We have seen how developing countries’ financial systems tend to be domin- ated by the banking sector. Clearly, therefore, a well-functioning and efficient banking sector is a prerequisite for successful financial sector development (FSD). While countries may eventually come to be more ‘market-based’ sys- tems, the role of banks will remain of crucial importance, just as is the case with the more market-based developed countries. What are the issues from a developing country perspective, however?
A number of key weaknesses with developing country banking systems can be identified:
1 As we saw in Chapter 4, banking systems in developing countries tend to be relatively highly concentrated – i.e. dominated by a few, large banks.
This uncompetitive situation leads to high costs, in terms of margins and spreads (Knight, 1999).
Figure 5.4 Regional growth rates, 1985–2005.
Source: IMF WEO.
2 This situation is compounded by the often-large degree of state owner- ship, which, in general terms, reduces the efficiency of the system, and by restrictions on foreign banks entering the domestic financial sector (ibid.).
3 This lack of integration with the international financial system may restrict banks’ access to international finance, particularly longer-term financing on favourable terms.
4 Private sector banks are often owned and controlled by family and/or business groups, which engage in ‘connected lending’ – i.e. lending to other parts of the family or business empire – rather than basing decisions on purely commercial grounds. Lending on non-commercial terms is obviously likely to increase the quantity of non-performing-loans (NPLs) in the loan portfolios of the banking sector (World Bank, 2001).
5 Due to the general lack of capital market development, corporate bor- rowers in developing countries are often entirely reliant on the banking system for finance. This leads to a lack of risk-sharing (as provided by equity markets) and a dependence on debt-financing. The resulting highly leveraged borrower base raises the risk of the banking sector’s loan portfolios, increasing the chances of systemic risk in the banking sectors spreading throughout the economy. As Keynes remarked: ‘If you owe your bank a hundred pounds, you have a problem, but if you owe a million, it has.’
6 Higher information asymmetries, coupled with a lack of shareholder control due to weak equity markets, can result in little or no monitoring of banks to prevent excessive risk taking (Mishkin, 1991). There is an extensive literature on why banks may tend to take on excessive amounts of risk. In part this is a moral hazard issue: if the bank’s ‘bet’ comes off it gains; if it does not come off, the depositors lose. This is likely to be exacerbated by the sense that banks in trouble will be rescued by the state due to their importance to the financial system.12 In a more balanced financial system, the behaviour of banks will be monitored and restrained by large, institutional shareholders, for example.
7 Banks may also suffer from poor-quality human resources. The World Bank (1989) identifies the following specific problems that result from this problem:
• Technical mismanagement refers to a lack of technocratic banking skills, which is often a result of previously repressed financial sys- tems, where such skills were less needed. Thus lending may be sub- optimal due to: deficient internal controls, inadequate analysis of creditworthiness, political pressure or pressure from parent groups.
• Cosmetic mismanagement refers to the situation where banks attempt to hide or disguise losses. For example, to avoid alerting shareholders to problems, banks may keep dividend payments high despite losses, thus worsening an already difficult situation. Also various ‘creative accounting’ techniques may be employed.
• Desperate mismanagement refers to a situation where losses are too large to enable cosmetic mismanagement to be possible. In such a situation, bankers may ‘gamble for survival’ by lending to very risky projects at high rates of interest, or speculating on national or international capital or real estate markets.
• Fraud becomes more likely to occur in a context where the pre- vious forms of mismanagement have taken place and failed to rescue the situation. In circumstances like this, where it is clear that the ‘end’ is coming for the bank, senior managers may be tempted to grant themselves loans that they will not have to repay, for example.
8 ‘Credit rationing’ may be more of a problem in banking sectors in devel- oping than in developed markets, though it is clearly an issue in both.
Stiglitz and Weiss (1981) argue that banks cannot easily ascertain which borrowers have high probabilities of default due to problems of asym- metric information. In such an environment, if banks raise interest rates, they may simply encourage borrowers to switch to riskier projects, which if they succeed will allow the interest payments to be met (i.e. moral hazard effects). Furthermore, high interest rates may lead to ‘adverse selection’
where it is precisely those borrowers who do not intend to repay their loans who are willing to accept very high interest rates. Given that it is finan- cially prohibitive to monitor these effects in practice, profit-maximising, rational banks will not raise interest rates to market clearing levels, but instead will ‘ration credit’, when faced with excess demand for loans.
