Journal of Economic Literature Vol. XLV (December 2007), pp. 887?935 Capital Account Liberalization: Theory, Evidence, and Speculation PETER BLAIR HENRY? Research on the macroeconomic impact of capital account liberalization finds few, if any, robust effects of liberalization on real variables. In contrast to the prevailing wis- dom, I argue that the textbook theory of liberalization holds up quite well to a critical reading of this literature. Most papers that find no effect of liberalization on real vari- ables tell us nothing about the empirical validity of the theory because they do not really test it. This paper explains why it is that most studies do not really address the theory they set out to test. It also discusses what is necessary to test the theory and examines papers that have done so. Studies that actually test the theory show that lib- eralization has significant effects on the cost of capital, investment, and economic growth. 887 1. Introduction Acapital account liberalization is a deci-sion by a country?s government to move from a closed capital account regime, where capital may not move freely in and out of the country, to an open capital account system in which capital can enter and leave at will. Broadly speaking, there are two starkly dif- ferent views about the wisdom of capital account liberalization as a policy choice for developing countries. The first view, Allocative Efficiency, draws heavily on the predictions of the standard neoclassical growth model pioneered by Robert M. Solow (1956). In the neoclassical model, liberalizing the capital account facili- tates a more efficient international allocation of resources and produces all kinds of salu- brious effects. Resources flow from capital- abundant developed countries, where the return to capital is low, to capital-scarce developing countries where the return to capital is high. The flow of resources into the ? Henry: Stanford University, Brookings Institution, and National Bureau of Economic Research. I gratefully acknowledge financial support from a NSF CAREER Award, the John and Cynthia Fry Gunn Faculty Fellowship, and the Freeman Spogli Institute for International Studies, and the Stanford Center for International Development. I thank Anusha Chari, Roger Gordon, Diana Kirk, John McMillan, Diego Sasson, and two anonymous referees for extensive comments on earli- er drafts. Barry Eichengreen, Kristin Forbes, Michael Hutchison, Don Lessard, Ross Levine, Karen Lewis, Rick Mishkin, Paul Romer, Andrew Rose, Jay Shambaugh, Ren? Stultz, Romain Wacziarg, and seminar participants at the Brookings Institution, the Federal Reserve Bank of San Francisco, the International Monetary Fund, Stanford, Washington University in St. Louis, U.C.-Santa Cruz, and Yale also provided helpful comments. Sandra Berg provided stellar editorial assistance. Finally, I thank Sir K. Dwight Venner and the Eastern Caribbean Central Bank for hosting me in the summer of 1994, when I first started thinking about the issues addressed in the paper. dec07_Article1 11/28/07 12:01 PM Page 887 developing countries reduces their cost of capital, triggering a temporary increase in investment and growth that permanently raises their standard of living (Stanley Fischer 1998, 2003; Maurice Obstfeld 1998; Kenneth S. Rogoff 1999; Lawrence H. Summers 2000). Motivated in part by the prospective gains from incorporating Allocative Effciency arguments into their economic policies, dozens of developing- country governments from Santiago to Seoul implemented some form of capital account liberalization during the past quarter century. The alternative view regards Allocative Efficiency as a fanciful attempt to extend the results on the gains to international trade in goods to international trade in assets. The predictions of Allocative Efficiency hold only when the economy suffers from no distor- tions other than barriers to free capital flows. Because there are many other distortions in developing countries, skeptics argue that the theoretical predictions of the neoclassical model bear little resemblance to the reality of capital account policy. Dani Rodrik?s (1998) provocatively titled article ?Who Needs Capital Account Convertibility?? best characterizes this view.1 Rodrik?s empirical analysis finds no correlation between the openness of countries? capital accounts and the amount they invest or the rate at which they grow. He concludes that the benefits of an open capital account, if indeed they exist, are not readily apparent, but that the costs are manifestly evident in the form of recur- rent emerging-markets crises. In the wake of Rodrik?s polemic, evidence seemingly continues to mount in support of the view that capital account liberalization has no impact on investment, growth, or any other real variable with significant welfare implications. For example, in his survey of the research on capital account liberaliza- tion, Barry Eichengreen (2001) concludes that the literature finds, at best, ambiguous evidence that liberalization has any impact on growth. In another review of the litera- ture, Hali J. Edison et al. (2004) survey ten studies of liberalization and document that only three uncover an unambiguously posi- tive effect of liberalization on growth. Finally, in their comprehensive survey of the research on financial globalization, Eswar S. Prasad et al. (2003) extend the Edison et al. (2004) survey to fourteen studies, but still find only three that document a significant positive relationship between international financial integration and economic growth. Prasad et al. (2003) conclude that ?. . . an objective reading of the vast research effort to date suggests that there is no strong, robust, and uniform support for the theoret- ical argument that financial globalization per se delivers a higher rate of economic growth.? In contrast to existing surveys, this article demonstrates that a critical reading of the literature reveals that the textbook theory of liberalization stands up to the data quite well. It is true that most papers find no effect of liberalization on growth. But these papers tell us nothing about the empirical validity of the theory. They perform purely cross-sec- tional regressions that look for a positive cor- relation between capital account openness and economic growth, implicitly testing whether capital account policy has perma- nent effects on differences in long-run growth rates across countries. The funda- mental problem with this approach is that the neoclassical model provides no theoreti- cal basis for conducting such tests. The model makes no predictions about the cor- relation between capital account openness and long-run growth rates across countries, and certainly does not suggest the causal link needed to justify cross-sectional regressions. What the neoclassical model does predict is that liberalizing the capital account of a capital-poor country will temporarily increase the growth rate of its GDP per capita. The temporary increase in growth matters, because it permanently raises the 888 Journal of Economic Literature, Vol. XLV (December 2007) 1 See also Jagdish Bhagwati (1998) and Joseph E. Stiglitz (1999, 2000, 2002). dec07_Article1 11/28/07 12:01 PM Page 888 country?s standard of living. However, as it is the increase in the level of GDP per capita that is permanent?not its rate of growth? theory dictates that one tests for either a per- manent level effect or a temporary growth effect. Testing for a permanent growth effect makes no sense because capital accumula- tion, which is subject to diminishing returns, is the only channel through which liberaliza- tion affects growth in the neoclassical model. A small but growing branch of the litera- ture takes the time series nature of the neo- classical model?s prediction seriously. It does so by investigating whether countries invest more and grow faster in the immediate aftermath of a discrete change in their capi- tal account policy. In contrast to work that looks for a permanent impact of capital account openness on growth across coun- tries, papers in the policy-experiment genre find that opening the capital account within a given country consistently generates eco- nomically large and statistically significant effects, not only on economic growth but also on the cost of capital and investment. While the policy-experiment literature removes any serious doubts about whether liberalization has real effects, its findings raise more questions than they resolve. Are the magnitudes of the documented effects plausible? Do the transmission mechanisms emphasized by the theory really drive the results or are other forces at work? With their focus on aggregate data, papers in the policy- experiment genre do not have enough empir- ical power to be of any use. Fortunately, a new wave of papers demonstrates how to make progress on these and other important questions by using firm-level data. Analyzing capital account liberalization at the level of the firm instead of the country provides greater clarity about the channels through which liberalization affects the real economy. Disaggregating the data also brings clarity to the contentious debate on liberalization and crises. Because there are many different ways to liberalize the capital account, when trying to determine whether liberalizations cause crises, it is critical to specify exactly what kind of liberalization you mean. At a minimum, the distinction between debt and equity is critical. Recent research demonstrates that liberaliza- tion of debt flows?particularly short-term, dollar-denominated debt flows?may cause problems. On the other hand, all the evidence we have indicates that countries derive sub- stantial benefits from opening their equity markets to foreign investors. The rest of this paper proceeds as follows. Section 2 presents an organizing theoretical framework. Section 3 reviews the literature on the cross-sectional approach to the macro- economic effects of capital account liberaliza- tion. Section 4 uses the framework from section 2 to explain why the cross-sectional lit- erature finds no real effects of liberalization. Section 5 discusses the policy-experiment approach to the macroeconomic effects of lib- eralization. Section 6 discusses problems with the policy-experiment approach. Section 7 reviews recent advances using firm-level data. Section 8 examines whether liberalizations cause crises. Section 9 concludes. 2. Capital Account Liberalization and the Neoclassical Growth Model This section illustrates the fundamental predictions of the neoclassical growth model about the impact of capital account liberal- ization on a developing country. The frame- work is not novel, but it brings clarity to the discussion of the empirical literature that follows in section 3. Assume that output is produced using cap- ital, labor, and a Cobb?Douglas production function with labor-augmenting technological progress: (1) . Let k = A?KL be the amount of capital per unit of effective labor and y = AY?L the amount of output per unit of effective labor. Using this notation and the homogeneity of the production function we have: Y F K AL K AL= ( ) = ( ) ?, ? ?1 889Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:01 PM Page 889 (2) . Let s denote the fraction of national income that is saved each period and assume that capital depreciates at the rate ?, the labor force grows at the rate n, and total factor productivity grows at the rate g. Saving each period builds up the national capital stock and helps to make capital more abundant. Depreciation, a growing popula- tion, and rising total factor productivity all work in the other direction making capital less abundant. The following equation sum- marizes the net effect of all these forces on the evolution of capital per unit of effective labor: (3) k.(t) . When k .(t) = 0, the economy is in the steady state depicted by point A in figure 1. At point A, the ratio of capital to effective labor (k) is constant. In contrast, the steady-state level of capital (K) is not constant, but grow- ing at the rate n + g. Output per worker (Y?L)grows at the rate g. Finally, the steady state marginal product of capital equals the interest rate plus the depreciation rate: sf k t n g k t( )= ( ) ? + +( ) ( )? y f k k= ( ) = ? (4) . Equation (4) gives a general expression of the equilibrium condition for investment. This equation has important implications for the dynamics of a country?s investment and growth in the aftermath of capital account liberalization because the impact of liberal- ization works through the cost of capital. Let r? denote the exogenously given world inter- est rate. The standard assumption in the lit- erature is that r? is less than r because the rest of the world has more capital per unit of effective labor than the developing country. It is also standard to assume that the devel- oping country is small, which means that nothing it does affects world prices. Under these assumptions, when the developing country liberalizes, capital surges in to exploit the difference between the world interest rate and the country?s rate of return to capital. The absence of any fric- tions in the model means that the country?s ratio of capital to effective labor jumps immediately to its postliberalization, steady- state level. Figure 1 depicts this jump as a rightward shift of the vertical dashed line ? = +f k r s state( ). ? 890 Journal of Economic Literature, Vol. XLV (December 2007) Figure 1. Capital Account Liberalization in the Neoclassical Growth Model ks.state kk?s.state A sf(k) (n + g + ?)k Output Per Unit of Effective Labor dec07_Article1 11/28/07 12:01 PM Page 890 from ks.state to k?s.state. In the postliberaliza- tion steady state, the marginal product of capital equals the world interest rate plus the rate of depreciation: (5) . On the one hand, the immediate jump to the new steady state illustrates the problem with cross-sectional regressions writ large: instantaneous convergence generates a one- time spike in the growth rate of output per worker in the year of liberalization followed by a return to normalcy in all subsequent years. Cross-sectional regressions designed to measure long-run differences in growth rates across countries are ill-suited to detect the short-lived changes implied by instantaneous convergence. On the other hand, it is reasonable to regard instantaneous convergence as an unattractive feature of the model rather than a serious practical challenge to the cross-sectional regression approach. Section 4.1 demonstrates in detail that the inability of cross-sectional studies to detect the impact of liberalization on growth remains a problem under slower, more realistic speeds of transition. For now, the vital point to note is that, during the country?s transition to the postliberalization steady state, the capital stock grows faster than it did before the tran- sition. To see why, recall that in the preliber- alization steady state the ratio of capital to effective labor (ks.state) is constant and the stock of capital (K) grows at the rate n + g. In the postliberalization steady state, the ratio of capital to effective labor (k?s.state) is also constant and the capital stock once again grows at the rate n + g. However, because k?s.state > ks.state, it follows that the growth rate of K exceeds n + g during the transition. The temporary increase in the growth rate of capital has implications for economic growth through the formula for the growth rate of output per worker: k . ?Y = ? + g. L? k ? ( ) = +f k r s state. ? ? ? Since the growth rate of K exceeds n + g during the transition, k . ?k must be greater than 0 during the corresponding interval of time. Therefore, the growth rate of output per worker also increases temporarily. Figure 2 illustrates the hypothetical response of the interest rate, the growth rates of capital and output per worker, and the natural log of output per worker under the more palatable assumption that the interest rate converges immediately upon liberalization but the ratio of capital to effective labor does not. Figures 3 through 5 (see pages 906?08), which display the real-life responses of the cost of capital, investment, and growth during eighteen capital account liberalization episodes from 1986 to 1991, bear a striking resemblance to the panels in figure 2. Section 5 discuss- es the construction of figures 3 through 5 in detail, but even this preliminary glance at the data raises an obvious issue. If the simple time series evidence lines up with the theory, then why does a large literature find no real effects of liberalization? I have already hinted at part of the answer, but the full story requires a more thorough account of the previous literature. 3. The Cross-Sectional Approach to Measuring the Impact of Liberalization Does capital account liberalization pro- mote a more efficient allocation of resources across countries? This section describes the methodology, data, and results associated with the traditional, cross-sectional approach to this question. Attempts to measure the impact of capital account liberalization on economic perform- ance require information on countries? capi- tal account policies. The International Monetary Fund?s (IMF) Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER) provides the most readily available, standardized source of such information. Available since 1967, the AREAER provides a list of the rules and reg- ulations governing resident and nonresident 891Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:01 PM Page 891 892 Journal of Economic Literature, Vol. XLV (December 2007) Figure 2. The Impact of Liberalization on the Cost of Capital, Investment, and Growth Interest Rate Growth Rate of K r r? to Panel A: The Cost of Capital t to t to to t t Panel B: Investment n + g Panel D: Level of GDP Per Worker Panel C: Growth Rate of GDP per Worker Growth Rate of Y?L ln (Y?L) g Perma nent e ffect o f libera lization dec07_Article1 11/28/07 12:01 PM Page 892 capital-account transactions in each country, a table that summarizes the presence of restrictions, and a qualitative judgment on whether the country has an open or closed capital account. For the editions of AREAER published between 1967 and 1996, the sum- mary table contains a single line (line E2) entitled, ?Restrictions on payments for capi- tal transactions.? The presence of a bullet point in line E2 indicates that the country has some form of restrictions on capital account transactions. In effect, line E2 deliv- ers a binary judgment that takes on the value one if any capital controls are in place and zero otherwise. The typical study maps the qualitative information from line E2 into a quantitative measure of capital account openness by tal- lying the number of years that each country was judged to have an open capital account. That tally is then divided by the total num- ber of years in the period to produce a num- ber called SHARE?the fraction of years over a given period that the IMF judged the country to be free from ?restrictions on pay- ments for capital account transactions.? For example, if a country had an open capital account for three of the ten years from 1986 to 1995, then SHARE is equal to 0.3. 3.1 Empirical Methodology Papers that use the variable SHARE assess the real effects of capital account liberaliza- tion by running some variant of the following cross-country growth regression: (6).........GROWTHi = a0 + a1 ? SHAREi + CONTROLSi + ?i, where GROWTHi denotes the average growth rate of real GDP per capita in coun- try i over the time period in question?typi- cally about twenty years?and the variable CONTROLSi serves as a compact stand-in for the usual set of control variables used in cross-country growth regressions in the tradi- tion of Robert J. Barro (1991). With the vari- able SHARE in hand, estimating equation (6) is a relatively straightforward exercise because data on the left-hand-side variable are readily available from standard sources such as the IMF?s International Financial Statistics. Because equation (6) does not contain time subscripts, it makes no use of the varia- tion in growth rates or openness over time within a given country. In other words, it identifies the impact of capital account poli- cy exclusively through the cross-country variation in average growth rates and frac- tion of years open. In effect, the estimation procedure tests for a permanent impact of liberalization on growth, measuring whether countries with open capital accounts have higher long-run growth rates than countries whose capital accounts are closed. The problem with this test is that we have already seen (in section 2) that the neoclassi- cal model does not predict that countries with open capital accounts will have higher long-run growth rates than countries with closed capital accounts. In the neoclassical model, differences in long-run growth rates across countries are driven exclusively by differences in their growth rates of total fac- tor productivity (TFP). Because it contains no channel through which capital account liberalization affects TFP growth, strictly speaking, the model provides no theoretical basis for estimating equation (6).2 In other words, papers that estimate the effects of capital account openness on growth do not provide a test of any causal theory. What the theory does predict is that capital-poor coun- tries will experience a permanent fall in their cost of capital and a temporary increase in the growth rate of their capital stock and GDP per capita when they liberalize. Testing whether countries with more open capital accounts invest more or grow faster 893Henry: Capital Account Liberalization 2 Foreign direct investment (FDI) is one form of capi- tal inflow that could bring technology in addition to capi- tal. If FDI brings new technology, it could generate knowledge spillovers that result in TFP growth. Section 5.4 considers the empirical relevance of this consideration as well as other extensions of the neoclassical model in which liberalization might have an impact on TFP growth. dec07_Article1 11/28/07 12:01 PM Page 893 than countries with closed capital accounts is not logically equivalent to testing whether countries that liberalize experience a tempo- rary increase in investment and growth. Moreover, the following example demon- strates that there are nontrivial examples in which the purely cross-sectional regression framework leads to specious conclusions about the data. 3.1.1 Numerical Example of the Pitfalls of the Cross-Sectional Regression Approach Consider two countries (A and B) that are in steady state and have the same level and growth rate of total factor productivity. Country A has a closed capital account while country B has no capital account restrictions whatsoever. Assume that both countries are small and that the world interest rate r? is lower than r, the domestic interest rate in country A. Since country B is open and faces the world interest rate, r?, country B will have a higher ratio of capital to effective labor than country A. Because both coun- tries are in steady state and have identical growth rates of total factor productivity, they will also have identical growth rates of GDP per capita (although B has a higher level of GDP per capita). Now consider the path of these two economies over the next twenty years under the following scenario: country A keeps all of its restrictions on capital in place for the first ten years but removes all restrictions on cap- ital inflows and outflows for the second ten years while country B, on the other hand, makes no change in its capital account poli- cy and remains completely open for the entire twenty-year period. The transition dynamics in this example follow immediately from the discussion in section 2. Country B continues to grow at its steady-state rate because nothing there has changed. To understand what happens to country A when it liberalizes, recognize that as soon as it does so it will face the same cost of capital as country B. Therefore, when country A reaches its new steady state, it will have the same ratio of capital to effective labor as country B. For country A to reach its new, higher, steady-state ratio of capital to effective labor, the growth rate of its capital stock must exceed its growth rate of effective labor throughout the transition to its new equilibrium. Since A began in a steady state where capital and effective labor were grow- ing at the same rate, the growth rate of A?s capital stock must increase temporarily. The temporary increase in the growth rate of A?s