Capital Market Liberalization, Globalization, and the IMF
3.4 Concluding Comments
3.3.6 Justifying Interventions
The fact that short-term capital flows have potentially such large adverse effects on others—beyond those directly involved in the flows—implies that there is anexternality, and as always, when there is an externality, there is a prima facie case for government intervention. The question is only whether there are interventions that can address the adverse consequences of the externality, without more than offsetting ancillary negative side effects and, if so, what is the best form of intervention. The experiences of Chile and Malaysia suggest that there are such interventions.21
Even the IMF recognizes the importance of externalities in this arena—
witness their concern about contagion and their use of contagion as a jus-tification for bail-outs. But if crises justify government actions, then it makes sense to address the underlying causes. (One should not just build a bigger hospital to address public health problems.) One of the causes of crises are destabilizing short-term capital flows. Accordingly, it makes sense to try to stabilize such flows. Even if interventions areimperfect (i.e., they are ‘leaky’), there may be a large social benefit from the reduction in the overall magnitude of the volatility of short-term capital flows.
flows, or even better, ensure that they move counter cyclically. It should be working harder to address the underlying failures in capital markets, devising ways by which more of the risk of interest rate and exchange rate fluctuations can be shifted to developed countries and international financial institutions.
And in the future, it should rely more on evidence and less on ideology in developing its policy agenda. The IMF’s stance on capital market liberalization has, in many circles, undermined its credibility; in spite of its authors’ claims that the paper represents ‘an evolution, not a revolution’ in IMF thinking, this paper confirms what many in the developing world have long known:
IMF advice in this area confronted countries with risk without reward.
3.5 2007 Postscript: It’s Hard to Change One’s Mindset
An August 2006 IMF working paper, by the IMF’s former Chief Economist and his colleagues, reinforces the conclusions that changing the mind set of the IMF about capital market liberalization—whatever the evidence—will not be easy (Kose et al. 2006). Again, they are to be commended for reporting honestly that ‘there is still little robust evidence of the growth benefits of broad capital account liberalization . . . ’ Yet they conclude ‘our critical reading of the recent empirical literature is that it lends some qualified support to the view that developing countries can benefit from financial globalization.’
The reason, they assert, is that there are ‘potential collateral benefits,’ such as financial market development, institutional development, better governance, and macroeconomic discipline, which themselves lead to higher growth and lower consumption volatility.
They never explain, however, why, if these collateral benefits exist, they do not show up in the reduced form regressions, regressions that should show growth benefits regardless of the channel through which they come.
And, equally remarkably, they never test—indeed hardly mention—the poten-tial ancillary channels through which capital market liberalization might adversely affect growth and volatility. Consider, for instance, one of the channels, ‘market discipline.’ In support of that, they mention my own earlier World Developmentpaper (Stiglitz 2000). But as I point out, the problem with capital markets as disciplinarians is that they are myopic—they enforce a dis-cipline that is not oriented towards long-term growth. Indeed, they typically focus more on a country’s liabilities than on its assets, and therefore, CML may actually impede growth. (Even more so, they ignore other aspects of societal well-being, such as those reflected in health and the environment.) Moreover, capital markets are anunreliabledisciplinarian, subjecting even well-behaved and well-performing countries to ‘discipline’ when the market appetite for developing country risk suddenly changes.
Indeed, what the IMF should have done was to look at the capital flow implications of their ‘theory.’ Their theory has it that CML is stabilizing because money flows into a country when it needs it, i.e., is counter cyclical, not pro-cyclical, and that it is growth enhancing, i.e., countries that liberalize have more investment, because of the greater availability of capital from the outside and/or increased domestic investment. Both theory and evidence cast doubt on both of these propositions, and the IMF study does little to dispel these doubts. It is largely because of the exposure to the risk of pro-cyclical capital movements that CML has been so roundly criticized.
The authors of the IMF study also remarkably fail to address directly most of the other criticisms of CML, either within their model or in more general models. They do not challenge the conclusion that capital market liberaliza-tion has exposed countries to a new set of shocks and that these external shocks are among the major sources of volatility in those countries that have become financially integrated. Moreover, they deal only partially with the claim that, while CML exposes countries to more shocks, it also makes it more difficult for countries to engage in countercyclical macroeconomic policy.22 After noting the several ways in which liberalization has made the conduct of monetary policy more difficult, they note that ‘the fact that so many emerging markets have successfully instituted more independent, inflation-focused central banks, is quite noteworthy.’ They fail to observe that there is little evidence that independent, inflation-focused central banks lead to faster growth. (This is one of several instances where what should be treated as intermediate variables, of concern only to the extent that they promote growth or stability or a better distribution of income, are treated almost as ultimate objectives.) One of the main intentions of capital market restrictions (e.g., in the case of Chile) was to enhance the scope of monetary policy. They never even discuss this issue.
The underlying problem, of course, is that the authors continue to be wedded to the perfect capital markets view of the world (with infinitely lived individuals). Thus, they report, ‘Lucas’s (1987) claim that macroeconomic sta-bilization policies that reduce consumption volatility can have only minimal welfare benefits continues to be influential . . . ’ without noting that Lucas’
result depends critically on individuals not facing borrowing constraints.
Unfortunately, the empirical behavior of consumption can only be explained within models that assume that at least lower income individuals do face borrowing constraints. They conclude (notwithstanding the analysis of the models presented in this chapter, which they do not note): ‘there is a strong presumption in theory that financial integration is good for growth and . . . it
22 Of course, trade openness may have reduced inflationary pressures, and this may have allowed monetary authorities more scope to act. But we are concerned here with financial integration, not trade integration.
should unambiguously lead to reductions in the relative volatility of con-sumption.’
