Financial Globalization, Capital Account Liberalization and Risk Sharing Mayank Gautam∗ (Job Market Paper) November 2007

Abstract Conventional wisdom suggests that countries that are financially integrated should be better insured against macroeconomic risk. In this paper, I study whether international risk sharing has improved with greater financial integration during the period 1973-2003. Using panel data regressions, I demonstrate that risk sharing improves with greater financial integration for industrial countries, especially during the globalization period. However, developing countries do not show any discernible pattern in risk sharing. Emerging economies, which accounted for a huge share of capital flows in the recent past, show only a marginal change in risk sharing. Capital flows come in different types such as portfolio equity flows, debt flows, foreign direct investment among others and each affects risk sharing differently. I examine the unique role of different types of capital flows in risk sharing and find that risk sharing benefits are higher for industrial countries for each of these components. On the contrary, greater reliance on debt flows appears to have reversed the potential risk sharing gains for the emerging economies. JEL Classification: F15, F36, G15 Keywords: Financial markets integration, income insurance, consumption risk sharing, international capital markets

∗ PhD candidate, Department of Economics, University of Houston, Houston, TX, 77204 (e-mail: [email protected]). I thank Bent Sørensen and Sebnem Kalemli-Ozcan for constant support and encouragement throughout this work.

1

Introduction

One of the most important potential benefits of international financial integration is the opportunity that it provides for international diversification of macroeconomic risks. Holding foreign assets and liabilities is a mechanism that contributes to delinking of domestic income (and eventually consumption) and output. When risk is diversified across national borders, holding foreign assets and liabilities can hedge against idiosyncratic output volatility and ensure a steady stream of income flows. Economies around the world are not synchronized and at a point in time, some countries expand while others fall into recession. In a non-synchronized world, international financial integration ensures that domestic income is no longer determined by domestic output, but it becomes contingent on “world” output growth. Since the collapse of the Bretton Woods system in 1973, capital account liberalization by many countries spurred a gradual integration of international capital markets. Especially in the last two decades, the world has witnessed an astronomical increase in the volume and the composition of capital flows among countries.1 In this paper, using panel data regressions, I examine whether international portfolio diversification is the major channel through which risk sharing takes place. I also examine the evolution of risk sharing patterns and channels in the light of changes in the volume and the composition of international financial flows. Using a panel of 68 countries, I demonstrate that industrial countries make unequivocal risk sharing gains, while developing countries and emerging market economies do not show any discernible pattern. Among the types of capital stock—Portfolio equity, portfolio debt, FDI etc.—industrial countries derive risk sharing benefits from each of these types. On the contrary, developing countries significantly relied on less stable capital flows such as short-term debts which appears to have reversed some potential risk sharing advantages for them. There is a substantial literature that tests the predictions of risk sharing under complete markets with contingent claims.2 Not only is complete risk sharing rejected, but also a very low degree of risk sharing is observed for countries. The International Real Business Cycle literature, most notably Backus et al. (1992), and others like Baxter and 1 Obstfeld and Taylor (2005) report that global capital markets have grown rapidly in the 1980s and especially the 1990s. They also note that the bulk of this growth has taken the form of “diversification finance”, captured by increase in external assets and liabilities while small net position as a ratio of national output. The “asset swapping for the purpose of mutual diversification” is considered more important for its risk sharing benefits. 2 See Obstfeld and Rogoff (1996) for a textbook treatment of the theory of international trade in assets.

1

Crucini (1995) and Stockman and Tesar (1995), find that inter-country consumption correlations are lower than output correlations. Other studies involving various regression techniques examine the pattern of risk sharing (some of the notable contributions include Obstfeld 1994; Lewis, 1996, 1999; Sorensen and Yosha, 1998), and conclude that there is only limited evidence for risk sharing. More recently, some studies document that risk sharing has increased with recent spurt of financial globalization (e.g., Sørensen, Yosha, Wu, and Zhu, 2007; Artis and Hoffman, 2006, 2007).3 Another noticeable feature of the risk sharing literature is that it mostly examines the advanced industrial countries. Starting in the mid-1980s, developing countries led by emerging market economies made huge strides in the global financial market. Capital-poor developing economies, in theory, can obtain greater risk sharing than capital-rich developed countries from increased financial flows. Developing economies tend to have volatile output growth given their factor endowment structure and specialized nature of their output. Potentially, they have better scope to insulate domestic income from domestic output fluctuations by generating steady income streams from cross holding of assets and liabilities. In turn, a smooth income stream may be helpful in stabilizing consumption and facilitating desired investment plans. The empirical risk sharing literature seems to has overlooked the unfolding of this new era in international financial integration. Obstfeld (1994) and Lewis (1999) in their studies have some developing countries, but the sample ends too soon to capture this new trend. Kose et al. (2007) undertake a study involving a broader set of countries for longer time period, but it only considers consumption risk sharing.4 My study aims to fill this gap in the literature. Using alternative measures of risk sharing, I extend the analysis to a larger group of countries and also trace the patterns and channels of risk sharing. International financial integration and risk sharing may be expressions of the same underlying behavior: if agents diversify their portfolios internationally, countries become more financially integrated. Agents will likely obtain smoother income (and eventually smoother consumption) streams as domestic shocks partially will be offset by foreign asset income. In turn, countries share macroeconomic risk in response to greater financial 3 Sørensen, Yosha, Wu, and Zhu (2007) run risk sharing regression using growth rates of the variables while Artis and Hoffman (2006, 2007) run using levels. Both studies arrive at the same conclusion that countries with higher degree of financial integration appear to share income risk to a greater extent. 4 The ex-ante “income”-based risk sharing is effective for smoothing both permanent and transitory shocks. The ex-post “consumption”-based risk sharing smoothes only transitory shocks. I return to this issue shortly.

2

integration. Consider the (simplified) identities:5 GNI

=

GDP

+

X

(rij Aij − rji Aji ) ,

(1)

i6=j CONS = GNI −

Gross National Saving

,

(2)

where GNI is Gross National Income (formerly referred to as Gross National Product), GDP is Gross Domestic Product, Aji is the value of the stock of foreign assets in country i owned by residents in country j, rji is the rate of return on these assets and, Aij and rij are the value of the total stocks in country j owned by residents in country i and the return on those, respectively. CONS is total consumption, including government as well as private consumption. If the term rij Aij − rji Aji is not perfectly correlated with GDP, the GNI of a country may be less variable than it would be in the absence of international assets. Consumption (CONS) may be stabilized relative to GDP because GNI is stabilized, or because pro-cyclical saving helps insulate consumption from shocks to GDP that are not stabilized in GNI. Previously, very little systematic empirical evidence has been brought to bear on this issue. Lane (2001) concludes that “positive gross international investment positions in general are not associated with income-smoothing at business-cycle frequencies”, although Lane (2000) finds that the international equity positions do contribute to Ireland’s risk sharing with other European countries. However, international security holdings have been rapidly increasing throughout 1990s, therefore, any impact on risk sharing should now be easier to detect. In a recent study, Sørensen, Yosha, Wu, and Zhu (2007) document for a sample of OECD countries that during the period 1993-2003, increase in portfolio holdings by residents increased risk sharing. In this paper, I employ two alternative measures of risk sharing to examine whether risk sharing increases with greater international financial integration (measured by the amount of international financial assets that they trade) and also explore the channels of risk sharing by looking at various components of capital flows.6 I demonstrate that 5 The equations displayed highlight the major components of the national accounts and ignore less important parts. In the national accounts, GNI equals GDP (the value of domestic production) plus net factor income from the rest of the world. Net factor income from the rest of the world is net asset income plus domestic residents’ income from foreign countries minus income of foreign residents from the domestic country. Since the latter type of factor income is based on residency rather than citizenship, it is typically small. The major part of the difference between GNI and consumption is gross saving which consists of depreciation and net saving (by governments, corporations, and individuals). Sørensen and Yosha (1998) examine the contribution of the various components of GDP to international risk sharing in much more detail. 6 Sørensen, Yosha, Wu, and Zhu (2007) argue that it is advisable to look at both income and consumption measures of risk sharing. Consumption risk sharing is important because, ultimately, economic agents care about consumption and the macroeconomic literature focuses on consumption risk sharing. However, consumption data are affected by taste shocks

3

for industrial countries, greater financial integration corresponds to higher risk sharing. Among the channels of risk sharing, in terms of equation (1), a large domestic stock of foreign assets (Aij ) (variously transformed) predicts higher risk sharing. Developing countries, although as theory predicts should potentially gain the most from financial integration have been unable to realize any risk sharing gains. More importantly, emerging economies, which accounts for a lion’s share of capital flows among developing countries show only marginal change in risk sharing. A notable result for the emerging economies is a decrease in the risk sharing in response to increased holding of debt instruments. The reliance on less stable capital flows such as short-term debt appears to have reversed some potential risk sharing advantages for these countries. The rest of the paper is organized as follows. Section 2 presents the appropriate measure of financial integration for the purpose of risk sharing, looks at the capital flows data, and points to literature covering the issues related to capital flows and international financial integration. Section 3 starts by characterizing the full risk sharing allocation, and its relation to consumption and income smoothing. It draws empirical implications of the theory, comparing various approaches in the risk sharing literature, and also lays out the alternative measures of risk sharing used in this study. Section 4 asks if countries with greater financial integration obtain better income and consumption risk sharing and in more detail, if risk sharing is correlated with the amount of foreign assets or liabilities held as a fraction of GDP. Section 5 discusses the results and section 6 concludes the paper with final remarks and policy implications.

2

International portfolio holdings and financial integration

2.1

Measuring Financial Integration

Financial openness is traditionally measured on the basis of legal restrictions on crossborder capital flows. This measure of financial integration uses IMF’s widely used Annual Report on Exchange Arrangement and Exchange Restrictions (AREAR) involving over 60 different types of control and called de jure measures.7 De jure measures summarize (broadly defined) and because net foreign capital income, such as dividends and interest from foreign assets, directly affects GNI, “income”-based risk sharing based on GNI captures better “signal-to-noise” ratio. On the other hand, consumption data may be preferable if the returns to foreign assets are dominated by yet-to-be-realized capital gains which will affect consumption but not be recorded in net foreign asset income. 7 Kose et al. (2006) discuss at length the evolution of de jure measure of financial openness in literature.

4

a host of information on capital controls from AREAR reports, but it is not an ideal measure for the study of risk sharing among countries. Shortcomings of this measure include its reliance on various restrictions associated with foreign exchange transactions and its failure to capture the degree of enforcement of capital controls. China, to take the most obvious example, controls capital movements, but has become a major exporter of capital and also remains a favored capital destination. Therefore, in order to decipher the link between financial integration and risk sharing, we need a measure of financial integration that captures the cross holding of foreign portfolios and works in practice. The measure of financial integration that is ideal for this study is the de facto measure of integration based on actual capital flows (advocated, for example, in Prasad, Rogoff, Wei and, Kose, 2003). De facto measures take into account how much a country is integrated into international capital markets in practice. The flows could be measured as gross flows and net flows, and one issue is whether to measure integration using gross or net flows. Gross flows seem to be an ideal measure for risk sharing since it captures the two-ways flows. For instance, domestic risks can be laid off by issuing state-contingent foreign liabilities. In other direction, holding foreign assets provide diversification, since the return on these foreign equities will be determined by external event. Even if net positions are small, large gross positions have important implications for risk sharing.8 For instance, if countries have large gross stocks of assets and liabilities, small exchange rate changes can have large valuation effects and serve as a mechanism for risk sharing even if net positions are small. In this paper, I rely on de facto measures of financial openness which is the sum of gross inflows and outflows. These stocks are essentially just a refined cumulative version of the underlying flows corrected for valuation effects. Kose et al. (2006) note that this measure preserves the spirit of measuring integration and obviates many of problems associated with flow data. Nonetheless, such annual flows tend to be volatile and are prone to measurement error. Additionally, foreign investment stocks are commonly measured at a point in time in current nominal terms. Based on the growth in these nominal variables, it would be falsely claimed that the market integration has increased. A way around this problem is to normalize foreign capital at each point in time by some measure of the size of the economy. A readily available measure of the size of an economy is its output level measured in current prices in a common currency unit. Thus, my measure of financial integration focuses on capital-to-GDP ratios of the form: 8 Net flows could be small if the countries undertake capital flows for the mutual diversification motive. Obstfeld and Taylor (2005) characterize today’s capital flows driven by asset “swapping” by rich countries for mutual diversification.