Moreover, some sectors may be entirely excluded on the grounds that monitoring their activities is particularly difficult – such as small farmers, for example.
9 Asymmetric information leading to credit rationing is one example of banks in developing countries (though also more broadly) behaving in ways contrary to what orthodox economic theory would suggest. Simi- larly, Honohan and Stiglitz (2001) argue that many of the modern risk management practices in the banking sector that international regulators use to inform their activities simply do not occur in many developing countries to any meaningful extent.
In this regard, the authors argue strongly against adopting a ‘minimal- ist’ approach to banking regulation and supervision, which is premised on the existence of a largely self-regulating system employing modern risk management practices. Regardless of one’s view as to the effective- ness of such quantitative means of controlling risk in developed coun- tries, it is clearly the case that their absence is likely to lead to banks taking on excessive risk in an environment of ‘minimalist’ regulation:
The conventional retreat into a minimalist regulatory strategy is danger- ously complacent. For example, it neglects just how imperfectly bank
capital is measured and the fact that bank management may have an inventive to increase the measurement difficulties. It over-rates the accur- acy of the risk-adjustments, potentially encouraging banks to increase their assumption of under-priced risk. Finally, it over-emphasizes accounting measures of capital, neglecting the economically relevant aspects of franchise value.
(Honohan and Stiglitz 2001: 1) The combination of these weaknesses in the domestic banking systems in developing countries leads to (a) suboptimal performance by banks in terms of their key functions within the financial system, and (b) a fragile system, prone to banking crises as banks take on excessive risk. These two outcomes are of course strongly related. A fragile banking system is one where many banks may be technically insolvent, which in turn may be a result of their failing to follow best practice in commercial banking.
From the perspective of regulation and supervision of the banking system, the key aim is therefore to ensure the solvency of banks. This focus on solv- ency was not always the core purpose of bank regulation and supervision, however. Before financial liberalisation, for example, the core function of regulators and supervisors was to ensure that banks were acting in accord- ance with government policy and directives regarding the setting of interest rates and the allocation of credit to the appropriate sectors. The transition to the focus on bank solvency has, perhaps not surprisingly, been a difficult process given the very different expertise required on the part of bank supervisors.
Based on formal calculations of insolvency,13 whole banking systems in many developing countries may be insolvent, despite the fact that they are liquid.14 It was concerns on the solvency of banking systems which led to the 1988 Basel Capital Accord.15 The Accord established a minimum ‘capital adequacy ratio’ for banks, which stipulated that banks must hold a particular level of capital, measured as a percentage of risky assets. The figure was set at 8%, i.e. banks were (and are) required to hold 8% of risky assets as
‘regulatory capital’. Therefore 8% is the capital adequacy ratio.
The impact of the Basel Accord has been dramatic. In 1988 many develop- ing countries did not have capital requirements at all, and of those that did, many were well below the 8% level. Furthermore, it was often the case that no prudential supervision was undertaken to verify the figures reported by banks, so that banks had an incentive to ‘massage’ the figures by, for example, reducing levels of provisioning. By 1999, in contrast, only 7 out of 103 coun- tries had capital ratios under 8%, with almost 30 countries having ratios greater than 10%. Interestingly, only one of these countries was an OECD member.
Despite the fact that it was originally devised with banks from developed markets in mind, by the end of the 1990s more than 93% of all countries (88%
in emerging markets) claim to adjust capital ratios for risk in line with the
guidelines set by the Basel Accord. Recent years have seen a fundamental restructuring of the Basel Capital Accord. ‘Basel II’, as it is known, was due to come on stream in 2007/8, and the implications of its introduction for developing economies are discussed in Box 5.3 below.
Box 5.3 Basel II and developing countries
Prior to 1988 banks were largely free to choose how much capital they should set aside to protect themselves from future losses. However, growing international competition in the banking sector saw the amount of risk capital held by banks progressively fall. Concerned about systemic risk to the banking system, regulators decided to act: the Basel Capital Accord of 1988 stipulated that banks should hold a min- imum of 8% of capital, i.e. for every £100 that was lent, £8 should be set aside in case of default. This 8% figure did vary, according to whether the loan was short- or long-term, or was made to an OECD member country or not, but these distinctions were rather crude and, it was argued, led to distortions in the banking system.
