Current Developments in Monetary and Financial Law, Vol. 2

Chapter 14 A Dozen Things to Love (or Hate) About Capital Flows

International Monetary Fund
Published Date:
October 2003
  • ShareShare
Show Summary Details

Free and open international trade in goods and services generally receives widespread support among economists. It might seem paradoxical that there is less confidence in the profession that free and open international trade in financial capital yields benefits everywhere and at all times that exceed the costs. This paper argues that the difference of opinion producing this disparate policy advice is really quite narrow. Most economists would accept a handful of basic propositions about capital flows—reflecting both good and bad and explaining properties of basic economic theory and stylized empirical regularities. Analysts differ, however, in the weight they attach to each.


Despite the sense of disarray and dissension sometimes conveyed to those outside the profession, economists by and large tend to agree on many fundamental propositions. Chief among them is the belief that substantial benefits accrue from free and open international trade in goods and services.1 Indeed, the progress in opening trade in the latter half of the last century, although uneven, could be viewed as the singular triumph of an idea over short-term political expediency. It might seem paradoxical, then, that there is less confidence in the profession that free and open international trade in financial capital yields benefits everywhere and at all times that exceed the costs.

Over the years, many economists, including James Tobin (1978), Rudiger Dornbusch (1986), and Jagdish Bhagwati (1998), have suggested slowing the trade in assets across borders.2 In addition, institutions previously associated with advocacy of untrammeled capital flows—including the IMF—have begun to question just how far to push the opening up of trade in financial assets. Representative of that strain of thought was the product of a recent working group of the Financial Stability Forum, whose membership spans finance ministries, central banks, and securities regulators. In a March 2000 report on short-term capital flows, the working group reported:

The use of controls on capital inflows may be justified for a transitional period in the face of very strong inflows or as countries strengthen the institutional and regulatory environment in their domestic financial systems, especially if the process of liberalization had not been carried out in a well-sequenced manner. In other words, some measures to discourage capital inflows may be used to reinforce or complement prudential requirements on financial institutions and other resident borrowers.3

Such support for limiting the international mobility of capital is by no means universal, with advocates of the free flow of capital including Martin Feldstein (1999) and Michael Dooley (1996), among others.

What this paper argues is that the difference of opinion producing this disparate policy advice is really quite narrow. Most economists would accept a handful of basic propositions about capital flows—reflecting both good and bad and explaining properties of basic economic theory and stylized empirical regularities. Analysts differ, however, in the weight they attach to each. While it is convenient to lay many of these propositions out in terms of simple tools of the economics trade, that technique should not be viewed as an obstacle to understanding. In reality, they are quite intuitive.

These 12 propositions logically fall into three groups, which correspond to the next three sections of this paper. In the next section, I will consider some implications of a fundamental tenet of economics: that people enter into trade because they expect to become better off as a result. International capital flows are but one example, essentially involving the exchange of goods across time. That simple proposition has two important implications. For one, the decisions of economic agents will be sensitive to market prices, giving policymakers a means to influence those outcomes through taxes or transfers in a way that is superior to quantitative controls. For another, a choice between current and future consumption ultimately has to be anchored by its income prospects. Simply, fundamentals matter.

The next set of propositions (in Section 3) are empirical in nature and are based on repeated observations that asset prices can sometimes show a remarkable amount of comovement across markets, that it can be difficult to relate those price movements to fundamentals, and that flows can be quite changeable. Moreover, these properties appear with more force in emerging market economies. Simply put, episodes of contagion, in which financial prices move sharply across financial centers in ways at odds with underlying economic developments and capital floods in or dries up, seem more prevalent in markets that are not as far up the ladder of economic progress.

The fourth section addresses some policy implications of simple economic models. Chief among them is the realization that policies are interrelated, known in the literature as the “impossible trinity” or less felicitously as the “trilemma.”4 National authorities can choose only two items from a menu of three: fixed exchange rates, independent monetary policy, and freely mobile capital. Thus, policymakers may, on occasion, be inclined to impose capital controls by an unwillingness to sacrifice the other two devotions of the impossible trinity. The remaining propositions about policy reflect other limitations on what authorities can do: it is easier to slow the entry of capital in good times than delay the exit of capital in bad times; constraining the volatility of one asset price may make other asset prices more volatile, and controls tend increasingly to be evaded over time.

By understanding these 12 propositions, it should be easier to put into perspective the sometimes conflicting advice on policymaking, which is dealt with in more detail in the last section.

