RDP 9213: The Impact of Financial Intermediaries on Resource Allocation and Economic Growth 3. The Role of the Financial System in Knowledge Accumulation and Growth

The depth of a country's financial markets depends on a variety of factors. These include the nature and extent of controls on financial intermediation, the nation's legal and accounting rules and the level of per-capita income. As many economists have noted, there is a positive relationship between the level of per-capita income and the extent of financial intermediation. The causation runs two ways; wealthier countries can afford more sophisticated financial systems and the depth of the financial system helps an economy to grow. This section discusses various aspects of this two-way relationship.

The following five issues are discussed in some detail:

1. the effect of interest rate ceilings on resource allocation.
2. the role of financial intermediaries in screening loan proposals.
3. the role of the financial system in providing liquidity insurance.
4. the implications of differences in private and social returns.
5. the risks of financial liberalisation.

In each case, the discussion focuses on how the financial system affects resource allocation. Given the lessons from the Romer model, particular attention is given to the interaction between the financial system and the accumulation of knowledge. The discussion is heavily biased towards domestic financial markets and away from international financial markets. This is not to say that access to world capital markets is not important for resource accumulation and economic growth, for indeed it is. At a very basic level, access to international capital markets permits national investment to exceed national savings and thus it allows the achievement of a given level of national income earlier than could have been achieved through reliance solely on domestic savings. Integration into world capital markets can also make it more attractive for foreign capital to undertake investment domestically. This foreign capital often brings access to technology that previously was unavailable to the country. Thus, the extent to which a country has access to both world financial and goods markets has a potentially large impact on economic growth. To keep what is already a wide range of issues from increasing further, many of these important international issues are pushed into the background in the following discussion. The focus is primarily on the behaviour of domestic financial institutions.

3.1 Interest Rate Ceilings

The work of McKinnon (1973) and Shaw (1973) identified interest rate controls as one of the most important causes of financial repression. Since then, there has developed a voluminous literature on the effects of these controls. This literature has largely focused on the implications for national savings of ceilings on loan and deposit rates. It has been argued that these controls reduce the incentive to save. As a result, national savings are lower and, without access to world capital markets, actual investment is lower. Whether or not lower interest rates actually lead to reduced savings and capital formation has been the topic of much empirical research. Lee (1991) provides a useful summary of this work. He concludes that while the evidence on the existence of a positive relationship between interest rates and savings is mixed, there is probably a quite weak but positive relationship. Perhaps more importantly, he concludes that “there is overwhelming evidence supporting the positive effects of real interest rates on the volume of financial savings” (page 8). A similar conclusion is reached by the World Bank in the 1989 World Development Report.

It is possible that repression of the banking sector does not have adverse implications for resource allocation if other forms of finance can substitute for bank loans. Typically, controls on interest rates lead to the establishment of some type of unofficial or ‘kerb-market’ that is able to supply intermediated credit to borrowers. In many cases, this market acts on the fringe of the official sector and because of the risks involved, it has to incorporate a large risk premium into its borrowing and lending rates. As a result, access to this market is often limited. Equity finance is also often not an option. The information problems involved in the issue of equity and the inability of equity investors to obtain widely diversified portfolios mean that stock markets are typically underdeveloped in financially repressed economies.

Given these constraints on other forms of finance, interest rate controls do have important implications for allocation of savings. Provided that the interest rate ceilings are binding, the financial intermediaries must ration credit. This need to ration has important consequences for the type of projects that get financed and thus on the growth rate of the economy.

When banks are required to ration credit they choose which projects receive the limited supply of available finance. Shaw (1973) argues that the rationing process is expensive to administer, is vulnerable to corruption and intensifies the risk aversion and liquidity preference of financial intermediaries. As a result, the banks' loan allocation process is unlikely to be socially optimal. There will be a bias against projects with high expected returns but with relatively high risk. While the expected return of a project is important from a social perspective, it is irrelevant to the bank provided that the bank is assured of receiving its principal and interest. Banks will thus favour projects with low risk.

The following example clarifies this point. Suppose there are two borrowers that require one unit of funding at the maximum interest rate that can be charged by the bank. The interest rate ceiling, however, means that the bank can only raise sufficient deposits to fund one of the loans. Further, assume that the first project has an expected rate of return of R₁, that , and that there is no uncertainty concerning the project's return. The second project is assumed to have a higher expected rate of return than the first project, however it is risky. Suppose that its expected rate of return is given by R₂ > R₁ and that there is a 50 per cent chance that the project will yield R₂+α and a 50 per cent chance that it will return R₂−α. Lastly, assume that neither borrower has any collateral, private returns equal social returns and that bank is risk neutral.