5

Foreign assets-to-GDP ratioit =

X

(Aijt /Yit )

(3)

i6=j

Foreign liabilities-to-GDP ratioit =

X

(Ajit /Yit )

(4)

i6=j

Financial openness-to-GDP ratioit =

X

((Aijt + Ajit )/Yit )

(5)

i6=j

Here Aijt measures the total stock in country j that is owned by country i at a time t. The measure of financial integration draws upon the work of Lane and Milesi-Ferretti (2006). The authors have constructed an extensive dataset of stock of gross assets and liabilities for 145 countries covering the period 1970-2004.9 Their dataset contains information about the composition of international financial positions, including portfolio equity investment, external debt, foreign direct investment etc. In addition, the dataset has the virtue of accounting for valuation effects and other problems that typically plague raw country-level data, and also corrects for cross-country differences in data definitions and variable construction.

2.2

A first look at the data

Table 1 displays the ratio of foreign assets, liabilities and the financial openness to GDP for 1973, 1987, and 2003 for the country groups that consist the sample.10 Data sources and definition are in the appendix. The assets and liabilities refer to value of gross foreign assets and liabilities divided by GDP respectively. Financial openness is measured as the ratio of sum of assets and liabilities to GDP. The financial openness across the world has quadrupled in the span of 30 years from 1973-2003. The most noticeable being the case of industrial countries where the financial openness has increased almost seven fold from 0.68 to 4.62. Developing country averages should be approached with caution because the country sample is small until the 1990s. Developing countries also register a modest but similar trend in financial integration compared to the industrial countries. 9 In this paper, the authors substantially extend their widely-used External Wealth of Nations database (Lane and MilesiFerretti (2006) by employing a revised methodology and utilizing a larger set of sources. While their benchmark series is based on the official estimates from the International Investment Positions, they compute the stock positions for earlier years using data on capital flows and account for capital gains and losses. 10 The country groups consist of industrial countries, emerging market economies and the developing countries. Emerging markets are a subset of developing economies. Morgan Stanley Capital International Inc. (MSCI) maintains a list of 26 “more financially integrated” countries, which we are calling emerging markets. The appendix has the list of countries included in the sample in different country groups.

6

Emerging market economies dominate the capital flows in the group of developing countries.11 Breaking the financial openness into assets and liabilities, it is observed that industrial countries witnessed a drastic increase in both assets and liabilities during the globalization period.12 Table 2 displays the components of assets and liabilities into the ratio of equity, debt and FDI to GDP for the country groups. Until 1987, developing countries and especially emerging economies held almost negligible shares of equity and FDI assets to GDP. Debt flows heavily dominated the asset portfolio of developing countries in the entire span of the sample. Industrial countries too have a higher share of debt, however, unlike developing countries they registered a marked increase in both equity and FDI asset holdings since 1987. The portfolio liabilities trend broadly echoes the portfolio asset holdings for almost all the country groups. However, as a break from the past, emerging economies, experienced an upward bulge in FDI liabilities during 1990s. A spate of capital account liberalizations appears to have altered the capital flows experiences of emerging market economies. As a group, they experienced a surge in FDI from the level of 7% of GDP to 17% of GDP from 1987 to 2003, and also registered a modest increase in equity liability position from 1% of GDP to 7% of GDP. The varied assets and liabilities position for industrial countries are widely known. The case in point is Ireland, which held amounts of foreign equity and debt far exceeding the level of Irish GDP. Among other industrial countries, the capital flow experiences are quite varied. For equity holdings in 2003, Japan held an amount of foreign equity equal to only 6% of GDP in 2003, where as Switzerland had the same ratio as high as 91%.13 Average debt asset holdings are four times larger than foreign equity holdings and average FDI assets are slightly larger than portfolio equity holdings. Liabilities outstanding in industrial countries are quite similar to asset holdings although debtor nations, such as Australia, hold fewer assets than liabilities and vice versa for creditor nations, such as Switzerland. Table 3 displays the ratio of foreign equity, debt and, FDI holdings to GDP for 1987 and 2003 for 21 emerging economies. Average equity and FDI assets are lower compared 11 Emerging markets are a subset of the developing countries in our sample. Emerging markets are also defined as the market of a developing country with high growth expectations and where the capital markets are at an early stage of development. The countries included in the list of emerging markets are Argentina, Brazil, Chile, China, Colombia, Egypt, India, Indonesia, Israel, Jordan, Korea, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, South Africa, Thailand, Turkey, and Venezuela. 12 Obstfeld and Taylor (2005) note that today’s global capital market is “North-North” (that is, flows between rich countries) with comparatively little movement of capital (net or gross) in the “North-South” direction (from rich to poor countries). 13 For a detailed layout of assets and liabilities position of industrial countries, see Sørensen, Yosha, Wu, and Zhu (2007).

7

to industrial countries. Emerging economies held only 4% of equity assets as a ratio of GDP compared to 31% held by industrial countries during 2003. Similarly, for the FDI assets, the corresponding averages are 40% and 6% for the industrial countries and the emerging economies respectively. Liabilities outstanding on average are also lower for emerging market economies when compared to corresponding numbers from industrial countries. The capital flow experiences among the emerging economies has remained varied. For instance, foreign equity holdings in Chile increased from 1% of GDP to 25% of GDP. While this partly may be due to a run-up in the value of foreign equity holdings, we observe the same pattern, although slightly less pronounced, for international holding of debt and FDI—debt holdings relative to GDP almost doubled and FDI holdings trebled in the span of 16 years from 1987 to 2003. There are large differences across countries. For the ratio of equity holdings to GDP, Chile and South Africa held 25% and 29% respectively, while China and Mexico held as low as 1%. The more noticeable trend among emerging market economies comes from their holdings of outstanding liabilities. For the emerging economies, the average liability holdings as a ratio of GDP far outweigh their average asset holdings. The total equity outstanding to GDP for the emerging economies on an average is 8% in 2003. The higher ratio for equity liabilities for emerging economies confirm their status as favorite destination for capital flows seeking to cash on equity market buoyancy. Israel, South Africa, and Malaysia are the leaders where as countries like China and India are still seeking firm grounds in spite of their strong equity markets. Other important development in last two decades for the emerging economies has been their propensity to offload debt liabilities. Chile reduced their debt outstanding as the ratio of GDP from a high level of 94% to 56%. Average debt liability has also decreased from 58% to 49% for the emerging economies as a group. Another important development has been the emergence of the emerging economies as the favorite FDI destinations. For example, China experienced a ten fold increase in FDI in last two decades and increased it from a modest level of 3% to 32%. Chile and Malaysia too proved to be attractive markets for FDI. On the other hand, India, Indonesia, and Turkey still have one digit share of FDI to GDP. The large variation across time and across countries delivers the variation that allows to test econometrically if countries with large amounts of foreign assets and liabilities obtain more risk sharing.

8

2.3

Theoretical background and previous literature

Theoretically, international financial integration is a celebrated concept—much like international trade in goods—with a variety of benefits for the countries that engage in them. A more open and competitive international market induces a more efficient international allocation of capital. From the view point of growth and development, the impact of capital flows from developed (capital-abundant) to less developed (capital-scarce) countries is important for the issues of convergence. However, Gourinchas and Jeanne (2006), using a calibration of simple neo-classical growth model, document that the countries have much more to gain from upgrading their domestic engines of growth and development than from attracting large quantities of capital. For trade in goods and services, most economists agree that benefits of unfettered trade far exceeds the cost. However, in the matter of capital flows, researchers reluctantly agree to its benefits from qualitative perspective, but it is hard to quantitatively show its importance. Starting in the mid 1980s, many countries undertook liberalization of capital controls that spurred a wave of financial globalization leading to rising cross-border financial flows among industrial countries and between industrial and developing countries. The new degree of international capital mobility since mid 1980s restored the trend of financial opening that eluded the world economy for almost a century.14 Nevertheless, capital transactions today are mostly concentrated in industrial countries and few emerging economies, and the capital flows to developing countries remain quite limited by the standards of pre-1914 economy (Obstfeld and Taylor, 2005). The new wave of financial globalization also led to burgeoning empirical literature on the spectacular growth in international financial integration and its implications for the macroeconomic behavior of open economies.15 The literature pursues various strands in explaining the astronomical increase in capital flows beginning 1970s. For example, Edwards (1991) shows that government size and openness are important determinants of inward FDI from OECD to developing countries during the period 1971-1981. Using data on bilateral portfolio flows from a set of 14 industrialized countries during 19891996, Portes and Rey (2005) find evidence that imperfections in the international credit markets can affect the amount and direction of capital flows. Among the set of developing countries, Lane (2004) also finds credit market frictions to be a determinant of debt flows 14 Obstfeld and Taylor (2005) compare the new wave of financial openness with the its pre world war one counterpart. They note that prior to world war one, a vibrant, free-wheeling capital market linked financial centers in Europe, the Western Hemisphere, Oceania, Africa, and the Far East. 15 see Prasad, Rogoff, Wei, and Kose (2003) for an extensive review of literature.

9

during 1970-1995. Calvo, Leiderman, and Reinhart (1996) focus on role of external (push) and internal (pull) factors as potential determinants of foreign investments using a crosssection of developing countries. They find that low interest rates in United states played an important role in accounting for the renewal of capital flows to these countries in 1990s. Importance of institutions have been widely studied for its role in driving various components of capital flows in the recent wave of financial globalization. In Alfaro, KalemliOzcan, and Volosovych (2005), the authors document that institutional quality is an important determinant of capital flows in the period 1970-2000. Wei (2000) and Wei and Wu (2002) use data on bilateral FDI from 18 industrialized source countries to 59 host countries during 1994-1996 and find that corruption reduces the volume of FDI and affects the composition of flows by increasing the loan-to-FDI ratio during the period. Faria and Mauro (2005) find that better institutional quality helps tilt country’s capital structure towards FDI and portfolio equity flows which tend to bring more collateral benefits of financial integration. Ju and Wei (2006) set to resolve the puzzle of low capital flows to developing countries on the basis of respective country’s strength of property rights. Keeping the inefficient financial system constant and using a model which features financial contracts and firm heterogeneity, the authors show that risk of appropriation determines the inflow of FDI. One component of capital flows in the current period of financial globalization that warrant mentioning is the enormity of debt flows. Tables 1, 2, and 3 display the importance of debt flows for all the country groups in the sample. Among the debt instruments, shortterm debt flows are generally considered less stable than its other counterparts like equity and FDI flows. In this context, prompt liberalization by governments in liberalizing debt inflows over equity market liberalization remains a puzzle.16 Henry (2007) notes that international financial system protects the rights of debt holders more vigilantly than those of equity holders. Additionally, a lack of transparency in equity markets of emerging economies makes foreigners reluctant to invest. Also, weak protection of property rights of equity investors reinforces the tendency of capital suppliers to purchase debt instead of equity (Henry and Lorentzen, 2003). 16 Henry

(2007) has an extensive review of literature on this issue.

10

3

International Risk Sharing

I start by characterizing the full risk sharing allocation, and its relation to consumption and income smoothing. I draw empirical implications of the theory, comparing various approaches in the risk sharing literature, and also lay out the alternative measures of risk sharing.

3.1

Theoretical background

A basic function of world capital market is to allow countries with imperfectly correlated income risks to trade them. The cross holding of assets and liabilities thereby reduces the global cross-sectional variability in income (and eventually consumption) levels. The situation where consumption growth rates in all countries are identical is denoted as “full (or perfect) consumption risk sharing”. This insures consumption from fluctuations in domestic output and makes consumption contingent on “world” output, meaning that there is no systematic uncertainty about “world” output, only idiosyncratic uncertainty about national output. The perfect risk sharing will be an equilibrium allocation if consumers have identical constant Relative Risk Aversion (CRRA) utility functions and access to a complete ArrowDebreu markets.17 Under a hypothetical complete-markets regime with free international asset trade, risk is pooled on anonymous financial markets with individuals in distant locations. The relative prices of Arrow-Debreu securities are common to all countries and, individuals trade so that to equate his or her marginal rate of substitution between consumption in different states of nature to a common relative-price ratio. Similarly, we say that there is “full (or perfect) income risk sharing” when the growth rate of GNI is identical in all countries. In this case, we would expect consumption growth rates to be similar, at least if taste shocks are not too large. If countries are fully integrated financially, they would hold the same portfolio of assets, their income–and therefore their consumption–would be closely correlated. Yet, this is far from being the case. The implications of theory are only partially corroborated when taken to the data. There is a substantial literature that looks into this anomaly, for example, Stockman and Tesar (1995) emphasize the role of non-traded goods and taste shocks as possible reasons for low cross-country consumption correlation as predicted in theory. The assumption of complete markets also appears far fetched in real world. 17 See

Obstfeld and Rogoff (1996) for a textbook treatment of the theory of international trade in assets.

11

Theoretically, it is safe to assume that GDP of a country is a homogeneous tradable goods, and the country sells rights to these goods and buys insurance against all future contingencies. Nevertheless, moral hazard issues thwarts an attempt to trade financial instruments linked to a broad measure of national output.

