Printer Friendly

Export instability and long-term capital flows: response to asset risk in a small economy.


An open capital account allows long-term capital flows to automatically mitigate adverse effects of export instability on domestic saving and investment. An application of portfolio management theory shows that risks that are systematic to the domestic market are diversified internationally. This may help explain why foreign investment finances many high-risk investments in export sectors of LDCs and why results of studies of the effects of export instability show inconsistent results. This theory is presented and tested empirically.


Through the early 1960s, most economists had an unquestioning faith in the proposition that instability of the export prices (and earnings) of less developed economies was one of the factors that retarded their development. In the absence of any definitive empirical work, the chain of reasoning used to support this conclusion began with the observation that the exports of most less-developed countries (LDCS) are concentrated in a relatively small number of agricultural or mineral products, with each country facing inherently unstable export prices and earnings. This extreme variability, and the resulting risk, it was argued, distorted resource use and depressed growth. Batra [1975], for example, argued that if export prices are more unstable than domestic, resources will tend to gravitate to the less risky sector, making the return on capital unequal in the sectors, producing under-investment in the export sector, and making the export sector inefficiently small relative to the domestic. Others conjectured that export earnings instability creates uncertainty in the supply of foreign exchange needed for capital imports, as in Hawkins, Epstein, and Gonzales [1966] and Glezakos [1984]; or disrupting and discouraging investment, as in McNicol [1978] and MacBean [1966]. The weight of these arguments led to a number of proposals for stabilizing export prices or earnings of LDCS, some of which have been implemented (e.g., the STABEX Fund of the EEC and the IMF's Compensatory Finance Facility) and some of which have not (e.g., the Integrated Program for Commodities proposed by the UNCTAD). The empirical work done since 1960, however, has been far from unanimous in associating export instability with its alleged adverse effects. Some studies have found the expected negative correlation between instability and investment, savings, or growth, some have found no relation, and some have even found a positive correlation. This discrepancy between theory and empirical findings has been somewhat of a paradox.

This paper resolves the discrepancy between theory and empirical findings by appealing to a simple theory of portfolio management in which there are two groups of savers. One group invests in a large, diversified capital market, and the other in a smaller, riskier market where diversification is more difficult. The analysis also sheds light on the often observed phenomenon that in LDCs with more-or-less open capital accounts, many risky investment projects-especially in the export sectors-have historically been undertaken by foreign investors, rather than being financed by domestic savings. Section II below explains this theory, section III presents the results of some empirical tests, and section IV draws some conclusions.



The capital market in a "typical" LDC with a closed capital account could be described by the following stylized model: The market is divided into two sectors, an export sector (X) and a sector that produces for the domestic market (D). The D sector is large, relative to the X sector. A risk-free asset (e.g., government bonds) may or may not be available. The returns on the asset in the D sector rD) are fairly stable (with variance [[[alpha].sub.D].sup.2]) relative to the highly variable returns on the asset in the X sector ([r.sub.x], with variance [[[alpha].sub.X].sup.2] ). The availability of foreign exchange from the X sector determines the availability of imported inputs for the D sector, while income in the X sector-has some impact on the demand for D-sector products, implying that cycles in the two sectors are strongly correlated. That is, the covariance of returns ([[sigma].sub.XD]) is high.

In such a context it is clear that export instability has the kind of adverse effects predicted in the literature cited in section I. The riskiness of returns in the X sector discourages investment in that sector, and the high covariance of returns with the D sector means that much of the risk is systematic or non-diversifiable. In equilibrium, savers will hold relatively little of their portfolio in the form of X assets. Capital will exit the X sector until the marginal product of capital in X exceeds that in D by a premium sufficient to compensate for the extra risk. Or, from the viewpoint of portfolio management, the price of stock in X will fall until its return provides the required risk premium to induce savers to hold it in their portfolios. Assuming that there exists a risk-free asset with a fixed yield [r.sub.F], and that sector D assets are large enough to play the role of the market relative to individual stock X, this equilibrium risk premium is well-known as:

[Mathematical Expression Omited] (1)

where bars denote the equilibrium expected value, and [lambdal, the "market price of risk," is [Mathematical Expression Omited]. The greater the risk, the greater will be this risk premium. The exit of capital from the export sector and the resulting risk premium implies that prospective projects will not be done, even though their expected return is high relative to that on alternative projects in the D sector. As predicted in the literature, too few resources will be devoted to the export sector relative to the domestic sector, and the difference between the two will be positively correlated with their relative riskiness. Furthermore, increased export riskiness, even when compensated by portfolio shifts into domestic sector assets (or risk-free assets, if available) nonetheless decreases the welfare of savers, thereby discouraging savings and investment (which are equal if the capital account is closed), consistent with the theories in much of the literature.

