FRBSF Economic Letter

2000-30 | October 6, 2000

Knightian Uncertainty and Home Bias

Kenneth Kasa

More Americans than ever buy foreign goods, own foreign stocks, and work for foreign corporations. The U.S. economy is undoubtedly becoming more “globalized.” What is often not appreciated, however, is that it is still not as globalized as it should be, at least according to traditional economic theories. It is ironic that at the same time as politicians and journalists are debating the costs and benefits of globalization, most economists are struggling to explain why the world isn’t more globalized.

This Letter reviews recent attempts to explain the apparent lack of globalization. It starts by outlining the various ways in which the world economy falls short of the globalization predicted by standard economic models. It then discusses previous attempts to explain this lack of globalization by incorporating trading costs and information asymmetries into traditional frictionless models of international economics. Unfortunately, each of these attempts is deficient in some way, and so I conclude by discussing a new theory based on the concept of “Knightian Uncertainty.” This theory encapsulates the vague, but intuitive, notion that foreign asset markets are less “familiar” than domestic asset markets.

Manifestations of home bias

In a recent paper, Obstfeld and Rogoff (2000) point to six puzzles confronting international economists. These are: (i) investors hold a disproportionate share of domestic assets in their portfolios, apparently forgoing diversification benefits, (ii) consumption is far less correlated across countries than traditional models of risk-sharing predict, (iii) national savings rates are highly correlated with national investment rates, apparently contradicting the idea that globalization severs the link between them, (iv) when expressed in common currency units, prices of the same good differ wildly and persistently across countries, contradicting the idea of “Purchasing Power Parity,” (v) exchange rates are far more volatile than their underlying macroeconomic determinants, and (vi) people trade a lot more with residents of their own country than with residents of adjacent countries, i.e., political borders by themselves play a significant role in determining trading volumes; as one example, Helliwell (1998) finds that after you control for distance and population, trade between Canadian provinces is about ten times greater than trade between provinces and U.S. states.

Obstfeld and Rogoff argue that each of these puzzles reflects or results from a sort of “home bias” in consumption and investment. Of course, saying there is a bias in anything implicitly presumes a standard against which this bias is measured. In models of international trade and finance, this standard is based on the predictions of models that make several key simplifying assumptions. Specifically, it is typically assumed that individuals in different countries have (nearly) identical tastes and information and they can freely trade all goods and securities, i.e., there are no taxes or transportation costs. Naturally, attempts to resolve these puzzles have focused on relaxing these assumptions.

Potential explanations and their drawbacks

It has long been known that barriers to international investment, like taxes and capital controls, can produce a home bias in portfolios, which then weakens the correlation of consumption across countries (see, e.g., Black 1974). Unfortunately, there are empirical problems with this explanation. First, transaction costs on international investment are no longer very large, at least in the major markets. Second, Tesar and Werner (1995) show that turnover rates on the foreign component of portfolios are actually greater than on the domestic component. These facts cast doubt on simple transaction cost explanations of portfolio home bias.

Rather than incorporate costs to trading assets, Obstfeld and Rogoff (2000) incorporate costs to trading goods and services. They argue that these costs potentially provide a “Grand Unified Theory” of international economics. It is not surprising that trade barriers, like tariffs and shipping costs, can simultaneously explain the lack of trade between countries, weak consumption correlations, and high correlation between national saving and investment rates. What’s more surprising is that they can rationalize portfolio home bias. Because of trading costs, individuals consume different goods, and, as a result, they use assets to hedge different risks. Obstfeld and Rogoff show that the desire to hedge different risks can lead to portfolio home bias. What’s most surprising about the implications of trading costs, however, is their potential role in explaining exchange rate volatility. In traditional models, commodity arbitrage and the law-of-one-price rein in exchange rate movements. By weakening the forces of arbitrage, trading costs permit speculative forces to play a greater role in determining exchange rates.

While no doubt promising, a trading cost explanation has a couple of drawbacks. First, Engel (2000) points out that it predicts a counterfactually high correlation between consumption and real exchange rates. Second, explicit trading costs appear to be too small to explain the magnitude of home bias. Obstfeld and Rogoff are not unaware of this, so they emphasize that their trading costs should be interpreted metaphorically as capturing all the reasons why it is more costly to trade across borders. Unfortunately, appealing to intangible costs detracts from the power of the theory, and raises obvious questions about the nature of these costs.

A third explanation, which gets closer to these intangible costs, appeals to information asymmetries. It seems quite natural to assume that domestic residents know more about local investment opportunities than foreigners do. Gehrig (1993) shows that this advantage produces portfolio home bias. In addition, Brennan and Cao (1997) show that if foreigners have less information about domestic market conditions, their investment decisions will be more sensitive to new information, and as a result, they will be net buyers during good times and net sellers during bad times. This accords with conventional wisdom regarding international capital flows.

Unfortunately, explanations based on asymmetric information also have their drawbacks.

Based on the strong regional component of capital flows, Froot, et al. (1998) argue that the correlation between asset returns and capital flows is best explained by global demand shocks, not shocks to country-specific market conditions. Also, while it is natural to assume domestic residents know more about local conditions, the exact nature and source of this asymmetry are left unspecified and unexplained. Home bias is a pervasive phenomenon, which seems to go beyond knowledge of investment opportunities. Instead, there appears to be a more fundamental factor involved.