3.2

Empirical studies on international risk sharing

The empirical literature on risk sharing can be broadly categorized into two strands. The first strand emphasize that in a complete market set up, consumption growth rates across countries should be highly correlated. When risk is fully shared, the consumption of a country co-moves with world consumption, but does not co-move with idiosyncratic (country-level) shocks. However, this is far from being true, based on findings in literature, it is now a stylized fact that country consumption correlations are no higher than country output correlations. The phenomenon has come to be known as international consumption correlation puzzle or the “quantity puzzle”. The international business cycle literature, most notably Backus et al. (1992), Baxter and Crucini (1995), and Stockman and Tesar (1995) among others have taken this prediction to international macroeconomic data, finding that international consumption correlations are lot less than what theory predicts. In fact, these consumption correlations are not higher than country output correlations.18 Pakko (1998) and Canova and Ravn (1997) come to a similar conclusion, but both studies acknowledge the sensitivity of their results to the method of detrending. The second strand of empirical implications of risk sharing focuse on regression based measures. Studies in the regression based measure emphasize that under full risk sharing, the consumption of an economic agent does not respond to idiosyncratic shocks, in particular income shocks. The proposition has been tested on micro-data by, e.g., Cochrane (1991), Mace (1991), and Townsend (1994).19 These studies perform cross-sectional or panel regressions of individual consumption on sources of idiosyncratic risk (mainly on income but also on variables such as sickness or layoffs). In many of these studies full risk sharing is rejected (Cochrane, 1991; Townsend, 1994; Hayashi et al., 1996) 18 Kose et al. (2007) report the correlations for different country groups. They find that the correlation coefficient between country level output and world output level is higher than respective consumption correlation. 19 Other significant studies using micro data includes Attanasio and Davis (1996) and Hayashi et al. (1996). There is a substantial literature on intra-country risk sharing for developing countries as well. This literature considers the motives and methods of risk sharing among individuals localized in a geographical unit like village. Townsend (1994) looks at villages in India and documents that although complete risk sharing is rejected there is evidence of substantial consumption smoothing at village level. In a series of papers in micro literature, researchers have set out to explore the reasons for consumption smoothing, ranging from communal networks, ethnic ties to hoarding for adverse shocks. Morduch (1995) documents that in an environment of poor credit markets; households are more likely to choose lower mean, lower variance production methods.

12

Obstfeld (1994) tests for full risk sharing among nine large OECD countries by running, for each country, the time series regression ∆logcit on ∆logcw t . In the absence of world wide taste shocks, the coefficients in these regressions are estimated consistently, and if there is full risk sharing, they should be unity. The coefficients of may of Obstfeld’s regressions are positive but smaller than unity, which suggests that there is partial risk sharing but not full risk sharing. Even if the risk sharing is rejected it is important to quantify the extent to which risk is shared with a group of economic agents. It is also interesting to identify the exact channels through which risk is shared, and to quantify the amount of risk sharing obtained via each channel. Asdrubali, Sørensen, and Yosha (1996) develop a method for answering these questions. In the paper, they measure the fraction of idiosyncratic GDP shocks absorbed through various channels of cross-regional (or cross-national) insurance, and the fraction of shocks that is not smoothed. They distinguish between income insurance through cross-regional holding of debt and equity (“capital market income smoothing”) and via the super-regional tax-transfer system (“federal government income smoothing”).20 They document that 40% of shocks to gross state product are insured by capital markets, 13% by federal government, and 23% by capital markets. Sørensen and Yosha (1998) using the same methodology analyze the pattern of international risk sharing among OECD and sample of European countries. The study also looked at the contribution of crosscountry factor income flows and claim that it is not significantly different from zero.21 The evidence for consumption smoothing via savings are rather strong, about 40% of shocks to GDP are insured for OECD countries. Using similar methodology, recent studies have looked at the risk sharing experiences of countries in the wake of increased financial integration. The varied experience of number of countries with the capital account liberalization and their subsequent financial openness provides a fertile ground for the risk sharing studies. Given the fact that the capital flows in recent decade are more a result of “diversification finance”—which hedge against domestic fluctuations and has a risk sharing motive—studies quantifying risk sharing by various channels have become even more important. Sørensen, Yosha, Wu, and Zhu (2007) document that for a panel of OECD countries when home bias declines, risk sharing increases during the period 1993-2003. In the study, a large domestic stock of foreign 20 They also study, in a unified framework, income insurance (via markets and tax-transfer system) and consumption smoothing through saving and portfolio adjustment (e.g., borrowing and lending). M´ elitz and Zumer (1999) further extend their model. 21 This result is in line with the ‘home bias’ puzzle as documented by French and Poterba (1991) and Tesar and Werner (1995).

13

assets predicts higher risk sharing. Kose et al. (2007) in a similar study document that industrial countries unlike developing countries, share greater consumption risk with increased capital flows.

3.3

Measuring risk sharing

I consider two mechanisms for smoothing country-level output fluctuations. First, residents of a country can hold claims to output in other countries. If output across countries is imperfectly correlated, the out-of-country holding of assets will insure income. Second, a country’s residents can adjust their wealth portfolio in response to income fluctuations by buying and selling domestic assets and by borrowing and lending inside the country’s credit markets. The first mechanism—ex-ante income insurance—is effective for smoothing both permanent and transitory shocks. To illustrate: if in some year U.S. GDP is drastically reduced due to natural disaster, national income in U.S. will not fall by as much as output if many residents hold assets outside U.S. and receive interest and dividend income from out-of-country assets. This is true regardless of the persistence of the shock to U.S. GDP. The second mechanism—ex-post adjustment of wealth portfolios—can only smooth transitory shocks. This is a well understood implication of permanent income theory: facing an income shock, inhabitants of a region will adjust their stock of wealth in order to maintain their level of consumption only if the income shocks is perceived as transitory. In practice, macroeconomic shocks contain transitory and persistent components that are hard to identify empirically. Therefore, both income and consumption based risk sharing measures are considered for this study. Additionally, reporting consumption risk sharing is important because, ultimately, consumption is important for making welfare comparisons. Income and consumption smoothing are disjoint in other respects as well. Income smoothing is one mechanism by which fluctuations in consumption can be minimized. Alternative channels include international transfer schemes, international borrowing/lending etc. None of the alternative channels are fool-proof and devoid of limitations. Transfer schemes works best within the regions of a country, but its wide international application is ruled out. Limited enforcement of debt contract at the international level limits the use of borrowing/lending channel for consumption smoothing. Domestic savings could be run down to smooth consumption fluctuations, but it is a costly strategy as it requires altering investment. An alternative way out could be to accumulate a buffer stock of net 14

foreign equity which can be drawn up on in times of need. However, this is not a good option as it requires an extended period of postponed spending in order to accumulate the net foreign asset. International financial integration affects income and consumption risk sharing differently, therefore, it is important to know the mechanics of these measures. I measure the extent to which risk is shared within countries. The representative consumer of each country is a risk averse maximizer of life time expected utility from consumption. If utility functions are CRRA, and all countries have a common intertemporal discount factor, a perfect (Pareto efficient) risk sharing allocation satisfies xit = ki Xt for all and all realizations of uncertainty, where xit and Xt are generic variables representing aggregate income and consumption. The constant ki is independent of time and “state of the world”. If full risk sharing is achieved via income risk sharing, then xit and Xt represent both income and consumption (since, in this case, income equals consumption). If full risk sharing is achieved only after both income and consumption smoothing (in this case, income does not equal consumption), then xit and Xt represent consumption My empirical methodology closely follows that of Sørensen, Yosha, Wu, and Zhu (2007). The methodology quantifies deviations from perfect income and consumption risk sharing, respectively. Although I undertake a panel regression for this study, it is intuitive to look at the cross-section regressions and interpret the coefficients. Consider a group of countries and the following set of cross-sectional regressions: ∆ log GNIit − ∆ log GNIt = constant + βK,t (∆ log GDPit − ∆ log GDPt ) + it .

(6)

and GDPit are country i’s year t per capita GNI and GDP, respectively, and GNIt and GDPt are the year t per capita aggregate GNI and GDP, respectively. The coefficient βK,t measures the average co-movement of idiosyncratic GNI growth (i.e., the deviation from aggregate GNI growth) with idiosyncratic GDP growth in year t. Aggregate fluctuations cannot be eliminated by the sharing of risk, which is why the aggregate component is deducted from the growth rates. Under perfect risk sharing, the left-hand side of equation (6) will be zero implying that βK,t will be zero. The smaller the co-movement of idiosyncratic GNI with GDP, the more GNI is buffered against GDP fluctuations and the smaller the estimated value of βK,t . Since GNI equals GDP plus net factor income from abroad, this regression measures the amount of income risk sharing provided by net factor income flows—the lower βK,t , the higher is income risk sharing within the group in year t. The βK,t coefficients measure the evolution of risk sharing over time. Often it is more instructive to look at the equivalent series 1 − βK,t . This series will take the value 1 if risk GNIit

15

sharing is perfect and the value 0 if GNI moves one-to-one with output. In a similar manner, another set of regression for consumption risk sharing ∆ log Cit − ∆ log Ct = constant + βC,t (∆ log GDPit − ∆ log GDPt ) + it ,

(7)

where Cit is country i’s year t per capita final consumption, and Ct is the year t per capita aggregate final consumption for the group. The coefficient βC,t measures the average co-movement of the countries’ idiosyncratic consumption growth with their idiosyncratic GDP growth in year t. The smaller the co-movement, the more consumption is buffered against GDP fluctuations. Therefore, this regression provides a measure of the extent of consumption risk sharing.

4

Does greater financial integration predict better income and consumption risk sharing ?

4.1

Panel data regressions: specification

I estimate panel data regressions of the form: git GNI

git + it . = constant + κ GDP

(8)

git corresponds to ∆ log GNIit −∆ log GNIt ( similar for GDP git ). In the above regression, GNI ∆ log GNIit is the growth rate of real per capital GNI, ∆ log GDPit stands for the growth rate of real per capita GDP, and ∆ log GNIt and ∆ log GDPt represent their “world” counterparts over time t. This regression is similar to (6) except that it is now estimated as a panel pooling the years in the sample. In this specification, suggested by Asdrubali, Sørensen, and Yosha (1996), 1−κ is a scalar that measures the average amount of income risk sharing during the time-period considered. The coefficient κ measures the average co-movement of the countries’ idiosyncratic GNI-growth with their idiosyncratic GDP-growth over the sample period. In this regression, subtracting from each variable the aggregate value is crucial because aggregate GDP-growth of the group is not insurable. M´elitz and Zumer (1999) impose structure on κ such that κ = κ0 + κ1 γi , where γi is an “interaction” variable that affects the amount of risk sharing that country i obtains. 1 − κ0 − κ1 γi then measures the average amount of income risk sharing obtained by country i during the time-period in question. Sørensen et al. (2007) enhance this method

16

by allowing κ to change over time as follows: κ = κ0 + κ1 (t − t¯) + κ2 (IFIit −

IFIt ) ,

(9)

where IFIit ≡ International Financial Integrationit is the Financial Openness measure for country i at time t. t¯ is the middle year of the sample period, and IFIt is the (un-weighted) average across countries of IFIit at time t. The estimated value of 1 − κ0 corresponds to the average amount of income risk sharing within the group. 1 − κ0 − κ1 (t − t¯) − κ2 (IFIit − IFIt ) then measures the amount of income risk sharing obtained in period t by country i with International Financial Integration IFIit . I include a time trend to capture trend changes that may be caused by other developments in the international markets which may spuriously affect the measure.22 The parameter −κ1 captures the average year-by-year increase in income risk sharing. In this respect, the specification implied by (8) and (9) is a “middle-of-the-road” specification between the specification in (6)—where the amount of income risk sharing can change freely from period to period—and the specification in (8) where the amount of income risk sharing does not change over time. In the specification implied by (8) and (9), the amount of income risk sharing is allowed to change over time with the trend and with International Financial Integration. The parameter −κ2 (which will typically be negative) measures how much higher than average International Financial Integration increases the amount of income risk sharing obtained. In fact, −κ2 can be interpreted as an “exchange ratio” that translates fractions of International Financial Integration to percentage points of idiosyncratic shocks absorbed via income risk sharing. I perform an analogous analysis using total foreign assets and liabilities holdings as a ratio to GDP. In this case, the specification remains κ = κ0 + κ1 (t − t¯) + κ2 [IFIit − IFIt ], where IFIit measures International Financial Integration in terms of total foreign assets and liabilities holdings of country i at time t. If total asset portfolios are small relative to GDP the ratio of foreign holdings to GDP may be more relevant for macroeconomic income and consumption risk sharing. Further, we can consider liabilities in the same way as assets. If pay-outs from domestic liabilities are (roughly) proportional to output, 22 I experimented with using both time trend and time-fixed effects in the regression. The time trend is rarely significant and time fixed effect changes the results very little—this is because the aggregate value of the variables have been subtracted leaving little variation to be captured by time dummies. A quadratic term in time as an interaction term is also allowed but the quadratic term was rarely significant and the estimated coefficient to measures of financial integration was robust to this alternative.