If the bank lends to the first project, its rate of return on the loan will be . If the bank lends to the second project, its expected rate of return depends upon the size of α. If α is large enough to make then there is a 50 per cent chance that the borrower will be unable to pay the bank the complete principal and interest. In this case, the bank's expected rate of return on this second project is less than . Because of the interest rate control, the bank is unable to charge the higher return-higher risk borrower an interest rate sufficiently high to compensate for the risk. The lower risk-lower return borrower thus obtains the rationed credit.

In the above example, the problem could be mitigated by the second borrower having collateral to back the loan. This, however, just pushes the problem one step back. Banks extend the rationed credit to those with the greatest amount of collateral and not to those whose projects have the highest expected rates of return.

Clearly, the way in which the bank allocates the country's savings is sub-optimal in the above example. This has implications not only for the level of national income but also for the growth rate of income. Research and development are typically risky activities. Sometimes they will lead to a break-through, other times they will yield nothing. These activities also involve up-front costs that are only later recouped. Thus, they must be financed. If the activities that lead to knowledge accumulation are riskier than other types of activities then the need to ration credit is likely to result in under-investment in these relatively risky activities.

In the Romer model presented above, the growth rate is slower than is socially optimal since inventors do not take into account the positive externality that their research generates. When interest rate ceilings exist (and banks can effectively screen potential borrowers) the problem is compounded because the financial intermediaries ignore the social returns of the projects for which they lend. If banks cannot charge different borrowers different interest rates then risky, but high expected return activities will be under-funded by financial intermediaries.

3.2 Ability to Screen

In the above discussion it was assumed that banks knew both the expected return and the distribution of the returns for each project and that the bank could distinguish between borrowers. It was argued that in this case interest rate ceilings were likely to lead to the financing of safer projects with lower expected returns. The assumption of perfect information is a strong one. In fact, much recent research concerning financial markets has focused on the asymmetric information between the borrower and the lender. Potential borrowers are assumed to know much more about their project than anybody else. This asymmetry in information is at the core of one of the most important reasons for the existence of financial intermediaries. That is, financial intermediaries exist because of their ability to cost-effectively screen proposed investment projects, to assess the collateral of various loan proposals and to monitor the performance of projects for which funds have been lent^[2]. The efficiency with which the banks perform this screening and monitoring function has important implications for the efficiency with which the country's savings are allocated.

Perhaps the most influential work on the effects of asymmetric information on credit markets is the work of Stiglitz and Weiss (1981). They show that if banks are not able to screen projects then increasing the rate of interest that borrowers are required to pay may actually reduce the return to the bank. This result reflects the fact that as the interest rate increases the quality of the pool of customers seeking loans from the bank deteriorates. This occurs through investors with safe projects deciding not to borrow (adverse selection) or through each investor choosing a more risky project (moral hazard).

In the Stiglitz and Weiss model, all projects have the same expected return but projects vary in terms of their riskiness. The more risky the project, the lower is the sensitivity of the borrower's return to the interest rate charged. If the borrower has limited liability then for outcomes in which the project is a failure, the interest rate charged is irrelevant because no interest is paid. Since the loan is only paid off when the project is a success, it is the return in the successful state that is important for the borrower when considering whether or not to borrow. Since risky projects have higher probabilities of failure, they must have better outcomes in the good states. As a result, borrowers with risky projects are prepared to pay higher interest rates.

The bank recognises this problem and may be unwilling to increase the loan interest rate to clear the market for fear of attracting risky projects. As the result of setting a non-market clearing rate the bank must ration credit. Since it cannot distinguish between projects, the allocation of the rationed funds is random. This situation is commonly referred to as equilibrium credit rationing. In contrast, Section 3.1 was concerned with disequilibrium credit rationing.

This discussion has assumed that all projects have the same expected value. While this is a useful assumption for highlighting the effects of asymmetric information on bank lending, it is not realistic. An economy has a range of potential projects that can be undertaken. Some projects have high expected rates of return, others have low or negative expected rates of return. The ability of banks to screen the quality of projects plays a role in determining the return received by the country on its savings.