The Basel Committee for Banking Supervision (BCBS) has since spent more than a decade developing and refining the long-awaited reform to the 1988 Accord. The headline objective is simple: to align the regulatory capital with the actual risks inherent in lending and other activities. Basel II has evolved through a number of incarnations, and in 2008 is in the process of fine-tuning prior to full-implementation.
Key features
• Countries can choose from a ‘menu’ of options ranging from the
‘standardised approach’ (which is broadly similar to the 1988 Accord with some refinements and extensions to better align capital with risk) to the ‘Internal Ratings Based’ (IRB) approaches.
• The IRB approaches, as the name suggests, enable banks to deter- mine their own regulatory capital requirements, through an assessment of the risk inherent in any particular activity. Although the BCBS currently stipulates the modelling process through which this should be undertaken, the intention is to ultimately move to a situation where the banks use their own internal risk management models.
The table below gives some estimates of the differential impacts of these changes – though these are in terms of orders of magnitudes rather than precise estimates due to the evolving nature of the reforms. As we see, compared to the original 8% borrowers rated BBB or above will see lower capital requirements associated with their loans, while those
below this will see higher requirements. The first impact is therefore likely to be an increase in the cost of borrowing for lower-rated bor- rowers, who are disproportionately located in developing countries. Of course, this is to be welcomed to the extent that it accurately reflects risk. However, there are reasons to suggest this will not be the case, particularly as the Accord does not take into account the benefits of international diversification, which lower the risk of lending to develop- ing countries in the context of an internationally diversified portfolio.
Some other implications are:
• Increased ‘procyclicality’ of lending: as capital requirements vary with a bank’s perception of risk, they will move with the business cycle, encouraging lending in boom times and discouraging it in bad times.
• Countries where large international banks dominate their banking system (as in many Eastern European and Baltic states) may see the supply of credit to lower (or unrated) borrowers decline and/or become more expensive, as banks using the IRB approach reassess the regulatory capital implications of lending.
Lesson: the BCBS has no formal representatives from developing countries, yet Basel II will certainly become a global standard; if devel- oping countries are expected to implement international standards they should (a) be involved in their formulation, and (b) have this formula- tion take account of their own circumstances, and not just those in developed countries.
Whilst it is relatively straightforward to set headline figures for things such as capital adequacy ratios, as pointed out above, this in itself does not ensure that the banking sector is solvent in reality. Indeed, as Liliana Rojas-Suarez (2001) demonstrates, levels of regulatory capital held by banks in developing
Estimated changes in capital requirements under Basel II Borrower’s
credit rating
Probability of default (PD)
1988 Accord Standard approach
IRB foundation
AAA 0.03 8 1.6 1.13
AA 0.03 8 1.6 1.13
A 0.03 8 4.0 1.13
BBB 0.20 8 8.0 3.61
BB 1.40 8 8.0 12.35
B 6.60 8 12.0 30.96
CCC 15.00 8 12.0 47.04
Source: Bank of England’s Spring Quarterly Bulletin, 2001.
countries have had no relationship to actual riskiness. That is to say, the numerous banking crises that have occurred in developing countries since the advent of the Basel Accords could not have been predicted by observing holdings of regulatory capital, which also seem to have had no effect in restraining excessive risk-taking behaviour. As with the Honohan and Stiglitz (2001) argument discussed above, where international regulatory forms devised to be suitable for developed economies are simply transferred whole- sale to the very different financial systems in developing countries, the results are likely to be very different.
If banks cannot ‘police themselves’ the onus on strong regulation and supervision becomes all the stronger. However, as illustrated in Figure 5.5, there is considerable evidence that, in practice, supervision is less stringent in lower income countries. However, this should be contrasted with the fact that, in many instances as seen in Chapter 4, headline regulations are at least as stringent in developed countries as is the case in developed ones.
To determine the effectiveness of these regulations Barth et al. (2004) examine regulatory and supervisory practice in 107 countries to assess the relationship between specific practices and banking sector development and soundness. They try to answer seven specific questions:
1 Should banks be prevented from providing other financial services?
From an economic perspective, many have argued that banks should be pre- vented from doing so. In part this reflects their special status as deposit- taking institutions, and the systemic risks of bank failures, but is also the result of concerns about the level of power and influence that such integrated financial giants could yield.