17

as often assumed in theoretical models of international risk sharing, liabilities would be effective in smoothing output shocks. I also consider the ratio of components of foreign assets and liabilities to GDP in order to decipher the channels of risk sharing. Returns on equity and FDI liabilities may be more correlated with output and therefore, may provide more risk sharing than returns on debt liabilities. For equity assets I let κ = κ0 + κ1 (t − t¯) + κ2 (Eit −

Et ) ,

(10)

where Eit ≡ log[(foreign equity holdings)it /GDPit ] is the ratio of (gross) foreign equity holdings to GDP for country i in year t. I use a similar formulation for debt and FDI asset holdings and explore similar specifications using liabilities. I also allow risk sharing to increase proportionally with the total amount of foreign portfolio asset holdings (of equity plus debt holdings) relative to GDP. In this case, I let κ = κ0 + κ1 (t − t¯) + κ2 [EDit − EDt ] where EDit ≡ log[(foreign equity+ debt holdings)it /GDPit ] is the log-ratio of foreign debt+equity holdings to GDP for country i in year t. I include several interaction terms or explore, say, the sum of equity, debt, and FDI—as extensions which are simple permutations of the formulas already described above. I also estimate the contribution of International Financial Integration to consumption risk sharing using regressions of the form: f C it

git + it , = constant + η GDP

(11)

where Cfit corresponds to ∆ log Cit − ∆ log Ct , and ∆ log Cit is the growth rate of real per capita consumption, ∆ log Ct is its “world” counterpart, and η = η0 + η1 (t − t¯) + η2 (IFIit − IFIt ) . In the same manner as the analysis performed for income risk sharing, I allow for interaction terms based on the ratio of foreign assets holdings to GDP, foreign liabilities holdings to GDP, the ratio of foreign equity holdings to GDP, etc.

4.2

Panel data regressions: results

Table 6 displays result for income and consumption risk sharing for industrial countries for two sub samples—1973-2003 and globalization period (1987-2003). The independent variable in the regression are the various interactions of the broad measures of financial integration with idiosyncratic component of output. Column (1) presents the co-movement 18

of idiosyncratic component of output and GNI implying income risk sharing according to the regression specification. A higher number for the coefficient means that co-movement is nearly perfect and there is no income risk sharing. In column (2), the first row result means that there is an average risk sharing of 1% in the presence of interaction term. Although average income risk sharing is low, the interaction term has much more meaningful coefficient. In column (2), the point estimate for the interaction of financial openness with GDP is negative and significant implying that income risk sharing increases with increased financial openness. According to the regression specification, a negative coefficient for the interaction variable means that risk sharing within a country increases with an increase in financial integration. The point estimate is also important in economic terms: the coefficient for financial openness is –0.05%, which implies that a country increasing its financial openness by 0.1 will increase income risk sharing by about 1.5%. Decomposing financial openness into total assets and total liabilities, the coefficients are negative and significant for both at the 5% level of significance displayed in columns (3) and (4) respectively. For the full period sub-sample, the assets have a greater role in risk sharing compared to the liabilities. In the right panel of table 6, the average co-movement of consumption and output is lower compared to the income and output co-movement displayed in the first row implying that average consumption risk sharing is higher than average income risk sharing. Lower value for income risk sharing in the left panel confirms that net factor income from abroad doesn’t help much in delinking output and income. On the contrary, savings play a greater role in decoupling output and consumption levels within a country. Column (1) in the right panel confirms this assertion, a point estimate of 0.66 implies that 34% of consumption risk is being shared within the country via savings channel. Nevertheless, the financial integration has not impacted the consumption risk sharing the same way as income risk sharing. The point estimate for all the interaction terms in columns (2), (3), and (4) are not significant at the conventional levels of significance. However, the coefficients are still negative supporting the hypothesis that increased financial integration increases risk sharing, but the impact is small enough to be detected. In the lower panel of table 6, the results are displayed for the globalization period. Capital flows between industrial countries reached an astronomical limit in this period and the regression results corroborates its impact on risk sharing. The impact of financial integration is also more pronounced in this period—the point estimates for financial openness, assets, and liabilities are all higher compared to the full period sub-sample.

19

All the coefficients for the interaction terms are significant at a much stricter 1% level of significance. Consumption risk sharing is once again not impacted by the greater financial flows in the globalization period. Nonetheless, the coefficients for interaction terms are higher and have expected signs. Table 7 displays the results for emerging market economies for the two sub-samples. In the top panel of table 7, GNI co-moves more than one-to-one with the output level, implying that there is dis-smoothing of income.23 Average consumption risk sharing does better, but it falls short of the level of industrial countries. For the interaction terms as well, unlike industrial countries, the coefficients are positive denoting that increased financial integration led to lowering of risk sharing. In the left panel, in column (2), the coefficient for financial openness is positive and significant. Out of assets and liabilities, the liabilities prove to be significant implying that portfolio flows and FDI outstanding reduces income risk sharing for emerging economies. Later, I will analyze the break-up of assets and liabilities to better understand the role of different components in risk sharing. Table 7 also displays the results for globalization period using the same specification. The results are starkly different compared to first period, for instance average co-movement of output and GNI implying average income risk sharing in column (2) and liabilities in column (4) have significant and positive coefficient. The point estimates are also higher compared to the full period. In addition to that, for the co-movement of output and consumption as well, the point estimates are higher and significant denoting lowering of consumption risk sharing with increased financial openness. A positive and significant coefficient for liabilities imply that an increase in liability holdings decreased the consumption and income risk sharing for the emerging market economies. These findings suggest that the spurt of financial integration has not augured well for the emerging economies in terms of risk sharing advantages. Table 8 displays the results for developing countries using the same specification as the last two tables. The result are similar to that of emerging economies except for the co-movement of output and GNI in the globalization period. In the bottom panel of table 8, the output and the liabilities have negative coefficients denoting that income risk sharing increased with increased liability holdings. The results are contrary to the findings from emerging economies. Since the emerging economies make up for the bulk of countries in the developing countries sample, the result suggests that liability holdings 23 There is no priori restriction on the sign of the empirical estimation of the coeffecient. In the context of our regression specification, a negative risk sharing implies that idiosyncratic GNI responds more than one-to-one to GDP shocks—there is dis-smoothing of income relative to output.

20

help other developing countries share income risk better. However, later tables looking at the components of financial flows tell a different story. For the co-movement of consumption and output displayed in column (4) of the bottom panel, the coefficient for the liabilities term are positive and significant implying decrease in consumption risk sharing. Therefore, the results are ambiguous for the developing countries as income risk sharing and consumption move in opposite direction. Next, we turn to more comprehensive set of regressions using different components of financial flows. I resort to alternative measures of financial integration and look at total equity, debt, and FDI (assets and liabilities combined for all three). In table 9, I use total equity, debt, FDI, and sum of equity and debt as the interaction variables for the next set of regression. The purpose of combining equity and debt is to focus on aggregate portfolio flows and to distinguish it from the foreign direct flows. The first row in the top panel of table 9 displays the co-movement of output and GNI, the coefficient imply that average risk sharing is not more than 1% in any of the specification for the industrial countries. In columns (1) and (3), the equity and FDI interaction terms have negative and significant coefficients. This means that an increase in total equity and FDI in the portfolio helps a country better share its income risk. In column (4), the aggregate of equity and debt holdings have a larger point estimate meaning that aggregate portfolio flows play an important role in income risk sharing. When FDI and portfolio flows are regressors in a multivariate regression, the portfolio flows loose its significance and FDI comes out as a significant variable in sharing income risk. In the bottom panel of table 8, the results get reversed and only debt holdings remain significant in sharing consumption risk. In column (2) in the bottom panel, the interaction term with total debt has a negative and significant coefficient, and in column (5), portfolio flows are significant even in the presence of FDI denoting the importance of debt flows in sharing consumption risk. Table 10 displays the results from same specification as the previous table for the globalization period. The average income risk sharing is higher than the full period and the components of financial flows have higher point estimates too. All the three components, equity, debt, and FDI have significant coefficients for their corresponding interaction terms. Column (5) is noteworthy, wherein portfolio flows retain its significance and FDI becomes insignificant in the regression. These results confirm the important role played by all the components of financial flows in sharing income risk. Notably, portfolio flows seem to be more important for risk sharing in the globalization period.

21

Tables 11 and 12 display the results for emerging economies with portfolio flows and FDI for the full period and globalization period respectively. In both the tables, debt flows have a significant coefficient with a positive sign. For the emerging economies, it implies that increase in total debt flows led to decrease in income risk sharing for both the sub-samples. The equity and FDI flows have no role in either income or consumption risk sharing in any of the periods. In the bottom panel of both the tables, consumption risk sharing increases with increase in debt flows during the globalization period. Debt flows in emerging economies appears to have reversed risk sharing gains in both income and consumption risk sharing during globalization period. The results for other developing countries are displayed in tables 13 and 14 for the full period and globalization period respectively. In the bottom panels of both the tables, consumption risk sharing trends are similar to that of the emerging economies. The debt flows once again have positive and significant coefficients meaning that debt flows negatively affects consumption risk sharing.24 During the globalization period, debt holdings seem to have a positive effect on income risk sharing. Interestingly, in the multivariate regression displayed in column (5), the coefficient for portfolio holdings are positive and significant, a reversal of sign from column (4). This result implies that when all the variables are considered together, portfolio holdings does seem to decrease income risk sharing. More of this result will be probed in later tables where I will pinpoint the impact of assets and liabilities separately for both equity and debt flows. The next set of tables explored the channels of risk sharing in further detail where I look at the components of assets and liabilities separately. Table 15 considers the categories of asset holdings for the industrial countries for the full period. Among the class of assets, equity and debt have contributed positively to income risk sharing. In columns (1) and (2), the coefficients are negative and significant at 5% level of significance. These results are more detailed in the sense that it is now easy to pinpoint that portfolio assets perform better in income risk sharing compared to debt assets. In column (4), for both the sub-samples, when portfolio flows are combined together, the result still holds. For the consumption risk sharing, only debt holdings play an important role in risk sharing and equity and FDI turn out to be insignificant. However, when combined together in column (4), the interaction term has usual negative sign implying higher risk sharing. In table 16, outstanding liabilities too have positive role in income risk sharing in both the sub-samples. Unlike FDI assets, FDI liabilities now turn out to be negative and significant. 24 Because the difference in income and consumption is savings it appears that developing countries and especially emerging economies with high outstanding debt had relatively less pro-cyclical savings during our sample period.

22

For the consumption risk sharing only debt liabilities are significant and equity and FDI do not play any role in risk sharing.25 Table 17 displays the results for asset categories in the globalization period for both emerging markets and other developing economies. In the top panel of table 17, for the emerging economies, portfolio assets reversed some income risk sharing gains. In columns (1) and (2) in the top panel, equity assets and debt assets have positive and significant point estimates implying lower income risk sharing. A significantly positive coefficient for debt assets means that an increase in debt holding did not help emerging economies share consumption risk. For other developing countries, the results broadly echo the pattern of emerging economies. In the lower panel, the interaction term with equity assets have positive and significant point estimates for both income and consumption risk sharing. Table 18 displays the result for liabilities outstanding for emerging markets and developing economies in the globalization period. For the emerging economies, only the interaction term for debt liability has a positive and significant coefficient implying lower income risk sharing in response to greater debt holdings. For other developing countries, once again debt liability is significant but only for consumption risk sharing. Interestingly, for other developing countries, in column (4), the interaction term with portfolio equity and debt liabilities combined together have negative point estimate significant at 10% level of significance. Although none of the equity and debt liabilities are significant on their own, but when combined together the interaction term comes out significantly negative. This effect is also translated to previous results where we initially observed that total liabilities and total debt interaction terms have negative coefficient reflecting increase in debt and liability holdings increasing risk sharing.

5

Discussion

Following the risk diversification argument, the effects on the domestic economy of various shocks could be mitigated through international transactions. Accordingly, growth variations are expected to be smaller in consumption and GNI than in GDP. Nevertheless, the precise mechanics of risk sharing will depend on the composition of international balance sheet as between debt and equity-type investments and also on the mix between assets and liabilities that are denominated in domestic currency versus foreign currency. 25 It is speculated that FDI liabilities might smooth consumption—without affecting international factor income flows—if owners of international corporations smooth wages across country-borders. Evidence of such behavior is presented in Budd and Slaughter (2000).

