While a domestic institutional investor sector is a fundamental component of well-developed financial markets, the presence of one can be a mixed blessing in some emerging markets, in which governments are often financially strained and look for “captive” investors with whom to place their debt.[1] In particular, governments can be tempted to obtain financing from institutional investors through moral suasion or by twisting regulations in situations in which market access by normal means becomes scarce. Scarcity of financing in a country may arise from concerns about the soundness of its public finances. If those concerns are well founded, the government’s use of institutional investors to obtain financing will allow the country’s debt to grow, and the eventual debt crisis to become more severe, while the losses that institutional investors eventually suffer may compromise the whole financial system. Conversely, a financing shortage in a country may arise from disruptions in the country’s financial markets, with little justification in terms of economic fundamentals. This may be the result of poorly informed investors reacting as a “herd” and magnifying a small financial disturbance. In such a case, local institutional investors with better information and longer investment horizons can help increase stability in the market.
A group of large, well-managed institutional investors is the anchor of many advanced domestic capital markets. Complex problems generally arise when a country’s domestic capital market is still relatively small, and when governments have large debts and are subject to frequent liquidity shortages.
THE GROWTH OF INSTITUTIONAL INVESTORS
Institutional investors are an important source of financing for central governments. In 2000, pension funds, insurance companies, and mutual funds held about one-quarter of total central government debt in emerging markets. By 2005, the share of government debt held by these institutional investors had grown to almost one-third of total central government debt (IMF, 2006b).
Although institutional investors are critically important for the functioning of a country’s domestic government debt market, they are not a homogenous group with similar investment objectives. On the contrary, different types of institutional investors follow their own investment guidelines, and as a result the demand for government bonds ranges from short-term treasury bills to long-term instruments. Pension funds and life insurance companies have a predictable funding flow and fairly predictable liabilities for long periods of time. As a consequence, they have a long-term planning horizon and look for assets that generate a stable flow of real income. By contrast, mutual funds and investment companies focus on the current market value of their portfolios, which is their main indicator of performance. Moreover, as they could face redemptions from shareholders at almost any time and are required to mark all their assets to market, they pay close attention to the liquidity of the financial instruments in which they invest.
Banks are different from standard institutional investors because of the nature of their liabilities. As they have short-term deposits that are fixed in nominal terms and are redeemable on demand, banks differ from mutual funds (which also have short-term liabilities but whose value fluctuates with the market value of their assets) and from insurance companies and pension funds, which face long-term liabilities.
The growth in the assets held by institutional investors has been remarkable in all segments of the global economy (Figure 8.1). In the advanced economies, the assets of pension funds and mutual funds increased from approximately 80 percent of GDP in 1997 to 112 percent of GDP in 2003.[2] Institutional investors are less important in emerging markets, but the growth of their assets in these markets has been very rapid as well, from 18 to 30 percent of GDP over the 1997–2003 period. In the mid-1990s, Latin American institutional investors held assets equal to approximately 10 percent of regional GDP, and hence their assets accounted for a much smaller share of GDP than those of average institutional investors in the emerging markets. Over the 1997–2003 period, the size of the assets held by Latin American institutional investors grew faster than that of institutional investors located in other emerging markets and, by 2003, the aggregate assets of Latin American institutional investors were almost identical in size to those of institutional investors in the emerging markets. This rapid growth in the asset size of Latin American institutional investors was mainly due to the creation of private pension funds that took place in many Latin American countries in the mid-1990s.
In the advanced economies, pension funds and insurance companies have traditionally been the largest institutional investors, although the amounts invested by investment companies (essentially, a variety of mutual funds) have recently been growing at a faster pace. The relative importance (in terms of total assets) of the different types of institutional investors varies from country to country. Insurance companies are relatively more important in the United Kingdom and in Japan, and pension funds in the United Kingdom, while investment companies prevail in the United States.
Within Latin America, the countries with the largest presence of institutional investors are Chile and Brazil (Table 8.1). Chile was the first country in the region to privatize its pension system, and the assets of its institutional investors now amount to 88 percent of GDP (with pension funds managing assets equivalent to 60 percent of GDP). In Chile, insurance companies grew together with pension funds, primarily because they provide both retirement income and life insurance to pension fund contributors. Mutual funds, though much smaller, still hold almost 9 percent of the Chilean GDP in assets under management. In Argentina, pension funds are the largest group of institutional investors, with assets amounting to 12 percent of GDP, while insurance companies manage assets equivalent to 5 percent of GDP, and mutual funds hold assets representing only 1 percent of GDP. Brazil is a unique case, as mutual funds are the largest institutional investors in that country, and their assets represent almost 30 percent of GDP.[3] Asset holdings of mutual funds are also substantial in Colombia, where they are equivalent to 23 percent of GDP, more than twice the amount managed by pension funds.