Consider the following example. There are two types of projects that exist in equal quantities and each requires 100 units of financing. One project is a “good” project; it has a high expected return and a low variance. The second project is a “bad” project; it has a low expected return and a high variance. For simplicity, assume that the borrowers undertaking both types of projects have no collateral to offer. There are two possible outcomes for each project. In the unfavourable outcome the good project returns 80 and the bad project returns 67. In this case neither type of borrower is able to repay the full loan. For the borrower with the good project there is a 10 per cent chance that the unfavourable outcome will occur. For the borrower with the bad project this probability is 40 per cent. In the favourable outcome the return to the good project is 123 and the return to the bad project is 130. Finally, it is assumed that private returns equal social returns. The following table summarises the possible outcomes.

Since borrowers only pay interest in the good state, they are prepared to borrow at interest rates that exceed the social return on their projects. Borrowers with good projects are prepared to borrow at rates up to 23 per cent while borrowers with bad projects are prepared to pay up to 30 per cent. The social rates of return on these projects are 18.7 per cent and 4.8 per cent respectively.

Suppose that the bank cannot distinguish between the two projects but knows that the two projects exist in equal quantities and knows the distribution of returns for each project. Further, initially assume that interest rates are such that both types of borrowers apply for loans. Thus, there is a 50 per cent chance that a project is a good type and a 50 per cent chance that it is a bad type. If the loan rate is denoted by R_L then the expected return to the bank from a loan is:

If the bank is risk neutral, the loan rate must be set so that the expected return to the bank equals the bank's cost of providing the funds (R_c). This implies that:

Thus, if the cost of providing loans is 5 per cent then the loan rate will equal 16.8 per cent. At this rate, both types of projects apply for and receive finance. This is despite the fact that the social return of the bad project (4.8 per cent) is less than the social cost of funds (5 per cent). Now suppose that the cost of providing a loan increases to 10 per cent. According to the above formula the loan rate should increase to 23.47 per cent. However, at this interest rate those with the good projects no longer find it profitable to borrow. Since the bank knows the distribution of returns, it would know that at the high deposit rate only those with poor projects would be applying for loans. Thus, at a deposit rate of 10 per cent the bank would need to charge a loan rate of 38.66 per cent. At this rate, those with bad projects would no longer wish to borrow. As a result, if the cost of providing a loan is 10 per cent there will be no financial intermediation^[3].

This example highlights two important points. First, it highlights a critical problem for countries with under-developed financial systems. If the financial intermediaries lack the skill needed to assess the quality of projects, it may be difficult for an intermediation industry to be profitable unless interest rates are controlled. The inability of banks to screen, leads to those with good projects subsidising those with bad projects. The higher the cost of providing loans, the greater is this subsidy. As interest rates rise it becomes unprofitable for those with good loans to pay the subsidy. Once the subsidy is not being paid, those with bad projects may be unwilling to borrow at loan rates that are needed to ensure bank profitability. All financial intermediation may cease.

The cost of providing a loan may be high for a variety of reasons. Intermediaries may have a shortage of capital and may be seen as relatively risky. Thus, investors will demand a risk premium before they are willing to deposit funds with the intermediary. This risk premium increases the cost of providing funds to borrowers. Second, the intermediary may be operated inefficiently. Poor internal controls and procedures and high operating costs may mean that while deposit rates are low, loan rates may be quite high. Third, if there is reasonably free access to international capital markets it is difficult to sustain risk-adjusted interest rates that are significantly different from world interest rates. High world interest rates are thus likely to lead to high domestic deposit interest rates. If, for any of these reasons, the cost of providing a loan is high, a strong financial intermediation industry may find it difficult to be profitable unless there are some interest rate controls.

The second general point made by the above example is that the inability of banks to screen causes savings to be allocated poorly. With a moderate cost of providing loans, the financial intermediary provides funds to both types of projects. That is, it finances projects with both positive and negative net social value. It is even possible that projects that have a negative expected gross rate of return get financed. In terms of the growth model presented in Section 2, these distortions can be thought of as reducing the efficiency with which the economy translates its existing resources into new knowledge. That is, they slow the rate at which the economy accumulates the key resources needed for economic growth.

The ability of banks to screen loan applicants is central to the efficacy of the financial system. An inability to screen can result in either the failure of financial intermediaries to develop or, if they do develop, in an inefficient allocation of savings. The ability to screen has three essential components. First, the intermediary must be able to assess the quality of the project, second it must be able to assess the value of collateral offered as security and third it must have the ability to monitor the project through time.