capital stock will also generate a tempo- rary increase in the growth rate of its GDP per capita. Since GDP per capita in A and B were growing at the same rate before the shock, it follows that A grows faster than B throughout A?s transition to its new steady state. Because A grows faster than B throughout the transi- tion, it also follows that A will have a higher average growth rate for the twenty-year peri- od under consideration than country B. What would we conclude about the impact of capital account openness on growth if, using data from multiple countries like A and B, we ran a regression of GROWTH on SHARE? The answer is that we would conclude that capital account openness has a negative impact on growth. To see why, remember that A?s capital account was open for ten years and closed for ten. This means that SHARE for country A is 0.5, a smaller number than the value of SHARE for country B, which is equal to 1. Since A has faster growth and a lower value of SHARE over the period than country B, it follows that regressing GROWTH on SHARE would produce a negative coeffi- cient on SHARE, with the attendant (and specious) conclusion that liberalization exerts a negative impact on growth. Yet in this example it is clear that country A?s deci- sion to liberalize capital inflows does lead to a temporary increase in growth. The strictly cross-sectional regression framework is sim- ply powerless to detect the effect. Although the cross-sectional approach does not test the predictions of the neoclassical 894 Journal of Economic Literature, Vol. XLV (December 2007) dec07_Article1 11/28/07 12:01 PM Page 894 growth model, a large number of papers use equation (6), or a close cousin, as their benchmark of analysis. Of the fourteen stud- ies of liberalization and growth summarized in table 3 of the survey by Prasad et al. (2003), twelve perform cross-sectional regressions. The corresponding numbers from other surveys are nine of ten in table 6 of Edison et al. (2004); eleven of twelve in table 1 of C?sar Calder?n, Norman Loayza, and Klaus Schmidt-Hebbel (2004); and twenty-two of twenty-five in M. Ayhan Kose et al. (2006). Given the prevalence of the cross-sectional regression approach in the literature on the macroeconomic effects of capital account liberalization, it is important to spell out what these papers do and the results they find. 3.2 Evidence Attempts to quantify the real effects of capital account liberalization using SHARE have not been kind to the Allocative Efficiency view. Most of the published papers in the cross-sectional genre find no effect of liberalization on investment or growth. In the most prominent example of such papers, Rodrik (1998) examines the relation between liberalization and growth in a sample of 100 developed and developing countries from 1975 to 1989. Specifically, Rodrik estimates regressions with average growth rates of GDP per capita on the left- hand side and SHARE on the right-hand side. His regressions show no statistically sig- nificant correlation between growth and SHARE. Since the theoretical channel from liberalization to growth operates through capital accumulation, Rodrik also regresses the average ratio of investment-to-GDP on SHARE and again finds no statistically sig- nificant correlation. He concludes that there is no evidence to suggest that countries with fewer restrictions on capital movements grow faster, or invest more, than countries with greater restrictions. One problem with the variable SHARE is that it is constructed from a binary measure of capital controls that treats all countries as either completely closed or completely open. As such, it does not allow for varying degrees of intensity across countries that are classified as open. In an attempt to address this prob- lem, Dennis Quinn (1997) uses the informa- tion in the AREAER to create a measure of the intensity of capital account openness. Quinn?s measure splits capital account trans- actions into two categories: capital account receipts and capital account payments. For each of the two categories, he assigns a coun- try a score between 0 and 2 based on a read- ing of the qualitative description of the AREAER reports. A score of 0 indicates the most restrictive of regimes, where payments for capital transactions are completely forbid- den, while a score of 2 indicates that capital transactions are not subject to any taxes or restrictions. Intermediate scores?broken into increments of 0.5?indicate intermediate levels of restrictiveness. Quinn then adds the scores for the two categories to create a vari- able called CAPITAL, which ranges from 0 to 4. Quinn?s indicators are available annually from 1950 to 1997 for twenty-one Organisation for Economic Co-operation and Development (OECD) countries. For the forty-three non-OECD countries in his sam- ple, they are available for four different years: 1958, 1973, 1982, and 1988. Quinn runs cross-sectional regressions of average annual growth rates on changes in CAPITAL and finds a positive and significant correlation. Quinn?s data provide a novel attempt at quantifying the intensity of capital account openness, but there are two reasons why they may be of limited use for understand- ing the impact of capital account liberaliza- tion on developing countries. First, Quinn?s data for non-OECD countries end in 1988 and, therefore, do not cover the most rapid period of capital account liberalization in the developing world (the late 1980s and early 1990s). Given the time span of his data, Quinn?s results may reflect the impact of the move from closed to open capital markets among the developed nations. For 895Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:01 PM Page 895 example, in 1973 eleven OECD countries and twenty-six developing countries scored 2 or lower on Quinn?s index. In 1988, the number of OECD countries with a score of 2 or lower fell to one. In contrast, the num- ber of developing countries with a score of 2 or lower rose to thirty-two. Second, Quinn, like Rodrik, also uses a strictly cross-section- al regression framework, so his results imply that easing restrictions on capital account transactions have permanent effects on growth. Since theory predicts only a tempo- rary effect, it is not clear how to interpret this result (more on this in section 4.1). Whatever the correct interpretation of Quinn?s results may be, in finding a positive effect of liberalization on growth, his study is an outlier among the most heavily cited papers that assess the macroeconomic impact of capital account liberalization using cross-sectional regressions. For instance, using data on a sample of devel- oped and developing countries from 1976 to 1995, Ross Levine and Sara Zervos (1998b) examine whether capital account liberalization has an effect on investment. Like Rodrik and Quinn, they also test for permanent effects, but use a different measure of liberalization. Specifically, Levine and Zervos ask whether countries experience a permanent increase in the growth rate of their capital stocks when their stock markets become more integrat- ed with the rest of the world. They argue that emerging stock markets became more integrated in the 1980s and use 1985 as a common break point across all countries. Next, they examine whether the growth rate of the capital stock of the countries in their sample increased after 1985. They find no evidence that this is the case and conclude that capital account liberalization has no effect on investment. Among the entire class of papers that test whether countries with more open capital accounts have permanently higher growth rates than those that do not, the bulk of evi- dence supports the findings of Rodrik (1998) and Levine and Zervos (1998b) over Quinn (1997).3 For example, of the ten studies of liberalization surveyed in Edison et al. (2002), only three find an unambiguously positive effect of liberalization on growth (see their table 6). Prasad et al. (2003) extend the Edison et al. survey to fourteen studies, but still find only three papers that uncover a significant positive relationship between international financial integration and economic growth. A number of papers debate whether the absence of significant evidence is a conse- quence of trying to find an unconditional correlation between openness and long-run growth when the relationship between the two variables is more complicated. Because countries with strong institutions stand to gain more from capital account liberaliza- tion, the relationship between growth and capital account policy may be a conditional one that depends on the strength of a coun- try?s institutions.4 Michael W. Klein and Giovanni P. Olivei (2007) explore this hypothesis using data on sixty-seven coun- tries from 1976 to 1995. They find a signifi- cant correlation between SHARE and growth, but the developed countries in the sample drive the results. There is no signifi- cant correlation between growth and SHARE for the non-OECD countries, lead- ing the authors to conclude that capital account liberalization only promotes finan- cial development when done in an environ- ment that is supportive in terms of its overall institutional development. Sebastian Edwards (2001) uses Quinn?s CAPITAL measure and the variable SHARE to conduct formal tests for a condi- tional relationship between capital controls and long-run growth rates in a sample of twenty industrial and emerging economies during the 1980s. To test whether the 896 Journal of Economic Literature, Vol. XLV (December 2007) 3 See also Alberto Alesina, Vittorio Grilli, and Gian Maria Milesi-Ferretti (1994), Grilli and Milesi-Ferretti (1995), Aart Kraay (1998), and Edison et al. (2002). 4 See Kose et al. (2006) for an extensive review of the literature on the conditional effects of liberalization. dec07_Article1 11/28/07 12:01 PM Page 896 impact of capital account policy varies according to a country?s level of develop- ment, Edwards interacts each country?s capital account openness measure with its level of GDP per capita in 1980. He finds evidence that an open capital account posi- tively affects growth after a country achieves a certain degree of economic development. Carlos Arteta, Eichengreen, and Charles Wyplosz (2001) take exception with Edwards?s results on the grounds that he (1) uses Quinn?s measure in 1973 and 1988 only; (2) weights observations by GDP per capita in 1985, thereby giving richer countries a higher weight; and (3) uses instruments that may not be appropriate. Using data on sixty- one countries from 1973 to 1992, Arteta, Eichengreen, and Wyplosz perform cross- sectional regressions using Ordinary Least Squares and Instrumental Variables. They find a fragile association between growth and capital account liberalization and conclude that there is little evidence to suggest that capital account liberalization has a larger impact in high- and middle-income countries than in poor developing ones. Like Arteta, Eichengreen, and Wyplosz (2001), Kraay (1998) also finds no evidence of a conditional relationship between capital account liberalization and growth.5 In fact, while future work may prove otherwise, it seems fair to say that the absence of signifi- cant results in the cross-sectional literature is not a consequence of failing to condition their tests on the level of institutional develop- ment.6 Rather, the next section demonstrates that the ostensible lack of an effect of liberal- ization on real variables stems from a number of more basic issues common to papers in the cross-sectional literature?those that test for the conditional and unconditional impact of liberalization on growth alike. 4. The Reasons Why Cross-Sectional Studies Find No Impact of Liberalization on Growth There are at least three reasons why stud- ies that use purely cross-sectional regres- sions find no impact of liberalization on growth (although not all three reasons apply to all studies). First, as mentioned earlier, cross-sectional studies test for a permanent effect of liberalization on growth when the- ory says that the impact should be tempo- rary. Second, many of the papers in this literature include both developed and developing countries in the sample when theory argues for a separate examination of the two types of countries.7 Third, the vari- able SHARE contains important sources of measurement error that hinder empirical attempts to capture the impact of liberaliza- tion on the real economy. I now discuss the empirical consequences of each of these three points. 4.1 What Does ?Temporary? Mean in the Context of the Neoclassical Model? The extent to which it really matters whether you test for a permanent versus temporary growth effect of liberalization depends on exactly what ?temporary? means in the context of the neoclassical model. Testing for a permanent effect of liberaliza- tion probably does little harm to the data if the growth rate of output remains substan- tially above its steady-state value for rough- ly the same length of time as the interval over which cross-sectional studies compute the growth rates used in their regressions (thirty years or so). If the period of elevated economic growth is significantly shorter, 897Henry: Capital Account Liberalization 5 Kraay?s paper also conducts within-country analyses similar in spirit to the policy-experiment approach dis- cussed in section 5. He finds no impact of liberalization, but the dates in his analysis come from a different means of identifying liberalization episodes. 6 This is not to deny the importance of institutional considerations. 7 Some of the papers in this literature do acknowledge differences between countries (see the discussion in sec- tion 3.3). But papers that do acknowledge the importance of country heterogeneity also use the SHARE variable to test for permanent effects of liberalization. dec07_Article1 11/28/07 12:01 PM Page 897 however, then the distinction between per- manent and temporary matters a great deal. The critical question, then, is the follow- ing: For how many years following a liber- alization is it reasonable to expect capital-deepening-induced growth that is large enough to detect with a cross-section- al growth regression? There are two cases to consider. The first case?instantaneous convergence?was dismissed in section 2 as an artifact of the neoclassical model. We now consider the second case?when the speed of transition is slower, in accordance with reality. There are a variety of formal ways to slow down the speed of transition. In the discus- sion which follows, I do so by treating the transition from ks.state to k?s.state as if it came about from a permanent increase in a closed country?s savings rate. The savings-rate assumption offers several advantages in determining how long the growth effect of liberalization remains detectable. First, an increase in the savings rate has the same qualitative implications as a liberal- ization: for a given r and r?, there exists a permanent increase in the savings rate such that the country?s new steady state ratio of capital to effective labor following the increase would be exactly the same as under a liberalization. Second, the savings-rate assumption facilitates a more concise exposi- tion than introducing adjustment costs (or some other theoretical friction) to slow down the installation of capital. Third, and most importantly, the speed of convergence under the savings-rate assumption is actual- ly slower than it would be in the canonical model of a small open economy with adjust- ment costs (see Barro and Xavier Sala-i- Martin 1995, chapter 2). A slower speed of convergence implies a longer period of time during which the economy?s output growth remains elevated above its steady-state value. This means that relative to other ana- lytical methods of slowing down the transi- tion dynamics, the savings-rate assumption employed in the next several paragraphs actually understates the importance of testing for a temporary effect and therefore over- states the statistical power of cross-sectional regressions. By how much does the country?s growth rate in each year following the liberalization exceed the growth rate that would have occurred had the liberalization never taken place? To answer this question, recall that the growth rate of output per worker at any point in time is: k .(t) .A(t)?Y(t) =  ?? + ?? . L? k(t) A(t) Before liberalization, the economy is in steady state with k .(t) ?k(t)? = 0 and .A(t) ?A(t)? = g. Be- cause liberalization has no impact on total factor productivity in the model, it follows that the deviation of the growth rate of out- put per worker from its steady-state value (g) in the aftermath of liberalization equals k .(t) ?k(t)?. To calculate the size of the out-put growth deviation at any point in time, we therefore need to know the entire time path of k(t) during the economy?s transition to its new steady state. Since we are treating con- vergence from ks.state to k?s.state as though it comes from an increase in the savings rate, standard arguments show that: (7) , where ? = (1 ? )(n + g + ?). Using equation (7) to compute the time path of k(t) requires values of three key parameters: ks.state?the steady-state ratio of capital to effective labor before liberaliza- tion, k?s.state?the postliberalization steady- state ratio of capital to effective labor, and ? ?the speed of convergence. Pinning down the parameters ks.state and k?s.state requires assumptions about the dif- ference between the autarky interest rate in the developing country and the interest rate in the rest of the world. I begin by assuming that the autarky interest rate in the developing country is twice as high as the world interest rate, with r = 0.16 and k t k k k e s state s state s state t( ) ( )= + ? ? . . . ? ? ? 898 Journal of Economic Literature, Vol. XLV (December 2007) dec07_Article1 11/28/07 12:01 PM Page 898 r? = 0.08. At these rates, ks.state = 2.15 and k?s.state = 4.63.8 The simplest way to pin down ? is to plug reasonable values for the capital share (), population growth (n), total factor produc- tivity growth (g), and the depreciation rate (?) into the formula ? = (1 ? )(n + g + ?). The combination of a capital share of one- third, population growth and TFP growth between 1 and 2 percent, and depreciation of 4 percent, places ? at about 4 percent per year. Alternatively, one can obtain the speed of convergence from cross-country growth regressions. Early estimates implied speeds of convergence of about 2 percent per year (N. Gregory Mankiw, David Romer, and David N. Weil 1992; Barro and Sala-i-Martin 1992). Subsequent work argues that the 2 percent estimates are inaccurate for a variety of methodological reasons (Francesco Caselli, Gerardo Esquivel, and Fernando Lefort 1996). Caselli, Esquivel, and Lefort find that the speed of convergence is about 10 percent per year when estimated appro- priately. In what follows, I perform calcula- tions with a speed of convergence in the mid-range of those used in the literature (? = 0.04). The basic conclusion?that there is no statistically detectable impact of liber- alization on growth over long horizons? remains intact for speeds of convergence at the high (? = 0.1) and low (? = 0.02) ends of the spectrum. With ? = 0.04, ks.state = 2.15, and k?s.state = 4.63, equation (7) implies that the average growth rate of k(t) in the first five years fol- lowing liberalization is large: 3.8 percent per year. Multiplying 3.8 by the elasticity of output with respect to capital (1/3), gives the average deviation of the growth rate of output per worker from its long-run steady- state value during the first five years postlib- eralization: 1.27 percentage points per year. After year 5, the impact of liberalization on growth drops off considerably. From years six through thirty, the average annual growth rate of k(t) is 0.0154 or 1.5 percent per year. This translates into an average annual growth rate of output per worker that is 0.0051, or 0.51 percentage points above its steady-state value. I focus on the growth impact in years six and beyond because this period constitutes the differ- ence between the policy experiment approach to measuring the impact of liber- alization (see figures 3 through 5) and the cross-sectional regression approach. Given that difference, the key question is the fol- lowing: Can a cross-sectional regression dis- tinguish a liberalization-induced growth effect of 0.51 percentage points per year from the usual noise in the data? Comparing the magnitude of the hypothetical growth effect (0.0051) to that of the standard errors we typically see in regressions of growth on SHARE provides a rough and ready answer to the question. For instance, Rodrik (1998) reports standard errors of 0.006 on the vari- able SHARE. Dividing 0.0051 by 0.006 gives a t-statistic of 0.85?less than half the magnitude required for statistical signifi- cance. In contrast, the t-statistic on the average effect over the first five years is 2.11 (.0127 divided by 0.006). Changing the speed of convergence has no material impact on the calculations. When ? = 0.02, the average annual growth effect of liberalization between years six and thirty is 0.0041, with a corresponding t-sta- tistic of 0.69. For ? = 0.1, the average annu- al growth effect over the same horizon is 0.0045, with a t-statistic of 0.75. Similarly, increasing the autarky rate interest rate in the developing country to three times the level in the rest of the world also does not alter the results. Now, it is far from clear that the rate of return to capital in developing countries is even twice as high as in the 899Henry: Capital Account Liberalization 8 From the first order condition for investment, equa- tion (4), it follows that . Plugging in r = 0.16, ? = 0.04, and  = 1/3 gives ks.state = 2.15. The analogous set of calculations gives the value for k?s.state. k r s state. = +? ?? ? ?? ?? ? ? 1 1 dec07_Article1 11/28/07 12:01 PM Page 899 developed world (see section 6.1.2), so play- ing with a threefold differential begins to strain credulity. Nevertheless, I do so to illustrate the likely futility of the cross- sectional approach. Holding ? constant at 0.04 and increasing r to three times the level of r?, the average annual growth effect rises to 0.008. With standard errors on the order of 0.006, a growth effect of 0.008 is still too small to attain statistical significance in a cross-sectional regression. One potential concern about the calcula- tions in this subsection is that they focus on growth in GDP per worker while cross- sectional studies of liberalization focus on growth in GDP per capita. If the growth rate of the labor force exceeds the growth rate of the population, then calculations of statisti- cal significance using growth in GDP per worker will be smaller than calculations based on growth in GDP per capita, thereby understating the true statistical power of cross-sectional regressions. Data on the growth of the labor force and population in developing countries speak to the issue. Between 1960 and 1997, the pop- ulation in low- and middle-income develop- ing countries grew at an average annual rate of 2.2 percent per year while the labor force grew at 2.1 percent per year (World Bank 1999). Since the growth rate of the popula- tion actually exceeded the growth rate of the labor force, if anything, calculations using output per worker actually overstate the impact of liberalization on the growth rate of GDP per capita. In other words, the calculations in this subsection underscore the empirical rele- vance of the critique that cross-sectional regressions are unlikely to capture the true impact of capital account liberalization because they test for permanent instead of temporary effects. Although the impact of liberalization on growth remains positive for a long time, the statistically significant por- tion of that impact occurs in the immediate- to-near aftermath of liberalization. Trying to detect this near-term change by running cross-sectional regressions on average growth rates over long time horizons is not an effective empirical strategy. 