PENSION FUNDS
The development of the pension fund industry in Latin America is relatively recent. In most Latin American countries, the industry started to grow in the mid-1990s as a result of the creation of private pension fund management companies when these countries started to move from pay-as-you-go pension systems to fully funded pension schemes. On average, the advanced economies tend to have larger pension funds than Latin American countries, but this is mostly because the United States, Great Britain, and Canada have very large pension funds. Once these three countries are dropped from the sample, the size (expressed as a share of GDP) of Latin American pension funds is not too different from (and, if anything, larger than) that of those in the advanced economies (Figure 8.2).
Pension funds located in the advanced economies hold about one-quarter of their assets in government bonds (Figure 8.3). There are, however, substantial differences among countries in this group. In countries with large pension funds like the United Kingdom and United States, funds hold a relatively low proportion of government bonds. Countries with smaller pension funds (such as Austria and Italy) are characterized by much larger holdings of government bonds. Pension funds of emerging market countries located outside Latin America hold more than 50 percent of their assets, on average, in government bonds.[4] The average share of government paper held by Latin American pension funds is 44 percent of total assets, which is larger than the prevailing average share in pension funds in the advanced economies but smaller than the prevailing average share in emerging markets. Again, there are large differences within the region. Government debt is particularly important (close to or greater than 50 percent of total assets) in Mexico, Argentina, Uruguay, and Colombia, but relatively unimportant in Peru, Brazil, and Chile.
Usually, holdings of public sector debt are particularly high when a country’s private pension system is initially established. This is in part a result of the design of pension reforms, which often have the objective of helping governments to finance the costs of the transition out of a state-managed social security system, with its remaining liabilities to the retiring population. As high as these requirements may be, pension funds have at times held even higher shares of public debt owing to limited attractive investment opportunities in the private sector and legal limits on foreign asset holdings. Several Latin American countries (including Chile and Mexico) established special guidelines when their private pension systems were initially set up, allowing pension funds to hold a large fraction of their assets in government bonds in order to reduce the financing risks of the transition from a pay-as-you-go system to a fully funded one.[5] The idea was that limits on holdings of government debt would be reduced over time to ensure that pension funds diversified their assets and did not concentrate their exposure in the public sector. In other cases (e.g., Argentina and Uruguay), there were strict limits from the very beginning on the pension system’s holdings of government bonds. The existing regulations are diverse, but most countries now impose limits (which are not always enforced) on the holding of government bonds or on overall exposure to the public sector (Table 8.2).
Thus, in Chile, pension fund exposure to the public sector remained at around 50 percent of total assets through most of the 1980s and started to fall gradually only in the 1990s, before dropping sharply in the last six years. One possible reason for this reduction in pension funds’ holdings of public debt is that, as Chilean pension funds were confronted with a shortage of public debt (they currently hold roughly 85 percent of the total), they started to find alternative investments in Chile and abroad (aided by a gradual relaxation of the foreign asset share limit).
In Mexico, pension funds’ holdings of public debt started at very high levels (97 percent of assets) and remain at very high levels, though they are gradually being reduced. Argentina is clearly an outlier, as the holdings of government bonds significantly increased almost seven years after the inception of the reformed pension system, when the government faced the 2001–2002 debt crisis. This increase in the holdings of government paper was, by and large, not voluntary and was driven by the need to ensure financing prior to the crisis. A similar scenario is observed in Uruguay following the recent debt crisis (Figure 8.4).
The only large Latin American country in which private pension funds hold a small share of total public debt is Brazil. Incidentally, Brazil is also the only large Latin American country that did not implement pension reform in the last few decades and hence has voluntary pension funds.[6] Note that this low level of public debt holdings is due not to the fact that Brazil has small pension funds (in 2004, the assets of Brazilian pension funds were about 16 percent of GDP), but rather to the fact that these funds hold a small amount of government securities. In 2004, only 12 percent of the assets of Brazilian pension funds were invested in government securities (Cowan and Panizza, 2006).[7]
Although their exposure to the public sector is likely to decline in the coming years, pension funds are still likely to be important participants in the public debt market.[8] Moreover, the cited preference for real returns makes them a natural investor base for local currency markets (Levy Yeyati, 2004). While governments should try to take advantage of the needs of pension funds and other institutional investors to fulfill their financial programs, a prudential regulatory framework needs to ensure that the government does not force them to hold more government bonds than these investors consider optimal. In most cases this objective is facilitated by requiring pension funds to mark bonds to market and by limiting their ability to book them as loans or long-term investments that could be considered at technical values until their maturity (for a discussion of this issue, see the last section of this chapter).