For intermediaries to be able to perform these tasks effectively the nation must have a strong legal system. For intermediaries to be able to assess the value of a firm's collateral, the system of property rights must be clearly defined and widely accepted. If there is uncertainty regarding the current ownership of the particular assets offered for collateral or if there is uncertainty about future changes in property rights then banks will have difficulty is assessing the security behind their loans. Given that, by its very nature, collateral is only used in the “bad case” scenarios, banks must value the collateral at its value in those bad cases. Uncertainty about property rights therefore prejudices the value of collateral.

A second aspect of the legal system that is important for financial intermediaries to effectively perform their tasks concerns the laws regarding truthful reporting and accounting standards. If widely accepted accounting standards do not exist, it becomes difficult to asses the current position of the firm and the likely outcome of the project. Banks become uncertain as to whether a firm's financial statements reflect the true position of the firm or just “creative accounting”. This makes the screening task more difficult. The task is also made more difficult if the country does not have strong penalties for deliberate misreporting by a firm of its current financial position and its likely future developments. If such laws do not exist then the intermediary needs to screen projects, not just on the basis of expected outcomes but also on the probability that the borrower is lying. This makes the screening process more costly and in all probability, less effective.

Given that a widely accepted and enforced system of property rights and accounting rules has been established, the structure of the banking system can play an important role in minimising the costs arising from asymmetric information. In particular, the adoption of a universal, as opposed to an arms-length, banking system is likely to reduce the costs associated with screening problems. The universal system involves the providers of finance having an active role in the internal management of the firm. For instance, a representative of the bank may sit on the board of the firm and the bank may own shares in the firm. This allows the bank to actively monitor the performance of the firm's managers and their investment decisions. It also serves as a signal to other investors that the firm is being soundly managed.

The German and Japanese banking systems are examples of the universal system while the Australian and US systems are examples of the arms-length system. However, DeLong (1991) argues that prior to World War I the US banking system was more like the universal system. He suggests that this was partly responsible for the rapid growth in the United States. Financial institutions, by having their representatives on the boards of firms were able to make managers more accountable and ensure that long-term, but high-return investments were undertaken.

While the universal system should help minimise distortions arising from asymmetric information and conflicts between the incentives of managers and owners it is not without its costs. Amongst these is the potential for intermediaries to use their market power over firms with which they are associated to extract monopoly returns. The system may also be more fragile to large shocks. If banks hold significant share holdings in firms and the share market declines, the capital of the banks is reduced. This may make the banks less willing to lend. Notwithstanding these problems, the universal system does offer significant gains in terms of monitoring, provided that the financial institutions have the capability and knowledge to screen projects once the asymmetry in information is removed.

3.3 Risk Sharing and Liquidity Insurance

Financial intermediaries are able to improve the allocation of a nation's savings, not only through their ability to screen projects, but also through their ability to aggregate idiosyncratic risk and thus reduce the value of liquid balances that individuals wish to hold.

Investment is a risky activity. It typically leads to a stochastic flow of revenue. While a particular investment may have positive net present value, the stochastic nature of the returns means that there may be periods when revenue unexpectedly fails to cover costs. If there is no external source of finance and the investor does not hold sufficient liquid balances, the project would have to be liquidated in such a period. This liquidation would have to occur even though the project still had positive net present value. As a result of this risk, individual investors are required to hold savings in liquid balances as a form of liquidity insurance. The size of these liquid balances is reduced if financial intermediaries exist.

The above argument has been formally modelled by Bencivenga and Smith (1991). They show that the introduction of a financial intermediation industry permits the economy to reduce the fraction of its savings that it holds in unproductive liquid assets. Banks borrow from, and lend to, a large number of individuals and as a result face a fairly predictable withdrawal pattern. Thus, they can economise on liquid asset holdings and are able to provide temporary finance to projects that have a short-term liquidity problem. An economy with financial intermediaries is able to economise on liquidity insurance and therefore is able to devote more of its savings to higher return illiquid assets. This, in turn, increases the steady-state growth rate of the economy.

In the Bencivenga and Smith (1991) model, the financial intermediaries either exist or they do not exist. There is no concept of them operating more or less efficiently. However, the more effective are the financial intermediaries in insuring against liquidity risk, the greater is the share of savings that can be devoted to illiquid high yield assets. Obviously, the greater is the stability and confidence in the banking sector, the lower is the need to hold liquid balances. Similarly, liquid balances can be reduced if the macroeconomy is sound and there are instruments available that allow financial intermediaries to minimise risks.