4.2 Country Heterogeneity The second reason why cross-sectional studies may fail to find significant effects of liberalization is that they lump both devel- oped and developing countries together in their samples. Including both sets of coun- tries increases sample size, but doing so without employing an empirical methodolo- gy that explicitly recognizes the fundamen- tal theoretical difference between developed and developing countries under- mines a study?s ability to interpret the data. The neoclassical model predicts that capital account liberalization will have a different impact on a developing country than on a developed one. Developing countries are capital poor and should experience net cap- ital inflows, a permanent fall in their cost of capital, and a temporary increase in growth when they remove all restrictions on capital flows. In contrast, capital-rich, developed countries should experience exactly the opposite effects.9 For instance, with both developing and developed countries includ- ed in their samples, the Rodrik (1998) and Levine and Zervos (1998b) results may sug- gest that capital account liberalization has no effect on investment and growth, but the results may also reflect the opposing effects of liberalization in the developing- and developed-country subsamples. Developing and developed countries also implemented the process of capital account liberalization at different times. Developed countries liberalized in the late 1970s and early 1980s (Helmut Reisen and Bernhard Fischer 1993). Most developing countries did not liberalize until the late 1980s and early 1990s (Donald J. Mathieson and 900 Journal of Economic Literature, Vol. XLV (December 2007) 9 Although the rich country?s GDP falls because it exports capital, its GNP will rise as it receives the income that accrues to that capital. I discuss the Lucas (1990) paradox in section 5. dec07_Article1 11/28/07 12:01 PM Page 900 Liliana Rojas-Su?rez 1993). The differences in liberalization dates highlight an addition- al limitation of the Rodrik paper. Since developing-country liberalizations com- menced in the late 1980s and early 1990s, empirical attempts to determine whether liberalization boosted investment and growth should use data that extend into the 1990s. Because Rodrik?s data end in 1989, they exclude developing-country economic performance in precisely the period when we would expect to see the effects of liber- alization. Therefore, it is not clear that Rodrik?s data tell us anything about the impact of capital account liberalization on growth in developing countries.10 4.3 Measurement Error Problems with Liberalization Indicators Cross-sectional studies conclude that capi- tal account liberalization has no real effects because they do not find any correlation between economic growth and broad meas- ures of capital account openness such as SHARE. As discussed in section 3.1, studies construct the variable SHARE using the judg- ment on capital account openness contained in line E2 of the IMF?s AREAER. In turn, the judgment in the AREAER is based on the restrictions applied to an exhaustive list of all possible capital account transactions.11 When the IMF changes its assessment of a country?s capital account openness, the AREAER pro- vides no information on the specific aspect of the capital account the country liberalized. Because the underlying data provide no indi- cation of what has been liberalized, neither can any index that is based on such data. Without any indication of what drives the variation in SHARE, it is also unclear how to map that variation to a well-articulated model for the purpose of empirical estimation. For example, AREAER does not tell whether the IMF made a change in the judgment about the openness of a country?s capital account because of an easing of restrictions on capital inflows or capital out- flows. This distinction matters. Theory pre- dicts that when a capital-poor country liberalizes capital inflows it will experience a permanent fall in its cost of capital and a temporary increase in the growth rate of its capital stock and GDP per capita. In princi- ple, if that same developing country were to liberalize capital outflows, nothing would happen to its cost of capital, investment, or GDP. If the goal is to understand whether capi- tal account liberalization has real effects, then theory tells us that the right question to ask is the following: Do we see a fall in the cost of capital and a temporary increase in investment and growth after developing countries liberalize restrictions on capital inflows? The most direct and transparent way of answering this question is to examine whether the behavior of the cost of capital, investment, and growth in the immediate aftermath of liberalization differs from its behavior in the immediately preceding peri- od. We now discuss a class of papers which attempt to do just that. 5. The Policy-Experiment Approach to Capital Account Liberalization Examining the behavior of real variables before and after the removal of restrictions on capital inflows requires information about the date on which the restrictions were lifted. The policy-experiment approach grapples with the complexity of identifying liberalization dates by narrowing the scope of the problem. Instead of trying to deter- mine the date on which the entire capi- tal account was judged to be open, as in the cross-sectional approach, the policy- experiment literature tries to identify the first point in time that a country liberalizes a specific aspect of its capital account policy. 901Henry: Capital Account Liberalization 10 Figure 5A of Kose et al. (2006) looks at the data from 1985 to 2004. Like Rodrik, they find no significant rela- tionship between growth and capital account openness. The discussion in section 4.1 of this paper explains why. 11 For an extensive enumeration of these transactions, see the appendix of Kose et al. (2006). dec07_Article1 11/28/07 12:01 PM Page 901 One example of liberalizing a specific aspect of the capital account is a decision by a country?s government to permit foreigners to purchase shares of companies listed on the domestic stock market. Liberalizing restric- tions on the ownership of domestic shares enables foreign capital to flow into a part of the country?s economy from which it was pre- viously prohibited. Just such a policy change occurred repeatedly in the late 1980s and early 1990s as a number of developing coun- tries opened their stock markets to foreign investors for the first time (Stijn Claessens and Moon-Whoan Rhee 1994). These stock market liberalizations constitute discrete removals of barriers to capital inflows that bear close resemblance to the model in sec- tion 2. As the closest empirical analogue to the textbook example, stock market liberal- izations provide real-world policy experi- ments that allow researchers to evaluate the empirical validity of the two competing views of capital account liberalization outlined in the introduction. Since the stock market is forward looking, the critical question is when does the market first learn of a credible, impending liberal- ization? In principle, obtaining this informa- tion simply involves identifying the date on which the government declares that foreign- ers may purchase domestic shares. In prac- tice, the liberalization process is not so transparent. The literature uses official poli- cy decree dates when they are available, but in many cases there is no obvious govern- ment declaration or policy decree to which one can refer (see, for example, Levine and Zervos 1994). When policy decree dates are not available, papers in this area typically pursue two approaches. First, many countries initially permitted foreign ownership through closed-end coun- try funds. Since one presumably needs gov- ernment permission to establish a fund, the first country fund establishment date serves as a proxy for the official implementation date. The second way of indirectly capturing official implementation dates is to monitor the International Finance Corporation?s Investibility Index (Claessens and Rhee 1994). The investibility index is the ratio of the market capitalization of stocks that for- eigners can legally hold to total market capi- talization. As such, a large jump in the investibility index provides indirect evidence of an official liberalization. In the end, the policy-experiment literature defines the date of a country?s first stock market liberalization as the first month with a verifiable occur- rence of any of the following: liberalization by policy decree, establishment of the first country fund, or an increase in the investibil- ity index beyond a certain threshold.12 An important limitation of the policy- experiment approach is that relatively few developing countries have a stock market, publish reliable stock market data, and also implemented a liberalization. Hence, the number of countries that can feasibly be included under the stock-market-liberaliza- tion-based implementation of the policy- experiment approach is much smaller than in purely cross-sectional studies (an issue I discuss further in section 6.1.3). Table 1 presents a list of eighteen devel- oping countries that liberalized their stock markets between 1986 and 1993. The third column of the table indicates that the modal indicator of liberalization for this sample is the establishment of a closed-end country fund. Relative to the broader indicators employed elsewhere in the literature, the establishment of a country fund, in particu- lar, and stock market liberalizations in gen- eral may seem like a narrow way to define capital account liberalization. But again, it is precisely the narrowness of stock market liberalizations that make them more useful than broad indicators of capital account liberalization for testing the implica- tions of the theory. For all of the reasons explained in section 4.3, broad indicators, 902 Journal of Economic Literature, Vol. XLV (December 2007) 12 See Geert Bekaert and Campbell R. Harvey (2000) and Peter Blair Henry (2000b) for further details on determining the dates of stock market liberalizations. dec07_Article1 11/28/07 12:01 PM Page 902 such as the IMF?s, introduce significant measurement error. Since measurement error reduces the statistical power of any regression, it is important to focus on policy experiments where the true variation in the data is large relative to the noise. Stock mar- ket liberalizations provide just such experi- ments, because they constitute a radical shift in the degree of capital account openness (Jeffrey A. Frankel 1994). In addition to identifying episodes of large changes in capital account openness, focus- ing narrowly on stock market liberalization offers another advantage. In contrast with the unspecified easing of restrictions indi- cated by movements in the variable SHARE, there is no theoretical ambiguity about the expected impact of lifting restric- tions on the flow of capital into the stock market of a developing country. Since stocks are risky, however, using stock market liber- alizations to examine the impact of capital account liberalization on economic per- formance raises issues of uncertainty not addressed by the deterministic framework outlined in section 2. 903Henry: Capital Account Liberalization TABLE 1 DATES OF STOCK MARKET LIBERALIZATIONS AND MAJOR ECONOMIC REFORMS Country Year of Means of Stabilization Trade Privatization Brady Plan Liberalization Liberalization Program Liberalization Debt Relief Argentina November 1989 Policy Decree November 1989 April 1991 February 1988 April 1992 Brazil March 1988 Country Fund January 1989 April 1990 July 1990 August 1992 Chile May 1987 Country Fund August 1985 1976 1988 NA Colombia December 1991 Policy Decree NA 1986 1991 NA India June 1986 Country Fund November 1981 1994 1991 NA Indonesia September 1989 Policy Decree May 1973 1970 1991 NA Jordan December 1995 Policy Decree May 1994 1965 January 1995 June 1993 Malaysia May 1987 Country Fund NA 1963 1988 NA Mexico May 1989 Policy Decree May 1989 July 1986 November 1988 September 1989 Nigeria August 1995 Policy Decree January 1991 NA July 1988 March 1991 Pakistan February 1991 Policy Decree September 1993 2001 1990 NA Philippines May 1986 Country Fund October 1986 November 1988 June 1988 August 1989 South June 1987 Country Fund July 1985 1968 NA NA Korea Taiwan May 1986 Country Fund NA 1963 NA NA Thailand September 1987 Country Fund June 1985 Always Open 1988 NA Turkey August 1989 Policy Decree July 1994 1989 1988 NA Venezuela January 1990 Policy Decree June 1989 May 1989? April 1991 June 1990 Zimbabwe June1993 Policy Decree September 1992 NA 1994 NA Notes: The capital account liberalization dates identified in this table are the dates on which the eighteen countries in Column 1 eased restrictions prohibiting foreign ownership of domestic stocks. The liberalization dates in Column 2 are an amalgamation of those in Henry (2000a), Bekaert and Harvey (2000), Kim and Singal (2000), and Levine and Zervos (1994). Column 3 indicates the form of the liberalization, policy decree or country fund opening; Section 5 of the paper contains a detailed description of both forms. Columns 4 through 7 list the dates of major economic reforms that occurred around the same time as the capital account liberalizations. The stabilization pro- gram dates in column 4 come from Henry (2002) and various issues of the IMF Annual Reports. Column 5 lists trade liberalization dates from Sachs and Warner (1995). The privatization dates in column 6 come from the Privatization Data Base maintained by the World Bank. Finally, column 7 lists the month and year that each country received debt relief under the Brady Plan. The debt relief dates come from Cline (1995), Lexis Nexis, and various issues of the Economist Intelligence Unit. ?Venezuela reversed its trade liberalization reforms in 1993. dec07_Article1 11/28/07 12:01 PM Page 903 5.1 Extending the Theoretical Framework to Incorporate Uncertainty This subsection shows that the funda- mental predictions of the deterministic neoclassical model of capital account liber- alization extend to the risky-asset setting of the policy-experiment approach.13 In the absence of uncertainty, it is optimal to invest until the marginal product of capital equals the interest rate. When the payoff from investing in capital is risky, the inter- est rate is no longer the relevant hurdle rate for investment. Optimality in a risky world requires that investment take place until the expected marginal product of capital equals the interest rate plus a risk premium to compensate for the uncertain return to capital. For a country that has not yet opened its stock market to the rest of the world, this means that the first-order condition for investment is f ?(k)e = r + ? + ?, where f ?(k)e denotes the expected marginal product of capital, r is the domestic interest rate, ? is the aggregate equity premium, and ? is the rate of depreciation of the capital stock. In the context of the capital asset pricing model (CAPM), ? is equal to the price of risk, ?, times the variance of the market return, so that the first-order condition for investment under uncertainty is: (8) (r?M) + ?, where r?M is the risky return to investing in the market. Starting from the equilibrium in equation (8), the issue is whether liberalization reduces the cost of capital as it does when there is no uncertainty.14 The first compo- nent of the cost of capital is the domestic ? = +f k r Vare( ) ? interest rate (r), which responds the same way to liberalization as in the case of perfect certainty?it falls to the world interest rate (r?). The second component of the cost of capital is the aggregate equity premium. Whereas the equity premium before liberal- ization is equal to the price of risk times the variance of the local market return, the equi- ty premium after liberalization is equal to the price of risk times the covariance of the local market with the rest of the world (see Stulz 1999b). The change in both components of the cost of capital means that following liberal- ization the first-order condition for invest- ment becomes: (9) (r?M,r?W) + ?. Because the world interest rate (r?) is less than the domestic interest rate (r), compar- ing the right-hand-side of equation (9) with that of (8) indicates that liberalization will reduce the cost of capital if the variance of the domestic market return is greater than its covariance with the world market. Historical stock returns show that this condi- tion holds for emerging stock markets (Stulz 1999b). Stock market liberalization there- fore reduces the cost of capital, and the attendant predictions about investment and growth follow accordingly. 5.2 Evidence: The Cost of Capital, Investment, and Growth We do not observe the cost of capital directly, but we do observe stock prices. Because stock prices and the cost of capital move inversely, we should observe a one- time revaluation of the stock market if lib- eralization reduces the cost of capital. In a sample of twelve emerging economies that liberalized between 1986 and 1991, the average country experienced a revaluation of 26 percent in real dollar terms (Henry 2000b). E. Han Kim and Vijay Singal (2000) and Rodolfo Martell and Stulz (2003) also document a sharp rise in stock prices in ? = +f k r Cove( ) ? ? 904 Journal of Economic Literature, Vol. XLV (December 2007) 13 The discussion in this section borrows heavily from Jeremy I. Bulow and Summers (1984) and Ren? M. Stulz (1999b). 14 For expositional convenience, I assume that liberal- ization has no effect on the marginal product of capital. Section 7.1 discusses what happens when this assumption does not hold. dec07_Article1 11/28/07 12:02 PM Page 904 conjunction with liberalizations.15 Consis- tent with these revaluations, liberalizations also coincide with an average fall in dividend yields of 5 to 75 basis points (Bekaert and Harvey 2000).16 While the magnitude of the effect differs slightly depending on the sam- ple of countries and the exact liberalization dates applied, there is broad agreement across all policy-experiment studies that lib- eralization reduces the cost of capital. Since liberalizations reduce the cost of capital, some projects that were negative net present value before liberalization become worthwhile to undertake once the govern- ment permits foreigners to hold domestic shares. Consequently, the rate of capital accumulation should rise until the marginal product of capital is driven down to the country?s new, lower cost of capital. If invest- ment increases temporarily, so too should the growth rate of GDP per capita. The policy-experiment approach tests the investment and growth predictions by esti- mating panel regressions of the following form: (10) . The variable GROWTH generically denotes the growth rate of the variable of interest: the capital stock in the case of the investment regressions and GDP per capita. The variable LIBERALIZE is a country-spe- cific dummy variable that takes on the value one in the year that country i liberalizes (year [0]) and each of the subsequent five years (year [+1] through year [+5]). Equation (10) highlights two important dimensions along which the policy-experi- ment approach differs from the traditional a LIBERALIZEit it+ +1 ? ? GROWTH ait = 0 cross-sectional regression approach typified by equation (6). First, equation (10) uses every annual observation of each country?s growth rate to exploit the variation in growth rates within countries over time. In contrast, equation (6) uses a single number on the left-hand-side?the average growth rate over the time period in question. Second, the definition of the variable LIB- ERALIZE ensures that equation (10) tests for a temporary effect of liberalization on growth instead of a permanent one as in equation (6). If liberalization causes a tem- porary increase in growth, then estimating equation (10) should produce a positive and significant estimate of the coefficient on the variable LIBERALIZE. In other words, the coefficient on LIBERALIZE measures the average annual abnormal growth during the five-year window subsequent to the imple- mentation of a liberalization. It is natural to ask why policy-experiment studies choose a five-year liberalization window and how sen- sitive the results are to that choice. It follows from the discussion in section 4.1 that the results should be sensitive to the choice of window length and that a short window of five years or less is theoretically appropriate. The behavior of both investment and GDP in the aftermath of liberalizations sup- ports the predictions of the neoclassical model. For the eighteen countries in table 1, the raw growth rate of the capital stock rises from an average of 5.4 percent per year in the five years preceding liberalization to an average of 6.5 percent in the five-year postliberalization period (Henry 2003). Similarly, in a sample of eleven emerging economies, stock market liberalization leads to a 22-percentage-point increase in the growth rate of real private investment (Henry 2000a). Of the eleven countries in the sample, only two do not experience abnormally high rates of investment in the first year after liberalization, and only one in the second year after liberalization. Bekaert, Harvey, and Christian Lundblad (2005) find that the growth rate of GDP per capita 905Henry: Capital Account Liberalization 15 A related literature shows that firms in emerging markets experience stock price increases when they list their shares on stock exchanges in developed countries (see G. Andrew Karolyi 1998 and Stephen R. Foerster and Karolyi 1999). 16 See also Vihang R. Errunza and Darius P. Miller (2000). dec07_Article1 11/28/07 12:02 PM Page 905 increases by a percentage point per annum. An important question that I defer until section 6.3 is whether the growth and investment numbers are mutually consistent. Figures 3 through 5 summarize the prin- cipal findings of the policy-experiment liter- ature by plotting the cost of capital, investment, and growth in event time for the eighteen countries listed in table 1.17 When developing countries liberalize capi- tal flows into their stock markets, the cost of capital falls and investment increases along with the growth rate of GDP per capita. In contrast to the cross-sectional studies reviewed in section 3, the evidence from the policy-experiment approach demonstrates that the real-world impact of capital account liberalization is quite consistent with the Allocative Efficiency view. 5.3 Economic Reforms The policy-experiment evidence is also consistent with interpretations in which lib- eralization plays no causal role. Figure 3, which uses the dividend yield as a proxy for the cost of capital, helps to demonstrate the point. The theoretical justification for inter- preting a fall in the dividend yield as a reduc- tion in the cost of capital comes from Myron J. Gordon (1962). In the Gordon model, the 906 Journal of Economic Literature, Vol. XLV (December 2007) 17 The data used to construct figure 4, the average growth rate of the real capital stock for the eighteen coun- tries in listed in table 1, come from Barry P. Bosworth and Susan M. Collins (2003). Figure 3. The Cost of Capital Falls When Countries Liberalize the Capital Account Notes: Figure 3 plots the average value of the aggregate dividend yield in event time for the eighteen countries listed in table 1. The x axis denotes time in years relative to the liberalization year: ?0? denotes the year in which the liberalization took place, ??1? denotes the year prior to liberalization, ?+1? denotes the year after liberalization, and so on. 6 5.5 5 4.4 4 3.5 3 2.5 2 ?5 ?4 ?3 ?2 ?1 0 1 2 3 4 5 Year Relative to Liberalization D iv id en d Yi eld (P erc en t) dec07_Article1 11/28/07 12:02 PM Page 906 dividend yield equals the cost of capital, ?, minus the (constant) expected future growth rate of dividends: D /P = ? ? ge.18 If the growth rate of dividends does not change with liberalization, then a fall in the dividend yield implies a fall in the cost of capital. Because liberalizations are part of a general process that involves substantial macroeco- nomic reforms, however, there is a strong possibility that large changes in expected future growth rates do occur at the same time that countries liberalize. Columns 4 through 7 of table 1 demon- strate that the timing of economic reforms? inflation stabilization, trade liberalization, and privatization?poses a significant chal- lenge to the central policy-experiment find- ings reviewed in section 5.2. Stabilizing inflation drives up asset prices, investment, and output (Michael Bruno and William Easterly 1998; Guillermo A. Calvo and Carlos A. V?gh 1999; Easterly 1996; Fischer, Ratna Sahay, and V?gh 2002; Henry 2000a, 2002). When a country liberalizes trade, domestic production shifts toward the coun- try?s comparative advantage, thereby increas- ing its output for a given stock of capital and labor. Accordingly, trade liberalization also generates a temporary increase in the growth rate of GDP and a permanent increase in the level (Anne O. Krueger 1997, 1998; Jeffrey D. Sachs and Andrew M. Warner 1995; Frankel and Romer 1999; Romain Wacziarg and Karen Horn Welch 2003). Privatization raises the efficiency and value of formerly state-owned enterprises (William L. Megginson and Jeffry M. Netter 2001). Finally, in return for adopting reforms, many countries received debt relief under the Brady Plan. Debt relief also has a positive impact on stock prices, investment, and out- put in certain emerging economies (Serkan Arslanalp and Henry 2005, 2006a, 2006b). 