Abstract
Cross-country capital flows have been widely studied in the literature; however, why some countries may form more similar foreign investment portfolios than others has not been investigated. Using data for a broad panel of countries during the period 2002–2015, we adopt gravity equations to estimate cross-country foreign portfolio investment patterns. The main empirical results reveal that countries are more likely to form similar foreign portfolio investment patterns if: (i) countries are geographically closer; (ii) countries share the same official language; and (iii) countries adopt fixed exchange rate regimes.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Data Availability Statement
The data that supports the findings of this study are available in the supplementary material of this article.
Notes
The control variables are included based on the following rationale. Real GDP per capita and population size are important measures of countries’ fundamentals that may affect foreign investment decisions. For instance, the role of economic growth in the stock market performance is well established in the literature (King and Levine 1993; Levine 1991; Levine and Zervos 1998). Regarding population size, Abel (2000) adopts an overlapping generations (OLG) model to study the relationship between demographics and stock market and finds that a baby boom can increase the price of capital. As a result, we control the real GDP per capita and the population size of the source country and the host country in the regressions.
The reason we control for the absolute differences in per capita incomes and populations sizes is to follow the Linder hypothesis. As analyzed by Frankel (1997), the Linder (1961) hypothesis predicts that countries with similar per capita incomes will have similar preferences and similar but differentiated products, and thus, will trade more with each other. The Linder hypothesis is usually described as predicting the effect of the absolute difference of per capita GDPs on trade (or capital flows in this paper). In the spirit of Linder (1961), Gruber and Vernon (1970) and Thursby and Thursby (1987) include the absolute difference in per capita incomes in the standard gravity estimation to analyze the differences in countries’ consumption patterns. In line with these studies, we control for the absolute differences in log per capita GDPs and log population sizes in the estimations.
Foreign financial asset data end in the year 2015.
According to the IMF, the top 10 economies for foreign investment are the US, Japan, Luxembourg, the UK, Germany, France, Ireland, Cayman Islands, the Netherlands, and Hong Kong. In this paper, we construct the total FPI for each economy by summing up the investments spent on these economies’ financial assets.
The kspeg data are available on https://www2.gwu.edu/~iiep/about/faculty/jshambaugh/data.cfm .
One potential problem in the kspeg index is that there are a large number of missing values (in 434 country pairs). This is mainly because the kspeg index considers only the direct peg between one country to a base country (such as the US). For instance, two countries i and j may peg to the US dollar; however, the relationship between countries i and j may be missing in the kspeg index. Thus, we revise the index by adding more information to pairs (i,j) that have missing values. For instance, for the pair (i,j), if they peg to some third country, we set kspeg to one; otherwise, it is set to zero.
To construct a binned scatter plot between y and x, we first regress the y variable and the x variable on the same set of control variables, respectively, and collect residuals from two regressions. In the examples, the set of regressors we use to construct scatter plots is based on Column (1) in Table 1, after we exclude the two main variables y and x. We denote the residuals from the y variable regression and the x variable regression by \(e_y\) and \(e_x\), respectively. Then, we divide \(e_x\) into 100 equal-sized bins and every \(e_y\) must fall into one bin. In each bin, we compute the means of \(e_y\) and \(e_x\). Finally, we plot all the means of \(e_y\) against the means of \(e_x\).
For the Contiguity index, countries that share the same border may not necessarily have similar cultures or economic situations. A country such as Russia that is geographically large is bordered by many countries. For instance, China is one of the neighboring countries of Russia; however, the cultures and economies of China and Russia are very different. This is one reason why we consider Distance is a better measure to capture the similarities between a pair of countries. According to the definition, Distance is the weighted bilateral distance (population weighted) between two countries. In this way, the measure puts greater weight on the bilateral distance between bigger cities (usually with larger population sizes). Note that trade and capital flows between big cities usually play very important roles in a country pair; thus, the Distance measure may well capture economic and culture exchanges across countries. As shown in the regression tables, the Distance index shows a more robust pattern than the Contiguity index.
For the Common colonizer index, we consider that it may affect the regression results, but again, we do not think it is a perfect measure for capturing cultural similarities. Especially after controlling for indices such as common language, the effect of Common colonizer may become weaker. We also can see from the data that there are countries that have never been fully colonized. For pairs that include such countries, the Common colonizer index takes the value of zero; however, this does not mean the cultures of the countries in those pairs are not similar. For instance, China is considered one country that has not been fully colonized, but China and many Asian countries have cultural similarities. Thus, we include the Common colonizer index in the regressions only to control for the potential effect that might come from it; we do not treat this index as one of the key explanatory variables. In fact, previous studies such as Coeurdacier and Guibaud (2011) also find that the Common colonizer index does not have a robust effect on bilateral cross-border equity holdings, and its sign varies across model specifications.
For the Common religion dummy, based on the definition, it is calculated by adding the products of the shares of Catholics, Protestants, and Muslims in a country pair. One reason that the Common religion dummy shows various effects when we focus on the foreign investment in different destination countries is that this index may not completely reflect the religious similarity between the countries in a pair. For instance, this dataset contains a number of East Asian countries, which have very small religious population based on this definition. Taking Korea as an example. According to Korea’s 2015 national census, more than half of Koreans consider themselves not to be religious, and among the rest of the population, a large percentage are members of other religions that are not considered in the Common religion index. In fact, the value of the Common religion dummy is almost close to zero for country pairs that include Korea; however, Korea and other East Asian countries such as China share cultural similarities. To find the true effect of Common religion on foreign investment between countries, a better measure is needed. Unfortunately, at this stage, we are unable to find the perfect measure for common religion. Thus, we control only the index we have in regressions but do not treat it as one of the main variables that determine the foreign investment pattern across countries.