The Fundamentals of Capital Flows

Underlying the trade theory that explains the flow of goods and services internationally is the notion of equal exchange. Residents of one country bring goods to the world market, say the wine of Portugal in David Ricardo’s famous example, and in return receive goods of equal value to them from residents of another country, say English cloth to complete the Ricardian example.5 That both parties enter voluntarily into the transaction leads to a natural conclusion: the trade must have made each participant no worse off and must have made at least one better off. Had that not been true, the wine would have been drunk in Portugal and the cloth cut to cover the residents of England.

This property can be no less true for the free trade in assets. The difference, though, is that a financial asset adds an intertemporal dimension to the problem, in that its issuance represents a current receipt of revenue that obligates repayment at a later date or dates. A country borrowing on world markets today can engineer consumption outstripping current income as long as it can credibly establish its willingness to spend less than its income in order to repay the debt at a future date.

This point is established more formally in consumer choice theory and expressed clearly in Obstfeld and Rogoff (1997). The simplest example traces back to Irving Fisher’s work (Fisher, 1930), which can be applied to two countries instead of two households. Consider a country with income of y1 and y2 in terms of a single good in periods 1 and 2. Consumption in both periods, at c1 and c2 respectively, is valued, but future benefits are discounted at a constant rate, ∃. Expressing that benefit in terms of an invariant utility function allows us to write what consumers ultimately care about as:

As is obvious, different combinations of consumption today and tomorrow can attain the same level of utility, producing a family of indifference curve, one of which is depicted as “consumption tradeoff' in Fig. 1.

Figure 1:The benefit of intertemporal trade

At the same time, a country must live within its means when viewed over the two periods. That is, any consumption in excess of its income in the first period must be repaid with interest, at the rate r, in the second period. This tells us:

which gives the linear “budget line” that slopes down in the figure.

The combination of consumption today and tomorrow where the consumption tradeoff is just tangent to the budget line offers the country the highest attainable level of welfare. In the particular example of the figure, this country would borrow on world financial markets in period 1 to consume more than its income and repay in the second period by consuming less than its income.6 That outcome, in general, depends on the shape of the utility function, the time profile of income, and the real interest rate. The important point is a comparative one: without trade, the country would be limited to consuming exactly its income in each period, which would not make it as well off as under free trade. Intertemporal trade makes possible the option of consuming any point along the budget line. Unless circumstances satisfy the special case where the tangency of the indifference curve and the budget line occurs at the initial endowment [y1, y2], the country will choose a consumption path that involves some borrowing or lending. And that trade makes it better off.7 The country on the other side of the transaction performs a similar calculation, implying

Proposition 1: The voluntary trade in capital internationally is done only when both parties in a transaction believe it does not make them worse off and at least one better off.

This is not to say that those participants could be wrong in their beliefs or may not be including in their calculations costs to other members of society. It more narrowly says that the fact that trade is voluntary creates a presumption that participants are not knowingly harming themselves.

While most economists would accept this proposition, their judgments about the magnitudes of the benefits accruing from free intertemporal trade differ. To see why, note that the consumption decision could be thought to have two dimensions. Households might choose to consume more or less than their incomes in the first period because their endowments are uneven across the two periods (call this income smoothing) or because the rate at which they internally discount the future differs from the market interest rate (call this interest-rate arbitrage). For the simplest form of utility, where

the interior condition for optimal consumption takes the form:

This simple case can be used to measure separately the benefits of income smoothing and interest rate arbitrage.

To begin this calculation, note that if the world interest rate equaled the rate of time preference (r = ∃), a country would want to consume the same amount in each period (c1 = c2). International trade in assets would permit a country to borrow or lend to smooth uneven endowments to engineer that optimal consumption path, as in Fig. 2. Note that the point where households consume their endowments in each period lies below the indifference curve available when there is free trade in capital—that is the welfare loss associated with halting the mobility of capital. The distance from that point (y1, y2) to the preferred indifference curve therefore represents the income compensation required to make the country just willing to accept an uneven consumption pattern. Thus, it is a way of measuring in terms of income the value put on open capital flows. Fig. 3 places numbers to this example. The horizontal axis measures the potential differences between income in the first and second periods as a share of second-period income, as in

Figure 2:Benefit of smoothing income

Figure 3:Benefit of smoothing income

in percent terms.8

The vertical axis plots the overall percent increase in income in both periods necessary to make the country indifferent between consuming exactly those uneven endowments without international trade and smoothing consumption with international trade.9 It is not until income is substantially uneven over time—more than a one-quarter difference between the two periods—that the required income compensation necessary to offset the welfare loss from restricting capital mobility exceeds 1 percent. But in situations where income is very uneven over time, the compensation can be large—as much as 5 percent.