907Henry: Capital Account Liberalization Figure 4. Investment Booms When Countries Liberalize the Capital Account Notes: Figure 4 plots the average growth rate of the aggregate capital stock in event time for the eighteen countries listed in table 1. The x axis denotes time in years relative to the liberalization year: ?0? denotes the year in which the liberalization took place, ??1? denotes the year prior to liberalization, ?+1? denotes the year after liberalization, and so on. 7.5 7 6.5 6 5.5 5 4.5 4 G ro wt h Ra te o f t he C ap ita l S to ck (P erc en t) Year Relative to Liberalization ?5 ?4 ?3 ?2 ?1 0 1 2 3 4 5 18 Recall that ? = r + ?Var(r~M) before liberalization and ? = r? + ? COV(r~M,r~W) after liberalization. dec07_Article1 11/28/07 12:02 PM Page 907 Papers in the policy-experiment literature approach adopt different methods of trying to disentangle the impact of liberalization from that of other economic reforms. Some papers treat reforms and liberalization symmetrical- ly, constructing dummy variables that take on the value one in the year a reform program of a particular type begins and each of the five subsequent years.19 Using the variable REFORMS as a compact stand-in for the full complement of the reform dummies, the key regression then simply becomes an extension of equation (10): (11) + . Other papers, instead of conducting a before-and-after analysis of reforms, use REFORMSit i+ ? t GROWTH a a LIBERALIZEit it= +0 1 ? continuous proxies such as the level of infla- tion and trade openness.20 Kose et al. (2006) discuss the relative merits of both approach- es on page 19 of their survey. 5.4 Policy Endogeneity The impact of stock market liberalization on stock prices, dividend yields, investment, and growth remains statistically and eco- nomically significant after controlling for reforms (Bekaert and Harvey 2000; Bekaert, Harvey, and Lundblad 2005; Henry 2000a, 2000b). But an even more difficult interpre- tation problem remains. Do liberalizations drive up stock prices, investment, and eco- nomic growth or does causation run the other way round? Because liberalizing the stock market during bad times may draw 908 Journal of Economic Literature, Vol. XLV (December 2007) Figure 5. The Growth Rate of GDP Per Capita Rises When Countries Liberalize the Capital Account Notes: Figure 5 plots the average growth rate of GDP per capita in event time for the eighteen countries listed in table 1. The x axis denotes time in years relative to the liberalization year: ?0? denotes the year in which the liberalization took place, ??1? denotes the year prior to liberalization, ?+1? denotes the year after liberalization, and so on. 5 4.5 4 3.5 3 2.5 2 1.5 1 .5 0 G ro wt h Ra te o f G D P Pe r C ap ita (P erc en t) Year Relative to Liberalization ?5 ?4 ?3 ?2 ?1 0 1 2 3 4 5 19 See, for example, Henry (2000b). 20 See, for instance, Bekaert and Harvey (2000) and Bekaert, Harvey, and Lundblad (2005). dec07_Article1 11/28/07 12:02 PM Page 908 criticism for selling off the country at fire sale prices, politicians may be may be more inclined to open up when times are good (Summers 1994). Liberalization may also be the country-level equivalent of an initial public offering (Martell and Stulz 2003). You don?t have an IPO when things are going poorly; you have one when valuations are high because of exceptional growth opportu- nities that require financing or because the shares are simply overvalued. To the extent that the economy?s current and expected future performance influence policymakers? decision to liberalize, the endogeneity prob- lem is quite real. In trying to think of a legitimate instru- mental variable for this problem, one is struck by the distinct lack of variables that are correlated with the decision to liberal- ize but uncorrelated with the stock market or macroeconomic fundamentals. Changes of political regime are a natural candidate but, in addition to being correlated with lib- eralizations, political changes also bring new economic programs that influence market expectations. IMF programs do not make good instruments because, contrary to popular wisdom, the IMF did not insist on capital account liberalization as an ele- ment of conditionality during its programs at the time (see Fischer 1998). Because none of the papers in the policy-experiment literature are fully able to dispel with con- cerns about endogeneity, the economically large and statistically significant correla- tions between liberalization and real vari- ables uncovered by the policy-experiment literature require a measured interpreta- tion. 6. Problems with the Policy-Experiment Approach The strength of the policy-experiment approach is that it provides a clear picture of what happens to developing countries when they liberalize capital inflows. The weakness of the policy-experiment approach is that it raises more questions than it answers (even putting aside the issue of endogeneity raised at the end of section 5). This section of the paper addresses three of those questions. 6.1 Why Is the Financial Impact of Liberalization So Small? Liberalization drives up the value of the stock market by roughly 30 percent in real dollar terms. This is a large number relative to most event studies of the stock market conducted using U.S. data (see A. Craig MacKinlay 1997). But it is small in compari- son with theoretical predictions about the impact of capital account liberalization. Casual observation suggests that developing countries have much lower ratios of capital to labor than rich countries. If the neoclassical model provides a reasonable description of the world, then we should observe a much larger financial impact when developing countries permit capital flows into their stock markets (Stulz 1999a, 1999c, 2005). There are at least three possible explana- tions for the relatively modest increase in stock prices (decrease in the cost of capital) associated with liberalizations. The first is that the policy-experiment approach treats liberal- ization as a one-shot event when it is really an incremental process. The second explanation is that the rate of return to capital in develop- ing countries may not really be that much higher than in the developed world (Robert E. Lucas 1990; Mankiw 1995). The third is that developing countries do have substantial- ly higher rates of return but the presence of capital market imperfections causes much of the return differential to persist, even in the face of free capital flows. The next three sub- sections elaborate on each of these points in turn. 6.1.1 Treating Liberalization as a Binary Process May Understate Its True Impact The policy-experiment literature esti- mates the change in the cost of capital that occurs the first time a country opens its stock 909Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 909 market to foreign investors. This approach presents three problems that could poten- tially explain the relatively small valuation effects. First, stock price effects at the date of liberalization may be small if the liberal- izations were anticipated. Stock prices do increase in anticipation of liberalization but the preliberalization run-up in valuations is not large enough to account for the puzzle (Henry 2000b). Second, investors may perceive that the liberalization will be reversed. If foreign investors believe that the government will reverse the policy at some point in the future, then they will value domestic shares less highly than if the policy change was fully credible. Although Malaysia in 1997 is the only case of an actual reversal (see section 6.2), liberalizations may suffer from a sort of peso problem that reduces the magnitude of the revaluation.21 The third problem is that countries sel- dom move from having a completely closed stock market to one that is fully open. Most countries undergo several stock market lib- eralizations subsequent to the first. South Korea provides a good example. South Korea began allowing foreigners very limit- ed access to its stock market through closed- end country funds as early as 1982, but did not start lifting its statutory ceiling on for- eign investment until 1992.22 Even then, South Korea did not finish lifting its regula- tory ceiling until 1998 (Inseok Shin and Changyun Park forthcoming). If a country takes several episodes to open its stock market completely, then look- ing exclusively at the stock market response to its initial opening may understate the total financial impact of the liberalization process (Stulz 1999b). One way of address- ing the issue is to estimate the market response to every opening, sum up all of the responses, and use the resulting number as a proxy for the total liberalization effect. It turns out that stock market responses to lib- eralizations subsequent to the first are fairly small, so the incremental-opening hypothe- sis cannot account for the small financial impact of liberalization. Another alternative is to estimate the total financial impact using continuous measures of liberalization such as the IFC?s investible index (Edison and Francis E. Warnock 2003). Again, the investible index indicates the fraction of the domestic stock market capitalization that foreign investors may legally own. In contrast to the policy-experi- ment approach, which estimates the one- time response of the stock market at the time of a large, discrete change in the investible index, the continuous approach regresses returns or dividend yields on the level of the investibility index over the entire sample. The continuous approach also reveals a small change in the cost of capital. The estimated cumulative fall in the divi- dend yield is about 140 basis points (Frank De Jong and Frans A. De Roon 2005). All in all, the evidence suggests that the dichotomy between the binary measure of liberalization used by policy-experiment papers and the more gradual nature of the process cannot explain the relatively small financial impact of liberalizations. 6.1.2 Poor Institutions in Developing Countries May Reduce the Return to Capital Large differences in capital-to-labor ratios across countries imply large differ- ences in rates of return to capital only if all countries have the same level of total factor productivity. In standard production theory, the level of total factor productivity is cap- tured by the parameter A, which is usually taken to be an index of technology or ideas. But it is important to remember that, for a 910 Journal of Economic Literature, Vol. XLV (December 2007) 21 See Rawi Abdelal and Laura Alfaro (2002, 2003) for a detailed discussion of the imposition of capital controls in Malaysia. The reversal of open capital markets by Thai authorities in December 2006 lasted less than 24 hours, but the sharp impact it had on valuations reinforces the point. 22 See Keith Park and Antoine W. Van Agtmael (1993), Margaret M. Price (1994), and Ian M. Wilson (1992). dec07_Article1 11/28/07 12:02 PM Page 910 given stock of technology, total factor pro- ductivity measures the general efficiency with which an economy transforms capital and labor into output (Solow 2001). Holding technology constant, many factors may influence the general level of efficiency. For instance, weak institutions and inap- propriate government regulation can distort economic decision making, reduce total fac- tor productivity, and lead to lower rates of return than would otherwise prevail (Arnold C. Harberger 1998). The rate of return to capital in emerging economies may not be much higher than in rich countries if emerg- ing-market governments fail to create ?an environment that supports productive activ- ities and encourages capital accumulation, skill acquisition, invention and technology transfer? (Robert E. Hall and Charles I. Jones 1999). To illustrate the point, table 2 demonstrates the gap between the G-7 countries and the eighteen emerging economies that liberalized their stock markets using some frequently employed measures of the quality of econom- ic institutions. Column 2 shows that, accord- ing to the Hall and Jones (1999) measure of social infrastructure, the median G-7 country ranks fourteenth of 130 countries, while the median emerging economy ranks sixty- fourth. Columns 3 through 5, which present analogous measures using the Heritage House?s Index of Economic Freedom, the World Bank?s Ease of Doing Business Index, and the World Economic Forum?s Global Competitiveness Index, display a similar pattern. In other words, the relatively small finan- cial effects of liberalization may simply indi- cate that the return to capital in developing countries is not that much higher than in the developed world. Recent work produces indirect estimates of rates of return that sup- port this view (see, for example, Caselli 2005 and Caselli and James Feyrer 2007). There are still relatively few studies that attempt to measure the rate of return to capital in developing countries directly. In my view, this continues to be an important gap in the literature. 6.1.3 Capital Market Imperfections A third explanation for the relatively small financial impact of liberalization is that developing countries really do have higher rates of return to capital but the presence of capital market imperfections, such as agency problems, asymmetric information, and poor investor protection, leads to a persistent return differential between rich and poor countries (Stulz 2005). Investor protection matters. Where accounting standards and enforcement bod- ies do not exist to restrain insiders, resources invested in a company may be wasted on superfluous managerial perks or even stolen outright (Michael C. Jensen 1986). The insiders may be controlling shareholders, such as a founding family, a firm?s top man- agers, or both. Because outside investors know less than insiders about the firm?s prospects and the behavior of its managers, they will demand higher returns or simply not invest. Strong investor protection can help to mit- igate problems of agency and asymmetric information by providing minority sharehold- ers with mechanisms to restrain insiders. A large body of research demonstrates that the degree to which a country?s laws protect the legal rights of minority shareholders exerts a significant influence on the size and robust- ness of capital markets (Rafael La Porta et al. 1997, 1998, 2002). Firms located in countries with strong investor protection have greater access to external finance, invest more, and have higher valuations than their counter- parts in countries with weak investor protec- tion (La Porta et al. 1997, 1998, 2002; Andrei Shleifer and Robert W. Vishny 1997; Raghuram G. Rajan 2000).23 911Henry: Capital Account Liberalization 23 In a related paper, Simon Johnson et al. (2000) show that measures of investor protection do a better job of explaining stock market declines during the Asian Crisis than do standard macro variables. dec07_Article1 11/28/07 12:02 PM Page 911 Conversely, investors will stay away from countries in which investor protection is weak (Rudiger Dornbusch 2000, p. 92). Table 3 shows that the eighteen emerging economies that liberalized their stock mar- kets rank lower than developed countries on five commonly used measures of investor protection: the rule of law, efficien- cy of the judicial system, contract repudia- tion, expropriation risk, and the accounting system. Weak investor protection decreases effective returns, thereby reducing the incentive for capital to flow from rich to poor countries (Shleifer and Daniel Wolfenzon 2002).24 Because investors in emerging economies receive less protection than their counterparts in rich countries, lifting restrictions on capital inflows may generate smaller changes in asset prices and capital flows than would occur if emerging economies gave investors the same average level of protection they receive in developed economies. 912 Journal of Economic Literature, Vol. XLV (December 2007) 24 See also the discussion in Charles P. Himmelberg, Robert Glenn Hubbard, and Inessa Love (2004). TABLE 2 THE ECONOMIC INFRASTRUCTURE OF EMERGING MARKETS IS WEAKER THAN THAT OF DEVELOPED COUNTRIES Hall and Jones Heritage House World Bank Doing World Economic Index Index Business Index Forum Index Argentina 77 23 101 69 Brazil 68 87 121 66 Chile 43 15 28 27 Colombia 79 73 79 65 India 87 126 134 43 Indonesia 45 76 135 50 Jordan 34 58 78 52 Malaysia 22 42 25 26 Mexico 64 91 43 58 Nigeria 116 97 108 101 Pakistan 113 107 74 91 Philippines 90 69 126 71 South Korea 33 33 23 24 Taiwan 28 16 47 13 Thailand 20 38 18 35 Turkey 71 49 91 59 Venezuela 53 99 104 88 Zimbabwe 94 122 153 119 Emerging Market Average 63 68 83 59 Developed Country Average 14 14 23 15 Notes: Table 2 compares the average quality of the economic infrastructure of the eighteen emerging economies that liberalized (see table 1) with that of the G-7 countries, using four commonly employed measures of the qual- ity of economic institutions. All four indices rank countries in descending order of quality (e.g., a ranking of 1 indicates superior quality, while a ranking of 130 indicates poor quality). The rankings in column 2 are based on the Hall and Jones (1999) measure of social infrastructure, which ranks 130 countries. Column 3 uses Heritage House?s Index of Economic Freedom (Heritage House 2006), which ranks 161 countries. Column 4 uses the World Bank?s Ease of Doing Business Index (World Bank 2006), which ranks 175 countries. The rankings in col- umn 5 are based on the World Economic Forum?s Global Competitiveness Index (World Economic Forum 2006), which ranks 125 countries. ?Average Emerging Market? is the average ranking, on each index, of the eighteen emerging economies that liberalized. ?Average Developed Country? is the average G-7 country ranking on each index. dec07_Article1 11/28/07 12:02 PM Page 912 Although table 3 corroborates the basic story, an important question for future research is the extent to which the capital- market-imperfections view truly explains the relatively small financial impact of liberaliza- tion. For instance, an empirical prediction of the market-imperfections view is that the impact of liberalization should be larger in countries that have stronger investor protec- tion. Unfortunately, the policy-experiment approach is not well-suited to test this predic- tion. With typically no more than twenty-five countries that meet the criteria for sample inclusion, there are simply not enough data points to conduct precise tests of the cross- sectional prediction that liberalization will have a larger impact in countries with stronger investor protection.25 In a related exercise that is not a study of liberalization per se, Alfaro, Sebnem Kalemli-Ozcan, and Vadym Volosovych (forthcoming) try to determine whether the capital-market-imperfections view or the weak-institutions view has greater empiri- cal relevance for the observed pattern of 913Henry: Capital Account Liberalization 25 Menzie D. Chinn and Hiro Ito (2006) do find that countries whose financial systems have a higher level of legal and institutional development benefit more from capital account openness but, like other cross-sectional studies, the paper examines the correlation between openness and financial development, not the impact of opening on financial development. TABLE 3 INVESTOR PROTECTION IN EMERGING MARKETS IS WEAKER THAN IN DEVELOPED COUNTRIES Rule of Judicial Contract Expropriation Accounting Law Efficiency Repudiation Risk Standards Argentina 5.4 6.0 4.9 5.9 4.5 Brazil 6.3 5.8 6.3 7.6 5.4 Chile 7.0 7.3 6.8 7.5 5.2 Colombia 2.1 7.3 7.0 7.0 5 India 4.2 8.0 6.1 7.8 5.7 Indonesia 4.0 2.5 6.1 7.2 NA Jordan 4.4 8.7 4.9 6.1 NA Malaysia 6.8 9.0 7.4 8.0 7.6 Mexico 5.4 6.0 6.6 7.3 6 Nigeria 2.7 7.3 4.4 5.3 5.9 Pakistan 3.0 5.0 4.9 5.6 NA Philippines 2.7 4.8 4.8 5.2 6.5 South Korea 5.4 6.0 8.6 8.3 6.2 Taiwan NA NA NA NA NA Thailand 6.3 3.3 7.6 7.4 6.4 Turkey 5.2 4.0 6.0 7.0 5.1 Venezuela 6.4 6.5 6.3 6.9 4 Zimbabwe 3.7 7.5 5.0 5.6 NA Emerging Market Mean 4.8 6.2 6.1 6.8 5.7 Developed Country Mean 9.1 9.2 9.2 9.5 6.4 Notes: Table 3 compares the average quality of investor protection in the eighteen emerging economies that lib- eralized (see table 1) with that of the G-7 countries, using five measures employed by LaPorta, L?pez-de-Silanes, Shleifer, and Vishny (1998). All scores are based on a scale of 1 to 10, with 10 indicating the highest quality and 1 the lowest. The source for all of the numbers in this table is LaPorta, L?pez-de-Silanes, Shleifer, and Vishny (1998). dec07_Article1 11/28/07 12:02 PM Page 913 asset prices and capital flows to emerging economies. They construct a measure of the strength of institutions for a large cross- section of countries using variables from Political Risk Service?s International Country Risk Guide. As a proxy for inter- national capital market imperfections, they construct a measure of economic ?remote- ness? that they interpret as capturing infor- mational frictions as in Joshua D. Coval and Tobias J. Moskowitz (1999, 2001) and Richard Portes and Helene Rey (2005). In running horse races between the two views, Alfaro, Kalemli-Ozcan, and Volosovych (forthcoming) find that their measure of institutions explains over half of the variation in FDI and portfolio flows as a fraction of GDP. They interpret their find- ing as indicative of the primacy of weak institutions over capital market imperfec- tions for explaining the paucity of capital flows to developing countries. One problem with their interpretation is that it is not clear that economic remoteness is the right proxy for capital market imperfections such as agency problems and investor protection. Furthermore, it is also difficult to say where institutional quality ends and capital market imperfections begin. Poor institutions (e.g., the absence of property rights) lead to mar- ket imperfections such as weak investor pro- tection. While it is not clear that one can successfully distinguish between the institu- tions view and the capital market imperfec- tions view using aggregate cross-country data, the Alfaro, Kalemli-Ozcan, and Volosovych (forthcoming) paper does pro- vide empirical support for the notion that the myriad of potential distortions not explicitly modeled in the neoclassical frame- work can account for some of the Lucas (1990) paradox.26 6.2 Can We Believe the Investment Results? Although the financial impact of stock market liberalization is small relative to pre- dictions, the evidence in section 5.2 demon- strates that it is apparently still large enough to have significant effects on real invest- ment. However, it is not clear how much confidence we should place in results that attribute economywide investment booms to a policy change that affects directly only those firms that are traded on the stock mar- ket. In most emerging economies, the eco- nomic activity of stock-market-listed firms accounts for a relatively small fraction of GDP, so it is natural to ask whether the aggregate investment booms that occur in the aftermath of liberalizations may be plau- sibly linked with events that partially open a single component of the capital account. For instance, the establishment of a coun- try fund is the modal means through which the policy-experiment approach identifies capital account liberalization dates. The size of a country fund is typically small relative to the size of the liberalizing country?s capital stock. The Chilean liberalization listed in table 1 illustrates the point. Chile liberalized its stock market in May 1987 and the vehicle through which it did so was the Toronto Trust Mutual Fund, a Canadian closed-end fund with a net asset value of 37.7 million U.S. dollars.27 A capital inflow of this size is not large enough to account for the size of Chile?s subsequent investment boom. Because the capital stock of publicly traded Chilean firms is a subset of the entire econ- omy?s, to illustrate the point it is sufficient to show that 37.7 million dollars is small rela- tive to the increase in the capital stock of the publicly traded firms. In 1987, Chile?s stock market capitalization was 5.34 billion U.S. dollars and the ratio of the market value of assets to book value was 0.7. This means that 914 Journal of Economic Literature, Vol. XLV (December 2007) 26 For more on the distinction between credit market and governance institutions, and an empirical attempt to disentangle them, see the discussion of Jiandong Ju and Shang-Jin Wei (2006) in Kose et al. (2006). 