Knightian Uncertainty: a new explanation

One of the striking things about home bias is that not only is there a bias between countries, but there also appears to be a bias within countries as well. Huberman (2000) discusses a wide range of evidence showing that individuals prefer to invest locally. For example, the domestic portfolios of U.S. money managers are biased toward firms that are located near their offices. Portes and Rey (1999) show that standard “gravity models” of international trade also explain international capital flows. That is, geographic distance helps to explain the magnitude of bilateral capital flows. This is surprising in an age of computer networks and telephones.

To explain the preference for local investment, Huberman appeals to the notion of “familiarity.” He argues that people simply prefer to deal with familiar situations. However, he doesn’t really define or explain what familiarity is. Fortunately, recent developments in decision theory potentially provide an answer.

In a classic book, Knight (1921) proposed a distinction between “risk” and “uncertainty.” Risk describes situations where an explicit probability distribution of outcomes can be calculated, perhaps on the basis of actuarial data. In contrast, uncertainty describes situations where probabilities are unknown, and more importantly, where they are impossible to calculate with any confidence due to the uniqueness or specificity of the situation. Knight argued that “profits” should be defined as the reward to bearing uncertainty.

The concept of “Knightian Uncertainty” has always been attractive to economists, but it has been difficult to formalize. For example, Savage (1954) pointed out that, even when it is impossible to construct an explicit probability distribution, people nonetheless manage to make decisions and resolve trade-offs, and under certain axioms, these decisions reveal the implicit probabilities they assign to potential outcomes. Whether these probabilities are “objective” and based on relative frequencies, or whether they are “subjective” and based on individual beliefs, was shown to be irrelevant to the analytics of decisionmaking. After Savage, economists lost interest in Knightian Uncertainty.

Recently, however, there has been a revival of interest in Knightian Uncertainty. This interest was sparked by research showing how to relax the Savage axioms in a way that delivers a workable model of Knightian Uncertainty. Epstein and Miao (2000) apply this research to international investment. They assume domestic investment involves risk, whereas foreign investment involves Knightian Uncertainty. This formalizes Huberman’s idea that international investment is less familiar.

Although resolving Obstfeld and Rogoff’s six puzzles is not the direct object of their study, Epstein and Miao’s results suggest that Knightian Uncertainty can potentially explain several of these puzzles. For example, their model predicts portfolio home bias, weak consumption correlations, and volatile asset prices.

At this point, using Knightian Uncertainty to explain home bias faces two challenges. The first is to quantify the degree of Knightian Uncertainty that is necessary to account for the magnitude of home bias. Epstein and Miao’s results are as yet purely qualitative. Second, current models of Knightian Uncertainty abstract from learning. One might expect that over time Knightian Uncertainty would dissipate with learning. On the other hand, Knightian Uncertainty might be self-perpetuating if lack of familiarity discourages the sort of contact and experience that would otherwise diminish it. If so, then home bias might be with us for a long time.

Kenneth Kasa
Senior Economist


References

Black, Fischer. 1974. “International Capital Market Equilibrium with Investment Barriers.” Journal of Financial Economics 1, pp. 337-352.

Brennan, Michael J., and H. Henry Cao. 1997. “International Portfolio Investment Flows.” Journal of Finance 52, pp. 1,851-1,880.

Engel, Charles. 2000. “Comments on Obstfeld and Rogoff’s `The Six Major Puzzles in International Macroeconomics: Is There a Common Cause'” NBER Working Paper No. 7818.

Epstein, Larry G., and Jianjun Miao. 2000. “A Two-Person Dynamic Equilibrium under Ambiguity.” Working Paper, Department of Economics, University of Rochester.

Froot, Kenneth A., Paul O’Connell, and Mark Seasholes. 1998. “The Portfolio Flows of International Investors.” NBER Working Paper No. 6687.

Gehrig, Thomas P. 1993. “An Information Based Explanation of the Domestic Bias in International Equity Investment.” Scandinavian Journal of Economics 21, pp. 97-109.

Helliwell, John. 1998. How Much Do National Borders Matter? Brookings Institution.

Huberman, Gur. 2000. “Home Bias in Equity Markets: International and Intranational Evidence.” In Intranational Macroeconomics, eds. Gregory Hess and Eric van Wincoop. Cambridge: Cambridge University Press.

Knight, Frank H. 1921. Risk, Uncertainty, and Profit. Boston: Houghton Mifflin.

Obstfeld, Maurice, and Kenneth Rogoff. 2000. “The Six Major Puzzles in International Macroeconomics: Is There a Common Cause?” Forthcoming in NBER Macroeconomics Annual.

Portes, Richard, and Helene Rey. 1999. “The Determinants of Cross-Border Equity Flows.” NBER Working Paper No. 7336.

Savage, Leonard J. 1954. Foundations of Statistics. New York: John Wiley and Sons.

Tesar, Linda, and Ingrid Werner. 1995. “Home Bias and High Turnover.” Journal of International Money and Finance 14, pp. 467-492.

Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Sam Zuckerman and Anita Todd. Permission to reprint must be obtained in writing.

Please send editorial comments and requests for reprint permission to
Research Library
Attn: Research publications, MS 1140
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120