The contribution to welfare of interest-rate arbitrage can be calculated analogously. To repeat, by the interior condition for consumption, spending today and tomorrow differ only when the real interest rate, r, differs from the rate of time preference, ∃. In Fig. 4, we ask how much extra income is needed to make a country consume exactly equal amounts in periods 1 and 2 (in the amount c) even though the world real interest rate is below the rate of time preference. The income compensation required to keep the country on the same indifference curve associated with the uneven consumption choice (that is produced with trade) but with equal consumption in both periods (if trade is not allowed) can be thought of as measuring the benefit of interest rate arbitrage—or borrowing or lending today to take advantage of the world real rate not equaling the rate of time preference.

Figure 4:Benefit of interest-rate arbitrage

Fig. 5 performs these calculations for the case where the world real interest rate equals 4 percent and the rate of time preference varies from just above zero to as high as 24 percent.10 As is evident, the income-equivalent benefits of interest rate arbitrage remain modest unless there is a substantial gap between how residents of the country discount the future relative to those in the rest of the world. This leads to a comparative statement most economists would accept:

Figure 5:Benefit of interest rate arbritrage

Proposition 2: In terms of consumption, the benefit from smoothing income is large compared to the benefit of arbitraging interest rate differentials.

Of course, the relative magnitude of the difference between the two aspects of a country’s dynamic choice importantly depends on the specification of welfare, including the assumption of a constant intertemporal elasticity of substitution. Indeed, focusing on the demand-side aspects of the effects of interest rates (by working with an endowment economy) may understate the welfare consequences of allowing trade. Production effects as firms equate the marginal product of capital to the real interest rate can be important, an issue discussed in Rogoff (1999).

More generally, an analyst’s assessment of whether income smoothing or interest rate arbitrage predominates in international capital markets importantly influences the resulting evaluation of the benefits of those markets. Quite obviously, support for capital mobility will be greatest among those who see world financial markets as fostering borrowing and lending to smooth uneven income prospects—either to build a nest egg for rainy days or a backstop when the rain falls. In contrast, those who read more of financial market activity as responding to differences in interest rates internationally tend to see less benefit in that activity.

But the interior condition explaining consumption behavior also has implications for constructing policies toward international capital flows. To put the issue simply, national authorities have to understand that their citizens desire to smooth consumption. Sometimes that desire implies a willingness to borrow now on world capital markets—in particular, when future income looks rosier or the world real interest rate looks low compared to that at home. If that borrowing is viewed unfavorably by officials, they have two means to constrain their citizens' efforts to trade future income for current consumption. Policymakers can impose a quota (or in some other manner restrict the flow of assets) or place tariffs on goods or a tax on capital flows. For instance, if, as in Fig. 6, the country expected income to rise in the future (seen as the uneven endowments), it would borrow on world markets, allowing it to spend more than current income with funding from abroad (seen by the dashed budget line just tangent to the dashed indifference curve). To stem that inflow of capital, the government could put an explicit limit on capital, represented as the vertical line in the figure drawn to represent an outright ban on new borrowing (thus mandating a consumption choice that lies along the indifference curve drawn with a solid line).

Figure 6:Restricting capital flows

In effect, the government imposes (y2/y1) as the consumption path. The problem is that, despite the quota, citizens of the country still want to obey the consumption-smoothing condition. The difference between what the public wants, consumption in the ratio (1+ r)/(l+∃), and what they are allowed, consumption in the ratio (y2/y1), defines an intervention wedge, T,

As long as transactions within the country are still voluntary, this wedge will be reflected in relative prices as people try to evade the quantitative restrictions. Either those who have access to foreign borrowing (presumably to roll over maturing holdings as net borrowing in the example was zero) could offload this right to do so and receive a return above r, or the price of consumption goods currently could rise relative to that in the future to induce an increased willingness to save.11

Alternatively, the government could apply a tax on foreign borrowing equal to T (or a tariff in an amount producing the same consequences for the intervention wedge) and thereby change incentives. A tax on foreign capital rotates the budget line in Figure 6, generating a voluntary outcome identical to the constrained one (and drawn by the more titled solid budget line). Unlike the quota, relative prices are directly affected and no private sector resources are wasted in the effort.

This leads naturally to proposition 3,

Proposition 3: Policy can influence the outcomes of individual choice by putting controls on quantities or by changing relative prices, but influencing prices tends to be less distorting than setting quantities.

Even more attractive to many economists is couching relative-price-based controls on capital in prudential terms and making them sensitive to the holding period of the instrument. For instance, as explained in World Bank (1997), the Chilean government taxes foreign interest income via a reserve requirement that effectively declines (in proportional terms) as the maturity of the instrument declines. Eichengreen and Wyplosz (1993) consider this issue from an industrial country perspective. A discussion of the various policy alternatives is given in Reinhart and Reinhart (1998) and a more specific treatment of reserve requirements in Reinhart and Reinhart (1999).