27 See Park and Van Agtmael (1993), Price (1994), and Wilson (1992). dec07_Article1 11/28/07 12:02 PM Page 914 the book value of assets was roughly 7.63 bil- lion U.S. dollars at the time Chile liberal- ized. In the five years after liberalization, Chile experienced abnormal capital stock growth of 2.2 percentage points per year. Applying this growth number to the capital stock in a continuously compounded fashion for the next five years adds an additional 890 million dollars of productive assets to the economy. The 37.7 million dollar capital inflow can account for less than 5 percent of this increase. The apparent incongruity of these num- bers forces harder thinking about the validi- ty of the empirical link between liberalization and real investment. In doing so, an important point to keep in mind is that the policy-experiment approach uses coun- try-fund openings as observable de facto indicators of harder-to-pinpoint de jure pol- icy changes.28 If country-fund dates are valid proxies for the occurrence of broader, undocumented liberalizations, then signifi- cant quantities of capital may flow in through stock markets that are not part of any particular country fund. If that is the case, then the size of the initial country fund will understate the quantity of resources that stock market liberalizations make available to finance an increase in the domestic capi- tal stock. Three facts suggest that closed-end country fund dates do indeed provide non- specious indicators of a larger move toward open capital markets. First, a steady stream of country funds and issuances of American Depository Receipts typically follow stock market liberalizations (Karolyi 2004; Juan Carlos Gozzi, Levine, and Sergio L. Schmukler forthcoming). In the case of Chile, six additional country funds with a cumulative net asset value of 991.8 million dollars were established between 1987 and 1992.29 More generally, table 4 shows that aggregate net equity inflows to emerging equity markets rise sharply following the median date of coun- try-fund openings. In principle, when capital flows into the stock market, the economy as a whole need not experience net capital inflows because foreign investors could be swapping debt for equity or the country could experience debt outflows that are larg- er than the inflow of portfolio equity. In practice, however, the epoch of stock market liberalization coincided with a period of strong net capital inflows to developing countries (Calvo, Leonardo Leiderman, and Carmen M. Reinhart 1996). Second, with the sole exception of Malaysia during the Asian Crisis of 1997?98, none of the stock market liberalization dates from table 1 were followed by a reversal of freedom of foreign access. Together with the first fact, this second fact confirms that stock market liberalizations indicate a discrete change in capital account policy that signi- fies the beginning of a steady march toward greater freedom of capital inflows. Third, stock market liberalizations coincide with a significant increase in the importation of capital goods. In a sample of twenty-five countries that liberalized their stock markets between 1980 and 1997, liberalization leads to a 6 percent increase in capital goods as a fraction of total imports and the share of total machine imports to GDP rises by 12 percent (Alfaro and Eliza Hammel 2007). Because developing countries do not produce a signif- icant portion of the capital goods they use, the observation that imports of capital goods rise in concert with the advent of portfolio equity inflows increases confidence in earli- er work on liberalization and aggregate investment. 6.3 Are the Growth Effects Plausible? The increase in economic growth due to liberalization is about one percent per year after controlling for a host of variables (Bekaert, Harvey, and Lundblad 2005). That estimate is inconsistent with the neoclassical 915Henry: Capital Account Liberalization 28 See Kose et al. (2006) for a detailed discussion of de facto versus de jure indicators. 29 See Park and Van Agtmael (1993), Price (1994), and Wilson (1992). dec07_Article1 11/28/07 12:02 PM Page 915 theory on which it is based.30 To see why, recall that the growth rate of capital increas- es by about one percent per year following stock market liberalizations and that the elasticity of output with respect to capital is roughly one-third. Therefore, a one-per- centage-point increase in the growth rate of the capital stock cannot raise the growth rate of GDP per capita by much more than one- third of a percentage point. Within the standard neoclassical frame- work, any increase in the growth rate of out- put not due to an increase in the growth rate of capital and labor must be the result of an increase in the growth rate of total factor pro- ductivity (TFP). For the eighteen countries in table 1, the growth rate of TFP does rise from an average of 0.19 percent per year in the five years preceding stock market liberal- ization to an average of 1.82 percent per year in the subsequent five. But one cannot glibly attribute the increase in TFP growth to stock market liberalization. The operative channel from liberalization to growth in the neoclassi- cal model runs strictly through capital accu- mulation. Total factor productivity does not enter into the story.31 Of course, if one is willing to step outside the neoclassical model, there are many sto- ries in which liberalization can raise total 916 Journal of Economic Literature, Vol. XLV (December 2007) 30 See the discussion on page 19 and page 22 of Kose et al. (2006). 31 Pierre-Olivier Gourinchas and Olivier Jeanne (2006) argue that the welfare effects of capital account liberaliza- tion in the neoclassical model are de minimus precisely because liberalization has no impact on total factor pro- ductivity. TABLE 4 COMPOSITION OF CAPITAL INFLOWS TO DEVELOPING COUNTRIES 1970?95 Average 1970?74 Average 1975?79 Average 1980?84 Average 1985?89 Average 1990?95 Millions Percent Millions Percent Millions Percent Millions Percent Millions Percent of dollars of total of dollars of total of dollars of total of dollars of total of dollars of total Net Resource 12529.1 32836.8 51604.7 32726.2 90184.1 Flows Net Flows 10121.3 80.8 27151.3 82.7 42374.6 82.1 20563.4 62.8 45316.2 50.2 on Debt (PPG+PNG) Public and 5628.0 44.9 18014.0 54.9 28383.4 55.0 14844.5 45.4 12820.5 14.2 Publicly Guaranteed Debt (PPG) Private 4493.4 35.9 9137.3 27.8 13991.2 27.1 5718.9 17.5 32495.8 36.0 Nonguaranteed Debt (PNG) Foreign Direct 1798.6 14.4 4247.2 12.9 6871.7 13.3 9006.5 27.5 24993.8 27.7 Investment Portfolio Equity 0.0 0.0 0.0 0.0 27.0 0.1 762.2 2.3 16855.0 18.7 Grants 609.2 4.9 1438.3 4.4 2331.3 4.5 2394.2 7.3 3019.0 3.3 Notes: This table presents data on the composition of aggregate net resource flows to the eighteen developing countries listed in table 1 (with the exceptions of South Korea and Taiwan, for which no data were available) from 1970 to 1995. All data come from the World Bank?s Global Development Finance Data Base. The first col- umn of the table lists all the components of net resource flows. Net resource flows are the sum of net resource flows on public and publicly guaranteed (PPG) debt, private nonguaranteed (PNG) debt, foreign direct invest- ment, portfolio equity, and official grants. Public and publicly guaranteed debt is an external obligation of a pri- vate debtor that is guaranteed for repayment by a public entity. Private nonguaranteed external debt is an exter- nal obligation of a private debtor that is not guaranteed for repayment by a public entity. Columns 2 through 6 display the data as averages over five-year intervals. dec07_Article1 11/28/07 12:02 PM Page 916 factor productivity. For instance, liberaliza- tion may enable firms to import more pro- ductive machinery (e.g., tractors instead of hoes) that effectively shift production in the country closer to the world technology fron- tier. Also, FDI in the form of green field investments or the purchase of a majority stake of a domestic firm may bring new technology and management techniques that increase the efficiency of the acquired firm and generate economywide knowledge spillovers. For instance, Frederic S. Mishkin (2006) argues that developing countries can import greater financial effi- ciency by allowing foreign investors to take controlling stakes in domestic financial firms and bring in state-of-the-art financial intermediation practices. More generally, he articulates how capital account liberal- ization could raise total factor productivity by imposing capital market discipline on governments, breaking up local monopo- lies, and broadly promoting a whole range of institutional improvements that Kose et al. (2006) refer to as ?collateral benefits? of financial globalization. These stories are theoretically plausible but empirically unsubstantiated. There was a significant increase in the TFP growth of Mexican plants that changed from domestic to majority foreign ownership in 1989 (Francisco P?rez Gonz?lez 2005). Similarly, FDI in Venezuela also increased TFP growth at the plant level (Brian J. Aitken and Ann E. Harrison 1999).32 But there is no evi- dence that plant-level productivity gains generate economywide knowledge spillovers that stimulate higher aggregate TFP growth (Mona Haddad and Harrison 1993; Aitken and Harrison 1999; Holger Gorg and David Greenaway 2004). Similarly, there is no rigorous evidence that capital account liberalization improves the allocative efficiency of the domestic financial system.33 It is true that capital account liberalization within a country tends to increase financial development (Levine 2001; Levine and Zervos 1998a, 1998b). It is also true that countries at high levels of financial development allocate capital more efficiently than countries at low levels of financial development (Love 2003; Jeffrey Wurgler 2000; Rajan and Luigi Zingales 1998; Eichengreen 2004). The temptation, of course, is to invoke transitivity and con- clude that capital account liberalization increases domestic allocative efficiency. The problem with such logic is that documenting a positive correlation between the efficiency of capital allocation and financial develop- ment in a cross section of countries does not permit us to infer that more financial devel- opment within a given country will improve its allocative efficiency. Without a convincing body of time series evidence that the quality of a country?s capital allocation improves as its level of financial development rises, no basis exists for concluding that liberalization indirectly improves the efficiency of domes- tic capital allocation through its impact on financial development.34 I am not arguing that one cannot tell sto- ries in which capital account liberalization induces higher TFP growth. Again, once you step outside the neoclassical model, there are many possible channels. Liberalization may ease liquidity constraints (Nandini Gupta and Kathy Yuan 2005; Todd Mitton 2006), thereby enabling firms to adopt tech- nologies that they could not finance prior to the liberalization. Liberalization also facili- tates increased risk sharing, which might encourage investment in riskier, higher growth technologies in the spirit of Obstfeld (1994), Levine (1997), or Robert G. King 917Henry: Capital Account Liberalization 32 For an analysis of the anticipated productivity gains from FDI using stock market data, see Anusha Chari, Paige P. Ouimet, and Linda L. Tesar (2004). 33 In fact, many economists argue for enhancing domestic financial efficiency before liberalizing the capi- tal account (see Dornbusch 1983; Edwards 1984; Ronald I. McKinnon 1991; Mishkin 2007). 34 Arturo Galindo, Fabio Schiantarelli, and Andrew Weiss (2007) provide some of the first evidence on this subject of which I am aware. dec07_Article1 11/28/07 12:02 PM Page 917 and Levine (1993). The point is that, while the developing countries in this sample may have increased their rate of adoption of new production technologies during the late 1980s and early 1990s, there is no sound way to draw such a conclusion from aggregate data. In contrast, the aggregate TFP growth numbers are consistent with the readily observable evidence on economic reforms. As discussed in section 6.1.2, economic reforms that raise the efficiency of a given stock of capital and labor will increase TFP and again table 1 shows that stock market liberalizations typically coincide with a raft of other such reforms. Of course, the only way to distinguish between competing theo- ries that can explain higher TFP in the after- math of liberalization is to confront these theories with data that have the power to make such fine distinctions. Recent studies of liberalization that move from aggregate to firm-level data show the way forward. 7. Firm-Level Perspectives on Capital Account Liberalization Perhaps the most important shortcoming of aggregate data is its limited ability to help us understand whether countries that liber- alize efficiently allocate the capital that flows in. While the rise in aggregate stock prices and investment documented in the literature suggests that liberalization promotes some efficient movement of capital between coun- tries, it says nothing about the efficiency of capital allocation within countries. Within- country efficiency would require that the stream of capital from developed countries get allocated to the highest return sectors within the emerging economies to which it flows. The efficiency of capital allocation is closely related to the cost of capital. For instance, the larger the fall in a country?s cost of capital, the more its investment should increase. In other words, cross-coun- try changes in investment in the aftermath of liberalizations should be negatively corre- lated with cross-country changes in the cost of capital on impact. The problem is that the small-sample-size problem inherent in the policy-experiment approach once again does not provide sufficient power to say anything meaningful about the empirical relevance of this hypothesis. Similarly, aggregate data are of little use in helping us understand the role of interna- tional risk sharing, a subject that also has important efficiency implications and is of broad interest in the literature.35 Because the cost of capital matters for investment and risk sharing matters for the cost of capital, it is natural to wonder about the importance of risk sharing relative to the fall in the interest rate for the overall change in liberalizing countries? cost of capital. Unfortunately, studies of liberalization that use aggregate data provide no help on this point. For each country, aggregate data supply one observ- able proxy for the change in the cost of capi- tal?the stock market revaluation discussed in section 5. But there are two forces at work: the change in risk sharing and the change in the interest rate. Since one data point is not sufficient to identify two effects, it is impossi- ble to disentangle them using aggregate data.36 Firm-level data, on the other hand, pro- vide more than sufficient degrees of free- dom with which to disentangle the impact of risk sharing from that of the interest rate. This section reviews how recent work uses such data to gain a better understanding of risk sharing and the allocation of capital within countries. More generally, this sec- tion explains how recent firm-level studies 918 Journal of Economic Literature, Vol. XLV (December 2007) 35 For surveys of the literature on international risk sharing, see Obstfeld and Rogoff (1996), chapter 5. 36 If we could observe internal, market-determined risk-free rates before and after liberalizations, disentan- gling the two effects would be straightforward. The prob- lem is that almost all of the countries listed in table 1 had some form of financial repression in place at the time they liberalized (see tables 1, 4, and 5 in John Williamson and Molly Mahar 1998). Hence, we do not observe market- determined, risk-free rates. dec07_Article1 11/28/07 12:02 PM Page 918 enhance our understanding of the real effects of liberalization. 7.1 Risk Sharing and Asset Prices The country-level organizing framework in section 5.1 extends naturally to a firm- level setting that lays bare the risk-sharing content of stock-price revaluations. The key idea is that the changes that occur in firms? stock prices at the time of liberalization embody signals about the impact of liberal- ization on the firms? fundamentals. If capital allocation is efficient, then firms? investment decisions should respond to these signals. An efficient allocation of capital is one that satisfies the first-order condition for invest- ment. Recall that, before liberalization occurs, the first-order condition for aggre- gate investment requires the expected mar- ginal product of the aggregate capital stock to be equal to the aggregate cost of equity capital plus the rate of depreciation. The dif- ference between the first-order condition for the economy as a whole and that for an individual firm lies in the equity premium. Whereas the aggregate equity premium is proportional to the variance of the return on the domestic market, Var(r?M), the equity premium for an individual firm is propor- tional to the covariance of the firm?s return with the local market. Written in symbols to express it more con- cisely, the first-order condition for an individ- ual firm?s investment before liberalization is: (12) (r?i,r?M) + ?, where the i subscripts indicate that we are now talking about a single firm instead of the entire economy. Liberalizing a country?s stock market changes the relevant source of systematic risk for pricing a firm?s shares from the local stock market index to a world stock market index. Due to this change, the first-order condition for investment after liberalization is: (13) (r?i,r?W) + ?,? = +f k r COVi e( )? ? ? ? = +f k r COVi e( ) ? where COV(r?i,r?W) is the covariance of firm i?s return with the world market. Subtracting equation (13) from equation (12) gives a use- ful expression for the change in the first- order condition from before-to-after liberalization: (14) , where DIFCOVi = COV(r?i,r?M) ? COV(r?i,r?W). The right-hand-side of equation (14) high- lights the two channels through which liber- alization changes a firm?s cost of capital. Moving from left to right, the first term is exactly the same as in the aggregate case: a fall in the risk-free rate as the country switch- es from financial autarky to financial integra- tion with the world market. This is a common shock to all firms in the economy. The second effect is unique to each firm: the greater the covariance of the firm?s stock return with the local market relative to the covariance of its return with the world market, the larger the change in the firm-specific component of its cost of capital. In the pristine world of theory, liberaliza- tion does not alter the firm?s expected future cash flow and the response of the firm?s stock price to the news of the liberalization mirrors exactly the change in its cost of capital. The firm?s stock price will increase if liberalization reduces the cost of capital and vice versa. In other words, equation (14) predicts that each firm?s revaluation will have an intercept term and a slope term. The intercept term should be the same across all firms within a given country. As for the slope effect, if risk sharing matters, then the revaluations that firms experience should be an increasing function of the variable DIFCOV. In the murky world of empirical work, lib- eralizations coincide with other economic reforms that undoubtedly affect expected future cash flows. Whereas aggregate studies attempt to disentangle the impact of liberal- ization from contemporaneous economic reforms by using rough macroeconomic indicators as proxies for cash flows (see  ? = ? +f k r r DIFCOVi e i( ) ( )? ? 919Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 919 equation (11) in section 5.3), firm-level studies control directly for changes in cash flows by using the real value of sales and earnings taken from firms? income state- ments. More generally, although the econo- my as a whole benefits from liberalization, firms that gain more will expand, while those that gain less (or lose) may contract.37 Because the process of liberalization gener- ates losers as well as winners, it raises important questions of political economy that aggregate data cannot answer. In con- trast, firm-level data allow researchers to control for the differential impact of liberal- ization across various sectors of the econo- my and therefore may help to answer some of these questions (see section 8.2). Recent work supports the two asset-pric- ing predictions of equation (14). First, firms experience significant stock price revalua- tions during liberalizations, controlling for the impact of contemporaneous reforms on expected future cash flows (Chari and Henry 2004; Dilip K. Patro and John K. Wald 2005). Second, the stock price change asso- ciated with liberalization is positively and significantly related to DIFCOV. In a sample of 430 firms from eight countries, the aver- age firm-level revaluation is about 15 per- cent in real dollar terms, changes in firm-specific covariances explain roughly one-third of the revaluation, and the com- mon shock is the same for all firms (Chari and Henry 2004). The finding that stock prices of firms in emerging markets move in accordance with changes in systematic risk provides some of the first empirical support for the CAPM in a domestic or international setting.38 It also runs counter to recent work which finds that stock returns in emerging economies gener- ally contain little firm-specific information (Randall K. Morck, Bernard Yeung, and Wayne Yu 2000). These two facts make you wonder whether the firm-level asset pricing results are spurious. Two observations sug- gest otherwise. First, the empirical design of the firm- level policy-experiment approach gives it power to uncover cross-sectional relation- ships between expected returns and system- atic risk that are difficult to detect in other settings. Covariances are measured with error (Fama and Kenneth R. French 2004), and measurement error reduces the statisti- cal power of any regression. Instead of test- ing for a relationship between levels of returns and levels of systematic risk, the firm-level policy-experiment approach focuses on episodes where there are large changes in both risk and returns. The mag- nitude of the liberalization-induced changes in expected returns and systematic risk asso- ciated with opening up the economy to for- eign capital flows may simply dominate the attenuating effects of measurement error that usually plague efforts to find cross-sec- tional pricing relationships. Similarly, changes in emerging-market stock prices may convey little firm-specific information in general, but they do convey such informa- tion during episodes like liberalizations when the magnitude of the information is sufficiently large.39 Second, using a similar approach to that of Chari and Henry (2004), Tomas Dvorak and Richard Podpiera (2006) also find that large changes in systematic risk have explanatory power for changes in the cross-section of expected returns. The policy experiment that provides the key source of variation for Dvorak and Podpiera?s exercise is the addi- tion of eight central and Eastern European 920 Journal of Economic Literature, Vol. XLV (December 2007) 37 See, for example, the impact of liberalization on the expansion of big versus small firms documented in Gozzi, Levine, and Schmukler (forthcoming). 38 Christopher Polk, Samuel Thompson, and Tuomo Vuolteenaho (2006) also argue that the CAPM matters for the cross section of stock prices. For surveys of the CAPM?s empirical record, see Eugene F. Fama (1991), John H. Cochrane (1999), and John Y. Campbell (2000). Karolyi and Stulz (2001) discuss the international CAPM. 39 The firm-specific information in stock prices also rises as countries adopt greater capital market openness (Kan Li et al. 2004). dec07_Article1 11/28/07 12:02 PM Page 920 countries into the European Union. Dvorak and Podpiera (2006) argue that accession into the European Union integrated the stock markets of the acceding countries with the rest of the world and thereby changed the source of systematic risk facing investors in the countries that gained membership. In their sample of seventy-four firms, the dif- ference between the beta of a firm?s stock return with the local market and its beta with the world market explains about 22 percent of the typical stock price revaluation that occurs with accession. 