The foreign investment in UK financial assets is different from assets from other countries on which the coefficient on kspeg is positive but statistically insignificant.
This variable varies from 0 to 12. We construct the variable based on standard time zones, abstracting from the issue of daylight savings. Data are obtained from https://www.timeanddate.com.
The real interest rate is the lending interest rate adjusted for inflation as measured by the GDP deflator (which takes a value from 0 to 100). We converted the real interest rate into a decimal term. Data are obtained from the WDI database. In the regression, we use the absolute difference in the real interest rates within each pair of countries.
Stock market return is the growth rate of the annual average stock market index. The data are obtained from the Global Financial Development database.
Capital account openness is measured with the Chinn and Ito (2006) index and is retrieved from http://web.pdx.edu/~ito/Chinn-Ito_website.htm.
References
Abel AB (2000) The effects of a baby boom on stock prices and capital accumulation in the presence of social security. The Wharton School of the University of Pennsylvania, Mimeo
Aggarwal R, Kearney C, Lucey B (2012) Gravity and culture in foreign portfolio investment. J Bank Finance 36:525–538
Aviat A, Coeurdacier N (2007) The geography of trade in goods and asset holdings. J Int Econ 71(1):22–51
Cerutti E, Claessens S, Puy D (2019) Push factors and capital flows to emerging markets: why knowing your lender matters more than fundamentals. J Int Econ 119:133–149
Chinn MD, Ito H (2006) What matters for financial development? Capital controls, institutions, and interactions. J Dev Econ 81(1):163–192
Coeurdacier N, Guibaud S (2011) International portfolio diversification is better than you think. J Int Money Finance 30(2):289–308
Davis S, Jeremy N, Paul W (2001) On the gains to international trade in risky financial assets. Chicago Graduate School of Business mimeograph
Forbes KJ (2010) Why do investors invest in US? J Int Econ 80(1):3–21
Forbes KJ, Warnock FE (2012) Capital flow waves: surges, stops, flight, and retrenchment. J Int Econ 88(2):235–251
Frankel JA (1997) The gravity model of bilateral trade. In: Frankel JA (ed) Regional trading blocs in the world economic system. Peterson Institute for International Economics, Washington, DC, pp 49–76
Fratzscher M (2012) Capital flows, push versus pull factors and the global financial crisis. J Int Econ 88(2):341–356
Griffin JM, Nardari F, Stulz RM (2004) Are daily cross-border equity flows pushed or pulled? Rev Econ Stat 86(3):641–657
Gruber WH, Vernon R (1970) The technology factor in a world trade matrix. In: Vernon R (ed) The technology factor in international trade. Columbia University Press, New York, pp 233–272
King RG, Levine R (1993) Finance and growth: Schumpeter might be right. Q J Econ 108:717–738
Klein MW, Shambaugh JC (2006) Fixed exchange rates and trade. J Int Econ 70(2):359–383
Lane P, Milesi-Ferretti GM (2008) International investment patterns. Rev Econ Stat 90(3):538–549
Levine R (1991) Stock markets, growth and tax policy. J Finance 46(4):1445–1465
Levine R, Zervos A (1998) Stock markets, banks, and economic growth. Am Econ Rev 88(3):537–558
Linder S (1961) An essay on trade and transformation. Wiley, New York
Okawa Y, van Wincoop E (2012) Grvaity in international finance. J Int Econ 87(2):205–215
Portes R, Rey H (2005) The determinants of cross-border equity flows. J Int Econ 65(2):269–296
Sarno L, Taylor MP (1999) Hot money, accounting labels and the permanence of capital flows to developing countries: An empirical investigation. J Dev Econ 59:337–364
Sarno L, Tsiakas I, Ulloa B (2016) What drives international portfolio flows? J Int Money Finance 60:53–72
Stein EH, Daude C (2007) Longitude matters: time zones and the location of foreign direct investment. J Int Econ 71(1):96–112
Thursby JG, Thursby MC (1987) Bilateral trade flows, the Linder hypothesis, and exchange risk. Rev Econ Stat 69(3):488–495
Vermeulen R (2010) Dynamic international portfolio adjustment: rational investors and the home bias. https://doi.org/10.2139/ssrn.1534626
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interest
The authors declare that they have no conflict of interest.
Ethical approval
This article does not contain any studies with human participants or animals performed by the authors.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors thank Giovanni Caggiano, Russell Smyth, Vinod Mishra, Ben Li, Jian Lu, Silvio Contessi and the seminar participants at Monash University for helpful comments. The authors are grateful to Joakim Westerlund (editor-in-chief) and three insightful and constructive anonymous referees. This research is part of Lei Pan’s PhD thesis at Monash University. The usual caveat applies.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Pan, L., Hu, R. & Du, Q. Foreign portfolio investment patterns: evidence from a gravity model. Empir Econ 63, 391–415 (2022). https://doi.org/10.1007/s00181-021-02133-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00181-021-02133-0