That said about the potential for policy, this simple framework also teaches an even more important lesson. Despite policy interventions that either shift or rotate the budget line, that constraint is anchored at one point—initial endowments. In a given period, a government imposing controls or taxes on capital does not alter what an economy has produced to trade on the world market, it has only altered the terms of that trade from the vantage point of its citizens. This is reflected in proposition 4:

Proposition 4: People have to live within their means—that is, fundamentals matter.

While policy can tilt returns in favor of consumption today or tomorrow, ultimately a nation’s choices are limited by current and future income prospects. Quotas, taxes, and transfers may disguise that for a time, but not forever. Therefore, a key lesson that is as much rooted in common sense as in high theory is do not expect from policy what it cannot deliver.

Table 1.Correlations Among Home-Currency Returns on Broad Equity Indexes(Weekly, January 1992 to March 2001)
Hong Kong
Hong Kong

Source: Bloomberg. United States (S&P 500), Germany (DAX), United Kingdom (FTSE), Canada (TSE300), Japan (Nikkei), Mexico (Bolsa), Brazil (Bovespa), and Hong Kong SAR (Hang Sen).
Source: Bloomberg. United States (S&P 500), Germany (DAX), United Kingdom (FTSE), Canada (TSE300), Japan (Nikkei), Mexico (Bolsa), Brazil (Bovespa), and Hong Kong SAR (Hang Sen).

Properties of Asset Prices and Flows

The next four propositions are empirical in nature and relate to the behavior of asset prices and flows. These properties are observed in many different markets, across many countries, and over long stretches of time. By way of example, Tab. 1 reports the correlation observed over the past nine years in the weekly home-currency returns from broad equity indexes in eight industrial and emerging market economies.

As is evident, equity prices covary closely, with correlation coefficients ranging from 0.2 to 0.71. Using more sophisticated techniques over a different sample, Calvo and Reinhart (1996) provide evidence that these comovements can be explained by a small number of common factors.

This simple evidence suggests an equally simple proposition:

Proposition 5: Asset prices move together internationally.

There are many potential explanations for these high correlations, including commonality in shocks in technology and confidence, the importance of commodities that are traded on world markets, the linkage of spending in each country through trade channels, and the presence of the same global investors in all these financial markets. Whatever the specific mechanisms explaining these correlations, policymakers have to take into account influences emanating from abroad and the potential that their own actions will echo offshore. More problematic, though, is the general property that these correlations, while high on average, are not stable over time.12 So while these financial linkages are important, they are hard to bet on.

Beyond observing the correlations among financial prices, empirical researchers have been unable to identify systematic relationships explaining their underlying behavior. The finance and economics literature is chock full of anomalies where facts fail to accord with theory, including the equity-premium puzzle (in which the return on stocks over the long run exceeds what interest rate arbitrage would suggest), home bias (in which investors tend to hold fewer foreign assets than efficient diversification of risk would seem to call for), and the failure of the expectations approach to the term structure of interest rates (in which the slope of the yield curve routinely fails to predict the future direction of interest rates).13 This work suggests the following:

Proposition 6: Asset prices are difficult to link to fundamentals.

Accumulating evidence of the inability of theory to fit the facts has led economists to develop models of contagion that put herding behavior by investors center stage (which also helps to explain proposition 5).

Data on either stocks or flows of financial assets are hard to come by and never sampled with the frequency of asset prices. But what is obvious in many different episodes and shows clear in the aggregate data is that capital flows can change quickly. By way of example, Reinhart and Reinhart (2001) examine movements in real capital flows to emerging market economies over the past 30 years. Figure 7 depicts capital flows to emerging markets (in billions of U.S. dollars in 1970 terms) in recession years versus recovery years for the 1970–1999 period. The picture shown in the first panel of Figure 7 reveals that net flows to emerging markets are almost twice as large when the United States is in expansion as when the United States is in recession. Furthermore, this vast gap between recession and expansion owes primarily to a surge in foreign direct investment (FDI) flows (which nearly go up fivefold from recession to expansion) and to portfolio flows (Figure 7, third panel). Indeed, other net inflows to emerging markets fall from about US$17 billion, when the United States is in recession, to about US$8 billion of net outflows in expansions.14

Figure 7.Real Capital Flows to Emerging Markets and the U.S. Business Cycle 1970–1999

The U.S. bank lending boom to Latin America in the late 1970s and early 1980s and the surge in Japanese bank lending to emerging Asia in the mid-1990s are two clear examples of this phenomena. From a compositional standpoint, the more stable component of capital flows, FDI, does seem to contract during downturns—suggesting that emerging markets may wind up during these periods relying more heavily on less stable sources of financing short-term flows. Taken together, this evidence supports the proposition that:

Proposition 7: Capital flows are changeable.