Consider now the effect of opening the capital account. Risk in the export sector that is not diversifiable ion the context of a closed economy (because of the high covariance between stocks in sectors D and X) is diversifiable to savers with portfolios in foreign capital markets. Since the risk premium in the closed-capital account equilibrium is high at the time of the opening, foreign savers will purchase X assets until the risk premium is reduced to a level based on the covariance between returns on X assets and returns on a market basket of foreign assets. That is, in the open-capital account equilibrium, the following equality will be met:

[Mathematical Expression Omited] (2)

where [carat over][lambdal] is [Mathematical Expression Omited] is the fixed risk-free rate in the world market, [[bar over]r.sub.w] is the expected return on a market basket of assets in the world market (exogenous to this model), [[sigma].sub.XW] is the covariance between returns in sector X and the world capital market, and [[[sigma].sub.W].sup.2] is the variance of returns in the world capital market. Investors in the domestic sector will also arbitrage returns there into line with foreign capital markets, so that:

[Mathematical Expression Omited] (3)

where [[sigma].sub.DW] is the covariance between the domestic sector and world capital markets. Equations (2) and (3) imply that the divergence between [[bar over]r.sub.D] and [[bar over]r.sub.X] is:

[Mathematical Expression Omited] (4)

Since [[sigma].sub.XW] and [[sigma].sub.DW] are both likely to be quite small (and almost equal) for a typical LDC, the "wedge" between the equilibrium rates of domestic and export sector expected returns is likewise small. Or, to put it another way, there is not under-investment in the export sector, as there was when the capital account was closed. This is true because much of the risk in the export sector that was systematic with a closed capital account becomes diversifiable for foreign savers when the account is opened.

In the open-capital account country, the total assets in each sector held by foreign and domestic savers together is determined in accordance with the equilibrium conditions (2)-(4). But what determines the composition of foreign and domestic savers' portfolios and the fraction of assets in each sector held by each class of saver? If domestic savers were perfectly integrated into world capital markets and had preferences identical to those of foreign savers, the portfolios of domestic and foreign savers would be identical. In this case, foreign savers would not hold a larger share of total "risky" (export sector) assets than of domestic sector or foreign assets.

However, this is not what is observed in the real world. As noted in section I, a disproportionately large share of risky export projects in LDCs have typically been undertaken by foreign investment. There are at least two general classes of market asymmetry that could be responsible for this. First, access to foreign asset markets may be difficult for domestic savers. Knowledge about the covariance of returns among different types of assets may not be readily available, transactions costs in dealing with brokers are high and savers may be tied to local markets by custom, making it costly to deal in foreign asset markets. As a result, part of the risk that is readily diversifiable for foreign savers is diversifiable for domestic savers only at some cost. For this reason, the relevant risk of investment in X as perceived by domestic savers is greater than [[sigma.sub.XW], and the equilibrium portfolio share of X assets is correspondingly smaller for them than for foreign savers, assuming that their preferences are the same. The greater the undiversifiable risk becomes, the more domestic savers reduce their holdings of X assets, and the greater is the share of investment in X done by foreign savers. Second, foreign savers in general would be expected to be more familiar with markets for internationally traded goods than with markets for goods produced for domestic consumption. For them, transaction costs in dealing with D assets may, therefore, be higher than for domestic savers. In this case, like that outlined above, foreign savers will in equilibrium hold a disproportionate share of X assets, and domestic savers a disproportionate share of D assets. But it should be emphasized that even in these two cases of market asymmetry, equilibrium equalities (2)-(4) still hold, ensuring that resources are allocated efficiently, that is, in such a way that rates of return between X and D are (approximately) equal. These asymmetries simply produce a corresponding asymmetry in the ownership of assets.


