Capital flows, push versus pull factors and the global financial crisis☆
Introduction
The 2007–08 global financial crisis had been preceded for many years by substantial global imbalances in trade and capital flows. It is in particular the United States which was not only the origin of the financial crisis, but which had been among the economies globally relying most heavily on capital inflows to finance a growing trade deficit. Many observers argued before the crisis that such a status quo was unsustainable and that ultimately deficit countries, such as the United States, would see capital inflows dwindle and exchange rates and asset prices fall during an adjustment process. However, the crisis played out very differently, with capital of domestic and foreign investors flowing massively into US assets between July 2008 and April 2009, and in particular after the collapse of Lehman Brothers. Yet what has been striking about the crisis was not only its global reach, but also the high degree of heterogeneity with which it affected different countries, both among advanced economies as well as among emerging market economies.
Moreover, it was not only the global transmission of the 2007–08 crisis, but also the recovery period since 2009 that has sparked a controversy about the drivers of global capital flows. Some emerging market economies (EMEs) have experienced massive portfolio capital inflows over the past two years, raising concerns about their viability and effects on domestic economies, exchange rates and capital markets. EME policy-makers have stressed the importance of “push” factors, i.e. in particular monetary and fiscal policies in advanced economies, as the main culprits behind this surge in capital flows. By contrast, others have emphasized “pull” factors, such as real divergences between EMEs and advanced economies (AEs), as the main driver of the current pattern of capital flows. In fact, this controversy has become one of the core issues of debate in international fora, such as the G20 which is considering a code of conduct for capital flow management, including the imposition of capital controls to deal with volatile capital flows.
The objective of the paper is to analyze the role of different drivers of global capital flows during the crisis and the subsequent recovery. The focus is on two questions: first, how important have been common, global shocks for capital flows? And second, how relevant have been macroeconomic policies, institutions and financial policies in helping countries shield themselves from such global shocks? The first of the questions is informative about push factors — if common global shocks explain a large part of the dynamics of global capital flows, then push factors are important. The second question allows gauging the relevance of pull factors — if capital flows are highly heterogeneous across countries, and if this differential pattern is accounted for by country-specific characteristics, then pull factors are relevant.
The paper's focus is at the micro level, i.e. at the level of individual mutual funds across a broad geographic coverage of 50 countries and markets worldwide. The paper uses a novel dataset which stems from EPFR Global, which contains weekly portfolio investment flows by more than 14,000 equity funds and more than 7000 bond funds, with USD 8 trillion of capital under management. The strength of EPFR data is not only its disaggregated information at the fund level, but crucially also the high time frequency. Compared to the pre-crisis period, the data shows remarkably strong divergences in capital flows across countries during the crisis, and more precisely a massive reallocation of capital from emerging markets (and a few advanced economies) to the US.
The second part of the paper aims to explain this heterogeneity of global capital flows during the financial crisis and the subsequent recovery. A factor model for the determinants of capital flows is formulated, distinguishing between different factors as well as allowing for a distinction of drivers during non-crisis times and those during the crisis. The focus is on the effect of a set of common global shocks – with a specific emphasis on liquidity and risk shocks as well as macro news shocks – as well as a set of idiosyncratic, country-specific shocks on capital flows. The findings show that the global factors account for a large share of the global capital flow pattern during the crisis. Importantly, the signs of the model parameters change substantially during the crisis episode. For instance, while an increase in risk before the crisis was associated with capital flows out of AEs and into EMEs, this effect reversed during the crisis inducing a substantial reallocation of capital from many EMEs into a few AEs. This evidence is thus consistent with the hypothesis that the dynamics of capital flows was primarily driven by safe-heaven flows during the crisis.
Another key finding of the analysis is the large degree of heterogeneity with which different countries were affected by the same global shocks, in particular during the crisis and also the subsequent recovery. The findings indicate that it has been the institutional quality, country risk together with the strength of macroeconomic fundamentals and policies that explain a large share of the heterogeneity of capital flows during the crisis. By contrast, countries' external (real and financial) exposure appears to have largely been irrelevant for understanding the global capital flow dynamics, including the retrenchment of capital during the crisis, in particular for emerging economies.
The final part of the paper attempts to quantify the relative importance of common shocks (“push” factors) and country-specific determinants (“pull” factors). The findings indicate that common factors were more important overall as a driver of net capital flows during the 2007–08 financial crisis. However, in the recovery period since March 2009, common factors appear to have become less important as drivers of global capital flows, whereas it is domestic pull factors that have come to dominate in explaining capital flows, in particular for countries in Emerging Asia and Latin America.