One major restriction on the portfolio of pension funds is the limit on the foreign asset share, which aims to ensure that savings are channeled into the domestic economy. The fact that this restriction is binding in most cases (Table 8.2), combined with the dearth of long-run private investment assets, has certainly contributed to pension funds’ marked concentration in government debt. A survey of institutional investors in six Latin American countries suggests that pension funds would like to hold more foreign assets but are prevented from doing so by existing constraints (Cowan and Panizza, 2006). Although there may be a prudential basis for the limit on foreign investment—to avoid a currency mismatch, as pension funds’ liabilities are denominated in domestic currency—it is unclear whether this reason justifies the imposed limit or whether the desire to create a captive demand for domestic financial instruments is the driving force of the regulations. Some recent developments may result in a relaxation of these constraints, with several countries opening up their markets to issuers from other countries in the region. For example, a Mexican company (América Móvil) is in the process of issuing long-term bonds in Chile which, under a newly implemented Chilean law, will be registered as domestic bonds and hence become exempt from restrictions based on foreign asset shares. Given the growth potential of the Latin American cross-border market, the IDB’s Private Sector Department is considering the possibility of promoting regional integration opportunities by providing guaranties to cross-border issuers.
INSURANCE COMPANIES
Insurance companies are the largest institutional investors in East Asia, but they are much less important players in Latin America (Figure 8.5). However, the increasing importance of pension funds has resulted in positive spillovers into the annuity market and contributed to the growth of the life insurance sector (IMF, 2004b). As a consequence, it is not surprising that Chile is the Latin American country with the largest insurance sector.
One positive aspect of having a large insurance sector is that, in the majority of countries, insurance companies are not required to mark their assets to market on a daily basis,[9] which allows them to face short periods of market volatility without having to book short-term losses. This, together with the fact that the majority of insurers do not benchmark their performance to any specific index, may limit “herding behavior” and is likely to contribute to the overall stability of domestic financial systems.
On a less positive note, in most emerging market economies, insurers are required to match assets and liabilities. As in several Latin American countries a large share of life insurance contracts are specified in foreign currency, and as insurance companies are often not allowed to hold a large share of foreign assets, these companies end up holding a large amount of sovereign dollar-denominated external debt (IMF, 2004b).
MUTUAL FUNDS
Assets of emerging market mutual funds grew rapidly in the second half of the 1990s and then stabilized over the 2000–2003 period. This trend was due to a contraction of assets held by mutual funds in emerging Asia and a continuous expansion in Latin America (Figure 8.6). Within Latin America, the countries that experienced the fastest growth were Brazil, Colombia, and Costa Rica.
One important difference between the asset composition of mutual funds located in the advanced economies and that of those located in emerging market countries is that in the former, equity funds tend to account for a larger share of the assets than bond funds, while the opposite is true for the latter (this is the case, for instance, in Brazil and Mexico) (IMF, 2004b). This difference is partly due to the fact that in most emerging market countries, stock markets are small, and government bonds are the most liquid instruments in the local capital market. But it is also due to the fact that, in an environment of low short-term interest rates, investors become interested in longer-term bonds and start switching from bank deposits to mutual funds that hold this type of asset. One source of concern with this investment strategy is that retail investors, which are reassured by the low default risk of these instruments, may not understand the market risk associated with their long-term nature, which may amplify market volatility (Box 8.1). In addition, local stock markets are often opaque, with imperfect monitoring and regulation, making them specialists’ markets. Finally, unlike those of developed countries, emerging markets’ stocks tend to be positively correlated with emerging market bonds, a fact that reduces the hedging benefits of stocks.
BANKS IN THE DOMESTIC GOVERNMENT BOND MARKET
Banks are unique players in the government bond market. On the one hand, they invest in government bonds as part of their regular asset management decisions and hold bonds in their portfolios just like other credits. On the other hand, banks are primary dealers and market makers of government bonds, which implies that they participate in regular treasury auctions and provide liquidity for these instruments in secondary markets.[10]
There are least three reasons that lead banks to hold government bonds in their balance sheet:
1. Banks hold government bonds (mainly short-term treasury bills) to manage their liquidity. Government bonds are ideal instruments for this purpose because they generally have a liquid secondary market and can be used for repos with the central bank or with other commercial banks.