7.2 Risk Sharing and Resource Allocation While the correlation between changes in returns and changes in systematic risk pro- vide some evidence of within-country effi- ciency at the level of asset pricing, the more pressing economic question is whether real resource allocation?investment?responds accordingly.40 As the benchmark for deter- mining the firm?s hurdle rate for investment switches from the local stock market index to a world stock market index, efficiency requires that allocation of the firm?s physical capital also shift in accordance with the change in the source of the country?s system- atic risk.41 In order to restore equilibrium, the increase in the firm?s capital stock must be large enough to drive the expected mar- ginal product down to the lower, postliberal- ization cost of capital. Equation (14) implicitly defines the size of the required increase and therefore delivers two testable predictions about investment. The first prediction is that the common shock to the cost of capital should cause the average investment rate of all firms to rise. The second prediction is that, given the common shock, the firm-specific shock (the change in covariance) implies that firms whose equity premia fall should invest even more than those whose premia rise. In other words, if physical capital is reallocated in line with the optimal smoothing of production risk, then high DIFCOV firms should expe- rience faster capital stock growth than low DIFCOV firms following liberalization. The first prediction about liberalization and firm-level investment finds empirical support in the literature. The growth rate of the average firm?s capital stock exceeds its preliberalization mean by an average of 3.8 percentage points per year (Chari and Henry forthcoming). This effect is much larger than the corresponding increase in the aggregate capital stock over the same time period (1.1 percentage points per year). Because stock market liberalizations most directly affect the investment incentives of publicly traded firms, the firm-level invest- ment results are more credible than the aggregate results and make a stronger case that liberalization does, indeed, have real effects.42 The second prediction about liberalization and firm-level investment enjoys rather less empirical success. There is no evidence that physical investment responds to changes in systematic risk, and firm-specific changes in equity premia (the DIFCOV variable) have an economically trivial and statistically insignificant effect on changes in invest- ment. The finding that firms? investment deci- sions are insensitive to firm-specific changes in their cost of capital delivers a powerful blow to the Allocative Efficiency view of lib- eralization. On the other hand, with little evidence from developed markets to suggest 921Henry: Capital Account Liberalization 40 The idea of trying to relate changes in investment to the liberalization-induced changes in stock prices follows in the spirit of earlier work that tries to relate changes in investment to changes in stock prices more generally (Olivier Blanchard, Changyong Rhee, and Summers 1993; Fischer and Robert C. Merton 1984; James Tobin and William C. Brainard 1977). 41 For a helpful exposition on the resource-allocation implications of changes in systematic risk, see Tesar and Ingrid M. Werner (1998), pp. 290?92. 42 The firm-level investment data used by Chari and Henry (forthcoming) comes from the International Finance Corporation?s Corporate Finance Database. For a detailed description of this dataset, see Asli Demirg??- Kunt and Vojislav Maksimovic (1998) and Laurence Booth et al. (2001). dec07_Article1 11/28/07 12:02 PM Page 921 that exposure to aggregate covariance risk drives expected returns, testing the hypothe- sis that firms in developing countries allocate physical investment in accordance with the CAPM may seem to fly in the face of all common sense. But if the risk-sharing-resource-allocation hypothesis is a straw man, then it is a very popular one. Virtually all studies of interna- tional risk sharing lean heavily on the intuition that the gains to trade in risky assets stem from the difference between the variance of domestic returns (consumption) and the covariance of domestic returns (consump- tion) with the rest of the world (see, for exam- ple, Stefano G. Athanasoulis and Eric Van Wincoop 2000; Obstfeld and Rogoff 1996, chapter 5; Karen K. Lewis 1999, 2000; and Van Wincoop 1994, 1999).43 Furthermore, the evidence shows that firm-level stock prices do move in conjunction with liberaliza- tion-induced changes in systematic risk. So the real issue is why the reductions in risk premia do not cause firms to adopt projects ? la Obstfeld (1994)?that were too risky to undertake in the absence of international risk sharing. 7.3 Other Firm-Level Approaches to Liberalization and Efficiency Testing whether liberalization-induced changes in risk drive changes in asset prices and investment provides one way of evaluat- ing efficiency, but it is also useful to adopt a broader perspective?one that is grounded in theory but not tied so restrictively to the CAPM. There is much to be learned about the impact of capital account policy on real variables by looking at firm-level data from a number of different perspectives. For instance, another way to tackle the question of liberalization and allocative efficiency is to turn the policy-experiment approach on its head. In addition to looking at how firms respond to liberalization, it may be just as instructive to study how firms respond when countries impose restrictions on capital inflows (Kristin J. Forbes 2007b). If capital account liberaliza- tion enhances efficiency, then imposing capital controls should diminish efficiency in at least two important ways. First, capital controls may reduce the supply of capital, thereby raising the cost of borrowing and tightening the financing constraints faced by domestic firms. Second, by reducing the supply of capital, capital controls can decrease competition and market disci- pline, permitting firms that might not sur- vive if their competitors had access to credit to flourish behind closed borders (Rajan and Zingales 2003; Johnson and Mitton 2003; Morck, David A. Stangeland, and Yeung 2000). A number of papers examine the extent to which the presence of capital controls exac- erbates the financing constraints faced by various kinds of firms. In the spirit of Steven M. Fazzari, Hubbard, and Bruce C. Petersen (1988), the basic test involves the examina- tion of the sensitivity of investment to cash flow.44 Forbes (2007a) uses the 1991?98 episode of El Encaje, the tax on capital inflows to Chile, as a policy experiment with which to assess the efficiency implications of imposing capital controls.45 Her analysis begins with the following observation. In developed countries, small, publicly traded firms exhibit higher investment growth than large, publicly traded firms. Forbes (2007a) finds that this is also the case in Chile before 1991 and after 1998. During El Encaje, how- ever, she finds that the investment growth of small firms drops below that of large firms. She goes on to show that El Encaje increased the sensitivity of investment to cash flow for 922 Journal of Economic Literature, Vol. XLV (December 2007) 43 An important difference between the aggregate studies to which I refer in this citation and the firm-level policy-experiment papers is that the aggregate studies focus on the implications of risk-sharing for consumption and welfare gains as opposed to investment and produc- tion. Section 7.4 discusses this distinction in further detail. 44 For a detailed discussion of the literature on financing constraints and investment, see Hubbard (1998). 45 Francisco A. Gallego and F. Leonardo Hernandez (2003) also conduct an analysis of El Encaje. dec07_Article1 11/28/07 12:02 PM Page 922 small, publicly traded firms whereas the sen- sitivity of large firms was unaffected. In economies in which small firms drive pro- ductivity increases, Forbes?s results might suggest one channel through which capital account liberalization could affect TFP. To know whether this is the case, we need more research on the extent to which large versus small firms drive TFP growth in developing countries. In contrast to Forbes?s case study, Harrison, Love, and Margaret S. McMillan (2004) use a cross-country, firm-level panel data set to study the impact of capital account restrictions. They take their meas- ures of capital account restrictions from the IMF publication AREAER discussed in sec- tion 3. Like Forbes (2007a), they find that the presence of capital account restrictions increases the sensitivity of firms? investment to cash flow. They also document that the sensitivity of domestically owned firms? investment to cash flow is greater than that of firms with foreign ownership or assets. In a paper that reverts back to the easing of capital controls, Luc Laeven (2003) con- structs a cross-country data set to address the question of whether financial liberalization eases financing constraints and increases competition. Financial liberalization typically refers to the removal of interest rate ceilings, directed credit, and other such distortions in the domestic financial markets.46 Laeven?s paper has implications for capital-account liberalization because his measure of finan- cial liberalization captures the impact of FDI (i.e., foreign entry) into the domestic banking sector. As such, it has some power to speak to the issue of whether foreign bank entry increases the supply of capital and makes the domestic lending market more competitive. He finds that the sensitivity of small firms? investment to cash flow falls by 80 percent as a result of financial liberalization. He also finds that the investment of large firms becomes more sensitive to cash flow after financial liberalization and interprets this finding as evidence that large firms may have had access to preferential credit before financial liberalization. One important weakness of firm-level papers that examine the efficiency implica- tions of capital controls is their interpretation of the sensitivity of investment to cash flow as a measure of firm-financing constraints. If firms face financing frictions, then their investment will be sensitive to cash flow. But the converse of the preceding statement need not be true: sensitivity of investment to cash flow does not imply that firms face financial constraints (Steven N. Kaplan and Zingales 1997, 2000; Jeremy C. Stein 2003). While there are multiple interpretations of the investment?cash-flow sensitivity of emerging-market firms and it remains to be seen how all of the results reviewed in this section will stand up to further scrutiny, there is little question that papers which use firm- level data provide two steps forward: (1) they increase our understanding of the transmis- sion mechanisms through which capital- account policy affects the real economy and (2) they provide a blueprint of how to con- struct firm-level identification strategies for disentangling those mechanisms. 7.4 Important Areas for Future Research: Capital Outflows and Consumption Data In spite of the progress made in this area, an important limitation of existing firm-level studies on the benefits of risk sharing is their almost-exclusive focus on (1) the investment and production side of the economy as opposed to consumption and welfare and (2) the removal of restrictions on capital inflows as opposed to capital outflows. The largest future welfare gains to capital account liberalization in developing coun- tries such as China and India are likely to come from the reduction in consumption and income volatility that will occur when their governments liberalize capital outflows, 923Henry: Capital Account Liberalization 46 See McKinnon (1973), Edward S. Shaw (1973), and Carlos Diaz-Alejandro (1985). dec07_Article1 11/28/07 12:02 PM Page 923 thereby permitting their citizens to hold assets abroad and unload some of the sys- tematic risk of the domestic economy. Beyond the headline-grabbing examples of China and India, there may also be large wel- fare gains to permitting capital outflows from countries such as Chile, Malaysia, and Taiwan that have government-managed or government-mandated funded retirement programs that are heavily tilted toward domestic assets. Because of developing countries? long his- tory of legal restrictions on capital outflows, past research in this area focused on ways that these countries could enjoy the welfare benefits of external portfolio diversification without permitting net capital outflows (Donald R. Lessard 1973, 1974, 1994; Zvi Bodie and Merton 2002). With the rise of sovereign wealth funds and the apparent impetus toward greater easing of restrictions on the investment of capital abroad, there would seem to be ample opportunities to identify relevant policy experiments in the area of liberalization of capital outflows.47 In light of these recent developments, the glaring absence of empirical research on the actual (as opposed to hypothetical) welfare impact of the liberalization of capital outflows from developing countries practically cries out for further research. Given the difficulties with aggregate data mentioned earlier, sub- stantial contributions to research in the area of risk sharing, welfare, and the liberalization of capital outflows are most likely to come from analyses of household-level consumption and portfolio data. Using the cross-sectional varia- tion in such data in conjunction with ma- jor policy changes may lead to useful and convincing empirical identification strategies. 8. Do Liberalizations Cause Crises? Stock market liberalizations deliver unde- niable benefits over the short-to-medium term but, viewed over a more extensive hori- zon, the data raise questions about the longer-run cost?benefit trade-off. For instance, extending figures 4 and 5 to ten years beyond the liberalization dates in table 1 would reveal the collapse in investment and output associated with the Asian Crisis. In addition to its effect on real variables, there is the additional question of whether liberalizations also amplify the cycle in asset prices (Graciela Laura Kaminsky and Schmukler 2002). For instance, in contrast to the efficient-markets prediction that there should be no revaluation of assets once mar- kets are open to foreign investors, equity prices continue to rise for two to three years after stock market liberalizations but then lose significant value over the longer term (Martell and Stulz 2003). At first glance, the boom?bust cycle in output, investment, and asset prices seems consistent with the view that liberalizations cause crises, but there are at least three important points to keep in mind. First, crises occur not only in countries that liberalize the capital account, but also in those where capital controls are in place (Forbes 2007b). There is, in fact, some sys- tematic evidence that the occurrence of crises and the imposition of capital controls are positively correlated (Reuven Glick and Michael Hutchison 2000; Eichengreen 2003). So it is equally plausible that bad pol- icy precipitates crises and countries impose capital controls as a way to postpone the financial-market consequences (such as depreciation of the currency) of weak or deteriorating fundamentals. In other words, at least some of the existing evidence sup- ports the notion that poor macroeconomic policies, not capital account liberalizations, cause crises. Second, when trying to determine whether liberalizations or fundamentals cause crises, it is important to think critically about timing. The median stock market liberalization date in table 1 is 1989?five years prior to the Mexican Crisis of December 1994 and 924 Journal of Economic Literature, Vol. XLV (December 2007) 47 For a discussion of the welfare implications of sover- eign wealth funds and capital outflows from emerging markets, see Summers (2006). dec07_Article1 11/28/07 12:02 PM Page 924 almost a decade before the Asian crisis of 1997. Given the length of time between the onset of stock market liberalizations and the rash of subsequent emerging-market financial crises, any causal link between the two is far from obvious. Of course, the seeds of a misguided policy may take a long time to yield their bitter fruit, so lags in tim- ing alone cannot dismiss the possibility that stock market liberalizations played a role. Nevertheless, the more proximate and plausible cause of the crises would appear to be a combination of fixed exchange rates and the build-up of short-term dollar- denominated debt: tesobonos in the case of Mexico and interbank loans in the case of Asia (Fischer 2001; Krueger 2000; Mishkin 1997, 2000; Obstfeld and Rogoff 1995). The question of whether stock market lib- eralizations or short-term debt caused the Mexican and Asian crises raises the third and most important point: cost?benefit analyses of capital account liberalization do not make sense without specifying exactly what is meant by the term ?capital account liberalization.? In its broadest form, a capital account liberal- ization can be any decision by a country?s gov- ernment that allows capital to flow more freely in and (or) out of that country. Allowing domestic businesses to take out loans from foreign banks, permitting foreigners to pur- chase domestic government debt instru- ments, and allowing foreigners to invest in the domestic stock market are but three exam- ples. At a minimum, we need to distinguish between two categories of liberalization: those that involve debt and those that involve equity. The distinction matters because the answer to the question?do liberalizations cause crises?depends critically on whether you are talking about the liberalization of debt or equity flows. 8.1 The Importance of Distinguishing Debt from Equity A debt contract requires regular payments regardless of the borrower?s circumstances. This means that, when bad news arises, creditors rush to get their money while they can. An equity contract, on the other hand, involves risk-sharing?large payouts for shareholders when times are good and little to nothing when times are bad. It may seem that foreign purchase of equities on the domestic stock market could be reversed if and when foreign investors become con- cerned about a country?s prospects, but share- holders cannot simply demand their money back. They have to sell their shares and prices will drop as soon as other market participants (domestic or foreign) anticipate the increase in supply. Furthermore, as prices fall, expect- ed returns rise so that the incentive to sell equity is no longer as strong (Hyuk Choe, Bong-Chan Kho, and Stulz 1999). In other words, servicing an equity contract involves procyclical payments that tend to stabilize the balance of payments, whereas debt service payments are countercyclical and have the opposite effect. Because it does not embody risk-sharing, excessive reliance on debt can cause finan- cial distress. In the 1970s, developing-coun- try governments obtained large quantities of floating-rate commercial bank loans. The Debt Crisis of the 1980s then demonstrated that debt contracts can induce large ineffi- ciencies when economic conditions turn out to be worse than anticipated at the time the contract was signed (Fischer 1987). Nor is it simply the flow of asset-service payments that is more likely to vary in a stabilizing way for equity than for debt. ?The debt crisis was caused most immediately and powerful- ly by the cutoff in new lending without any similar curb on the requirement to pay amortization? (Williamson 1997, p. 288). Williamson?s point about the cessation of new lending rings particularly true in the context of the further distinction between bonds and bank loans. The nature of bank lending makes that form of capital flows far more volatile than portfolio bond or equity flows. For instance, domestic banks borrow short-term in the foreign interbank market with the expectation that they will be able to 925Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 925 roll over the loans. The potential problems here are twofold. First, the decision by lenders to call in loans and cease rolling them over tends to be driven not by idiosyn- cratic shocks to individual borrowers but common shocks to the domestic economy (Wendy Dobson and Gary Clyde Hufbauer 2001). Second, common shocks dominate idiosyncratic shocks in emerging markets (Morck, Yeung, and Yu 2000). Taken togeth- er, these two facts create the potential for large-scale reversals of short-term interbank loans. Yet the vast majority of bank lending to emerging markets takes the form of short- term bank loans (loans with a maturity of less than a year).48 At the end of 1997, 55 per- cent of foreign bank loans worldwide were short-term and one third of these loans were of the interbank variety. For instance, for- eign capital flows to Thailand peaked at 25.5 billion dollars in 1995; 75 percent of this sum took the form of bank loans; two thirds of these loans had maturities of less than a year, and the majority went to Thai banks and nonbank financial intermediaries (Martin N. Baily, Diana Farrell, and Susan Lund 2000). The critical point about interbank loans is that they may be used by the recipient banks to make loans to finance domestic projects that may not be short-term, thereby creating a liability mismatch. These liability mis- matches wreak havoc on the domestic econ- omy in the face of external shocks. Excessive short-term borrowing in dollars from foreign banks by Asian banks, companies, and gov- ernments played a central role in the onset of the Asian Financial Crisis (Jason Furman and Stiglitz 1998; Steven Radelet and Sachs 1998). In essence, the mismatch between the term structure of Asian borrowers? assets (long term) and their dollar-denominated external liabilities (short term) meant that any bad news that made their lenders reluc- tant to extend new loans was bound to create a liquidity problem. The tendency toward sudden reversal of short-term bank flows underscores the cen- tral point of the discussion. Whereas portfo- lio flows adjust to shocks through changes in prices, short-term bank loans adjust through quantities (Dobson and Hufbauer 2001). Consequently, bank lending to emerging markets is far more volatile than portfolio investment in bonds or equities (Kose et al. 2006).49 Again, the Asian Crisis highlights the difference in volatility. In 1996, the five Asian Crisis countries (Indonesia, Malaysia, Philippines, South Korea, Thailand) received 47.8 billion dollars of capital inflows in the form of bank loans; in 1997, they experienced a collective outflow of 29.9 billion dollars?a reversal of almost 80 bil- lion dollars in a one-year period (Baily, Farrell, and Lund 2000). In contrast to the abrupt reversal in bank flows, portfolio flows remained positive in Asia throughout 1997. Baily, Farrell, and Lund (2000) document that aggregate port- folio flows to the crisis countries fell by about half but remained positive. Furthermore, the aggregate pattern holds at the individual country level for South Korea, Malyasia, and Thailand. In the case of Thailand, even as foreign banks were refus- ing to roll over loans, portfolio flows increased by 70 percent between the second and third quarters of 1997 and remained positive through the first half of 1998. The authors document that Indonesia did experi- ence net portfolio outflows in the fourth quarter of 1997 but they turned positive again by the middle of 1998. Beyond the Asian Crisis, heavy short-term borrowing in dollars played a central role in precipitating almost every emerging-market financial crisis during the 1990s (Dornbusch 926 Journal of Economic Literature, Vol. XLV (December 2007) 48 As I discuss shortly, it is important to note that bonds, which now comprise a much larger share of total debt flows to developing countries than they did in the 1970s and 1980s, are potentially less problematic than bank loans. 