Moreover, while capital flows are generally quite volatile, they are especially so to and from emerging market economies. The typical industrial country has a track record of credible monetary policymaking, respect for the rule of law, deep domestic financial markets, and no recent history of debt repudiation. Unfortunately, that has not always been the case for many emerging market economies. As a result, global investors are more likely to flood in or retreat quickly as assessments of and attitudes toward risk change.15 In this regard, size matters as well: the flow of capital to an emerging market economy in any given year can be quite large compared both to the pool of domestic saving and the outstanding stock of financial capital, especially relative to most industrial countries. This adds up to:

Proposition 8: Emerging market economies are different than the economies of industrial countries.

Nowhere is this more evident than in the work on the macroeconomic consequences of a sudden reversal of capital inflows. For instance, the case studies provided by the World Bank (1997) document much more wrenching adjustments to income associated with large reductions in capital inflows to emerging markets than in the industrial country experience detailed in Freund (2000).

These four regularities admit a variety of interpretations that map into differing attitudes toward capital flows. That is, these same observations could be used to advocate liberalizing or restricting capital transactions. As to the former, the volatility of prices and flows could be taken as evidence that financial markets readily reprice assets when the assessment of risk or income prospects change. In that environment, such changes in relative prices send signals to redirect resources toward more appropriate uses. That asset prices are difficult to explain may only indicate that the process is complicated. But as a corollary, government intervention would only interfere with the signals sent by the market and have effects that could be difficult to predict. As to the latter view representing greater wariness toward free-flowing international capital, the inability to explain asset prices by fundamentals would suggest that markets are not obviously directing resources toward their best uses. Instead, prices and flows bear the imprint of the herd behavior of global institutions, introducing unnecessary volatility into domestic financial markets.

Propositions About Policy

Perhaps the most important message the literature on open-economy macroeconomics sends to national authorities in both developed and developing countries is that their policies are interconnected. In particular, as explained by Jeffrey Frankel (1999) in his Graham lecture, policymakers have only two degrees of freedom in their choice among three main policies. That is, the choice of exchange rate system (fixed versus floating), monetary policy (independent to achieve domestic objectives or dependent on foreign considerations), and treatment of capital flows (unrestricted or restricted) are linked. In particular, authorities cannot pursue both an independent monetary authority and fixed exchange rates while allowing capital to be mobile (which is why Frankel calls this the impossible trinity). Simply put, some sand has to be thrown in the gears of international finance for a nation to have simultaneously a monetary policy that is sensitive to the domestic economy and that delivers a fixed exchange rate. If not, potentially large flows of international capital will swamp efforts to defend the currency if a wedge opens up between domestic and foreign interest rates. Thus, authorities have been driven to imposing capital controls on occasion because, to them, it was the better alternative than giving up monetary autonomy or a stable exchange rate.

As a general proposition, then, we have:

Proposition 9: The treatment of capital flows, monetary policy independence, and the exchange rate regime are inextricably linked.

Reinhart and Reinhart (1998) document the various efforts to deal with a surge in capital flows by authorities in emerging market economies.16 Disciplined by the impossible trinity, sometimes those efforts include imposing restrictions on the mobility of international capital so as to preserve some degree of flexibility in domestic monetary policy and fixity of the exchange rate.

The remaining propositions concern how a market economy responds to policy interventions over time. They are shaped by the judgments that policy action influences market expectations, that markets are interconnected, and that the private sector adapts to any given regime. As was related in the first few propositions of this paper, economists generally believe capital flows can serve useful purposes. In particular, a nation expecting its income to increase has an incentive to borrow now to smooth consumption. Indeed, in our example of optimizing behavior, if the real interest rate equals the rate of time preference, consumption will be the same in both periods, implying that the country will borrow on world markets in the first period.

The problem becomes more complicated, though, when we introduce the reality that asset trades involve risk and depend importantly on expectations. In an example owing to Bartolini and Drazen (1994) and Pakko (2000), suppose that doubts arise among global investors about the country’s ability or willingness to repay its loans next period. Presumably, the rate charged to borrowers will rise to reflect the assessment of a greater probability of default. Keeping the flow of foreign capital into the country at the same pace—that is, letting current consumption run above income—would require the government to subsidize foreign borrowing (or lower any existing tax). But the act of doing so may well make investors more worried about repayment prospects, which would perversely raise borrowing costs and lower the flow of capital. This signaling aspect of policy decisions may raise the overall amplitude of the swing in capital flows. International capital flows would seem to be drawn to a country in good times, when it enjoys the favor of Wall Street and the City of London. At such times, official discouragement will not be viewed unfavorably. Indeed, official discouragement might even be interpreted as a sign of strength that implied lower default risks going forward. In bad times, when investors want to slow capital inflows, or even withdraw capital, they will be very sensitive to policy action in forming their expectations. Official encouragement will look like weakness and may worsen the situation. This leads to:

Proposition 10: Controlling outflows is hard.