Putting these findings into perspective, the analysis of push factors versus pull factors as drivers of global capital flows in the present paper focuses on a relatively short period of time – a time span of about five years between end 2005 and end 2010 – while the discussion of push and pull factors has traditionally been made with reference to much longer cycles of capital flows. It is also important to highlight that this period of 2005 to 2010 has been in many ways extraordinary throughout for the dynamics of global capital flows, as a period of a sharp contraction of capital flows, in particular to some EMEs, during the 2007–08 crisis was followed by an equally extraordinary surge in capital flows to EMEs. Hence an important open issue is whether the current dynamics of global capital flows will continue well into the future, and what it implies for the risk of sudden stops and capital flow reversals with all its adverse implications for global growth and financial stability. But in particular because it is so important to understand better the dynamics and risks of periods of financial stress, the findings of the current paper may be instructive about how future crises may play out.
The paper has a number of implications for economic policy and for policy-makers. On the one hand, financial globalization and the exposure to common global factors have made countries more vulnerable to external and global shocks. Yet, on the other hand, the exposure to domestic risks has also been a relevant factor during the crisis and in the global capital flow surge thereafter, in particular those domestic risks related to poor macroeconomic fundamentals, policies and institutions. This implies that countries are far from innocent bystanders that are powerless in being exposed to volatile global markets, and that indeed they have tools to insulate to some extent their economies from adverse global shocks.
These findings have a bearing in particular on the current debate on how EMEs should deal with volatile capital flows. To the extent that capital flows are driven by global factors, some EME policy-makers have argued that this would justify the use of capital controls as well as policy interventions e.g. in FX markets. However, such policies may be misguided if the drivers of capital flows are mainly found in idiosyncratic, country-specific policies and conditions, which calls for policy-makers to rather focus on making their domestic economies more resilient by improving institutions, deepening financial markets and enhancing macroeconomic and macroprudential policies.
The paper is related to various strands of the literature. First, it relates to a growing literature on the 2007–2008 financial crisis, which has partly focused on the US and its policy responses (e.g. Calomiris, 2008), while the literature on the global transmission of the crisis has been more limited (e.g. Blanchard et al., 2010, Rose and Spiegel, 2011). Second, it links to the literature on capital flows during crises and other periods of extreme changes in capital flows, such as sudden stops, surges, retrenchments and capital flight.1 A recent focus in this literature has been on gross capital flows. Forbes (2010) distinguish between the four different types of extreme capital movements (surges, stops, flights and retrenchments), separating the activity by domestic and foreign residents, and their determinants. Milesi-Ferretti and Tille (2011) and Broner et al. (2010) analyze the link between gross capital flows and crises, in particular the 2007–08 episode.
While recent work has underlined the importance of analyzing gross flows and distinguishing across asset classes, relatively little work so far has been undertaken on investment decisions and capital flows at the micro level of individual investors and funds. Notable exceptions are Calvet et al. (2009), Froot and Ramadorai (2005), Hau and Rey (2008), and Jotikasthira et al. (2009).
A third related strand of the literature is the work on determinants of capital flows, and in particular the distinction between global factors and domestic factors and their transmission channels. There is a large literature on the global transmission of past financial crises, with a strong interest in the role of different channels (e.g. Forbes and Rigobon, 2002, Bekaert et al., 2005, Bae et al., 2003, Karolyi, 2003, Bekaert et al., 2011). Some of this work analyzes the role and mostly finds strong evidence for the transmission of global shocks to financial markets and capital flows (Bacchetta and van Wincoop, 2010), including shocks to liquidity (Brunnermeier, 2009, Calvo, 2009, Kalemli-Ozcan et al., 2010).
A specific angle of this literature on crisis focuses on the role of contagion, which is defined by most studies as transmission of shocks or crises above and beyond what domestic fundamentals as well as common shocks can explain. Specific examples including the implications for capital flows and asset prices, are Forbes and Rigobon (2002), Dungey et al. (2011), Broner et al. (2006) and Claessens and Forbes (2001). Bekaert et al. (2011) analyze the presence of contagion in equity markets during the 2007–08 financial crisis and find evidence that contagion was mostly domestic in nature, i.e. comovements mainly among domestic firms increased substantially.