To date, little attention has been given to the stock market. However, just as financial intermediaries reduce the need for the holding of liquid balances so does the existence of a stock market. In a recent paper developing the implications of the existence of equity markets for economic growth, Levine (1991) argues that a stock market will increase the steady-state growth rate by reducing the need to hold liquid balances and by facilitating the accumulation of human capital. He argues that the amount of human capital is related to the amount of physical capital. If idiosyncratic liquidity shocks cause an entrepreneur to cease or scale back production, some human capital is lost. The stock market allows the entrepreneur to sell his/her stock to cover a liquidity problem and thus allows the firm to continue without the loss of physical and human capital. As a result, human capital accumulates more quickly and the steady-state growth rate is faster.

A similar idea is developed by Greenwald and Stiglitz (1989). They argue that, in some cases, the asymmetric information between the owner/managers of a firm and outside potential investors are so great that a stock market will not develop. Individuals are not able to diversify away the idiosyncratic risk of their particular project. This leads firms to under-invest as a form of insurance. The smaller scale of operations means that human capital accumulation through on-the-job training is lower, as are productivity gains achieved through learning-by-doing. Similarly, investment in research and development will be lower. As a result, the financial market distortions that prevent socially optimal risk sharing from taking place lead to lower productivity improvements and a slower steady-state growth rate.

The models of Bencivenga and Smith (1991), Levine (1991) and Greenwald and Stiglitz (1989) take the economy's financial structure as exogenously given. In contrast, Greenwood and Jovanovic (1990) derive the degree of financial intermediation endogenously. They argue that financial intermediation is expensive and that low income economies have difficulty in paying the costs involved. With few intermediaries, project selection is poor and as a consequence growth is slow. As income gradually increases, the economy can afford more financial intermediation. As a result, the quality of project selection improves and growth accelerates. In maturity, the economy has a fully developed financial sector and grows faster than it did when income was lower. In this model the degree of financial intermediation both causes growth and is a function of growth. While the paper does not address the question of appropriate government policy, it is likely that policies aimed towards reducing the cost of financial intermediation would allow the achievement of higher growth rates and possibility higher utility.

3.4 The Implications of Differences in Private and Social Returns

In the growth model presented above, the social return to the development of a new blue-print exceeds the private return. As a result, the rate of development of blueprints is slower than is socially optimal. Thus, economic growth is also slower than optimal. In the model, subsidising the production of blueprints will speed both the rate at which they are developed and the rate at which the economy grows.

A difference between social and private returns characterises many activities in the real world. A popular example is the accumulation of human capital (or education). It is often argued that individuals are not able to capture for themselves, the full value to society of their education. Since education is an activity that typically involves incurring a cost before the benefits are obtained, finance is often required. Even if financial institutions worked perfectly and financed all “projects” with positive net present value, there would be insufficient resources devoted to education as individuals would not have the incentive to obtain the socially optimal amount of education. One policy response to this problem has been for the government to subsidise education either directly or through the financial system. If the financial system is used, governments can either subsidise education loans made by commercial banks or the government can take over the role of the financial intermediary completely and provide direct finance.

Some governments have used this distinction between private and social returns to justify widespread intervention in financial markets. Certain industries or sectors are seen as providing dynamic social gains that exceed the private gains. In some cases, this has led to financial markets and financial policy becoming an important tool of industrial policy.

There is no doubt that there is a strong theoretical argument for intervention when social and private returns diverge. The argument is particularly strong if the distortion not only affects the current level of income but also affects the growth rate of income. In addition, even when the private returns equal the social returns, there may be a role for government if the incentives of financial institutions do not permit projects to be financed which have high social returns. As the above discussion suggests, these problems may be particularly severe in developing countries.

In practice, however, government intervention introduces its own incentive problems. The allocation of credit may be made, not on the basis of economic considerations, but rather on political and social grounds. Further, it is often difficult for government officials to accurately assess the relative social and private returns of various projects. The assessment process can easily become derailed by non-economic considerations. If the allocation of credit is completely controlled by government then the discipline of the market may be lost and financial intermediaries may have difficulty in developing the skills needed for good credit assessment. The development of these skills is an integral part of the development process, for it is difficult for the government to continue controlling the allocation of financial resources as the level of national income increases. Notwithstanding these difficulties, there does exist an important role for the government in the credit allocation process, particularly in the early stages of development. The challenge for government is to ensure that this role is not abused^[4].

3.5 Costs and Risks of Financial Market Liberalisation

There is a general presumption that financial liberalisation, through its ability to deepen a country's financial markets, leads to faster growth. Nevertheless, liberalisation of financial markets can have a number of adverse effects on the macroeconomy. Blundell-Wignall and Browne (1991) provide a useful summary of these effects.