Testing of the theory is somewhat tricky, since to look at the issue empirically requires finding a reasonably lengthy time period when a sample of small economies simultaneously maintained open capital accounts. The last such period ended around 1960. This was determined by a search through the IMF's publication Exchange Restrictions, published yearly since 1950, which contains information-country by country-on all types of capital account restrictions, including those on long-term and direct investment. Thus, the data are unfortunately somewhat dated, but since they cast some light on a hypothesis with contemporary implications, they should be of more than purely historical interest.

Here, four predictions of the theory are tested. Three of the predictions are related to the interactions between risk (proxied here by instability), savings, and investment, and one to the resulting structure of the economy. The first three are examined using a sample of countries with open capital accounts from 1950 until sometime after 1960. The end result is examined by comparing the structures of these economies to the structures of a sample of countries with closed capital accounts. Since the sample of open countries is small and there is no presumption of linearity in the relations being examined, two non-parametric statistics, the Spearman rank-order coefficient and the modified Olmstead-Tukey [19471 statistic recommended by Conover [1962], are used in addition to regression analysis.

The analysis of Section II suggests that a cross-sectional analysis of countries with open capital accounts should uncover three relationships that run counter to the conventional wisdom embodied in the literature referenced in Section I. First, the analysis predicts that instability should have no negative effect on saving in an economy. Second, instability should have no negative effect on total investment. And third, instability should be positively correlated with the fraction of investment that is foreign in origin. The present the results of tests of these three intermediate links in the theory for the countries in the open sample.

Section I of the table shows the results of the test of the first link-the correlation between instability and the propensity to save, as estimated by Knudsen and Yotopoulos 1976]. Using two measures of instability and three statistical measures of correlation, there is no evidence of a significant negative correlation. Section II tests the effect of instability on the investment rate. The results indicate that, as expected, instability has no negative impact on investment. Finally, section III tests the hypothesis that instability has a positive effect on foreign investment, as a fraction of total investment. All of the measures of correlation with the standard instability index are positive, and the two non-parametric statistics show significance at confidence levels around 90 percent. In the test for positive correlation with the Knudsen-Yotopoulos index, two of the three statistics are of the "correct" sign, though the significance levels are not as convincing as in the test using the standard instability measure. The reason for the partial discrepancy in these two sets of results may be attributable to the difference in the two instability indexes. The Knudsen-Yotopoulos index is a measure of short-term instability, while the standard instability index is a better measure of long-term instability, as indicated by Nash [1982]. For a variety of reasons, short-term instability creates less risk than long-term, and furthermore, the kind of risk which it does create would be more easily compensated by domestic investors. Since the whole motivating force behind this theory is the displacement of domestic by foreign investment to compensate for the risk created by instability, it would be expected that this would not occur in response to instability that does not create so much risk, as indicated above.

Next, consider the implications regarding the investment rates and consequent sizes of the domestic and export sectors of the economy. The hypotheses cited in Section I hold that if instability and risk are greater in the export sector, this will cause investment to be low, driving up the rate of return on investment and causing the equilibrium size of this sector to be inefficiently small. Correspondingly, the relatively less risky domestic sector will have high rates of investment, with low rates of return and in equilibrium will be inefficiently large. These hypotheses do not, of course, assert that risk is the only, or even the primary, determinant of the relative sizes of the sectors; only that there is a systematic relationship. For an example of this type of analysis.) However, in those countries where capital is free to move, the theory would predict that there should be little if any inequality in the equilibrium rates of return in the sectors. This means that the investment flow and resulting size of the export sector vis-a-vis the domestic is not influenced by the relative risk in the two sectors, as generated by instability. Consequently, in countries with no restrictions on capital flows, the relative sizes of the export and domestic sectors should be uncorrelated with their relative indexes of instability, while in countries with such restrictions, the correlation should be negative. The results of such a test, comparing ratios of domestic to export sector instability and domestic to export sector size of countries with restricted and unrestricted capital accounts. Using each of the two measures of instability and using three statistical measures, there is clear evidence of a strong negative correlation for the closed capital account sample, and no such evidence for the open. The end result of an open capital account seems to be as predicted.


The main point of this note is that neither savings nor investment in LDCs need be stifled by export instability, nor need there be any systematic distortions between investment in the domestic sector and export sector, if the capital account is open.