49 It is important to note that investment in bonds can also be volatile if the bonds are sufficiently short-term. dec07_Article1 11/28/07 12:02 PM Page 926 2000; Martin Feldstein 2002). A general finding is that the ratio of short-term debt to reserves predicts crises, and greater short- term exposure predicts more severe crises when positive capital flows turn negative (Rodrik and Andres Velasco 2000). 8.2 If Debt Is So Risky, Why Do We See So Much of It? Table 4 illustrates the dominance of debt over equity by breaking the composition of net capital flows to the eighteen developing countries listed in table 1 into five major cat- egories: public and publicly guaranteed debt flows, private nonguaranteed debt flows, FDI, portfolio equity, and grants. The sum of the first two categories reflects all debt flows. From 1970 to 1984, debt typically accounted for about 80 percent of all capital flows to developing countries. If debt consti- tutes such a dangerous form of lending, why do we do we see so much of it relative to equity? One obvious reason follows directly from the discussion in section 5: prior to the latter half of the 1980s, developing countries largely banned foreigners from holding domestic shares. If debt is so risky relative to equity, then why do governments liberalize debt inflows by removing restrictions on offshore borrow- ing by domestic banks while retaining hard limits on foreign participation in the equity market? There are number of competing explanations for the relatively slow liberal- ization of equity markets, all of which deserve more serious consideration in future research than the speculative treatment I give them in the next few sentences. Domestic capitalists in the nonfinancial sec- tor of the economy may favor liberalization because it reduces their cost of capital. But these capitalists may also be reluctant to cre- ate the necessary preconditions for success- ful equity market liberalization (Shleifer and Wolfenzon 2002). Alternatively, there may be competing interests. Liberalization gen- erates aggregate welfare gains, but there may also be potential losers who oppose the process (Rajan and Zingales 2003). For example, domestic banks may lose monopoly rents because liberalizations provide domes- tic firms with alternative sources of financ- ing. In turn, large, nonfinancial firms may be worse off because new sources of financing for their smaller, more financially con- strained industry peers may increase prod- uct market competition (Chari and Gupta forthcoming). Finally, governments may delay equity market liberalizations for an entirely different set of considerations. Future research should address these issues using firm-level data. Whatever explains the delay, with the advent of stock market liberalization, portfo- lio equity as a fraction of total capital inflows rose from less than 0.1 percent in 1980?84 to 18.7 percent in 1990?95?an increase of almost two-hundredfold. Debt as a fraction of total capital flows fell from 82 percent in 1980?84 to 50 percent in 1990?95. FDI flows as a fraction of total capital flows increased from 13 percent in 1980?84 to 28 percent in 1990?95. However, the flow of portfolio equity to emerging markets has slowed since the boom of the early 1990s (Dilek Aykut, Himmat Kalsi, and Dilip Ratha 2003). And while developing coun- tries? ratios of external debt to equity fell between 1997 and 2001, it is still not clear that they have fallen to prudent levels (Philip Suttle 2003, p. 9). Hence, the recent increase in portfolio equity and FDI inflows notwithstanding, developing countries still lean heavily toward debt.50 The observed pattern of debt and equity flows is an equilibrium outcome, resulting from the optimal response of borrowers and lenders to a given set of institutional arrangements. Therefore, the critical issue is what distortions in the international financial system produce incentives that lead to so 927Henry: Capital Account Liberalization 50 Home bias, the fact that foreign investors hold fewer foreign securities than they should, does not explain why the composition of securities they do hold favors debt so strongly. Lewis (1999) surveys the home-bias literature. dec07_Article1 11/28/07 12:02 PM Page 927 much debt and so little equity. One promi- nent example of an incentive-distorting insti- tutional arrangement was the Basel Capital Accord of 1988 (Basel I). Drafted in response to the 1980s debt cri- sis, Basel I tried to reduce systemic risk by linking banks? capital adequacy ratios to the riskiness of their loans. Under Basel I, banks had to set aside capital equal to at least 8 percent of their risk-adjusted assets to pro- vide a buffer in case of a loss on those assets. The problem was that Basel I?s system for risk-adjusting assets created an unintention- al bias toward short-term lending to emerg- ing markets. Lending to banks in OECD countries received a 20 percent risk weight- ing regardless of the maturity structure of the loan. For loans to non-OECD countries, short-term loans received a 20 percent weight, long-term loans a 100 percent weight (Rudi Bonte 1999). As a consequence of Basel I?s weighting system, short-term, foreign currency lending to banks in emerging markets required only one fifth of the capital required of long-term loans to banks in emerging markets and no more capital than a long-term loan to a bank in an OECD country. Skewing the incentive structure for banks in G-10 countries to lend short-term to emerging markets had pre- dictable consequences. As discussed in sec- tion 8.1, short-term, dollar-denominated loans, most of which originated with European and Japanese banks, constituted the lion?s share of debt contracts in the East Asian Crisis. The eventual implementation of Basel II may help alleviate the bias toward short-term debt in Basel I, but Rogoff (1999) identifies three other sources of debt bias in the international financial system not directly related to capital adequacy ratios. First, by making it less risky to hold deposits, deposit insurance in creditor and debtor countries increases the deposit base and expands the size of the banking system. Implicit subsidies are also a problem. As the size of banking system expands, it becomes increasingly difficult for the government to credibly commit to not bail out the banking system in the event of a financial meltdown (George A. Akerlof and Pual M. Romer 1993). Somewhat perversely, deposit insur- ance may also raise the probability of bank failures by reducing the incentive for depos- itors to monitor the lending practices of the financial institutions where they keep their money.51 Second, the international financial system protects the rights of debt holders more vig- ilantly than those of equity holders. G-7 lenders to emerging-market countries can resort to G-7 courts in the event of a debt dispute, but there is no such avenue of recourse for G-7 holders of emerging-mar- ket equity (Bulow 2002; Rogoff 1999; Obstfeld 1998). Third, the underdevelopment of equity markets in emerging economies exacerbates the problem of preferential treatment of debtholders in G-7 courts. The point here is closely related to the discussion in section 6.1.3. A lack of transparency in emerging equity markets makes foreigners reluctant to invest, and weak protection of the rights of equity investors reinforces the tendency of capital suppliers to purchase debt instead of equity (Henry and Peter Lombard Lorentzen 2003). 9. Conclusion Writing about capital account liberaliza- tion in 1998, Bhagwati threw down the gauntlet, declaring ?It is time to shift the burden of proof from those who oppose to those who favor liberated capital? (Bhagwati 1998). The explosion of papers written on the subject since that time indicates the seri- ousness with which the profession has taken his challenge. There is little evidence that economic growth and capital account openness are 928 Journal of Economic Literature, Vol. XLV (December 2007) 51 Rogoff (1999) also argues that funds from interna- tional financial institutions aimed at helping distressed- country debtors also provide an implicit subsidy to G-7 debt holders that encourages debt financing over equity. dec07_Article1 11/28/07 12:02 PM Page 928 positively correlated across countries. But there is lots of evidence that opening the capital account leads countries to temporar- ily invest more and grow faster than they did when their capital accounts were closed. Why does so much of the literature focus on the relationship between openness and long- run growth when the predictions of the neo- classical model all point toward the short-run impact of a country opening up? Part of the answer is tradition. Cross-sec- tional regressions of national growth rates on policy variables have been around for a while, so the gravitational pull of that methodological approach is quite strong. But I also think that the answer has some- thing to do with a professional obsession. There has always been a great deal of inter- est in uncovering policies that increase the steady-state rate of growth. As a conse- quence, economists tend to ignore the importance of short-run increases that per- manently raise the path of national income to a higher but parallel trajectory (Harberger 2005; Solow 2000, pp. 182?83). REFERENCES Abdelal, Rawi, and Laura Alfaro. 2002. ?Malaysia: Capital and Control.? Harvard Business School Case, no. 702-040. Abdelal, Rawi, and Laura Alfaro. 2003. ?Capital and Control: Lessons from Malaysia.? Challenge, 46(4): 36?53. Aitken, Brian J., and Ann E. Harrison. 1999. ?Do Domestic Firms Benefit from Direct Foreign Investment? Evidence from Venezuela.? American Economic Review, 89(3): 605?18. Akerlof, George A., and Paul M. Romer. 1993. ?Looting: The Economic Underworld of Bankruptcy for Profit.? Brookings Papers on Economic Activity, 2: 1?60. Alesina, Alberto, Vittorio Grilli, and Gian Maria Milesi- Ferretti. 1994. ?The Political Economy of Capital Controls.? In Capital Mobility: The Impact on Consumption, Investment and Growth, ed. L. Leiderman and A. Razin. Cambridge; New York and Melbourne: Cambridge University Press, 289?321. Alfaro, Laura, and Eliza Hammel. 2007. ?Capital Flows and Capital Goods.? Journal of International Economics, 72(1): 128?50. Alfaro, Laura, Sebnem Kalemli-Ozcan, and Vadym Volosovych. Forthcoming. ?Why Doesn?t Capital Flow from Rich to Poor Countries? An Empirical Investigation.? Review of Economics and Statistics. Arslanalp, Serkan, and Peter Blair Henry. 2005. ?Is Debt Relief Efficient?? Journal of Finance, 60(2): 1017?51. Arslanalp, Serkan, and Peter Blair Henry. 2006a. ?Helping the Poor to Help Themselves: Debt Relief or Aid?? In Sovereign Debt at the Crossroads: Challenges and Proposals for Resolving the Third World Debt Crisis, ed. C. Jochnick and F. Preston. Oxford and New York: Oxford University Press, 174?93. Arslanalp, Serkan, and Peter Blair Henry. 2006b. ?Policy Watch: Debt Relief.? Journal of Economic Perspectives, 20(1): 207?20. Arteta, Carlos, Barry Eichengreen, and Charles Wyplosz. 2001. ?When Does Capital Account Liberalization Help More Than It Hurts?? NBER Working Papers, no. 8414. Athanasoulis, Stefano G., and Eric Van Wincoop. 2000. ?Growth Uncertainty and Risksharing.? Journal of Monetary Economics, 45(3): 477?505. Aykut, Dilek, Himmat Kalsi, and Dilip Ratha. 2003. ?Sustaining and Promoting Equity-Related Finance for Developing Countries.? In Global Development Finance 2003: Striving for Stability in Development Finance, Volume 2: Summary and Country Tables. Washington, D.C.: World Bank, 85?106. Baily, Martin N., Diana Farrell, and Susan Lund. 2000. ?The Color of Hot Money.? Foreign Affairs, 79(2): 99?109. Barro, Robert J. 1991. ?Economic Growth in a Cross Section of Countries.? Quarterly Journal of Economics, 106(2): 407?43. Barro, Robert J., and Xavier Sala-i-Martin. 1992. ?Convergence.? Journal of Political Economy, 100(2): 223?51. Barro, Robert J., and Xavier Sala-i-Martin. 1995. Economic Growth. New York; London and Montreal: McGraw-Hill. Bekaert, Geert, and Campbell R. Harvey. 2000. ?Foreign Speculators and Emerging Equity Markets.? Journal of Finance, 55(2): 565?613. Bekaert, Geert, Campbell R. Harvey, and Christian Lundblad. 2005. ?Does Financial Liberalization Spur Growth?? Journal of Financial Economics, 77(1): 3?55. Bhagwati, Jagdish. 1998. ?The Capital Myth: The Difference between Trade in Widgets and Dollars.? Foreign Affairs, 77(3): 7?12. Blanchard, Olivier, Changyong Rhee, and Lawrence H. Summers. 1993. ?The Stock Market, Profit, and Investment.? Quarterly Journal of Economics, 108(1): 115?36. Bodie, Zvi, and Robert C. Merton. 2002. ?International Pension Swaps.? Journal of Pension Economics and Finance, 1(1): 77?83. Bonte, Rudi. 1999. ?Supervisory Lessons to Be Drawn from the Asian Crisis.? Basel Committee on Banking Supervision Working Paper, no. 2. Booth, Laurence, Varouj Aivazian, Asli Demirg??- Kunt, and Vojislav Maksimovic. 2001. ?Capital Structures in Developing Countries.? Journal of Finance, 56(1): 87?130. Bosworth, Barry P., and Susan M. Collins. 2003. ?The Empirics of Growth: An Update.? Brookings Papers 929Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 929 on Economic Activity, 2: 113?79. Bruno, Michael, and William Easterly. 1998. ?Inflation Crises and Long-Run Growth.? Journal of Monetary Economics, 41(1): 3?26. Bulow, Jeremy I. 2002. ?First World Governments and Third World Debt.? Brookings Papers on Economic Activity, 1: 229?55. Bulow, Jeremy I., and Lawrence H. Summers. 1984. ?The Taxation of Risky Assets.? Journal of Political Economy, 92(1): 20?39. Calder?n, C?sar, Norman Loayza, and Klaus Schmidt- Hebbel. 2004. ?Openness, Vulnerability, and Growth.? Central Bank of Chile Working Paper. Calvo, Guillermo A., Leonardo Leiderman, and Carmen M. Reinhart. 1996. ?Inflows of Capital to Developing Countries in the 1990s.? Journal of Economic Perspectives, 10(2): 123?39. Calvo, Guillermo A., and Carlos A. V?gh. 1999. ?Inflation Stabilization and BOP Crises in Developing Countries.? In Handbook of Macroeconomics, Volume 1C, ed. J. B. Taylor and M. Woodford. Handbooks in Economics, vol. 15. Amsterdam; New York and Oxford: Elsevier Science, North-Holland, 1531?1614. Campbell, John Y. 2000. ?Asset Pricing at the Millennium.? Journal of Finance, 55(4): 1515?67. Caselli, Francesco. 2005. ?Accounting for Cross- Country Income Differences.? In Handbook of Economic Growth, Volume 1A, ed. P. Aghion and S. Durlauf. Amsterdam and San Diego: North Holland, 679?742. Caselli, Francesco, Gerardo Esquivel, and Fernando Lefort. 1996. ?Reopening the Convergence Debate: A New Look at Cross-Country Growth Empirics.? Journal of Economic Growth, 1(3): 363?89. Caselli, Francesco, and James Feyrer. 2007. ?The Marginal Product of Capital.? Quarterly Journal of Economics, 122(2): 535?68. Chari, Anusha, and Nandini Gupta. Forthcoming. ?Incumbents and Protectionism: The Political Economy of Foreign Entry Liberalization.? Journal of Financial Economics. Chari, Anusha, and Peter Blair Henry. 2004. ?Risk Sharing and Asset Prices: Evidence from a Natural Experiment.? Journal of Finance, 59(3): 1295?1324. Chari, Anusha, and Peter Blair Henry. Forthcoming. ?Firm-Specific Information and the Efficiency of Investment.? Journal of Financial Economics. Chari, Anusha, Paige P. Ouimet, and Linda L. Tesar. 2004. ?Acquiring Control in Emerging Markets: Evidence from the Stock Market.? NBER Working Papers, no. 10872. Chinn, Menzie D., and Hiro Ito. 2006. ?What Matters for Financial Development? Capital Controls, Institutions, and Interactions.? Journal of Development Economics, 81(1): 163?92. Choe, Hyuk, Bong-Chan Kho, and Ren? M. Stulz. 1999. ?Do Foreign Investors Destabilize Stock Markets? The Korean Experience in 1997.? Journal of Financial Economics, 54(2): 227?64. Claessens, Stijn, and Moon-Whoan Rhee. 1994. ?The Effect of Barriers to Equity Investment in Developing Countries.? In The Internationalization of Equity Markets, ed. J. A. Frankel. Chicago and London: University of Chicago Press, 231?71. Cline, William R. 1995. International Debt Reexamined. Washington, D.C.: Institute for International Economics. Cochrane, John H. 1999. ?New Facts in Finance.? Federal Reserve Bank of Chicago Economic Perspectives, 23(3): 36?58. Coval, Joshua D., and Tobias J. Moskowitz. 1999. ?Home Bias at Home: Local Equity Preference in Domestic Portfolios.? Journal of Finance, 54(6): 2045?73. Coval, Joshua D., and Tobias J. Moskowitz. 2001. ?The Geography of Investment: Informed Trading and Asset Prices.? Journal of Political Economy, 109(4): 811?41. De Jong, Frank, and Frans A. De Roon. 2005. ?Time- Varying Market Integration and Expected Returns in Emerging Markets.? Journal of Financial Economics, 78(3): 583?613. Demirg??-Kunt, Asli, and Vojislav Maksimovic. 1998. ?Law, Finance, and Firm Growth.? Journal of Finance, 53(6): 2107?37. Diaz-Alejandro, Carlos. 1985. ?Good-Bye Financial Repression, Hello Financial Crash.? Journal of Development Economics, 19(1?2): 1?24. Dobson, Wendy, and Gary Clyde Hufbauer. 2001. World Capital Markets: Challenge to the G?10. Washington, D.C.: Institute for International Economics. Dornbusch, Rudiger. 1983. ?Panel Discussion on the Southern Cone.? International Monetary Fund Staff Papers, 30: 164?84. Dornbusch, Rudiger. 2000. Keys to Prosperity: Free Markets, Sound Money, and a Bit of Luck. Cambridge and London: MIT Press. Dvorak, Tomas, and Richard Podpiera. 2006. ?European Union Enlargement and Equity Markets in Accession Countries.? Emerging Markets Review, 7(2): 129?46. Easterly, William. 1996. ?When Is Stabilization Expansionary? Evidence from High Inflation.? Economic Policy: A European Forum, 22: 65?98. Edison, Hali J., Michael W. Klein, Luca Antonio Ricci, and Torsten Sl?k. 2004. ?Capital Account Liberalization and Economic Performance: Survey and Synthesis.? IMF Staff Papers, 51(2): 220?56. Edison, Hali J., Ross Levine, Luca Antonio Ricci, and Torsten Sl?k. 2002. ?International Financial Integration and Economic Growth.? Journal of International Money and Finance, 21(6): 749?76. Edison, Hali J., and Francis E. Warnock. 2003. ?A Simple Measure of the Intensity of Capital Controls.? Journal of Empirical Finance, 10(1?2): 81?103. Edwards, Sebastian. 1984. The Order of Liberalization of the External Sector in Developing Countries. Essays in International Finance, no. 156. Princeton: International Finance Section, Department of Economics, Princeton University. Edwards, Sebastian. 2001. ?Capital Mobility and Economic Performance: Are Emerging Economies Different?? NBER Working Papers, no. 8076. Eichengreen, Barry. 2001. ?Capital Account Liberal- ization: What Do Cross-Country Studies Tell Us?? World Bank Economic Review, 15(3): 341?65. 930 Journal of Economic Literature, Vol. XLV (December 2007) dec07_Article1 11/28/07 12:02 PM Page 930 Eichengreen, Barry. 2003. Capital Flows and Crises. Cambridge and London: MIT Press. Eichengreen, Barry. 2004. ?Financial Instability.? In Global Crises, Global Solutions, ed. B. Lomborg. Cambridge; New York and Melbourne: Cambridge University Press, 251?80. Errunza, Vihang R., and Darius P. Miller. 2000. ?Market Segmentation and the Cost of Capital in International Equity Markets.? Journal of Financial and Quantitative Analysis, 35(4): 577?600. Fama, Eugene F. 1991. ?Efficient Capital Markets: II.? Journal of Finance, 46(5): 1575?1617. Fama, Eugene F., and Kenneth R. French. 2004. ?The Capital Asset Pricing Model: Theory and Evidence.? Journal of Economic Perspectives, 18(3): 25?46. Fazzari, Steven M., Robert Glenn Hubbard, and Bruce C. Petersen. 1988. ?Financing Constraints and Corporate Investment.? Brookings Papers on Econ- omic Activity, 1: 141?95. Feldstein, Martin. 2002. ?Economic and Financial Crises in Emerging Market Economies: Overview of Prevention and Management.? NBER Working Papers, no. 8837. Fischer, Stanley. 1987. ?Sharing the Burden of the International Debt Crisis.? American Economic Review, 77(2): 165?70. Fischer, Stanley. 1998. ?Capital Account Liberalization and the Role of the IMF.? In Should the IMF Pursue Capital-Account Convertibility. Essays in International Finance, no. 207. Princeton: International Finance Section, Department of Economics, Princeton University, 1?10. Fischer, Stanley. 2001. ?Exchange Rate Regimes: Is the Bipolar View Correct?? Journal of Economic Perspectives, 15(2): 3?24. Fischer, Stanley. 2003. ?Globalization and Its Challenges.? American Economic Review, 93(2): 1?30. Fischer, Stanley, and Robert C. Merton. 1984. ?Macroeconomics and Finance: The Role of the Stock Market.? Carnegie?Rochester Conference Series on Public Policy, 21: 57?108. Fischer, Stanley, Ratna Sahay, and Carlos A. V?gh. 2002. ?Modern Hyper- and High Inflations.? Journal of Economic Literature, 40(3): 837?80. Foerster, Stephen R., and G. Andrew Karolyi. 1999. ?The Effects of Market Segmentation and Investor Recognition on Asset Prices: Evidence from Foreign Stocks Listing in the United States.? Journal of Finance, 54(3): 981?1013. Forbes, Kristin J. 2007a. ?One Cost of the Chilean Capital Controls: Increased Financial Constraints for Smaller Traded Firms.? Journal of International Economics, 71(2): 294?323. Forbes, Kristin J. 2007b. ?The Microeconomic Evidence on Capital Controls: No Free Lunch.? In Capital Controls and Capital Flows in Emerging Economies: Policies, Practices, and Consequences, ed. S. Edwards. Chicago: University of Chicago Press. Frankel, Jeffrey A. 1994. ?The Internationalization of Equity Markets: Introduction.? In The International- ization of Equity Markets, ed. J. A. Frankel. Chicago and London: University of Chicago Press, 1?20. Frankel, Jeffrey A., and David Romer. 1999. ?Does Trade Cause Growth?? American Economic Review, 89(3): 379?99. Furman, Jason, and Joseph E. Stiglitz. 1998. ?Economic Crises: Evidence and Insights from East Asia.? Brookings Papers on Economic Activity, 2: 1?114. Galindo, Arturo, Fabio Schiantarelli, and Andrew Weiss. 2007. ?Does Financial Liberalization Improve the Allocation of Investment? Micro-Evidence from Developing Countries.? Journal of Development Economics, 83(2): 562?87. Gallego, Francisco A., and F. Leonardo Hernandez. 2003. ?Microeconomic Effects of Capital Controls: The Chilean Experience during the 1990s.? International Journal of Finance and Economics, 8(3): 225?53. Glick, Reuven, Xueyan Guo, and Michael Hutchison. 2006. ?Currency Crises, Capital-Account Liberali- zation, and Selection Bias.? Journal of Economics and Statistics, 88(4): 698?714. Glick, Reuven, and Michael Hutchison. 2005. ?Capital Controls and Exchange Rate Instability in Developing Economies.? Journal of International Money and Finance, 24(3): 387?412. Gordon, Myron J. 1962. The Investment, Financing, and Valuation of the Corporation. Homewood, Ill.: Richard D. Irwin. Gorg, Holger, and David Greenaway. 2004. ?Much Ado About Nothing? Do Domestic Firms Really Benefit from Foreign Direct Investment?? World Bank Research Observer, 19(2): 171?97. Gourinchas, Pierre-Olivier, and Olivier Jeanne. 2006. ?The Elusive Gains from International Financial Integration.? Review of Economic Studies, 73(3): 715?41. Gozzi, Juan Carlos, Ross Levine, and Sergio L. Schmukler. Forthcoming. ?Internationalization and the Evolution of Corporate Valuation.? Journal of Financial Economics. Grilli, Vittorio, and Gian Maria Milesi-Ferretti. 1995. ?Economic Effects and Structural Determinants of Capital Controls.? International Monetary Fund Staff Papers, 42(3): 517?51. Gupta, Nandini, and Kathy Yuan. 2005. ?On the Growth Effects of Liberalizations.? University of Michigan Working Paper. Haddad, Mona, and Ann E. Harrison. 1993. ?Are There Positive Spillovers from Direct Foreign Investment? Evidence from Panel Data for Morocco.? Journal of Development Economics, 42(1): 51?74. Hall, Robert E., and Charles I. Jones. 1999. ?Why Do Some Countries Produce So Much More Output per Worker Than Others?? Quarterly Journal of Economics, 114(1): 83?116. Harberger, Arnold C. 1998. ?A Vision of the Growth Process.? American Economic Review, 88(1): 1?32. Harberger, Arnold C. 2005. ?On the Process of Growth and Economic Policy in Developing Countries.? USAID PPC Issue Paper no. 13. Harrison, Ann E., Inessa Love, and Margaret S. McMillan. 2004. ?Global Capital Flows and Financing Constraints.? Journal of Development Economics, 75(1): 269?301. 931Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 931 Henry, Peter Blair. 2000a. ?Do Stock Market Liberal- izations Cause Investment Booms?? Journal of Financial Economics, 58(1?2): 301?34. Henry, Peter Blair. 