23

Among the components of capital flows, FDI and equity holdings are more stable forms of capital flows compared to debt instruments. Asset holdings provide great benefits in term of risk sharing—large payout for shareholders when times are good and little to nothing when times are bad. Investing in overseas assets that have a high payoff when the domestic economy is doing badly insulates domestic income from fluctuations in domestic output. On the other hand, debt instruments do not have the desirable risk sharing properties of FDI and equity, and it also makes countries susceptible to reversals of sentiment and insolvency problems. Industrial countries have derived risk sharing benefits from all the components of capital flows and more so in the globalized period. Equity assets and liabilities have an unambiguous effect on increase in income risk sharing and debt liabilities seems to matter more for consumption risk sharing. Although FDI has the greater potential for fostering risk sharing, its impact is not uniform through out the period. Debt flows are considered the most unstable form of capital flows, but its significance in the regression suggests that industrial countries used the debt instruments wisely that fostered increased risk sharing. The result strengthens the argument that not all debt instruments are bad, for instance, trade credits—lifeblood of international trade—are types of short-run flows but augers well for countries and help them in risk sharing. For the developing countries and emerging economies, there is no unambiguous impact of capital flows on risk sharing. Nevertheless, one impact that stands out is that of portfolio debt. Unlike industrial countries, increase in portfolio debt led to a decrease in risk sharing for developing economies. Servicing an equity contract involves procyclical payments, whereas debt service payments are countercyclical. A debt contract requires regular payments, regardless of borrower’s circumstances. Moreover, in case of a downturn, the countercyclical debt payment will impact the normal consumption and investment activity in the economy. A fall in the level of output will lead to a fall in consumption leading to a lower risk sharing measure as estimated by the regression specification. For the emerging economies, the debt position is exacerbated by both government and private investors. Governments sometimes borrow beyond their means, and they are at times willing to take on excessive risk to save on interest costs. Private investors are reluctant to hold instruments that would provide far more flexibility such as GDP-indexed bonds, domestic equity, and local currency debts—in part, because of poor policy credibility and weak domestic institutions. The result is an excessive reliance on “dangerous” forms of debt, such as foreign currency denominated debt, short-term debt etc. Bai and

24

Zhang (2007) also postulate that greater frequency of sovereign defaults leads to less than warranted capital flows and in turn lowers international risk sharing. Another result that merits attention is the low level of consumption risk sharing for all group of countries. Cross country flows of capital is perceived to directly affect the income risk sharing and consumption smoothing is only indirectly affected. However, the link between capital flows and consumption is not as indirect. In a closed economy, aggregate consumption smoothing is difficult to undertake. If many individuals or corporations within a closed country attempt to increase saving in a particular year, bond prices will rise (i.e. the real interest rate will fall) reducing the incentive to save. By contrast, if international credit markets are integrated internationally, savings will be channeled to countries where the supply of funds has fallen. Theoretically, the link between cross holding of assets and consumption risk sharing is straight forward, however, the empirical literature on consumption risk sharing has never been unambiguous. Asdrubali, Sørensen and Yosha (1996) find that for US states, consumption risk sharing varies mysteriously where income risk sharing has a smooth pattern and increases from decade to decade. In another study, Sørensen and Yosha document for European Union that consumption risk sharing peaked in 70s, fell sharply in 80s, and is slowly increasing now. The erratic nature of consumption risk sharing overtime fails to bring out a unified pattern in our sub-samples. In addition to that, allowing free movement of capital flows often result in temporary surge in output and consumption which is incorrectly measured as no risk sharing.

6

Policy implications and final remarks

The study finds that for industrial countries, all the components of capital flows are positively and robustly related to income risk sharing. Among the components of capital flows, the impact of equity and debt flows is much more discernible than FDI flows. For the consumption risk sharing, the effects are hard to comprehend. Emerging markets, which accounted for a huge share of capital flows among developing countries, did not translate that into risk sharing advantages. On the contrary, debt instruments seem to have reversed some potential risk sharing gains. In both sub-samples, debt flows negatively affect income and consumption risk sharing. For the developing countries, the effect of capital flows is anything but ambiguous. As the results in this paper suggest, capital flows are good for risk sharing at least for

25

the industrial countries. Risk sharing has also increased for these countries in the wake of increased financial integration. But there is no clear-cut measurable effect on developing countries and especially emerging economies. One explanation for the low risk sharing in developing countries is that these countries are more concerned with “catching-up” and attaining a steady state level of growth. Hedging income from fluctuations in domestic output and attaining risk sharing may not be falling in their current scheme of things and could be an advance issue that concerns these countries once they have achieved a desirable level of output and growth. As per the policy implications, the capital flow needs to be larger to have a more robust impact on risk sharing. Although FDI is presumed to be most stable among the components of capital flows, the results do not conclusively show that. The FDI flows will have to dominate the capital flows to have a significant impact on risk sharing for all the country groups. For the emerging economies, the focus needs to shift from “short-equity, long-debt” to more stable forms of capital flows. Domestic policies that encourage shortterm borrowing—either through tax incentives or low reserve requirements, or indirectly through pegged exchange rates and sterilization of inflows—should be avoided. In this regard, the developing countries need to strengthen prudential control on cross-border borrowing. Nevertheless, the lack of well-defined property rights and institutional weaknesses thwart these country’s attempt to shift to more stable forms of capital flows. A policy to create a credible institution that safeguards the interest of private capital flows in the form of more equity and FDI will go a long way in attaining a desirable risk sharing for the developing countries.

26

Data Sources: Foreign equity, debt, and foreign direct investment (FDI) assets and liabilities are from Lane and Milessi-Ferretti (2006). GDP, GNI, Consumption, and Population are from the World Bank (2004) World Development Indicators.26 The specific codes for the variables are: GDP: GNI: Population: Private Consumption:

(current LCU); series code NY.GDP.MKTP.CN (current LCU); series code NY.GNP.MKTP.CN (persons); series code SP.POP.TOTL (current LCU); series code NE.CON.PRVT.CN

Nominal exchange rates and consumer price index are from the IMF (2005) International Financial Statistics CD-ROM. The specific code for the variable are: Consumer Price Index: (2000 = 100); series code 64...ZF Nominal Exchange Rates: (units of LCU per US$); series code AE.ZF I compute Purchasing Power Parity (PPP) adjusted GDP, GNI, and Consumption by deflating all series with the Consumer Price Index normalized to 1 in 1995. I do not use quantity indices for real GDP because we want to measure how the purchasing power of GDP gets insured internationally.27 I translate to PPP-adjusted U.S. dollars values using 1995 exchange rates. Growth rates of GDP, GNI, and Consumption are the growth rates of per capita PPP-adjusted variables.

26 WDI data set (except for developed countries) is further processed to get Penn World Table (PWT) data set. The main focus of PWT data set is to create cross-sectional comparability in national accounts data. By computing PPP adjusted variables normalized to a single year, I seek to undertake a similar transformation as proposed by PWT data set. 27 In the context of risk sharing, benchmark is the autarky equilibrium when countries consume the volume of their GDP (=GNI) (see Kalemli-Ozcan, Sørensen, and Yosha, 2001). A natural measure of output is then GDP deflated by CPI, not by GDP deflator (Sørensen and Yosha, 2007). This translates GDP and GNI per capita to amount of individual consumption that they can buy.

27

Appendix: List of Countries The sample comprises 69 countries—21 industrial and 48 developing.28 Industrial Countries: Morgan Stanley Capital International Inc. (MSCI) maintains an index of industrialized countries based on float-adjusted market capitalization index. The group consists of 23 industrial countries.29 Australia (AUS), Austria (AUT), Belgium (BEL), Canada (CAN), Denmark (DNK), Finland (FIN), France (FRA), Germany (DEU), Greece (GRC), Ireland (IRL), Italy (ITA), Japan (JPN), Netherlands (NLD), New Zealand (NZL), Norway (NOR), Portugal (PRT), Spain (ESP), Sweden (SWE), Switzerland (CHE), United Kingdom (GBR), and United States (USA). Developing Countries: Developing countries sample consist of emerging economies and other developing countries. Emerging economies: For the emerging economies, I rely on MSCI’s list of “more financially integrated” countries.30 Argentina (ARG), Brazil (BRA), Chile (CHL), China (CHN), Colombia (COL), Egypt (EGY), Indonesia (IDN), India (IND), Israel (ISR), Jordan (JOR), Korea (KOR), Malaysia (MYS), Mexico (MEX), Morocco (MAR), Pakistan (PAK), Peru (PER), Philippines (PHL), South Africa (ZAF), Thailand (THA), Turkey( TUR), and Venezuela (VEN). Other Developing Countries: Algeria (DZA), Bolivia (BOL), Cameroon (CMR), Costa Rica (CRI), Cote d’ Ivore (CIV), Dominican Republic (DOM), Ecuador (ECU), El Salvador (SLV), Fiji (FJI), Gabon (GAB), Ghana (GHA), Guatemala (GTM), Haiti (HTI), Honduras (HND), Iran (IRN), Jamaica (JAM), Mauritius (MUS), Nicaragua (NIC), Papua New Guinea (PNG), Paraguay (PRY), Senegal (SEN), Sri Lanka (LKA), Togo (TGO), Trinidad and Tobago (TTO), Tunisia (TUN), Uruguay (URY), and Zimbabwe (ZWE). 28 I excluded from the analysis small countries (those with population below 1 million), major oil producers, and other countries with incomplete or clearly unreliable data. 29 Hongkong, and Singapore were excluded from analysis because these countries are financial hubs and are therefore outliers in terms of standard measures of de facto financial integration. 30 Czech Republic, Hungary, Poland and Russia were excluded since they belong to transition economies. Taiwan was excluded due to incomplete data.

28

References Alfaro, Laura, Sebnem Kalemli-Ozcan, and Vadym Volosovych, 2005, “Capital Flows in a Globalized World: The Role of Policies and Institutions,” NBER Working Paper No. 11696, National Bureau of Economic Research, Inc.. Artis, Micheal J., and Mathias Hoffman, 2006, “The Home Bias and Capital Income Flows Between Countries and Regions,” CEPR Discussion Paper No. 5691. Artis, Micheal J., and Mathias Hoffman, 2007, “Declining Home bias and the increase in international risk sharing: Lessons from European Integraion,” Working Paper, University of Zurich. Asdrubali, P., Bent E. Sørensen, and Oved Yosha, 1996, “Channels of Interstate Risk Sharing: United States 1963–90,” Quarterly Journal of Economics 111 (4), 1081–1110. Attanasio O., and S. J. Davis, 1996, “Relative Wage Movements and Distribution of Consumption,” Journal of Political Economy 104 (6), 1227–62. Backus, D., P. Kehoe, and F. Kydland, 1992, “International Real Business Cycles,” Journal of Political Economy 100 (4), 745–775. Bai, Yan, and Jing Zhang, 2007, “Financial Integration and International Risk Sharing,” Working Paper, University of Michigan. Baxter, M., and M. Crucini, 1995, “Business Cycles and the Asset Structure of Foreign Trade,” International Economic Review 36 (4), 821–854. Budd, J. W., and M. J. Slaughter, 2000, “Are profits shared across borders? Evidence on international rent sharing,” NBER Working Paper No. 8014, National Bureau of Economic Research, Inc. Calvo, Guillermo, Leonardo, Leiderman, and Carmen Reinhart, 1996, “Inflows of capital to Developing countries in 1990s,” The Journal of Economic Perspective 10, 123–139 Canova, Fabio, and Morten Ravn, 1996, “International consumption risk sharing,” International Economic Review 37 (3), 573–601. Cochrane, J. H., 1991, “A Simple Test of Consumption Insurance,” Journal of Political Economy 99, 957–976. 29

Edwards, Sebastian, 1991, “Capital Flows, Foreign Direct Investment, and Debt-Equity Swaps in Developing Countries” in Capital Flows in the World Economy, editors: Horst Siebert, J. C. B. Mohr, Tubingen. Faria, Andre, and Paolo Mauro, 2005, “Institutions and the External Capital Structure of Countries,” IMF Working Paper 04/236, International Monetary Fund. French, K. R., and J. M. Poterba, 1991, “Investor diversification and international equity markets,” American Economic Review: Papers and Proceedings 81 (2), 222–226. Gourinchas, Pierre-Olivier, and Olivier Jeanne, 2006, “The Elusive Gains from International Financial Integration,” Review of Economic Studies 73 (3), 715–741. Hayashi, F., J. Altonji, and L. Kotlikoff, 1996, “Risk Sharing Between and Within Families,” Econometrica 64, 261–294. Henry, Peter Blair, 2006, “Capital Account Liberalization: Theory, Evidence, and Speculation,” NBER Working Paper No. 12698, National Bureau of Economic Research, Inc. Henry, Peter Blair, and Peter Lombard Lorentzen, 2003, “Domestic Capital Market Reform and Access to Global Finance: Making Markets Work,”NBER Working Paper No. 10064, National Bureau of Economic Research, Inc. International Monetary Fund, 2005, “International Financial Statistics CD-ROM,” International Monetary Fund, Washington, DC. Ju, Jiandong, and Shang-Jin Wei, 2006, “A Solution to Two Paradoxes of International Capital Flows,” NBER Working Paper No. 12668, National Bureau of Economic Research, Inc. Kalemli-Ozcan, Sebnem, Bent E. Sørensen, and Oved Yosha, 2001. “Economic integration, industrial specialization and the asymmetry of macroeconomic fluctuations,” Journal of International Economics 55 (1), 107–137. Kose, Ayhan M., Eswar S. Prasad, Kenneth, Rogoff, and Shang-Jin Wei, 2006, “Financial Globalization: A Reappraisal,” NBER Working Paper 12484, National Bureau of Economic Research, Inc.