This simple observation may help to explain why the empirical studies of the effect of instability on savings, investment, and GDP growth have been inconsistent in their findings. The mechanism by which instability generates inefficiency in resource use and, therefore, depresses these variables has its greatest impact only in economies where capital account controls isolate the local market from world capital markets. Since empirical studies have grouped open and closed account countries together without recognizing this distinction, results of different studies have been contradictory. But equally important, it underlines the economic cost of restricting long-term capital flows. Restrictions on foreign investment in the domestic market may mean that some high-risk projects with high expected returns-particularly in the export sector-are not undertaken if the risk premium required by domestic savers with limited diversification opportunities is too high to make the project feasible. Restrictions on the accumulation of foreign assets may leave domestic residents with the option of investing their savings in high-risk national projects, or not saving as much. The first type of restriction thus depresses marginal investment and diverts it away from the export sector, while the second acts on marginal savings. Either type is harmful in the long run, not only because it discourages capital accumulation, but also because it encourages the resource misallocation that comes from differences in risk among sectors.

Given the pervasive nature of the capital account restrictions in less-developed countries, it is very important to weigh all of the costs. While some of the costs have been ably explained by other investigators-Mills [1975] and Haberler [1975], for example-the costs described here have not been previously recognized. For less-developed countries that feel that risk induced by export instability is a problem, this is a cost that should be seriously weighed.


Batara, Reveenra N. The Pure Theory of International Trade Under Uncertainty. London: The MacMillan Press Ltd., 1975

Conover, W. J. Practical Nonparametric Statistics. N.Y.: McGraw-Hill, 1962.

Coppock, J. D. International Economic Instability. N.Y.: McGraw-Hill, 1962.

Erb, G. F., and S. Schiavo-Campo. "Export Instability, Level of Development and Economic Size of Less Developed Countries." Bulletin Oxford University Institute of Economics and Statistics, November 1969, 263-83.

Glezakos, C. "Export Instability and Economic Growth: A Statistical Verification." Economic Development and Cultural Change, July 1973, 630-78.

Export Instability and Economic Growth: Reply. Economic Development and Cultural Change, April 1984, 615-23.

Haberler, G. "The Case Against Capital Controls for Balance of Payments Reasons," in Capital Movements and Their Control, edited by A. K. Swoboda. Geneva: Institut Universitaire de Hautes Etudes Internationales, 1975, 63-82.

Hawkins, R., J. Epstein, and J. Gonzales. "Stabilization of Export Receipts and Economic Development-International Commodity Agreements and Compensatory Financing Plan." The Bulletin of New York University, November 1966.

Kenen, P. B., and C. S. Voivodas. "Export Instability and Economic Growth." Kyklos, Fasc. 4, 1972, 791-804.

Knudsen, O., and A. Parnes. Trade Instability and Economic Development. Lexington: Lexington Books, 1975.

Knudsen O., and P. Yotopoulos. "A Transitory Income Approach to Export Instability." Stanford Food Research Studies, XXI 1976, 91-108.

MacBean, A. L. Export Instability and Economic Development. Cambridge: Harvard University Press, 1966.

Maizels, A. "Review of Export Instability and Economic Development." American Economic Review, June 1968, 575.

McNicol, D. L. Commodity Agreement and Price Stabilization. Lexington: Lexington Books, 1978.

Mills, R. H., Jr. "An Evaluation of Measures to Influence Volatile Capital Flows," in Capital Movements and Their Control, edited by A. K. Swoboda. Geneva: Institut Universitaire de Hautes Etudes Internationales, 1975, 143-62.

Nash, J. D. "Long-Term and Short-Term Export Instability, and an Extra Cost of Restrictions on Capital Movement." unpublished manuscript, 1982.

Olmstead, C. S., and J. W. Tukey. "A Corner Test for Association." The Annals of Mathematical Statistics, vol. 18, 1947, 495-513.

Savvides, Andreas, "Export Instability and Economic Growth: Some New Evidence." Economic Development and Cultural Change, April 1984, 607-14.

Swidrowski, J. Exchange and Trade Control. Epping, Essex: Gower Press Limited, 1975.

Voivodas, C. S. "The Effect of Foreign Exchange Instability on Growth." The Review of Economics and Statistics, August 1974, 410-12.
COPYRIGHT 1990 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Nash, John
Publication:Economic Inquiry
Date:Apr 1, 1990
Previous Article:Efficiency wage models of unemployment: a second view.
Next Article:Interest rate ceilings and the role of security and collection remedies in loan contracts.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters