmichael, 2002; de Luna Martínez and Rose, 2003; Llewellyn, 2001; Mwenda and Fleming, 2001; Reddy, 2001; Skipper, 2000; Taylor and Fleming, 1999). One is that simply changing the structure of regulation cannot guarantee effective supervision, and integrated regulation per se is not a solution to regulatory failure. Correcting regulatory failure requires first and foremost better regulation: that is, setting more appropriate prudential and market conduct standards, improving surveillance, and strengthening enforcement. Integrated regulation may help facilitate this process, but it cannot cause these changes to occur by themselves. Indeed, the countries that adopted the integrated financial sector supervisory approach did so to enhance the supervisory process.

The second lesson is that there is no single best form of integrated regulatory agency. Unified financial services supervision has been adopted differently in many countries; its application has varied from country to country and there is no single right way of introducing or implementing unified models of financial services supervision. Factors that accounted for the differences include differences in starting points, differences in industry structures, and differences in objectives.

While there is some support for consolidated financial sector supervision in developing countries, a more contentious issue is whether the unified regulator should be separate from the central bank. The latter, in turn, partly stems from the issue of whether central banks should be (or continue to be) involved in banking supervision. Barth et al. (2002) noted that there are reasonable arguments both for and against this structural issue in the literature. In particular, the main point of contention is its impact on the safety, soundness and systemic stability of the financial system. In particular, arguments for central bank supervision of banks point to the informational advantage that it affords the central bank, which facilitates its conduct of monetary policy. On the other hand, those who argue against central bank supervision of banks typically cite the resulting conflict of interests between its monetary policy function aimed at price stability and bank supervision function aimed at financial stability.

Table 11 [ PDF 75.6KB | 1 pages ] shows the distribution of 151 developed and developing countries according to the location of bank supervision as of 2002.

The table shows that banking supervision continues to be assigned to the central bank in most countries—about 60% of the countries surveyed, with around 45% assigning banking supervision solely to the central bank. This model of banking supervision is also more common in developing countries, including those in Asia.

Clearly, there are strong conceptual arguments both for and against the central bank’s combined functions of banking supervision and monetary policy, which can be supported by the diversity of global experience. However, the empirical work that has been done on this structural issue is still limited. Barth et al. (2002) noted several studies that support a narrower focus for the central bank that does not include bank supervision, but the results are far from conclusive. Again, the general consensus in the literature so far is that there is no "one right answer," and that the answer will largely depend on country-specific circumstances and capacities. These include prevailing conditions in the financial system, the political environment, and the preferences of the public (Haubrich, 1996). Thus, the effects of monetary policy on banking supervision and vice versa should be explicitly examined before a country decides on whether to retain or remove bank supervisory duties from its central bank.

Another key issue that relates to the central bank’s supervisory function is whether it should supervise other financial service sectors as well, such as securities and insurance. Overall, the arguments in the literature reviewed by Barth et al. (2002) weigh more heavily against it because: (i) it will lead to excessive concentration of power; (ii) the conflict of interests would be more extensive; and (iii) it could unduly extend the financial safety net if the central bank’s lender of last resort function is seen as extending across all financial institutions, thereby worsening the moral hazard problem.

Meister (2001: 1; in Barth et al., 2002) very rightly emphasized that the "… design of regulatory and supervisory responsibilities is one of the most important matters affecting the future course of financial market policy. There is, however, no universally valid answer to the question of how this should be done." Furthermore, he noted that the answer cannot be derived from theory. Unfortunately, while there is significant literature discussing the pros and cons of having a single versus multiple regulators, there is very little empirical analysis addressing this issue. Èihák and Podpiera (2006: 3) claim to be the first to come up with a "…comprehensive, cross-country analysis of the emerging experience with integrated financial supervision." Using cross-country data on quality of supervision and on supervisory staffing, they examine whether fully integrated supervisory agencies⁸ are better than other structures, in terms of quality of supervision across sectors and cost efficiency.

Quality of supervision is measured by the degree of compliance with internationally accepted standards in banking, insurance, and securities regulation, that is, the Basel Core Principles for Effective Banking Supervision (BCP), the International Association of Insurance Supervisors (IAIS) Insurance Core Principles (ICP), and the International Association of Securities Commissions (IOSCO) Objectives and Principles of Securities Regulation, respectively. They use a database of assessments derived from the IMF and the World Bank’s Financial Sector Assessment Programs, which have been made publicly available by the countries involved. In particular, the relationship between observance of international standards and the organization of the supervisor (whether fully integrated or not) is examined in two ways—based on BCP compliance, and based on compliance on all three standards.