30

Kose, Ayhan M., Eshwar S. Prasad, and Marco E. Torrens, 2007, “How Does Financial Globalization Affect Risk Sharing? Pattern and Channels,” IZA Discussion Paper. Lane, Philip, 2000, “International Diversification and the Irish Economy,” Economic and Social Review 31 (1), 37–54. Lane, Philip, 2001, “Do international investment income flows smooth income?,” Weltwirtschaftliches Archiv 137 (4), 714–736. Lane, Philip, 2004, “Empirical Perspectives on Long-Term External Debt,” Topics in Macroeconomics 4, 1–21. Lane, Philip, and Milesi-Ferretti, G. M., 2006, “The external wealth of nations mark II: revised and extended estimates of foreign assets and liabilities, 1970–2004,” IMF Working Paper 06/69, International Monetary Fund. Lewis, K., 1996, “What can explain the apparent lack of international consumption risk sharing?,” Journal of Political Economy 104 (2), 267–297. Lewis, K., 1999, “Trying to explain home bias in equities and consumption,” Journal of Economic Literature 37 (2), 571–608. Mace, Barbara J., 1991, “Full insurance in the presence of aggregate uncertainty,” Journal of Political Economy 99 (5), 928–956. M´elitz, J., and F. Zumer, 1999, “Interregional and international risk sharing and lessons for EMU,” Carnegie-Rochester Conference Series on Public Policy 51, 149–188. Morduch, Jonathan, 1995, “Income Smoothing and Consumption Smoothing,” Journal of Economic Perspectives 9 (3), 103–114. Obstfeld, Maurice, 1994, “Are industrial-country consumption risks globally diversified?,” in Capital Mobility: The Impact on Consumption, Investment and Growth, editors: Leiderman, L. and A. Razin, Cambridge University Press, New York, NY, pp. 13– 44. Obstfeld, Maurice, and Kenneth Rogoff, 1996, “Foundations of International Macroeconomics,” MIT Press, Cambridge, MA.

31

Obstfeld, Maurice, and Alan M. Taylor, 2004, “Global Capital Markets: Integration, Crisis and Growth,” Cambridge University Press, New York. Pakko, Michael, 1998, “Characterizing Cross-Country Consumption Correlations,” Review of Economics and Statistics 80 (1), 169–174. Portes, Richard and Helene Rey, 2005, “The Determinants of Cross-Border Equity Flows,” Journal of International Economics 65, 269-296. Prasad, Eswar S., Kenneth, Rogoff, Shang-Jin Wei, and M. Ayhan Kose, 2003, “Effects of Financial Globalization on Developing Countries: Some Empirical Evidence,” IMF Occasional Paper No. 220, International Monetary Fund. Sørensen, Bent E., and Oved Yosha, 1998, “International risk sharing and European monetary unification,” Journal of International Economics 45 (2), 211–238. Sørensen, Bent E., and Oved Yosha, 2007, “Producer Prices versus Consumer Prices in the Measurement of Risk Sharing,” Applied Economics Quarterly, forthcoming. Sørensen, Bent E., Oved Yosha, Yi-Tsung Wu, and Yu Zhu, 2007, “Home Bias and International Risk Sharing. Twin Puzzles Separated at Birth,” Journal of International Money and Finance, forthcoming. Stockman, A. C., and Linda L. Tesar, 1995, “Tastes and technology in a two-country model of the business cycle: explaining international comovements,” American Economic Review 85 (1), 168–185. Tesar, Linda L., and Ingrid M. Werner, 1995, “U.S. Equity Investment in Emerging Stock Markets,” World Bank Economic Review 9 (1), 109–129. Townsend, R., 1994, “Risk and Insurance in Village India,” Econometrica 62, 539–591. Wei, Shang-Jin, and Andrei Shleifer, 2000, “Local Corruption and Global Capital Flows,” Brookings Papers on Economic Activity 2, 303–354 . Wei, Shang-Jin, and Yi Wu, 2002, “Negative Alchemy? Corruption, Composition of Capital Flows, and Currency Crises,” in Preventing Currency Crises in Emerging Markets, editors: S. Edwards and J. Frenkel, The University of Chicago Press, Chicago. World Bank, 2004, “World Development Indicators CD-ROM,” The World Bank, Washington, DC. 32

Table 1 Country Group-level Foreign Assets, Liabilities, and Financial Openness (mean for each group of countries)

Year: Country Group: Industrial Emerging Developing All countries

Assets 1973 1987 2003

Liabilities 1973 1987 2003

Financial Openness 1973 1987 2003

0.34 0.16 0.13 0.20

0.34 0.33 0.38 0.37

0.68 0.49 0.52 0.57

0.68 0.20 0.20 0.35

2.26 0.53 0.45 1.00

0.81 0.67 0.83 0.82

2.36 0.84 0.90 1.34

1.49 0.87 1.03 1.17

4.62 1.37 1.35 2.34

Note. The term “Assets” and “Liabilities” refer to value of gross foreign assets and liabilities divided by GDP respectively. Financial openness is measured as the ratio of aggregate of gross foreign assets and liabilities to GDP. Total assets is the aggregation of portfolio equity assets, Foreign direct investment (FDI) assets, and debt assets. Total liabilities is the aggregation of portfolio equity liability, FDI liability, and debt liabilities. Portfolio equity holdings measure ownership of shares of companies and mutual funds that are below 10% threshold—the statistical convention of distinguishing between portfolio and direct investment. The FDI category includes controlling stakes in acquired firms in addition to greenfield investments. The debt category includes portfolio debt securities, plus bank loans and deposits and other debt instruments. Lane and Milessi-Ferretti (2007) employ a broadly uniform methodology to construct estimates of foreign assets and liabilities position. While their benchmark series is based on the official estimates from the international investment positions, they compute the stock positions of earlier years on using data on capital flows and account for capital gains and losses.

33

Table 2 Country Group-level Foreign Assets and Liabilities Holding of Equity, Debt, and Foreign Direct Investment Relative to GDP (mean for each group of countries) A. Assets Year: Country Group: Industrial Emerging Developing All countries

1973 0.02 0.00 0.00 0.00

Equity 1987 2003 0.05 0.00 0.00 0.02

0.34 0.04 0.02 0.14

1973

Debt 1987

2003

1973

FDI 1987

2003

0.21 0.07 0.06 0.11

0.47 0.12 0.12 0.23

1.36 0.21 0.21 0.56

0.05 0.00 0.00 0.02

0.10 0.02 0.02 0.04

0.46 0.06 0.05 0.17

B. Liabilities Equity Year: Country Group: Industrial Emerging Developing All countries

Debt

FDI

1973

1987

2003

1973

1987

2003

1973

1987

2003

0.03 0.00 0.00 0.00

0.05 0.01 0.00 0.02

0.42 0.08 0.05 0.16

0.26 0.25 0.24 0.24

0.64 0.58 0.70 0.68

1.45 0.49 0.54 0.82

0.06 0.07 0.14 0.11

0.12 0.07 0.13 0.12

0.45 0.27 0.30 0.35

Note. The rows display the country group-level value of foreign equity, debt, and foreign direct investment holdings divided by GDP in the same year. Portfolio equity holdings measure ownership of shares of companies and mutual funds that are below 10% threshold—the statistical convention of distinguishing between portfolio and direct investment. The FDI category includes controlling stakes in acquired firms in addition to greenfield investments. The debt category includes portfolio debt securities, plus bank loans and deposits and other debt instruments. Lane and Milessi-Ferretti (2007) employ a broadly uniform methodology to construct estimates of foreign assets and liabilities position. While their benchmark series is based on the official estimates from the international investment positions, they compute the stock positions of earlier years on using data on capital flows and account for capital gains and losses.

34

Table 3 Emerging economies-level Foreign Assets and Liabilities Holding of Equity, Debt, and Foreign Direct Investment Relative to GDP

Equity

Assets Debt

FDI

Equity

Liabilities Debt

FDI

Year: Country: Argentina Brazil Chile China Colombia Egypt India Indonesia Israel Jordan Korea Malaysia Mexico Morocco Pakistan Peru Philippin. South Afr. Thailand Turkey Venezuela

1987

2003

1987

2003

1987

2003 1987

2003

1987

2003

1987

2003

0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.03 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.01

0.09 0.01 0.25 0.00 0.01 0.01 0.00 0.00 0.07 0.01 0.01 0.02 0.01 0.00 0.00 0.06 0.02 0.29 0.00 0.01 0.02

0.09 0.05 0.20 0.05 0.11 0.05 0.00 0.08 0.28 0.36 0.07 0.12 0.17 0.06 0.03 0.08 0.13 0.04 0.03 0.07 0.38

0.63 0.07 0.23 0.18 0.18 0.29 0.02 0.08 0.42 0.30 0.13 0.30 0.06 0.16 0.07 0.05 0.17 0.16 0.12 0.13 0.67

0.00 0.04 0.00 0.00 0.00 0.00 0.00 0.00 0.03 0.00 0.02 0.07 0.00 0.02 0.00 0.00 0.00 0.16 0.00 0.00 0.01

0.11 0.10 0.19 0.02 0.05 0.01 0.01 0.01 0.15 0.00 0.07 0.21 0.02 0.02 0.01 0.01 0.01 0.17 0.03 0.01 0.11

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.06 0.00 0.00 0.00 0.00 0.01 0.04 0.03 0.01 0.00

0.01 0.12 0.05 0.03 0.01 0.02 0.08 0.08 0.21 0.01 0.18 0.19 0.08 0.01 0.02 0.05 0.05 0.20 0.17 0.04 0.04

0.50 0.40 0.94 0.11 0.39 0.58 0.20 0.66 0.86 0.96 0.35 0.72 0.74 1.12 0.44 0.59 0.90 0.30 0.40 0.47 0.72

1.15 0.46 0.56 0.14 0.48 0.39 0.20 0.65 0.62 0.84 0.26 0.47 0.25 0.43 0.49 0.50 0.80 0.20 0.36 0.66 0.42

0.05 0.07 0.21 0.03 0.09 0.10 0.01 0.03 0.04 0.08 0.02 0.30 0.08 0.04 0.02 0.03 0.04 0.12 0.06 0.01 0.04

0.34 0.26 0.76 0.32 0.26 0.26 0.07 0.04 0.27 0.27 0.08 0.48 0.26 0.39 0.10 0.33 0.19 0.28 0.29 0.09 0.40

Average

0.00

0.04

0.12

0.21

0.02

0.06

0.01

0.08

0.58

0.49

0.07

0.27

Note. The rows display the value of foreign equity, debt, and foreign direct investment holdings divided by GDP in the same year for the emerging economies. Total assets is the aggregation of portfolio equity assets, Foreign direct investment (FDI) assets, and debt assets. Total liabilities is the aggregation of portfolio equity liability, FDI liability, and debt liabilities. Portfolio equity holdings measure ownership of shares of companies and mutual funds that are below 10% threshold—the statistical convention of distinguishing between portfolio and direct investment. The FDI category includes controlling stakes in acquired firms in addition to greenfield investments. The debt category includes portfolio debt securities, plus bank loans and deposits and other debt instruments. Lane and Milessi-Ferretti (2007) employ a broadly uniform methodology to construct estimates of foreign assets and liabilities position. While their benchmark series is based on the official estimates from the international investment positions, they compute the stock positions of earlier years on using data on capital flows and account for capital gains and losses.

35

Table 4 Correlation Matrix of GDP, Consumption, GNI Growth Rates and Foreign Asset, Liability Ratios Interacted with GDP Growth: Industrial countries 1987–2003 GDP growth

GNI growth

Cons growth

Equity

GDP growth 1.00

0.96

0.84

GNI growth

1.00

Cons growth Equity asset

Assets Debt

FDI

–0.24

–0.40

–0.05

–0.06

–0.12

0.05

0.83

–0.28

–0.42

–0.06

–0.11

–0.15

0.01

1.00

–0.19

–0.41

–0.03

–0.05

–0.18

0.06

1.00

0.43

0.35

0.67

0.37

0.54

1.00

0.16

0.41

0.76

–0.03

1.00

0.43

0.28

0.29

1.00

0.42

0.47

1.00

0.33

Debt asset FDI asset Equity liability Debt liability FDI liability

Liabilities Equity Debt FDI

1.00

Note. The term “GDP growth” represents the data series (∆ log GDPit − ∆ log GDPt ), where GDPit is country i’s year t per capita GDP, and GDPt is the year t per capita aggregate GDP for the group. The series “GNI growth” and “Cons growth” are defined similarly. The term “Equity asset” refers to the data series [(Eit − Et ) ∗ (∆ log GDPit − ∆ log GDPt )], where Eit is the period t natural logarithm of the ratio of foreign equity owned to GDP for country i, and Et is the (un-weighted) average across countries of Eit . “Debt asset,” “FDI asset,” and liabilities are defined similarly. The countries included in the sample are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, and United States.