The costs and risks associated with liberalisation can be broadly categorised into three related areas. First, the removal of liquidity constraints may make monetary policy more difficult to implement and the resulting higher levels of debt may complicate the response of the economy to various types of shocks. Second, in liberalised financial markets, asset prices are often volatile and may become misaligned, leading to misallocation of resources. Third, if intermediaries use a poor screening technology then liberalisation may allow more poor projects to be financed. The implications of poor project selection were discussed in Section 3.2. The following discussion centres on the other two potential costs.

Domestic financial liberalisation removes the liquidity constraints faced by individuals. As a result, individual consumption becomes less sensitive to current income. This is also true for countries as a whole as the opening of the economy to international markets breaks the nexus between national savings and investment. While in general, the removal of the liquidity constraints should increase individual and national welfare, there are a number of risks and potential problems. First, the operation of monetary policy may become more complex. No longer can the monetary authorities run monetary policy through direct controls on interest rates and credit. Instead, monetary policy is forced to operate directly through market mechanisms; that is by changing financial prices to induce wealth effects and intertemporal substitution. The responses of the economy to these changes in financial prices are often slow and variable making the monetary policy problem more complex than in a highly and effectively regulated system. On the other hand, if the regulations are being circumscribed by other financial institutions then a price based monetary policy may make monetary policy more effective than direct controls over a limited range of financial intermediaries.

A second potential risk relating to the removal of liquidity constraints concerns the degree of fragility of the corporate sector to various types of shocks. When debt levels are high, adverse shocks to the economy may create more severe recessions and recessions with greater persistence^[5]. Higher debt levels imply lower firm collateral and thus higher agency costs. These costs are amplified if firms enter a recession with highly geared balance sheets. As a consequence, firms may find it difficult to obtain finance in recessions, even for projects with positive net present value. In addition, the managers of a highly leveraged firm that suffers an adverse shock, may be unwilling to take risky but profitable investment decisions. If the managers have significant firm-specific capital they may be unwilling to entertain the increased probability of bankruptcy that the risky investment entails.

There are also potential problems on the international front from the increase in debt made possible by financial liberalisation. McKinnon (1986) argues that many developing countries over-borrowed in the liberalised markets of the 1970s. This over-borrowing was made possible by governments guaranteeing repayment of international borrowing. These guarantees removed the incentives for lenders to properly screen projects. As a result, many projects were undertaken which probably should not have been undertaken. The resulting debt-overhang has necessitated tight macro-economic adjustment programs to have been put in place in a number of countries.

The second class of risk arising in liberalised financial markets centres on the risks of resource misallocation arising from the behaviour of asset prices. In liberalised markets, many key financial prices are set in auction markets. Theory suggests that these prices are determined by fundamentals and react quickly to news concerning changes in these fundamentals. For example, share market prices should be set on the basis of the present discounted value of expected future dividends. An increase in expected future dividends should lead to an increase in share prices. Because asset prices respond to expectations they are often volatile. In addition, asset prices may become misaligned and deviate from fundamentals for long periods of time. Blundell-Wignall and Browne (1991) discuss the costs of this volatility and misalignment. They argue that the uncertainty generated by volatile asset markets may contribute to a shortening of investment horizons. This may lead to projects being financed which generate returns quickly but whose returns are lower than those of more longer-term projects. This could be a particular problem for research and development projects, whose returns are often obtained only after a long gestation period. Misalignment of asset prices, particularly exchange rates, also generates considerable concern. If exchange rates move away from fundamentals for long periods of time, the allocation of resources between the traded and non-traded sectors becomes distorted. For example, an overvalued exchange rate reduces the competitiveness of a country's traded good sector. If gaining access to foreign markets is an important part of the process of leaming-by-doing, or in achieving economies of scale, this overvaluation can have extremely adverse consequences for growth. Similar problems can occur in other asset markets, particularly the stock market and the housing market.

Footnotes

Diamond (1984) develops a theory of financial intermediation based on intermediaries minimising the cost of monitoring. [2]

In this example there will be no financial intermediation if the cost of providing funds exceeds 9.65 per cent. By slightly altering the pay-offs it is possible to have some range of interest rates after which good borrowers drop out of the market, that bad borrowers are still both prepared and able to, obtain finance. [3]

Collier and Mayer (1989) discuss various aspects of the role of government in the financial system. [4]

Bernanke and Gertler (1990) provide a formal model of this effect. [5]