2. Banks hold bonds as part of their portfolio decisions. For this purpose they generally buy longer-term treasury bonds that they book in their investment account and hold to maturity. From an accounting point of view these bonds are considered a long-term investment and are included in the “banking” book at their purchase price.
3. Banks hold government bonds for trading and to be market makers in the secondary market for these bonds. These holdings are generally small and valued at market prices.
In the United States banks hold a stock of government bonds equivalent to 14 percent of domestic credit, while in the Euro Area, the average holding of government bonds is 20 percent of domestic credit (Figure 8.7). Banks in Latin America have an average exposure to government bonds of around 25 percent of domestic credit. Banks in Argentina had the largest exposure to the public sector in Latin America in 2003, at close to 50 percent of domestic credit, followed by Mexico, where banks’ holdings of government paper represent 42 percent of domestic credit. In contrast, Chilean banks had the smallest exposure to the public sector in the region, well below 10 percent of domestic credit.
There are a number of explanations for the large holdings of government bonds among Latin American banks. In some cases banks in the region hold these bonds as part of their reserve requirements or to comply with regulations—which explains, for instance, roughly one-quarter of banks’ holdings of government bonds in Brazil. In the cases of Argentina and Mexico, banks’ decision to hold these bonds was not part of a portfolio allocation model, but rather the outcome of the resolution of the banking crises that 
affected the two countries. In particular, banks exchanged defaulted loans for specially issued government bonds in order to keep operating with an adequate level of capital when the bonds were booked at their technical values. In Argentina banks were also “persuaded” to increase their holdings of government paper in 2001 in order to avoid a government default. So, in a situation in which private credit was shrinking, banks substantially increased their holdings of government assets (Figure 8.8). In other cases, banks might decide voluntarily to hold government bonds because they provide a high yield, are perceived to be less risky (and implicitly guaranteed by the government) (Box 8.2), and face lower capital requirements than private assets.
In general, it is difficult to know whether banks that hold government bonds are doing so voluntarily or whether they have instead been induced to hold them through regulation or moral suasion. For example, in some cases central banks allow part of a bank’s reserve requirements to be held in government instruments, which is one possible incentive to hold public debt.[11] In other cases, regulations can stimulate demand for government bonds by allowing bonds to be on a bank’s books at technical values or at the price at which they were originally purchased instead of at market values.
It would thus be useful to separate the portion of banks’ exposure to the public sector that is induced through regulation from the portion that arises from their portfolio decisions. One practical way of making this distinction is to force banks to mark to market their exposure to the public sector (Box 8.3).[12]
HOW TO MAKE GOVERNMENT BONDS
ATTRACTIVE TO INSTITUTIONAL INVESTORS
Countries with a large base of investors have a greater capacity to deal with reductions in external demand for domestic financial assets. Thus, the relatively small size and number of long-term institutional investors in Latin America may be one of the factors that contribute to the region’s vulnerability to external financial shocks.
But there is a two-way interaction between the importance of institutional investors and the functioning of domestic bond markets. While the growth of institutional investors is good for public debt management, a coherent debt management strategy and the development of a sound market microstructure in a country can also foster the growth of institutional investors (Vittas, 1998; Catalan, Impavido, and Musalem, 2000). It is therefore interesting to review how a country’s public debt management policies may affect the development of institutional investors.
The first set of policies has to do with the choice of financing instruments. With respect to the type of bonds to be issued, the government can choose between bullet or amortization bonds; bonds with floating interest rates or fixed interest rates or indexed bonds; bonds issued in domestic or foreign currency; bonds issued under domestic or foreign legislation; and bonds with short or long maturity. It is not clear which type of bonds is preferred by domestic institutional investors, but in practice most Latin American countries are moving towards issuing standardized bonds, which are bullet instruments (i.e., the whole principal is paid at maturity), with semiannual interest payments, and in domestic currency.
The second set of policies is related to the development of a yield curve. Investors and other issuers can benefit from a fully developed yield curve for government bonds that sets the “benchmark” interest rates for different maturities (usually ranging from 3 months to 5 or 10 years).
The third set of policies has to do with increasing the liquidity of government bonds. A government can increase the liquidity of its bonds by making large benchmark issues (the minimum size of these benchmarks varies across countries). Governments can also improve the liquidity of their bond markets by facilitating the development of the repo market (which allows borrowing against bonds) and by taking measures to reduce transaction costs.