The IMF study does include a discussion of a number of recent papers on the relationship between crises and CML. On the basis of these studies, the authors conclude ‘there is little formal empirical evidence to support the oft-cited claims that financial globalization in and of itself is responsible for the spate of financial crises that the world has seen over the past three decades’
(Kose et al. 2006). But the study does not cite several World Bank studies that show that countries are more likely to have a crisisafterliberalizing (financial and capital markets).23 Moreover, the issue is not so much whether CML by itself leads to a crisis but whether it exposes a country to risks, which, together with one of the many other problems facing many developing countries, make a crisis more likely. Most students of the East Asian crises have come to the conclusion that capital market liberalization was a central factor in the crises in these countries, with capital rushing in during the early 1990s and rushing out in 1997. Without CML, these countries would not have faced the crises they faced. The same is true for the crises in Latin America at the end of the 1990s and early years of this decade.
A wealth of case studies—supported by economic theories that recognize market imperfections, including shortfalls in economic performance below potential—have provided the basis of much of the critique of CML. The IMF study, by contrast, reviews the cross-country econometric literature.
This literature has been (rightly) criticized—its results often appear not to be robust, and the studies face numerous data and econometric problems.
Unfortunately, the discussion of the IMF paper does not deal sufficiently with these problems to remove any skepticism that critics might have. For instance, while it notes the possible problem of ‘reverse causality,’ this is not the only source of endogeneity. There is, for instance, a problem of ‘self-selection,’ those countries that have chosen to liberalize their capital markets are those for which the benefits are the greatest (the costs are the lowest), even accounting for all the ‘observed’ variables. Assume, for instance, that (holding everything else constant), there are two kinds of countries: those for whom CML is good, and those for which it is not. Assume, moreover, each country knows whether CML is good for itself, but we (as outside observers) do not. Then, in a regression relating growth to CML the coefficient on the CML variable will provide an upwardly biased estimate of what will happen to a country that has not liberalized, should it decide to liberalize. Thus, if the regression coefficient comes out zero (as appears to be the case in many of the regressions), it means that the likely impact of liberalizing is negative.
While the IMF study does not take a look at most of the important ancillary costs of CML (e.g., the adverse effects on growth through the impact on debt
23 See also Furman and Stiglitz (1998).
equity ratios resulting from the higher volatility of relevant variables), neither does it establish convincingly effective links through the alleged positive ancillary channels. This is illustrated by the discussion of governance. While no one would support the notion that it is a good thing for countries to have bad governance, the links between capital market liberalization and governance, on the one hand, and governance (as conventionally measured) and growth and stability on the other are tenuous at best. In some studies of corruption, for instance, China and Vietnam do not fare well; yet these countries have been among the best performing countries. Hoff and Stiglitz (2005) argue that capital market liberalization actually had an adverse effect on the development of the rule of law in the economies in transition. It enhanced incentives for asset stripping (relative to wealth creation) and thus undermined the constituency for the creation of a rule of law. With capital market liberalization, those in these countries could take their money outside the country, enjoying the protections of the rule of law in Western countries, and at the same time take advantage of the absence of the rule of law (e.g., the absence of good laws concerning corporate governance) to divert assets to their own benefit at home. Cross-country studies by Khan (2004), bolstered by theoretical analysis, suggest that there is, in fact, overall little relationship between growth and governance. He divides developing countries into two categories: fast growing countries (such as those in East Asia) and slow growing (which includes most of those in Africa). Within each category, differences in governance are unrelated to growth; by the same token, differences in governance have little to do with the category (fast or slow growing) into which the country falls.
The debate about capital market liberalization focuses on regulations affect-ing short-term capital flows. The IMF paper, by contrast, looks more broadly at financial globalization and, for the most part, does not separately address the issues of concern in this chapter. Even most critics of CML believe that Greenfield foreign direct investments can have a beneficial effect.24
By the same token, even had there been a finding that capital market liberalization on average was associated with lower growth, it would only prove that capital market liberalization can be—and in many countries has been—done poorly. It does not address the more relevant question, has it had these negative effects when it has been implemented well? Can countries with limited administrative capacity implement such controls well? Are there
24 Much of what is classified as FDI is really a financial transaction, the purchase of an existing asset. The failure to distinguish among different categories of FDI undermines the usefulness of studies, such as the IMF study, that look simply at the relationship, e.g., between FDI and growth. Similarly, one should distinguish between FDI in natural resources (which typically creates few jobs and exposes countries to the risk of the Dutch disease) and in, say, manufacturing.
some controls that are more easily implemented by countries with limited administrative capacity?
The point of looking at historical lessons is to learn from them—it is not to imitate the mistakes of the past but to avoid them. Of course, this kind of reasoning has been at the center of the argument for capital market liberalization. Observing the many crises that have followed, they conclude that ‘inadequate or mismanaged domestic financial sector liberalizations have been a major contributor to crises that may be associated with financial integration’ (Kose et al. 2006).25The obvious lesson drawn—by the advocates of CML—is to strengthen one’s financial institutions and manage financial sector liberalization well before liberalizing capital markets. But if countries have the capacity to do that, they may also have the capacity to administer an effective set of capital market regulations—regulations that can enhance growth and reduce volatility.
Appendix 3.1 A Simple Model of ‘Regime Change’ in which