The authors find that restricting banks in this regard has a negative effect on both bank stability and banking sector development.
Figure 5.5 Comparative stringency of banking supervision.
Source: Barth, Caprio and Levine database.
2 Should entry into the banking system be controlled by barriers?
Again, it has been argued that the importance of the banking sector to the economy is such that the most vital task is to prevent weak or failing banks undermining the integrity of the financial system.
However, Barth et al. find no evidence that such restrictions are positively related to banking stability. Instead, they find some evidence that sug- gests that restricting the entry of foreign banks reduces stability. This is sup- portive of the view that restricting entry – and therefore competition – is likely to encourage corruption and a general decline in the efficiency of the system.
3 Do regulatory capital requirements enhance the stability of banks?
In this regard, the authors concur with Liliana Rojas-Suarez (op. cit.) and find little in the way of clear relationships in either direction. This leads them to caution on the widely held view that regulatory capital requirements are effective in preventing banks taking on excessive risks.
4 Do deposit insurance schemes enhance the stability of banks?
Here, quite the opposite is found. Deposit insurance, particularly generous schemes, are negatively associated with bank stability, which supports the view that strong regulation may not be sufficient to offset the moral hazard- producing effects of deposit insurance. More generally, the World Bank (2001) points out that to be successful, deposit insurance schemes require a minimum level of infrastructure, and should therefore not be attempted in the least developed countries: this would be likely to lead to excessive risk taking and increased fragility of the banking sector.
5 What is the relationship between different specific powers of regulators and supervisors and bank performance and stability?
Interestingly, very little evidence of any statistically significant kind emerged in this respect: ‘Thus, measures of supervisory power, resources, independ- ence, loan classification stringency, provisioning stringency, and others are not robustly associated with bank development, performance or stability’
(Barth et al., 2002: 245).
6 Does regulation that encourages private sector monitoring of banks enhance performance and stability?
Here, Barth et al. give an unequivocal yes. Furthermore, they suggest that regulation of this sort may be important in preventing financial crises from developing. Examples of private monitoring include:
• Treatment of ‘insiders’: owners, directors and senior management work best with a sensible incentive structure. For owners, incentives should be based on franchise value and the threat of removal of the banking licence;
for the board, penalties should exist for failing to disclose information;
for managers, compensation should be performance-related, with no rewards for failure.
• Ratings agencies have the potential to play a key monitoring role of the behaviour of banks. However, it is not clear that their record stands up to scrutiny, as they have historically been very poor at predicting systemic failures.
• Markets such as creditors in corporate bond markets can send clear sig- nals on bank performance through the trading of the bank’s corporate debt in the secondary debt market.
7 Does government ownership of banks enhance performance and stability?
Here, the answer is an unequivocal no: ‘There is no evidence, even in weak institutional settings, that government-owned banks are associated with positive outcomes’ (Barth et al., 2004: 245).
For the authors of the study, these findings suggest a number of key con- clusions. First, regulation that forces the disclosure of information and encourages effective private sector monitoring of banks activities is likely to be most effective in producing both bank stability and development. And, second: ‘. . . these findings raise a cautionary flag regarding reform strategies that place excessive reliance on countries adhering to an extensive checklist of regulations and supervisory practices that involve direct, government over- sight of and restrictions on banks.’
In short, the findings of this study suggest that ‘market discipline’ is by far and away the most effective means of supervising the activities of the banking sector, and many of the myriad banking regulations that have been introduced in developing countries in the past decade have had little or no positive effect on the robustness or development of the banking system.
However, this ‘advice’ should also be treated with some caution. As both Honohan and Stiglitz (op. cit.) and Liliana Rojas-Suarez (op. cit.) point out, many of the strongest pillars of market discipline – such as a corporate bond market where the price of banks’ debt in the secondary market is both a barometer of their riskiness and an incentive to lower risk so as to improve borrowing terms – may simply not exist. Similarly, international ratings agen- cies have shown little interest in rating financial institutions in developing countries, particularly the poorer ones. In such circumstances, effective, direct supervision – or ‘robust restraint’ of the financial system – may be the best available alternative. Relying on market discipline where there is no such discipline to restrain the risk-taking impulses of banks would appear to be a