Policymakers also have to appreciate that portfolio decisions of both global and domestic investors are not made in isolation. In such circumstances, a decision to restrict the flow of any one particular asset will have consequences for the prices of other assets. If one class of assets is blocked off to a global investor by capital controls, demand for other, unrestricted assets, will shift. The price consequences of those demand shifts, in turn, can induce changes in the portfolios of domestic investors not directly restricted by capital controls.

Reinhart (2000) provides a specific example in which a transactions tax on foreign borrowing has the effect of raising the volatility of domestic equity prices. Simply, the pressure of a shift in overall asset demand has to be released through changes in asset prices. The fewer asset prices permitted to adjust, the more those allowed to adjust will have to change.17 This adds to our list of propositions:

Proposition 11: Controls in one market may shift volatility to other markets.

The obvious advice that flows from this is that policy makers need to be aware of the unintended consequences of their actions. Substitution across assets makes it difficult to intervene in financial markets with surgical precision.

The ability to shift across markets also suggests that the effectiveness of any particular set of restrictions can be eroded over time. As Alfred Marshall asserted in 1897 (in a quote repeated in Keynes, 1951),

When one person is willing to sell a thing at a price which another is willing to pay for it, the two manage to come together in spite of prohibitions of King or Parliament or of the officials of a Trust or Trade Union.

With financial transactions, the problem is harder still. In the face of controls on one particular asset, market participants might simply move trading offshore or create a new instrument that mimics its risk and return characteristics within the domestic market. Indeed, over time, resources are used in the attempt to evade the controls. Garber (1996), for instance, argues forcefully that one form of a capital control—a tax on transactions—would be evaded over time. As Garber relates, a general principle of public finance is that a tax applied on an elastic market (where participants are quite sensitive to prices) delivers little revenue because it crowds out that taxed activity. This is formalized as:

Proposition 12: Capital controls are evaded over time.

Given this property, policymakers are left in the uncomfortable position of either seeing the efficacy of their controls erode or making their regulations more complicated over time.

To be sure, the likelihood that capital controls interact with expectations, indirectly trigger responses in other markets, and directly lose their force over time is pointed to by advocates of free and open capital markets. But those propositions are convincing only to the extent that they can be read in the experience of countries imposing capital controls. And there the record is mixed. De Gregorio, Edwards, and Valdés (2000) find that the effects of Chilean capital controls were quite elusive and, at most, support the conclusion that controls shaped the composition, not the overall volume, of flows. Kaplan’s and Rodrick’s (2000) reading of Malaysian capital controls is considerably more favorable, while Haggard’s (2000) is not.

A Summing Up

Table 2 summarizes the place of common ground among economists working on capital flows, sorting the dozen basic propositions advanced in this paper according to whether they are about theory, facts, or policy. While most economists would accept these propositions, the emphasis put on them varies considerably, and, as a result, so too do their policy recommendations.

Table 2.A Dozen Properties of Capital Flows
About Choices:About Facts:About Policy:
Trade improves welfare.Asset prices move together.The treatment of capital flows, monetary policy, and the exchange rate regime are linked.
Smoothing income is more important than arbitraging interest rates.Asset prices are hard to link to fundamentals.Controlling outflows is hard.
Control through prices not quantities.Flows are volatile.Controls shift volatility.
Fundamentals matter.Emerging market economies are different.Controls are evaded.

Within that table are three points of truly common ground. As set out in bold type, most economists would accept that the marketplace provides benefits, that policies are interrelated, and that, if intervention is necessary, it should work through prices, not quantities. After that, consensus is harder to reach. The building blocks for those advising free-flowing international capital are the items in italics in the table, which combine to suggest that trade brings benefits and policy intervention has uncertain, and likely little long-run, net benefit. Economists in favor of limiting capital mobility put more stock in inconvenient facts of international finance—volatility and contagion—as implying that market inefficiencies abound and that the welfare gain to shaving interest-rate differentials may be small.