36

Table 5 Correlation Matrix of GDP, Consumption, GNI Growth Rates and Foreign Asset, Liability Ratios Interacted with GDP Growth: Emerging Economies 1987–2003 GDP growth

GNI growth

Cons growth

Equity

GDP growth 1.00

0.97

0.78

GNI growth

1.00

Cons growth Equity asset

Assets Debt

FDI

0.15

0.36

–0.30

–0.13

0.49

–0.24

0.78

0.15

0.38

–0.29

–0.08

0.50

–0.22

1.00

0.13

0.37

–0.19

–0.06

0.44

–0.15

1.00

0.58

–0.02

–0.14

0.31

0.11

1.00

0.14

–0.13

0.60

0.28

1.00

0.26

–0.22

0.48

1.00

–0.02

0.33

1.00

–0.05

Debt asset FDI asset Equity liability Debt liability FDI liability

Liabilities Equity Debt FDI

1.00

Note. The term “GDP growth” represents the data series (∆ log GDPit − ∆ log GDPt ), where GDPit is country i’s year t per capita GDP, and GDPt is the year t per capita aggregate GDP for the group. The series “GNI growth” and “Cons growth” are defined similarly. The term “Equity asset” refers to the data series [(Eit − Et ) ∗ (∆ log GDPit − ∆ log GDPt )], where Eit is the period t natural logarithm of the ratio of foreign equity owned to GDP for country i, and Et is the (un-weighted) average across countries of Eit . “Debt asset,” “FDI asset,” and liabilities are defined similarly. The countries included in the sample are Argentina, Brazil, Chile, China, Colombia, Egypt, India, Indonesia, Israel, Jordan, Korea, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, South Africa, Thailand, Turkey, and Venezuela.

37

Table 6 Measures of financial integration and co-movement of income and consumption: Industrial Countries 1973-2003 Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Full Period g it (κ0 ) GDP

1.00*** 0.99*** 0.99*** 1.00*** (0.01) (0.01) (0.01) (0.01) g GDPit *Openness (κ2 ) –0.05** (0.02) g GDPit *Assets (κ2 ) –0.05** (0.03) g it *Liab. (κ2 ) GDP –0.05* (0.02) N 632 632 632 632

0.66*** 0.66*** 0.65*** 0.67*** (0.06) (0.05) (0.06) (0.06) –0.08 (0.06) –0.05 (0.04) –0.08 (0.06) 632 632 632 632

B. Globalization Period g it (κ0 ) GDP

0.99*** 0.98*** 0.97*** 0.99*** (0.02) (0.02) (0.02) (0.02) g GDPit *Openness (κ2 ) –0.10*** (0.03) g GDPit *Assets (κ2 ) –0.07*** (0.02) g GDPit *Liab. (κ2 ) –0.11*** (0.03) N 352 352 352 352

0.71*** 0.69*** 0.68*** 0.70*** (0.10) (0.08) (0.06) (0.09) –0.12 (0.09) –0.10 (0.07) –0.09 (0.10) 352 352 352 352

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Rows in the left panel of the table are coefficients from panel-data regressions of the form ∆ log GNIit − ∆ log GNIt = constant + κ (∆ log GDPit − ∆ log GDPt ) + it where κ = κ0 + κ1 (t − t¯) + κ2 times “financial openness,” “total assets,” or “total liabilities.” For example, “total assets” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total foreign assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as total assets. The trend coefficients are rarely significant, hence not reported. The right panel of the table presents the parameters from panel-data regressions for consumption risk sharing of the form similar to those of the left panel with the dependent variable (∆ log Cit − ∆ log Ct ). Description of the term in interaction is as follows: “Openness” refers to the measure of financial openness as described in table 1. “Assets” and “Liabilities” refer to total assets (aggregate of portfolio equity, debt and fdi) and total liabilities (aggregate of portfolio equity, debt and fdi) of country i. See the text for further details.

38

Table 7 Measures of financial integration and co-movement of income and consumption: Emerging economies 1973-2003 Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Full Period g it (κ0 ) GDP

1.02*** 1.02*** 1.03*** 1.02*** (0.01) (0.01) (0.01) (0.01) g GDPit *Openness (κ2 ) 0.06*** (0.02) g GDPit *Assets (κ2 ) 0.02 (0.02) g it *Liab. (κ2 ) GDP 0.05*** (0.01) N 629 629 629 629

0.81*** 0.80*** 0.81*** 0.78*** (0.08) (0.07) (0.08) (0.06) 0.09 (0.10) –0.02 (0.07) 0.13 (0.11) 617 617 617 617

B. Globalization Period g it (κ0 ) GDP

1.02*** 1.02*** 1.02*** 1.01*** (0.01) (0.01) (0.01) (0.01) g GDPit *Openness (κ2 ) 0.08** (0.03) g GDPit *Assets (κ2 ) 0.04 (0.03) g it *Liab. (κ2 ) GDP 0.07** (0.03) N 357 357 357 357

0.86*** 0.84*** 0.86*** 0.84*** (0.09) (0.09) (0.09) (0.09) 0.18** (0.08) 0.10 (0.08) 0.18** (0.09) 356 356 356 356

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Rows in the left panel of the table are coefficients from panel-data regressions of the form ∆ log GNIit − ∆ log GNIt = constant + κ (∆ log GDPit − ∆ log GDPt ) + it where κ = κ0 + κ1 (t − t¯) + κ2 times “financial openness,” “total assets,” or “total liabilities.” For example, “total assets” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total foreign assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as total assets. The trend coefficients are rarely significant, hence not reported. The right panel of the table presents the parameters from panel-data regressions for consumption risk sharing of the form similar to those of the left panel with the dependent variable (∆ log Cit − ∆ log Ct ). Description of the term in interaction is as follows: “Openness” refers to the measure of financial openness as described in table 1. “Assets” and “Liabilities” refer to total assets (aggregate of portfolio equity, debt and fdi) and total liabilities (aggregate of portfolio equity, debt and fdi) of country i. See the text for further details.

39

Table 8 Measures of financial integration and co-movement of income and consumption: Developing countries 1973-2003 Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Full Period g it (κ0 ) GDP

0.99*** 1.01*** 0.99*** 1.01*** (0.02) (0.01) (0.02) (0.01) g GDPit *Openness (κ2 ) –0.05 (0.04) g it *Assets (κ2 ) GDP –0.02 (0.02) g it *Liab. (κ2 ) GDP –0.03 (0.03) N 1443 1436 1436 1440

0.82*** 0.77*** 0.82*** 0.77*** (0.08) (0.05) (0.09) (0.05) 0.16 (0.13) 0.02 (0.11) 0.15 (0.10) 1424 1417 1417 1421

B. Globalization Period g it (κ0 ) GDP

0.99*** 1.02*** 0.98*** 1.02*** (0.02) (0.02) (0.03) (0.02) g GDPit *Openness (κ2 ) –0.08** (0.03) g GDPit *Assets (κ2 ) –0.06 (0.05) g it *Liab. (κ2 ) GDP –0.06* (0.03) N 814 814 814 814

0.96*** 0.88*** 0.98*** 0.87*** (0.08) (0.06) (0.10) (0.09) 0.27*** (0.06) 0.09 (0.14) 0.22*** (0.06) 803 803 803 803

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Rows in the left panel of the table are coefficients from panel-data regressions of the form ∆ log GNIit − ∆ log GNIt = constant + κ (∆ log GDPit − ∆ log GDPt ) + it where κ = κ0 + κ1 (t − t¯) + κ2 times “financial openness,” “total assets,” or “total liabilities.” For example, “total assets” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total foreign assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as total assets. The trend coefficients are rarely significant, hence not reported. The right panel of the table presents the parameters from panel-data regressions for consumption risk sharing of the form similar to those of the left panel with the dependent variable (∆ log Cit − ∆ log Ct ). Description of the term in interaction is as follows: “Openness” refers to the measure of financial openness as described in table 1. “Assets” and “Liabilities” refer to total assets (aggregate of portfolio equity, debt and fdi) and total liabilities (aggregate of portfolio equity, debt and fdi) of country i. See the text for further details.

40

Table 9 Co-movement of income and consumption and alternative Measures of Financial Integration: Industrial Countries 1973-2003 (1)

(2)

(3)

(4)

(5)

0.99*** (0.01) –0.03*** (0.01)

0.99*** (0.02)

0.99*** (0.02)

1.00*** (0.02)

1.00*** (0.01)

–0.03* (0.02) –0.03 (0.03) 619

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP g it *Debt GDP

–0.04 (0.03)

g it *FDI GDP

–0.03** (0.02)

g it *(Equity+Debt) GDP

N

632

632

–0.05** (0.02) 619

0.66*** (0.05)

0.67*** (0.05)

0.66*** (0.05)

0.66*** (0.05)

–0.08** (0.05) 619

0.06 (0.04) –0.11** (0.05) 619

610

Panel B g it Dependent Variable: CONS g it GDP g it *Equity GDP

0.67*** (0.06) –0.01 (0.02)

g it *Debt GDP

–0.09* (0.04)

g it *FDI GDP

0.01 (0.04)

g it *(Equity+Debt) GDP

N

610

632

632

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

41

Table 10 Co-movement of income and consumption and alternative Measures of Financial Integration: Industrial Countries 1987-2003 (Globalization Period) (1)

(2)

(3)

(4)

(5)

0.98*** (0.02) –0.05** (0.02)

0.97*** (0.02)

0.99*** (0.02)

0.97*** (0.02)

0.98*** (0.02)

–0.02 (0.03) –0.08*** (0.02) 352

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP g it *Debt GDP

–0.07*** (0.02)

g it *FDI GDP

–0.06* (0.03)

g it *(Equity+Debt) GDP

N

352

352

352

–0.09** (0.02) 352

0.67*** (0.07)

0.71*** (0.10)

0.68*** (0.07)

0.68*** (0.07)

–0.13 (0.08) 352

0.09 (0.07) –0.17* (0.09) 352

Panel B g it Dependent Variable: CONS g it GDP g it *Equity GDP

0.71*** (0.10) 0.01 (0.05)

g it *Debt GDP

–0.15* (0.08)

g it *FDI GDP

0.02 (0.06)

g it *(Equity+Debt) GDP

N

352

352

352

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

42

Table 11 Co-movement of income and consumption and alternative Measures of Financial Integration: Emerging economies 1973-2003 (1)

(2)

(3)

(4)

(5)

0.99*** (0.02) 0.01 (0.01)

1.02*** (0.01)

1.03*** (0.01)

0.99*** (0.02)

0.99*** (0.02)

0.02 (0.02) 0.03 (0.02) 567

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP g it *Debt GDP

0.06** (0.02)

g it *FDI GDP

0.01 (0.01)

g it *(Equity+Debt) GDP

N

489

629

616

0.04** (0.02) 573

0.78*** (0.06)

0.81*** (0.08)

0.72*** (0.07)

0.72*** (0.07)

0.06 (0.11) 561

0.01 (0.05) 0.06 (0.12) 555

Panel B g it Dependent Variable: CONS g it GDP g it *Equity GDP

0.70*** (0.07) 0.04 (0.03)

g it *Debt GDP

0.12 (0.09)

g it *FDI GDP

–0.01 (0.05)

g it *(Equity+Debt) GDP

N

481

617

604

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

43

Table 12 Co-movement of income and consumption and alternative Measures of Financial Integration: Emerging economies 1987-2003 (Globalization Period) (1)

(2)

(3)

(4)

(5)

1.01*** (0.02)

1.03*** (0.01)

0.99*** (0.02)

0.99*** (0.02)

0.03 (0.04) 0.06 (0.04) 0.93 339

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP

0.98*** (0.02) 0.00 (0.02)

g it *Debt GDP

0.08** (0.03)

g it *FDI GDP

0.01 (0.02)

g it *(equity+debt) GDP

adjusted R2 N

0.92 322

0.95 357

0.95 357

0.06* (0.03) 0.92 339

0.82*** (0.09)

0.87*** (0.09)

0.82*** (0.09)

0.82*** (0.10)