There were 65 available assessments for the BCP, which consisted of 13 advanced economies, 19 emerging market countries, and 33 developing countries. Twelve of these countries had fully integrated supervisors at the time of the assessment. They constructed an index of overall BCP compliance and regressed this on an integrated regulator dummy (1 for fully integrated; 0 otherwise). They noted that integrated supervisors are more common in developed economies where overall regulatory environments are also better, which could bias the result. Thus, they added two other variables to the regression to correct for this bias—a variable to measure the quality of the general regulatory environment and per capita GDP. The variables were added one at a time and simultaneously, although they noted that the latter regression had to be interpreted carefully since quality of regulatory environment and GDP per capita are positively correlated. Their results showed that, with just the integrated regulator dummy as explanatory variable, the coefficient was positive and highly significant. That is, fully integrated supervisory agencies tend to have higher quality of banking supervision. However, its impact and level of significance were significantly reduced when overall quality of the regulatory environment was added into the regression, and more so when GDP per capita was added as an explanatory variable. In fact, income level was a more powerful explanatory variable, while the integrated regulator dummy became insignificant.

Indices for compliance based on all three standards were also constructed using a sample of 36 countries which had complete assessments for all three sectors. The sample consisted of 10 industrialized countries, 12 emerging market countries, and 14 developing countries. In particular, compliance in each sector was broken down into the four components of good regulation: (i) regulatory governance; (ii) prudential framework; (iii) regulatory practices; and (iv) financial integrity/safety. Indices were then constructed for each component. Per capita GDP was also included as a control variable. For the indices of BCP compliance, results were similar to the earlier results; that is, income level was a more powerful explanatory variable, while the integrated regulator dummy had the expected positive sign but was insignificant. For the indices of ICP and IOSCO compliance, income level was again highly significant. The integrated dummy variable also had the expected positive sign, but was significant in only two components for each sector—regulatory practices (ICP), prudential framework (IOSCO) and financial integrity/safety net for both.

They also tried to distinguish integrated regulators in central banks from those outside, but they noted that the estimated coefficients were insignificant. One of the supposed benefits of integrated supervision is more consistent regulation across financial sectors. Èihák and Podpiera (2006) tested this by regressing the variation coefficient across the three sectors of each of the four component indices on the integrated dummy variable. They again included per capita GDP as the control variable, which was also highly significant. Their results indicated that overall quality of supervision was more consistent across the sectors being supervised by integrated supervisory agencies. In terms of individual components, only the estimated coefficients for regulatory practices and financial integrity/safety net were significant.

To test the argument that integrated supervision may lower costs, they used available data on the number of supervisory staff. Their results showed that integrating supervision did not lead to substantial supervisory staff reduction. They then tried to explain the number of supervisory staff by country population, area, the level of development (approximated by per capita GDP), and the size of the financial sector (approximated by M2/GDP). They found that population and the country’s level of development mattered. However, the dummy variable for integrated regulators was not statistically significant, although it had the expected negative sign. They attributed the latter result to the following: (i) the time since integration was in most cases not yet sufficient for the cost savings to materialize; (ii) looking just at supervisory staff numbers and not total staff numbers might not have captured some savings in support staff; (iii) the integrated regulator could have taken on new responsibilities that resulted in little change in the number of supervisory staff; and (iv) there were no true synergies among the sectors that would lead to supervisory staff savings.

The authors deserve to be commended for coming up with the first systematic empirical assessment of the impact of integrated financial supervision. They did caution that their regression results should not be interpreted as indicating causality in any way. Thus, Corbett (2007) rightly notes that while their results are suggestive, they cannot be taken as robust since both their data quality and econometric methods are rudimentary, and the specification of the regressions is problematic. She concludes that, "The main impression is that we still know little about the effect of integrated supervisors. This partly explains why there is no consensus view about what system works best" (Corbett, 2007: 24).

With regard to the data, Das and Quintyn (2002) note that the Financial Sector Assessment Program’s assessment of financial sector governance issues is based mainly on qualitative measures of governance prescribed under the various financial sector standards. Also, a key message coming out of the FSAPs is that where regulation is failing, it is due to failure of implementation (Carmichael and Kaufmann, 2001). That is, having best practice prudential and market conduct standards on paper does not mean the agencies involved also have the wherewithal and the willingness to implement them.

Èihák and Podpiera (2006) noted the positive relationship/correlation between the general regulatory environment and the degree of compliance with internationally accepted standards, and between the quality of regulatory environment and per capita GDP. Add to this the overwhelming result of their paper, which is the strong positive relationship between the various measures of compliance and level of income. Compliance with internationally accepted standards can be seen as a subset of the overall regulatory environment, which in turn is a key component of a country’s institutional framework. Thus, their result of a strong correlation between compliance with internationally accepted standards and income level confirms the very basic argument of New Institutional Economics. That is, "Institutions matter for economic performance."

Download this Discussion Paper [ PDF 711.9KB| 68 pages ].

Post a Comment

We welcome your feedback on this publication. Post a comment. ADBI is not obliged to acknowledge or publish comments and may abridge or edit them before web posting.

Comment(s)

There are [0] comment(s) for this entry. Post a comment.