The current consensus, reflected among other places in the report by a working group of the Financial Stability Forum quoted in the introduction, recognizes that capital flows are volatile (proposition 7) and financial market prices can sometimes diverge from fundamentals (proposition 6). Moreover, it holds that there are probably some benefits to independent monetary policy and a smooth exchange rate, leaving the third in the impossible trinity at risk (proposition 9). But it also supports the view that open capital flows yield substantial long-run benefits (proposition 1), especially when compared to the resources that are wasted by the private sector as it tries to evade controls and misses important market signals (proposition 12). But that consensus appears willing to consider limited capital controls that are in price terms (proposition 3), on inflows but not outflows (proposition 10), and are not seen as a substitute for fundamental reform (proposition 4). Better still, those controls should be couched in terms of prudential terms and penalize short, rather than long, holding periods. That the consensus changed owes importantly to economists' willingness to learn from experience—and the experience of the past few decades has shown that capital flows can change abruptly and that financial market prices are often hard to explain.


    AlstonRichardM.J. R.Kearl andMichael B.Vaughan1992“Is There a Consensus Among Economists in the 1990s?”American Economic Review82 (May) pp. 203209.

    BhagwatiJagdish1998“The Capital Myth,”Foreign Affairs77 (May/June) pp. 712.

    BartolimLeonardo andAllanDrazen1994“Capital Account Liberalization as a Signal,”IMF Working Paper 94/9 (Washington: International Monetary Fund).

    BuiterWillemH.1981“Time Preference and International Lending and Borrowing in an Overlapping Generations Model,”Journal of Political EconomyVol. 89 pp. 76997.

    CalvoGuillermoLeonardoLeiderman andCarmenReinhart1996“Inflows of Capital to Developing Countries in the 1990s: Causes and Effects,”Journal of Economic PerspectivesVol. 10 pp. 12339.

    CalvoSarah andCarmen M.Reinhart1996“Capital Flows to Latin America: Is there Evidence of Contagion Effects?” inPrivate Capital Flows to Emerging Marketsed. byMorrisGoldstein (Washington: Institute for International Economics).

    De GregorioJoséSebastianEdwards andRodrigoO.Valdés2000“Controls on Capital Inflows: Do They Work?”Journal of Development EconomicsVol. 63 pp. 5983.

    DooleyMichaelP.1996“The Tobin Tax: Good Theory, Weak Evidence, and Questionable Policy” in The Tobin Tax: Coping with Financial Volatilityed. byMahbubal HaqIngeKaul andIsabelleGrunberg (New York: Oxford University Press).

    DombuschRudiger1986“Flexible Exchange Rates and Excess Capital Mobility,”Brookings Papers on Economic Activity: 1 pp. 209226.

    DombuschRudiger1998“Capital Controls: An Idea Whose Time Is Goneȍ (unpublished: Cambridge, Massachusetts: Massachusetts Institute of Technology).

    EichengreenBarry andCharlesWyplosz1993“The Unstable EMS,”Brookings Papers on Economic Activity: 1 pp. 51124.

    FeldsteinMartin1999“A Self-Help Guide for Emerging Markets,”Foreign Affairs (March/April) pp. 93109.

    Financial Stability Forum2000“Report of the Working Group on Capital Flows,”April.

    FisherIrving1930The Theory of Interest (New York: The Macmillan Company).

    FrankelJeffreyA.1999“No Single Currency Regime Is Right for All Countries or at All Times,”NBER Working Paper No. 7338 (Cambridge, Massachusetts: National Bureau for Economic Research).

    FreundCaroline2000“Current Account Adjustment in Industrial Countries,”Board of Governors of the Federal Reserve System International Finance Discussion Paper No. 692 (Washington: Federal Reserve System).

    GarberPeterM.1996“Issues of Enforcement and Evasion in a Tax on Foreign Exchange Transactions,” in The Tobin Tax: Coping with Financial Volatilityed. byMahbubal HaqIngeKaul andIsabelleGrunberg (New York: Oxford University Press).

    KaplanEthan andDaniRodrick2000“Did the Malaysian Capital Controls Work?”prepared for an NBER conference on currency crises (December).

    KeynesJohnMaynard1951Essays in Biography (New York: W.W. Norton).

    KrugmanPaul1996Pop Internationalism (Cambridge, Massachusetts: Massachusetts Institute of Technology).

    LoretanMico andWilliamB. English2000“Evaluating Correlation Breakdowns during Periods of Market Volatility,”Board of Governors of the Federal Reserve System International Finance Discussion Paper No. 658 (Federal Reserve System).

    ObstfeldMauriceA.1982“Aggregate Spending and the Terms of Trade: Is There a Laursen-Metzler Effect?,”Quarterly Journal of EconomicsVol. 97 pp. 251270.

    ObstfeldMauriceA. andKennethRogoff1997Foundations of International Economics (Cambridge, MA: MIT Press).