0.14 (0.10) 0.58 338

0.06 (0.12) 0.13 (0.11) 0.58 338

Panel B g it Dependent Variable: CONS

output g it *equity GDP

0.79*** (0.10) 0.02 (0.05)

g it *debt GDP

0.16* (0.08)

g it *fdi GDP

0.07 (0.10)

g it *(equity+debt) GDP

adjusted R2 N

0.53 321

0.69 356

0.69 356

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

44

Table 13 Co-movement of income and consumption and alternative Measures of Financial Integration: Developing economies 1973-2003 (1)

(2)

(3)

(4)

(5)

1.00*** (0.01) –0.01 (0.01)

1.01*** (0.01)

1.01*** (0.01)

1.00*** (0.01)

1.00*** (0.01)

0.01 (0.01) 0.02 (0.02) 1321

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP g it *Debt GDP

–0.04 (0.03)

g it *FDI GDP

0.01 (0.01)

g it *(Equity+Debt) GDP

N

783

1443

1399

–0.04 (0.03) 1358

0.76*** (0.05)

0.75*** (0.06)

0.75*** (0.06)

0.75*** (0.076)

0.17 (0.11) 1339

–0.04 (0.07) –0.05 (0.07) 1302

Panel B g it Dependent Variable: CONS g it GDP g it *Equity GDP

0.64*** (0.06) 0.03 (0.03)

g it *Debt GDP

0.17* (0.10)

g it *FDI GDP

–0.05 (0.07)

g it *(Equity+Debt) GDP

N

767

1424

1380

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

45

Table 14 Co-movement of income and consumption and alternative Measures of Financial Integration: Developing economies 1987-2003 (Globalization Period) (1)

(2)

(3)

(4)

(5)

0.99*** (0.02) –0.01 (0.01)

1.02*** (0.02)

1.02*** (0.01)

1.01*** (0.01)

1.00*** (0.01)

0.00 (0.01) 0.05** (0.02) 762

Panel A g it Dependent Variable: GNI g it GDP g it *Equity GDP g it *Debt GDP

–0.06** (0.03)

g it *FDI GDP

–0.01 (0.01)

g it *(Equity+Debt) GDP

N

558

814

797

–0.07** (0.03) 779

0.85*** (0.05)

0.88*** (0.07)

0.82*** (0.06)

0.84*** (0.07)

0.29*** (0.04) 768

–0.02 (0.08) 0.16*** (0.08) 751

Panel B g it Dependent Variable: CONS g it GDP g it *Equity GDP

0.82*** (0.06) 0.00 (0.04)

g it *Debt GDP

0.26*** (0.04)

g it *FDI GDP

–0.02 (0.07)

g it *(Equity+Debt) GDP

N

549

803

786

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of total equity (both asset and liabilities) to GDP for country i, and Et is the average of Eit . The other categories take the same format as total equity. Description of the term in interaction is as follows: “Equity” refers to the value of total equity as described in table 1. “Debt” and “FDI” refers to total debt and total FDI holdings of country i.

46

Table 15 Asset categories and co-movement of income and consumption: Industrial Countries 1973-2003 Globalization Period (1)

(2)

(3)

Full Period (4)

(1)

(2)

(3)

(4)

Panel A g it Dependent Variable: GNI g it GDP

0.98*** 0.97*** 0.99*** 0.96*** (0.02) (0.02) (0.02) (0.01) g it *Equity GDP –0.03** (0.01) g it *Debt GDP –0.04*** (0.02) g it *FDI GDP –0.02 (0.03) g GDPit *(Eq.+Db.) –0.06** (0.02) N 351 352 352 352

1.00*** 0.99*** 0.99*** 0.99*** (0.02) (0.02) (0.01) (0.01) –0.02*** (0.01) –0.03** (0.02) –0.01 (0.01) –0.03** (0.01) 610 632 632 632

Panel B g it Dependent Variable: CONS g it GDP

0.71*** 0.66*** 0.71*** 0.66*** (0.10) (0.07) (0.10) (0.07) g GDPit *Equity 0.01 (0.03) g it *Debt GDP –0.09* (0.05) g GDPit *FDI 0.01 (0.04) g it *(Eq.+Db.) GDP –0.09 (0.06) N 351 352 352 352

0.67*** 0.64*** 0.67*** 0.65*** (0.06) (0.05) (0.06) (0.05) –0.01 (0.01) –0.05 (0.03) 0.02 (0.02) –0.05 (0.03) 610 632 632 632

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity assets. Description of the term in interaction is as follows: “Equity” refers to the value of equity assets as described in table 1. “Debt” and “FDI” refers to debt assets and FDI assets of country i.

47

Table 16 Liability categories and co-movement of income and consumption: Industrial Countries 1973-2003 Globalization Period (1)

(2)

(3)

Full Period (4)

(1)

(2)

(3)

(4)

Panel A g it Dependent Variable: GNI g it GDP

0.99*** 0.98*** 0.99*** 0.98*** (0.01) (0.02) (0.02) (0.02) g GDPit *Equity –0.04** (0.02) g GDPit *Debt –0.09*** (0.04) g GDPit *FDI –0.03* (0.02) g GDPit *(Eq.+Db.) –0.10*** (0.03) N 352 352 352 352

1.00*** 1.00*** 1.00*** 1.00*** (0.01) (0.02) (0.01) (0.02) –0.03*** (0.01) –0.03 (0.03) –0.03*** (0.01) –0.05 (0.03) 604 632 632 619

Panel B g it Dependent Variable: CONS g it GDP

0.71*** 0.69*** 0.71*** 0.70*** (0.10) (0.08) (0.10) (0.08) g GDPit *Equity 0.01 (0.06) g it *Debt GDP –0.17 (0.10) g it *FDI GDP 0.03 (0.04) g it *(Eq.+Db.) GDP –0.13 (0.10) N 351 352 352 352

0.70*** 0.66*** 0.66*** 0.67*** (0.06) (0.05) (0.06) (0.06) –0.01 (0.02) –0.11* (0.05) 0.00 (0.03) –0.10* (0.06) 604 632 632 619

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity liabilities to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity liabilities. Description of the term in interaction is as follows: “Equity” refers to the value of equity liabilities as described in table 1. “Debt” and “FDI” refers to debt liabilites and FDI liabilites of country i.

48

Table 17 Asset categories and co-movement of income and consumption: Emerging economies and other developing countries 1987-2003 (Globalization Period) Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Emerging Economies g it GDP

1.03*** 1.02*** 1.03*** 1.02*** (0.02) (0.01) (0.01) (0.01) g GDPit *Equity 0.02** (0.01) g GDPit *Debt 0.05* (0.03) g it *FDI GDP –0.01 (0.01) g it *(Eq.+Db.) GDP 0.05* (0.02) N 290 357 346 342

0.88*** 0.83*** 0.87*** 0.85*** (0.10) (0.09) (0.10) (0.10) 0.06 (0.04) 0.15* (0.08) 0.03 (0.03) 0.14 (0.09) 289 356 345 341

B. Developing countries g it GDP

1.01*** 0.99*** 1.02*** 0.98*** (0.02) (0.02) (0.01) (0.02) g GDPit *Equity 0.02** (0.01) g it *Debt GDP –0.03 (0.03) g it *FDI GDP –0.01 (0.01) g it *(Eq.+Db.) GDP –0.03 (0.03) N 443 814 721 782

0.86*** 0.96*** 0.89*** 0.98*** (0.09) (0.08) (0.06) (0.10) 0.08*** (0.03) 0.05 (0.07) –0.03 (0.03) 0.05 (0.07) 439 803 710 771

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity assets. Description of the term in interaction is as follows: “Equity” refers to the value of equity assets as described in table 1. “Debt” and “FDI” refers to debt assets and FDI assets of country i.

49

Table 18 Liability categories and co-movement of income and consumption: Emerging economies and other developing countries 1987-2003 (Globalization Period) Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Emerging Economies g it GDP

0.99*** 1.01*** 1.03*** 0.99*** (0.02) (0.01) (0.01) (0.02) g GDPit *Equity 0.01 (0.01) g it *Debt GDP 0.07** (0.02) g GDPit *FDI 0.01 (0.01) g it *(Eq.+Db.) GDP 0.05** (0.03) N 330 357 357 354

0.77*** 0.83*** 0.87*** 0.82*** (0.10) (0.09) (0.10) (0.09) –0.01 (0.03) 0.14 (0.08) 0.05 (0.10) 0.13 (0.11) 329 356 356 353

B. Developing countries g it GDP

0.99*** 1.02*** 0.99*** 1.00*** (0.02) (0.02) (0.02) (0.02) g GDPit *Equity –0.01 (0.01) g it *Debt GDP –0.05 (0.03) g it *FDI GDP –0.01 (0.01) g it *(Eq.+Db.) GDP –0.05* (0.02) N 530 814 814 811

0.83*** 0.85*** 0.96*** 0.84*** (0.06) (0.06) (0.08) (0.06) –0.01 (0.03) 0.21*** (0.05) –0.04 (0.07) 0.22*** (0.05) 521 803 803 800

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity liabilities to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity liabilities. Description of the term in interaction is as follows: “Equity” refers to the value of equity liabilities as described in table 1. “Debt” and “FDI” refers to debt liabilites and FDI liabilites of country i.

50

Appendix Table 1 Asset categories and co-movement of income and consumption: Emerging economies and other developing countries 1973-2003 Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Emerging Economies g it GDP

1.03*** 1.02*** 1.03*** 1.02*** (0.01) (0.01) (0.01) (0.01) g it *Equity GDP 0.02** (0.01) g it *Debt GDP 0.03** (0.01) g it *FDI GDP –0.01 (0.01) g GDPit *(Eq.+Db.) 0.03** (0.01) N 451 629 527 590

0.82*** 0.082*** 0.89*** 0.81*** (0.07) (0.08) (0.07) (0.08) 0.06 (0.04) 0.01 (0.07) –0.01 (0.02) 0.01 (0.07) 443 617 518 578

B. Developing countries g it GDP

1.03*** 0.99*** 1.03*** 0.99*** (0.02) (0.02) (0.01) (0.02) g GDPit *Equity 0.02** (0.01) g GDPit *Debt –0.01 (0.01) g it *FDI GDP 0.01 (0.01) g it *(Eq.+Db.) GDP –0.01 (0.01) N 643 1437 1060 1369

0.76*** 0.82*** 0.78*** 0.83*** (0.07) (0.08) (0.07) (0.09) 0.08*** (0.02) –0.01 (0.01) –0.05 (0.03) –0.01 (0.06) 632 1418 1041 1350

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity assets to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity assets. Description of the term in interaction is as follows: “Equity” refers to the value of equity assets as described in table 1. “Debt” and “FDI” refers to debt assets and FDI assets of country i.

51

Appendix Table 2 Liability categories and co-movement of income and consumption: Emerging economies and other developing countries 1973-2003 Dependent Variable:

g it CONS

g it GNI

(1)

(2)

(3)

(4)

(1)

(2)

(3)

(4)

A. Emerging Economies g it GDP

0.99*** 1.02*** 1.03*** 1.00*** (0.01) (0.01) (0.01) (0.01) g it *Equity GDP 0.01 (0.01) g it *Debt GDP 0.05*** (0.01) g it *FDI GDP 0.01 (0.01) g GDPit *(Eq.+Db.) 0.03 (0.02) N 495 629 629 612

0.71*** 0.78*** 0.81*** 0.72*** (0.07) (0.06) (0.08) (0.06) 0.03 (0.02) 0.14 (0.09) –0.01 (0.05) 0.08 (0.10) 487 617 617 600

B. Developing countries g it GDP

1.00*** 1.01*** 0.99*** 1.00*** (0.01) (0.01) (0.02) (0.01) g GDPit *Equity –0.01 (0.01) g it *Debt GDP –0.03 (0.02) g it *FDI GDP 0.01 (0.01) g it *(Eq.+Db.) GDP –0.03 (0.02) N 736 1443 1443 1426

0.66*** 0.76*** 0.83*** 0.75*** (0.06) (0.05) (0.07) (0.05) 0.03 (0.03) 0.15** (0.08) –0.08 (0.06) 0.15 (0.08) 720 1424 1424 1407

Note. All regressions include a constant and a time trend. Panel data regression uses the fixed effect estimation method. Robust and Clustered standard errors are in parentheses. ***, **, * denote statistical significance at 1%, 5%, g it and G g and 10% level respectively. GDP NIit denotes the deviation in growth rate of country-level GDP and GNI from g it denotes the consumption counterpart. the respective growth rates in the world-level GDP and GNI. Similarly, CONS Regression specification is same as in table (6). “Equity” refers [(Eit − Et )], where Eit is the period t natural logarithm of the ratio of equity liabilities to GDP for country i, and Et is the average of Eit . The other categories take the same format as equity liabilities. Description of the term in interaction is as follows: “Equity” refers to the value of equity liabilities as described in table 1. “Debt” and “FDI” refers to debt liabilites and FDI liabilites of country i.

52