Go to Brad DeLong's Home Page
TTh 12:30-2:00 Evans 60
(Sections W 10-12 Evans 6, W 12-2 Evans 2)
How mobile is capital across national borders these days? Is it much more mobile than in the past? How does capital mobility affect the development and functioning of the world economy?
These are not questions to which economists can give good answers. The problem is that we have good theories of how international capital markets should function that tell us that a number of different measures of international capital-market integration should move together. If international capital markets are not integrated, then:
- Rates of profit--returns to investments in physical capital--should widely differ across countries.
- Rates of interest--returns to investments in financial assets--should widely differ once one crosses a national border.
- Portfolios--the asset holdings of individuals or households--should be heavily concentrated in the assets of the country of residence.
- National savings and investment rates should be highly correlated.
If international capital markets are integrated, then:
- Rates of profit--returns to investments in physical capital--should differ little across countries.
- Rates of interest--returns to investments in financial assets--should differ little as one crosses a national border.
- Portfolios--the asset holdings of individuals or households--should be widely diversified: since the value of your human capital depends strongly on how well your home economy does, there is even a case that your asset portfolio should be short domestic equities.
- National savings and investment rates should be relatively uncorrelated.
Yet we seem to live in a world in which:
- We know relatively little about rates of profit--returns to investments in physical capital.
- Rates of interest--returns to investments in financial assets--appear to vary little as one crosses a national border.
- Portfolios--the asset holdings of individuals or households--are heavily concentrated in the assets of the country of residence.
- National savings and investment rates are highly correlated, but less correlated than in previous decades.
So how do we assess this mixed and confused grab-bag of results? Do we look at the high correlation between national savings and investment rates, and conclude that international capital mobility is low? Or do we look at the co-movements of interest rates, and conclude that international capital mobility is high?
The natural way to resolve this is to say that international capital mobility is high for some purposes but not for others. But our models don't give us much guidance as to which--for our models tend not to make the distinctions that we have to to make sense of the world.
In the rest of this lecture, let me first review the theory of international capital mobility, then survey the empirical evidence, and end with a non-conclusion: we know lots less than we should about this.
Theory: Potential Benefits and Costs of International Capital Mobility
For developed-developing country capital movements: allowing poor countries access to cheaper capital to fund inudstrialization, and allowing rich countries access to higher rates of return than they might otherwise have been able to achieve.
For developed-developed country capital movements: allowing risk-sharing--trade "across" states of nature.
Potential costs of international capital mobility: (i) reduced ability to run national redistributie policies, hence "race to the bottom" in capital taxation and transfer of wealth from labor to capital; (ii) vulnerability to noise-driven financial crises.
Measuring Capital Mobility: Law of One Price
Tests of uncovered returns on different currencies are relatively uninformative: who is to say what risk premia are, or what exchange rate forecast errors are? The law of one price says that (nominal) interest rates in one country are equal to (nominal) interest rates in another plus the expected rate of change in the (nominal) exchange rate plus the relevant risk premium. But in the data we have such tests have low power, and require lots of auxilliary maintained hypotheses.
The most useful information about whether capital is "mobile" in the sense of the law of one price is found by examining onshore-offshore interest rate differentials.
Frankel: financially "open" and financially "closed" countries in the mid-1980s. Open: Austria, Belgium, Canada, Germany, Hong Kong, Japan, Netherlands, Singapore, Sweden, Switzerland, U.K. Closed: Greece, Mexico, Portugal.
If countries possess perfect international capital mobility, their consumption levels should be highly correlated: ex-post cross-country differences in intertemporal marginal rates of substitution will be uncorrelated with any random variables on which internatinal contracts can be written: asset trade allows the sharing of some (but not all) risks.
But imperfect correlations among country-level consumption are likely to be in large measure the result of generalized asset-market incompleteness rather than of international capital-market segmentation.
Extent of International Portfolio Diversification
French and Poterba (1990, 1991): there is a substantial "home bias" in the portfolios of industrial-country investors. French and Poterba: U.S., Japanese, and U.K. investors hold 94, 98, and 82 percent of their stock market portfolios in home-headquartered companies. (On the other hand, in December 1991 Germany's gross external assets were 73% and liabilities were 51% of its annual GDP; for the U.S., 34.5% and 40.9%; for Japan, 59% and 48%.)
There is no consensus on how large the benefits to developed countries from international consumption insurance would be. There is no doubt that developing countries would gain to the tune of several percent of GDP from eliminating consumption volatility.
Feldstein-Horioka and the International Correlation of Savings and Investment Rates
Is the high correlation of savings and investment rates explained by low capital mobility, or by other factors that simultaneously drive both?
The cross-section relationship: in cross-section, countries' savings and investment rates are highly correlated. There is some sign that the strength of the cross-sectional relationship has been dropping since the 1960s.
The time-series relationship: much harder to get a clear picture--a lot of cross-country variability.
Is the high cross-sectioal savings-investment relationship due to...
- ...demography? Probably not. (Although see Alan Taylor.)
- ...real interest rates? Frankel argues that increases in national savings rates depress the local real interest rate, and so generate high investment. This seems to be not a contradiction but a confirmation of Feldstein-Horioka.
- ...hysteresis in factor supplies? No.
- ...corporate financing frictions? No.
- ...current-account targetting by governments? Maybe.
- ...intertemporal budget constraints? Maybe--but this is in part a restatement of Feldstein-Horioka's point.
- ...under the gold standard, the R^2 from the savings-investment regression was less than half of what it has been since.
How does one square the generally smooth international interest-rate arbitrage with the low international consumption correlations and the home portfolio bias, or with the still-sizable cross-sectional coherence between savings and investment?
Professor of Economics J. Bradford
DeLong, 601 Evans
University of California at Berkeley; Berkeley, CA 94720-3880
(510) 643-4027 phone (510) 642-6615 fax
This document: http://www.j-bradford-delong.net/Teaching_Folder/Econ_202b/Handouts/mobility.html