    PakkoMichaelR.2000“Do High Interest Rates Stem Capital Outflows?”Economic LettersVol.67 pp. 187192.

    ReinhartVincentRaymond2000“How the Machinery of International Finance Runs with Sand in Its Wheels,”Review of International EconomicsVol. 8 (February) pp. 7485.

    ReinhartCarmenM. andVincentRaymondReinhart1985“Some Lessons for Policy Makers Who Deal with the Mixed Blessings of Capital Inflows,” in Capital Flows and Financial Crisesed. byMilesKahler (Ithaca, NY: Cornell University Press) pp. 93127.

    ReinhartCarmenM.1999“On the Use of Reserve Requirement in Dealing with Capital Flow Problems,”International Journal of Finance and EconomicsVol. 4 (January) pp. 2754.

    RicardoDavid1973The Principles of Political Economy and Taxation (New York: Dutton).

    RogoffKenneth1999“International Institutions for Reducing Global Financial Instability,”NBER Working Paper No. 7265 (Cambrige, Massachusetts: National Bureau of Economic Research).

    SamuelsonPaulA.1947Foundations of Economic Analysis (Cambridge, Massachusetts: Harvard University Press).

    TesarLinda1995“Evaluating the Gains from International Risk Sharing,”Carnegie-Rochester Conference Series on Public PolicyNo. 42 pp. 95143.

    TobinJames1978“A Proposal for International Monetary Reform,”Eastern Economic JournalVol. 4 (July/October) pp. 15359.

    World Bank1997Private Capital Flows to Developing Countries (Washington: Oxford University Press for the World Bank).

Note: The author has benefited from helpful commentaries made by Carmen Reinhart and Michael Gilbert.

This is a point that recurs in Paul Krugman’s popular writings, as in Krugman (1996), that is shown more formally in a survey of economists by Alston, Kearl, and Vaughan (1992). In that survey, more than 90 percent of the respondents held that tariffs and quotas reduce welfare.

For Dornbusch’s recent views on capital controls, see Dornbusch (2001).

Report of the Working Group on Capital Flows, Financial Stability Forum (2000), paragraph 115, p. 35.

For a discussion of the impossible trinity, see Jeffrey Frankel (1999).

See chapter 7 of Ricardo’s The Principles of Political Economy and Taxation, which was written in 1817 (Ricardo, 1973).

This, in effect, is a two-period simplification of the intertemporal approach to the current account associated with Buiter (1981) and Obstfeld (1982).

Even in the special case where the tangency of the indifference curve to the budget line occurs at the endowment pair, the possibility of trade makes the country no worse off; rather, it fails to make it better off.

When this measure is zero, income is equal in both periods; as it approaches minus one, all income is earned in the first period; as it gets arbitrarily large, all income is earned in the second period.

This example considers the case where income is deferred (y1 > y2), but the problem is symmetric in the case where income is front-loaded (y1 > y2).

In general, the effect on current consumption of a change in the world real interest rates is ambiguous. On the one hand, a higher interest rate encourages deferring current consumption. On the other hand, a higher interest rate raises the income from saving, which encourages greater consumption in both periods With the functional form assumed here, the first effect (the substitution effect) just cancels the second (the income effect). Thus, current consumption is a fixed share of present-discounted income. The question posed in the text is about how an interest rate differential influences relative consumption and welfare. Also note, the example varies the discount factor, not the world interest rate, implying that only the substitution effect is operative.

Those efforts may also involve resource cost as the private sector tries to work around the constraint, implying that some of the country’s endowment will be wasted. That is the essence of rent seeking in the Krueger sense (Krueger, 1974), which is typically applied to restrictions on the trade of goods but also applies with force to the trade in assets.

Loretan and English (2000) address this phenomenon, emphasizing the differences in measures of association in small samples versus the general population.

Representative of the work in these areas is Prescott and Mehra (1982) (on the equity-premium puzzle), Tesar (1990) (on home bias), and Shiller (1988) (on term structure anomalies).

This disparate behavior between FDI and portfolio flows importantly owes to bank lending, which accounts for a significant part of other flows. Apparently, banks tend to seek lending opportunities abroad when the domestic demand for loans weakens, as it usually docs during recessions.

This is described in the work of Calvo, Leiderman, and Reinhart (1996) on what they refer to as the capital inflow problem, or the high degree of comovement in regional capital flows.

A particularly relevant case after the writing of that paper is that of Malaysia, where, as explained in Haggard (2000), controls were explicitly imposed to preserve monetary autonomy and smooth exchange rates.

Put this way, the proposition is similar to the Le Chatelier principle, which was applied to economics by Samuelson (1947).

    Other Resources Citing This Publication