2000b. ?Stock Market Liberaliza- tion, Economic Reform, and Emerging Market Equity Prices.? Journal of Finance, 55(2): 529?64. Henry, Peter Blair. 2002. ?Is Disinflation Good for the Stock Market?? Journal of Finance, 57(4): 1617?48. Henry, Peter Blair. 2003. ?Capital-Account Liberali- zation, the Cost of Capital, and Economic Growth.? American Economic Review, 93(2): 91?96. Henry, Peter Blair, and Peter Lombard Lorentzen. 2003. ?Domestic Capital Market Reform and Access to Global Finance: Making Markets Work.? In The Future of Domestic Capital Markets in Developing Countries, ed. R. E. Litan, M. Pomerleano, and V. Sundararajan. Washington, D.C.: Brookings Institu- tion, 179?214. Heritage Foundation. 2006. Index of Economic Freedom. Washington, D.C.: Heritage House. Himmelberg, Charles P., Robert Glenn Hubbard, and Inessa Love. 2004. ?Investor Protection, Ownership, and the Cost of Capital.? World Bank Policy Research Working Paper Series, no. 2834. Hubbard, Robert Glenn. 1998. ?Capital-Market Imperfections and Investment.? Journal of Economic Literature, 36(1): 193?225. Jensen, Michael C. 1986. ?Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.? American Economic Review, 76(2): 323?29. Johnson, Simon, Peter Boone, Alasdair Breach, and Eric Friedman. 2000. ?Corporate Governance in the Asian Financial Crisis.? Journal of Financial Economics, 58(1?2): 141?86. Johnson, Simon, and Todd Mitton. 2003. ?Cronyism and Capital Controls: Evidence from Malaysia.? Journal of Financial Economics, 67(2): 351?82. Ju, Jiandong, and Shang-Jin Wei. 2006. ?A Solution to Two Paradoxes of International Capital Flows.? NBER Working Papers, no. 12668. Kaminsky, Graciela Laura, and Sergio L. Schmukler. 2002. ?Short-Run Pain, Long-Run Gain: The Effects of Financial Liberalization.? NBER Working Papers, no. 9787. Kaplan, Steven N., and Luigi Zingales. 1997. ?Do Investment?Cash Flow Sensitivities Provide Useful Measures of Financing Constraints.? Quarterly Journal of Economics, 112(1): 169?215. Kaplan, Steven N., and Luigi Zingales. 2000. ?Investment?Cash Flow Sensitivities Are Not Valid Measures of Financing Constraints.? Quarterly Journal of Economics, 115(2): 707?12. Karolyi, G. Andrew. 1998. ?Why Do Companies List Shares Abroad?: A Survey of the Evidence and Its Managerial Implications.? Financial Markets, Institutions and Instruments, 7(1): 1?60. Karolyi, G. Andrew. 2004. ?The Role of American Depositary Receipts in the Development of Emerging Equity Markets.? Review of Economics and Statistics, 86(3): 670?90. Karolyi, G. Andrew, and Ren? M. Stulz. 2003. ?Are Financial Assets Priced Locally or Globally?? In Handbook of the Economics of Finance, Volume 1B: Financial Markets and Asset Pricing, ed. G. M. Constantinides, M. Harris, and R. Stulz. Amsterdam; London and New York: Elsevier, North Holland, 975?1020. Kim, E. Han, and Vijay Singal. 2000. ?Stock Market Openings: Experience of Emerging Economies.? Journal of Business, 73(1): 25?66. King, Robert G., and Ross Levine. 1993. ?Finance and Growth: Schumpeter Might Be Right.? Quarterly Journal of Economics, 108(3): 717?37. Klein, Michael W., and Giovanni Olivei. Forthcoming. ?Capital Account Liberalization, Financial Deepness and Economic Growth.? Journal of International Money and Finance. Kose, M. Ayhan, Eswar S. Prasad, Kenneth S. Rogoff, and Shang-Jin Wei. 2006. ?Financial Globalization: A Reappraisal.? NBER Working Papers, no. 12484. Kraay, Aart. 1998. ?In Search of the Macroeconomic Effects of Capital Account Liberalization.? Unpublished. Krueger, Anne O. 1997. ?Trade Policy and Economic Development: How We Learn.? American Economic Review, 87(1): 1?22. Krueger, Anne O. 1998. ?Why Trade Liberalisation Is Good for Growth.? Economic Journal, 108(450): 1513?22. Krueger, Anne O. 2000. ?Conflicting Demands on the International Monetary Fund.? American Economic Review, 90(2): 38?42. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny. 1997. ?Legal Determinants of External Finance.? Journal of Finance, 52(3): 1131?50. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny. 1998. ?Law and Finance.? Journal of Political Economy, 106(6): 1113?55. La Porta, Rafael, Florencio Lopez-de Silanes, Andrei Shleifer, and Robert W. Vishny. 2002. ?Investor Protection and Corporate Valuation.? Journal of Finance, 57(3): 1147?70. Laeven, Luc. 2003. ?Does Financial Liberalization Reduce Financing Constraints?? Financial Manage- ment, 32(1): 5?34. Lessard, Donald R. 1973. ?International Portfolio Diversification: A Multivariate Analysis for a Group of Latin American Countries.? Journal of Finance, 28(3): 619?33. Lessard, Donald R. 1974. ?World, National, and Industry Factors in Equity Returns.? Journal of Finance, 29(2): 379?91. Lessard, Donald R. 1994. ?The Effect of Barriers to Equity Investment in Developing Countries: Comment.? In The Internationalization of Equity Markets, ed. J. A. Frankel. Chicago and London: University of Chicago Press, 273?75. Levine, Ross. 1997. ?Financial Development and Economic Growth: Views and Agenda.? Journal of Economic Literature, 35(2): 688?726. Levine, Ross. 2001. ?International Financial Liberali- zation and Economic Growth.? Review of International Economics, 9(4): 688?702. Levine, Ross, and Sara Zervos. 1994. ?Capital Control Liberalization and Stock Market Development: Data Annex of Country Policy Changes.? Unpublished. 932 Journal of Economic Literature, Vol. XLV (December 2007) dec07_Article1 11/28/07 12:02 PM Page 932 Levine, Ross, and Sara Zervos. 1998a. ?Capital Control Liberalization and Stock Market Development.? World Development, 26(7): 1169?83. Levine, Ross, and Sara Zervos. 1998b. ?Stock Markets, Banks, and Economic Growth.? American Economic Review, 88(3): 537?58. Lewis, Karen K. 1999. ?Trying to Explain Home Bias in Equities and Consumption.? Journal of Economic Literature, 37(2): 571?608. Lewis, Karen K. 2000. ?Why Do Stocks and Consumption Imply Such Different Gains from International Risk Sharing?? Journal of International Economics, 52(1): 1?35. Li, Kan, Randall K. Morck, Fan Yang, and Bernard Yeung. 2004. ?Firm-Specific Variation and Openness in Emerging Markets.? Review of Economics and Statistics, 86(3): 658?69. Love, Inessa. 2003. ?Financial Development and Financing Constraints: International Evidence from the Structural Investment Model.? Review of Financial Studies, 16(3): 765?91. Lucas, Robert E., Jr. 1990. ?Why Doesn?t Capital Flow from Rich to Poor Countries?? American Economic Review, 80(2): 92?96. MacKinlay, A. Craig. 1997. ?Event Studies in Economics and Finance.? Journal of Economic Literature, 35(1): 13?39. Mankiw, N. Gregory. 1995. ?The Growth of Nations.? Brookings Papers on Economic Activity, 1: 275?310. Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. ?A Contribution to the Empirics of Economic Growth.? Quarterly Journal of Economics, 107(2): 407?37. Martell, Rodolfo, and Ren? M. Stulz. 2003. ?Equity- Market Liberalizations as Country IPO?s.? American Economic Review, 93(2): 97?101. Mathieson, Donald J., and Liliana Rojas-Su?rez. 1993. ?Liberalization of the Capital Account: Experiences and Issues.? IMF Occasional Paper, no. 103. McKinnon, Ronald I. 1973. Money and Capital in Economic Development. Washington, D.C.: Brookings Institution Press. McKinnon, Ronald I. 1991. The Order of Economic Liberalization: Financial Control in the Transition to a Market Economy. Baltimore and London: Johns Hopkins University Press. Megginson, William L., and Jeffry M. Netter. 2001. ?From State to Market: A Survey of Empirical Studies on Privatization.? Journal of Economic Literature, 39(2): 321?89. Mishkin, Frederic S. 1997. ?Understanding Financial Crises: A Developing Country Perspective.? In Annual World Bank Conference on Development Economics, 1996, ed. M. Bruno and B. Pleskovic. Washington, D.C.: World Bank, 29?62. Mishkin, Frederic S. 2000. ?Financial Market Reform.? In Economic Policy Reform: The Second Stage, ed. A. Krueger. Chicago: University of Chicago Press, 511?50. Mishkin, Frederic S. 2006. The Next Great Globalization: How Disadvantaged Nations Can Harness Their Financial Systems to Get Rich. Princeton and Oxford: Princeton University Press. Mishkin, Frederic S. 2007. ?Is Financial Globalization Beneficial?? Journal of Money, Credit, and Banking, 39(2?3): 259?94. Mitton, Todd. 2006. ?Stock Market Liberalization and Operating Performance at the Firm Level.? Journal of Financial Economics, 81(3): 625?47. Morck, Randall K., David A. Stangeland, and Bernard Yeung. 2000. ?Inherited Wealth, Corporate Control, and Economic Growth: The Canadian Disease?? In Concentrated Corporate Ownership, ed. R. K. Morck. Chicago and London: University of Chicago Press, 319?69. Morck, Randall K., Bernard Yeung, and Wayne Yu. 2000. ?The Information Content of Stock Markets: Why Do Emerging Markets Have Synchronous Stock Price Movements?? Journal of Financial Economics, 58(1?2): 215?60. Obstfeld, Maurice. 1994. ?Risk-Taking, Global Diversification, and Growth.? American Economic Review, 84(5): 1310?29. Obstfeld, Maurice. 1998. ?The Global Capital Market: Benefactor or Menace?? Journal of Economic Perspectives, 12(4): 9?30. Obstfeld, Maurice, and Kenneth S. Rogoff. 1995. ?The Mirage of Fixed Exchange Rates.? Journal of Economic Perspectives, 9(4): 73?96. Obstfeld, Maurice, and Kenneth S. Rogoff. 1996. Foundations of International Macroeconomics. Cambridge, Mass. and London: MIT Press. Park, Keith, and Antoine W. Van Agtmael. 1993. The World?s Emerging Stock Markets: Structure, Development, Regulations and Opportunities. Chicago: Probus Publishing. Patro, Dilip K., and John K. Wald. 2005. ?Firm Characteristics and the Impact of Emerging Market Liberalizations.? Journal of Banking and Finance, 29(7): 1671?95. P?rez Gonz?lez, Francisco. 2005. ?The Impact of Acquiring Control on Productivity.? Unpublished. Polk, Christopher, Samuel Thompson, and Tuomo Vuolteenaho. 2006. ?Cross-Sectional Forecasts of the Equity Premium.? Journal of Financial Economics, 81(1): 101?41. Portes, Richard, and H?le`ne Rey. 2005. ?The Determinants of Cross-Border Equity Flows.? Journal of International Economics, 65(2): 269?96. Prasad, Eswar S., Kenneth S. Rogoff, Shang-Jin Wei, and M. Ayan Kose. 2003. ?Effects of Financial Globalization on Developing Countries: Some Empirical Evidence.? IMF Occasional Paper, no. 220. Price, Margaret M. 1994. Emerging Stock Markets: A Complete Investment Guide to New Markets Around the World. New York: McGraw-Hill. Quinn, Dennis. 1997. ?The Correlates of Change in International Financial Regulation.? American Political Science Review, 91(3): 531?51. Radelet, Steven, and Jeffrey D. Sachs. 1998. ?The East Asian Financial Crisis: Diagnosis, Remedies, Prospects.? Brookings Papers on Economic Activity, 1: 1?74. Rajan, Raghuram G. 2000. ?NBER Corporate Finance Program Report.? http://www.nber.org/programs/ cf/cf.html. Rajan, Raghuram G., and Luigi Zingales. 1998. ?Financial Dependence and Growth.? American 933Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 933 Economic Review, 88(3): 559?86. Rajan, Raghuram G., and Luigi Zingales. 2003. ?The Great Reversals: The Politics of Financial Development in the Twentieth Century.? Journal of Financial Economics, 69(1): 5?50. Reisen, Helmut, and Bernhard Fischer. 1993. Financial Opening: Policy Issues and Experiences in Developing Countries. Paris: OECD Development Centre. Rodrik, Dani. 1998. ?Who Needs Capital-Account Convertibility?? In Should the IMF Pursue Capital- Account Convertibility. Essays in International Finance, no. 207. Princeton: International Finance Section, Department of Economics, Princeton University, 55?65. Rodrik, Dani, and Andres Velasco. 2000. ?Short-Term Capital Flows.? In Annual World Bank Conference on Development Economics, 1999, ed. B. Pleskovic and J. E. Stiglitz. Washington, D.C.: World Bank, 59?90. Rogoff, Kenneth S. 1999. ?International Institutions for Reducing Global Financial Instability.? Journal of Economic Perspectives, 13(4): 21?42. Sachs, Jeffrey D., and Andrew M. Warner. 1995. ?Economic Reform and the Process of Global Integration.? Brookings Papers on Economic Activity, 1: 1?95. Shaw, Edward S. 1973. Financial Deepening in Economic Development. New York: Oxford University Press. Shin, Inseok, and Changyun Park. Forthcoming. ?Stock Market Opening and the Cost of Capital: The Case of Korea.? In International Financial Issues in the Pacific Rim: Global Imbalances, Financial Liberalization, and Exchange Rate Policy, ed. T. Ito and A. Rose. Chicago: University of Chicago Press. Shleifer, Andrei, and Robert W. Vishny. 1997. ?A Survey of Corporate Governance.? Journal of Finance, 52(2): 737?83. Shleifer, Andrei, and Daniel Wolfenzon. 2002. ?Investor Protection and Equity Markets.? Journal of Financial Economics, 66(1): 3?27. Solow, Robert M. 1956. ?A Contribution to the Theory of Economic Growth.? Quarterly Journal of Economics, 70(1): 65?94. Solow, Robert M. 2000. Growth Theory: An Exposition. Second edition. New York and Oxford: Oxford University Press. Solow, Robert M. 2001. ?Applying Growth Theory across Countries.? World Bank Economic Review, 15(2): 283?88. Stein, Jeremy C. 2003. ?Agency, Information and Corporate Investment.? In Handbook of the Economics of Finance, Volume 1A: Corporate Finance, ed. G. M. Constantinides, M. Harris, and R. Stulz. Amsterdam; London and New York: Elsevier, North Holland, 111?65. Stiglitz, Joseph E. 1999. ?Reforming the Global Economic Architecture: Lessons from the Recent Crises.? Journal of Finance, 54(4): 1508?21. Stiglitz, Joseph E. 2000. ?Capital Market Liberalization, Economic Growth, and Instability.? World Development, 28(6): 1075?86. Stiglitz, Joseph E. 2002. Globalization and Its Discontents. New York and London: Norton. Stulz, Ren? M. 1999a. ?Globalization, Corporate Finance, and the Cost of Capital.? Journal of Applied Corporate Finance, 12(3): 8?25. Stulz, Ren? M. 1999b. ?Globalizaton of Equity Markets and the Cost of Capital.? New York Stock Exchange Working Paper, no. 99-2. Stulz, Ren? M. 1999c. ?International Portfolio Flows and Security Markets.? In International Capital Flows, ed. M. Feldstein. Chicago and London: University of Chicago Press, 257?93. Stulz, Ren? M. 2005. ?The Limits of Financial Globalization.? Journal of Finance, 60(4): 1595?1638. Summers, Lawrence H. 1994. ?A Changing Course toward Privatization.? In Does Privatization Deliver? Highlights from a World Bank Conference, ed. A. Galal and M. Shirley. Washington, D.C.: World Bank, 11?17. Summers, Lawrence H. 2000. ?International Financial Crises: Causes, Prevention, and Cures.? American Economic Review, 90(2): 1?16. Summers, Lawrence H. 2006. ?Reflections on Global Account Imbalances and Emerging Markets Reserve Accumulation.? L. K. Jha Memorial Lecture, Reserve Bank of India. http://www.president.har- vard.edu/speeches/summers/2006/0324_rbi.html. Suttle, Philip. 2003. ?Financial Flows to Developing Countries: Recent Trends and Near-Term Prospects.? In Global Development Finance 2003: Striving for Stability in Development Finance, Volume 2: Summary and Country Tables. Washington, D.C.: World Bank, 7?14. Tesar, Linda L., and Ingrid M. Werner. 1998. ?The Internationalization of Securities Markets since the 1987 Crash.? In Brookings?Wharton Papers on Financial Services, ed. R. Litan and A. Santomero. Washington, D.C.: Brookings Institution Press, 281?349. Tobin, James, and William C. Brainard. 1977. ?Asset Markets and the Cost of Capital.? In Economic Progress, Private Values, and Public Policy, ed. B. Balassa and R. Nelson. Amsterdam: North-Holland, 235?62. Van Wincoop, Eric. 1994. ?Welfare Gains from International Risksharing.? Journal of Monetary Economics, 34(2): 175?200. Van Wincoop, Eric. 1999. ?How Big Are Potential Welfare Gains from International Risksharing?? Journal of International Economics, 47(1): 109?35. Wacziarg, Romain, and Karen Horn Welch. 2003. ?Trade Liberalization and Growth: New Evidence.? NBER Working Papers, no. 10152. Williamson, John. 1997. ?Prospects of Avoiding Crises with Liberalized Capital Flows.? Estudios de Economia, 24(2): 287?95. Williamson, John, and Molly Mahar. 1998. A Survey of Financial Liberalization. Essays in International Finance, no. 211. Princeton: International Finance Section, Department of Economics, Princeton University. Wilson, Ian M. 1992. The Wilson Directory of Emerging Market Funds. Saskatoon: Wilson Emerging Market Funds Research. World Bank. 1999. World Development Indicators: 934 Journal of Economic Literature, Vol. XLV (December 2007) dec07_Article1 11/28/07 12:02 PM Page 934 1999. Washington, D.C.: World Bank. World Bank. 2006. Doing Business in 2006: Creating Jobs. Washington, D.C.: World Bank. World Economic Forum. 2005. The Global Competi- tiveness Report 2005?2006: Policies Underpinning Rising Prosperity. Houndmills, U.K. and New York: Palgrave Macmillan. Wurgler, Jeffrey. 2000. ?Financial Markets and the Allocation of Capital.? Journal of Financial Economics, 58(1?2): 187?214. 935Henry: Capital Account Liberalization dec07_Article1 11/28/07 12:02 PM Page 935 This article has been cited by: 1. Horst Feldmann. 2013. Capital account liberalization and unemployment in industrial countries. Applied Economics Letters 20:6, 566-571. [CrossRef] 2. Alessandro Prati, Martin Schindler, Patricio Valenzuela. 2012. Who benefits from capital account liberalization? Evidence from firm-level credit ratings data. Journal of International Money and Finance 31:6, 1649-1673. [CrossRef] 3. Ren? Cabral, Andr? Varella Mollick. 2012. Convergence Rates to Output Growth in a Global World: The Roles of Openness and Government Size. The International Trade Journal 26:3, 201-222. [CrossRef] 4. Ellen R. McGrattan. 2012. Transition to FDI openness: Reconciling theory and evidence. Review of Economic Dynamics . [CrossRef] 5. Anusha Chari, , Peter Blair Henry, , Diego Sasson. 2012. Capital Market Integration and Wages. American Economic Journal: Macroeconomics 4:2, 102-132. [Abstract] [View PDF article] [PDF with links] 6. Hatra Voghouei, M. Azali, Mohammad Ali Jamali. 2011. A survey of the determinants of financial development. Asian-Pacific Economic Literature 25:2, 1-20. [CrossRef] 7. Alvaro Concha, Arturo Jos? Galindo, Diego Vasquez. 2011. An assessment of another decade of capital controls in Colombia: 1998?2008. The Quarterly Review of Economics and Finance . [CrossRef] 8. Kian-Ping Lim, Jae H. Kim. 2011. Trade openness and the informational efficiency of emerging stock markets. Economic Modelling 28:5, 2228-2238. [CrossRef] 9. Farhang Niroomand, Edward Nissan. 2011. Relating economic infrastructure indexes to investor protection for selected emerging economies. Journal of Economics and Finance . [CrossRef] 10. Dennis Quinn, Martin Schindler, A Maria Toyoda. 2011. Assessing Measures of Financial Openness and Integration. IMF Economic Review 59:3, 488-522. [CrossRef] 11. Jittima Tongurai, Chaiporn Vithessonthi. 2011. The Impact of Capital Account Liberalization Measures. Journal of International Financial Markets, Institutions and Money . [CrossRef] 12. Tiago V. de V. Cavalcanti, Kamiar Mohaddes, Mehdi Raissi. 2011. Growth, development and natural resources: New evidence using a heterogeneous panel analysis. The Quarterly Review of Economics and Finance . [CrossRef] 13. S. Meral Cakici. 2011. Financial Integration and Business Cycles in a Small Open Economy. Journal of International Money and Finance . [CrossRef] 14. S. Meral Cakici. 2011. Technology Shocks Under Varying Degrees of Financial Openness. International Review of Economics & Finance . [CrossRef] 15. H. Tong, S.-J. Wei. 2011. The Composition Matters: Capital Inflows and Liquidity Crunch During a Global Economic Crisis. Review of Financial Studies 24:6, 2023-2052. [CrossRef] 16. Carlos A. Ibarra. 2011. Capital Flows and Real Exchange Rate Appreciation in Mexico. World Development . [CrossRef] 17. Barry Eichengreen, Rachita Gullapalli, Ugo Panizza. 2011. Capital Account Liberalization, Financial Development and Industry Growth: A Synthetic View. Journal of International Money and Finance . [CrossRef] 18. Borja Larrain. 2011. World Betas, Consumption Growth, and Financial Integration. Journal of International Money and Finance . [CrossRef] 19. MICHAEL D. BORDO, CHRISTOPHER M. MEISSNER. 2011. Foreign capital, financial crises and incomes in the first era of globalization. European Review of Economic History 15:01, 61-91. [CrossRef] 20. Andr? Varella Mollick, Ren? Cabral. 2011. Government Size and Output Growth: the Effects of ?Averaging out?. Kyklos 64:1, 122-137. [CrossRef] 21. ANDREAS KERN, ALEXANDER SALHI. 2011. The Euro-Mediterranean Partnership: A Macroeconomic Governance Perspective*. JCMS: Journal of Common Market Studies no-no. [CrossRef] 22. LI SHENG. 2011. Theorising free capital mobility: the perspective of developing countries. Review of International Studies 1-16. [CrossRef] 23. Xinhua Gu, Bihong Huang. 2011. A new approach to capital flows: Theory and evidence. Economic Modelling . [CrossRef] 24. Karen K. Lewis. 2010. Global Asset Pricing. Annual Review of Financial Economics 3:1, 110301095929009. [CrossRef] 25. Li Sheng. 2010. Growth-volatility tradeoff in the face of financial openness: a perspective of developing economies. Cambridge Review of International Affairs 23:4, 609-622. [CrossRef] 26. Winston Ricardo Moore. 2010. Capital account liberalization and commercial bank interest rate margins. Applied Financial Economics 20:21, 1673-1685. [CrossRef] 27. Moritz Schularick, Thomas M Steger. 2010. Financial Integration, Investment, and Economic Growth: Evidence from Two Eras of Financial Globalization. Review of Economics and Statistics 92:4, 756-768. [CrossRef] 28. Andreas Steiner. 2010. CONTAGIOUS POLICIES: AN ANALYSIS OF SPATIAL INTERACTIONS AMONG COUNTRIES' CAPITAL ACCOUNT POLICIES. Pacific Economic Review 15:3, 422-445. [CrossRef] 29. Joseph P. Joyce. 2010. Financial Globalization and Banking Crises in Emerging Markets. Open Economies Review . [CrossRef] 30. Nauro F Campos, Yuko Kinoshita. 2010. Structural Reforms, Financial Liberalization, and Foreign Direct Investment. IMF Staff Papers 57:2, 326-365. [CrossRef] 31. Dilip K. Das. 2010. Financial globalization: a macroeconomic angle. Journal of Financial Economic Policy 2:4, 307-325. [CrossRef] 32. Claudia M. Buch, Paola Monti. 2010. Openness and income disparities: does trade explain the ?Mezzogiorno effect??. Review of World Economics 145:4, 667-688. [CrossRef] 33. Randall Morck, Bernard Yeung. 2009. Never Waste a Good Crisis: An Historical Perspective on Comparative Corporate Governance. Annual Review of Financial Economics 1:1, 145-179. [CrossRef] 34. Andrei A. Levchenko, Romain Ranci?re, Mathias Thoenig. 2009. Growth and risk at the industry level: The real effects of financial liberalization. Journal of Development Economics 89:2, 210-222. [CrossRef] 35. Maurice Obstfeld. 2009. International Finance and Growth in Developing Countries: What Have We Learned?. IMF Staff Papers 56:1, 63-111. [CrossRef] 36. Martin Schindler. 2009. Measuring Financial Integration: A New Data Set. IMF Staff Papers 56:1, 222-238. [CrossRef] 37. M Ayhan Kose, Eswar Prasad, Kenneth Rogoff, Shang-Jin Wei. 2009. Financial Globalization: A Reappraisal. IMF Staff Papers 56:1, 8-62. [CrossRef] 38. Dani Rodrik, Arvind Subramanian. 2009. Why Did Financial Globalization Disappoint?. IMF Staff Papers 56:1, 112-138. [CrossRef] 39. JAEJOON WOO. 2009. PRODUCTIVITY GROWTH AND TECHNOLOGICAL DIFFUSION THROUGH FOREIGN DIRECT INVESTMENT. Economic Inquiry 47:2, 226-248. [CrossRef] 40. Ayhan Kose, Eswar Prasad, Kenneth Rogoff, Shang-Jin Wei. 2009. Financial globalization: A reappraisal. Panoeconomicus 56:2, 143-197. [CrossRef] 41. A BONFIGLIOLI. 2008. Financial integration, productivity and capital accumulation#. Journal of International Economics 76:2, 337-355. [CrossRef] 42. Lars Jonung. 2008. Lessons from Financial Liberalisation in Scandinavia. Comparative Economic Studies 50:4, 564-598. [CrossRef] 43. 2008. Book Reviews. Journal of Economic Literature 46:4, 989-1041. [Abstract] [View PDF article] [PDF with links] 44. Jeffrey Frankel, Ben Smit, Federico Sturzenegger. 2008. Fiscal and monetary policy in a commodity-based economy 1. Economics of Transition 16:4, 679-713. [CrossRef] 45. 2008. Full Report - World of Work Report 2008: Income inequalities in the age of financial globalization. World of Work Report 2008:1, i-162. [CrossRef] 46. K MANOVA. 2008. Credit constraints, equity market liberalizations and international trade###. Journal of International Economics 76:1, 33-47. [CrossRef] 47. Gerald Epstein. 2008. An employment targeting framework for central bank policy in South Africa. International Review of Applied Economics 22:2, 243-258. [CrossRef]