As to the structure of the paper, the next section presents the data and aims at establishing a number of stylized facts about the high-frequency dynamics of global capital flows during the crisis. The third section of the paper outlines the empirical methodology. Section 4 presents the empirical findings on the determinants and the effects of common and idiosyncratic shocks on capital flows. Section 5 concludes.
Section snippets
Global fund-level data and stylized facts of capital flows
The paper uses a dataset on portfolio capital flows and performance at the fund level, compiled by EPFR. It is a fairly novel database, which to the best of my knowledge, has so far been only used by one paper (Jotikasthira et al., 2009) and in a very different context. It contains daily, weekly and monthly flows by more than 16,000 equity funds and more than 8000 bond funds; hence its focus being on different types of portfolio investment. There is some difference in coverage, with the data at
Empirical methodology
The objective of the empirical analysis is to explain the global dynamics of capital flows during and around the 2007–08 financial crisis, and in particular the heterogeneity in capital flows across countries, as highlighted in the previous section.
The empirical model to get at these questions is a factor model with two types of factors — a set of global/common factors, and a set of domestic/idiosyncratic factors, , in order to test whether portfolio capital flows during and
Empirical results
This section presents the empirical results of the factor model. It starts by first estimating models (1)–(3) and outlining the findings for the effects of shocks on global capital flows (Section 4.1) and the determinants that explain the cross-country heterogeneity in capital flows in response to these shocks (Section 4.2). The section concludes by gauging the economic relevance of the various drivers associated with push and pull factors in explaining net capital flows during the 2007–08
Conclusions
Global portfolio capital flows have exhibited a collapse and substantial shift during the height of the 2007–08 crisis, and an even more sustained surge since the beginning of 2009. The paper has analyzed the drivers behind this collapse and surge over the past few years, building on a novel, high-frequency database of fund-level flows in portfolio equities and bonds. It has attempted to shed light in particular on the debate whether it has been push factors or pull factors that have been the
References (38)
- et al.
Real-time price discovery in stock, bond and foreign exchange markets
Journal of International Economics
(2007) - et al.
Time-varying integration and international diversification strategies
Journal of Empirical Finance
(2009) - et al.
When in peril, retrench: testing the portfolio channel of contagion
Journal of International Economics
(2006) Why do foreigners invest in the United States?
Journal of International Economics
(2010)- et al.
Cross-country causes and consequences of the 2008 Crisis: international linkages and American exposure
European Economic Review
(2011) - Bacchetta, Ph., van Wincoop, E., 2010. On the Global Spread of Risk Panics....
- et al.
A new approach to measuring financial contagion
Review of Financial Studies
(2003) - et al.
Time-varying world market integration
Journal of Finance
(1995) - et al.
Market integration and contagion
Journal of Business
(2005) - et al.
Global crises and equity market contagion
NBER Working Paper No. 17121
(2011)
The initial impact of the crisis on emerging market countries
Brookings Papers on Economic Activity
Deciphering the liquidity and credit crunch 2007–8
Journal of Economic Perspectives
Market Liquidity and Funding Liquidity
The Review of Financial Studies
Fight of flight? Portfolio rebalancing by individual investors
Quarterly Journal of Economics
Capital flows and capital-market crises: the simple economics of sudden stops
Journal of Applied Economics
Financial crises and liquidity shocks: a bank-run perspective
NBER Working Paper No. 15425
Systemic sudden stops: the relevance of balance-sheet effects and financial integration
NBER Working Paper No. 14026
Cited by (556)
What explains foreign portfolio investment inflows to BRICS countries?
2024, Economic Analysis and PolicyRevisiting capital flow drivers: Regional dynamics, constraints, and geopolitical influences
2024, Journal of International Money and FinanceChinese monetary policy spillovers on its international portfolio investment flows
2024, Journal of International Money and FinanceUIP deviations: Insights from event studies
2024, Journal of International EconomicsDo foreign institutional investors improve board monitoring?
2024, Journal of International Financial Markets, Institutions and MoneyEconomic sentiment and foreign portfolio flows: Evidence from Türkiye
2024, Central Bank Review
- ☆
I would like to thank Charles Engel, Kristin Forbes, Jeff Frankel, my discussant Carmen Reinhart, two anonymous referees as well as the other participants at the 2010 pre-conference and the 2011 Bretton Woods conference of the NBER/Sloan Global Financial Crisis project for comments and discussion. I am grateful to Daniel Schneider for excellent research assistance. The views expressed in this paper are those of the author and do not necessarily reflect those of the European Central Bank or the Eurosystem.