The fourth set of policies has to do with coordination between the central government and other public sector issuers. In Latin America the main issuers of domestic debt are the treasury and the central bank, and in many Latin American countries there are explicit agreements between the central bank and the treasury regarding the division of the market. In Uruguay, for instance, the central bank issues mainly in pesos, while the treasury issues in foreign currency. In Argentina, the central bank taps the short end of the market, while the treasury issues at longer maturities.
The fifth set of policies has to do with providing information about the government’s financing strategy. When institutional investors know the amount of financing that the government needs and the type and timing of instruments to be issued, they can plan their purchases of bonds and ensure that they have the necessary funds to participate in the primary issuance of government bonds. Regular auctions of government bonds are thus desirable. In several countries there are weekly auctions for short-term treasury bills and monthly or quarterly auctions for longer-term treasury bonds.
Finally, governments can improve the attractiveness of their bonds by improving the market microstructure, especially settlement, clearing system, and custody. The operational and legal infrastructure that supports the issuance and trading of domestic government debt affects the depth and liquidity of the government bond market. Without the right settlement infrastructure there is a risk of failure to deliver either the cash or the securities in a large transaction, and this could have significant ripple effects on other settlements. Likewise, it is critical to ensure a high-quality custodian for the bonds, with high credit ratings and solid operational procedures. In some countries, the custodian is a public institution, such as the central bank, but in many others there are private custodians. Finally, to minimize credit risk in transactions, most countries have instituted delivery versus payment mechanisms for settling the transactions.
PROTECTING INSTITUTIONAL INVESTORS
FROM VORACIOUS GOVERNMENTS
While institutional investors are critical players in domestic government bond markets, they could become victims of their own strength, as financially constrained governments might attempt to capture investors’ resources through regulation and persuasion. Therefore, it is essential to have in place good institutional and regulatory frameworks aimed at reducing the risk that a government will pressure institutional investors to buy government bonds when it faces financial strains. Such a system would require an independent regulatory agency able to enforce limits on institutional investors’ holdings of government bonds and induce institutional investors to appropriately evaluate the risk-return ratio associated with buying and holding government bonds.
Governments have found creative ways to induce institutional investors to increase their holdings of public debt by offering terms that are more favorable than those prevailing in the markets. For instance, central banks can impose high reserve requirements and then allow banks to fulfill them with government bonds issued at below-market interest rates. This has been the case in Brazil, for instance, where around 25 percent of banks’ holdings of government bonds are induced by regulation.
Another way to induce institutional investors to increase their holdings of government bonds is to provide advantages in the way some instruments are valued in the investors’ balance sheets. In other cases, authorities create new instruments that allow banks and pension funds to exceed the limits imposed by regulations.
Many of these “innovations” were certainly at work during the recent Argentine financial crisis. Initially, in 2001, the Argentine government attempted to avoid default and “induced” banks and pension funds to increase their holdings of public sector debt. As a result, banks increased their exposure to the public sector from 16.2 percent of assets in 1999 to 26.3 percent in 2001, and pension funds from 48.3 percent to 67.2 percent of total assets. These institutional investors were willing to accept (perhaps reluctantly) an increase in their holdings because the new instruments had regulatory advantages over the existing ones, as they could be assessed on balance sheets at “technical” values that were much higher than market values.
In addition, institutional investors may have an incentive to collaborate with the government once they accumulate a large exposure to the public sector, as a sovereign default or a restructuring of the public debt would then have a significant impact on their balance sheets. Governments can use this “coincidence” of interests to obtain the assistance of such investors. This vested interest in avoiding a debt restructuring could explain the collaboration of banks and pension funds with the Argentine government.
While there is no easy way to insulate institutional investors from governments in desperate need of financing, there are at least some measures that can mitigate the chances of excessive pressure. The obligation to mark to market all government instruments would be a deterrent to excessive exposure to the public sector, even for pension funds and life insurance companies that invest with a long-term horizon. In addition, it would help to require institutions to report their consolidated exposure to the public sector, including indirect holdings of government paper through mutual funds and investment companies in which the institutions invest.
In the case of banks, in addition to the obligation to mark to market all government financial instruments, supervisors could include quantitative limits (as part of credit diversification requirements) and introduce capital requirements for holding government bonds. Unfortunately, the new international framework (commonly referred to as Basel II) seems to have missed the opportunity for introducing these kinds of safeguards (Box 8.4).
While implementation of these recommendations would provide useful safeguards, large domestic capital markets are definitely an important first line of defense against financial crises. In the end, however, strong fiscal and debt management policies are the only policy measures that can truly protect a country from a possible default and indirectly support the soundness of institutional investors.
