Hot money in disaggregated capital flows

ABSTRACT

We explore the possible existence and behavior of hot money in six categories of disaggregated bilateral capital flows (equity inflows, equity outflows, bond inflows, bond outflows, banking credit inflows, and banking credit outflows) for 12 emerging markets vis-à-vis the US from 1995 to 2012 and provides several new findings. First, we identify the existence of hot money in all six categories above and conclude that both gross inflows and gross outflows can be the sources of hot money. Second, hot money in equity inflows (outflows) engages in positive (negative) feedback trading regarding local stock market returns. Third, some categories of hot money have a temporary influence on local stock market returns while the others have a permanent influence, supporting the explanations of both price pressures and information advantage. Finally, local stock market returns in half of our sample countries, which have tightened capital controls during the late 2000s global financial crisis (GFC), are more affected by hot money than in the other half. Our findings confirm several popular conjectures of hot money, and endorse the use of capital controls to limit financial vulnerability in the run-up to and during the GFC.

KEYWORDS:

• hot money
• equity flows
• bond flows
• bank credit
• equity price
• emerging market

JEL CLASSIFICATIONS:

• E44
• F20
• F34
• G1

Acknowledgements

I thank the editor (C J Adcock), the associated editor for their management on this manuscript, and I am very grateful for the very helpful comments from two anonymous referees.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Cheng Yan http://orcid.org/0000-0002-3569-6859

Notes

1 The earliest literature, Gray (1964), refers hot money as foreign short-term funds. Hot money in portfolio (equity and bond) flows has been widely blamed in the 1997 Asian Financial Crisis (Kaminsky and Reinhart 1998, 1999; Chari and Kehoe 2003), while Fuertes, Phylaktis, and Yan (2016) show that the bank credit has become dominated by hot money as well in the run-up to the GFC. Unlike the earlier literature which finds portfolio (equity) flows most volatile (c.f., Sarno and Taylor 1999a, 1999b), recent studies find higher volatility in international bank lending than in other types of international capital flows in the 2000s.

2 We follow the literature on hot money (see, e.g., Sarno and Taylor 1999a, 1999b; Fuertes, Phylaktis, and Yan 2016; Yan, Phylaktis, and Fuertes 2016) and only consider portfolio flows (i.e. equity flows and bond flows) and bank credit. We exclude FDI, as it is unlikely hot money can be a serious problem for FDI, given the intrinsic irreversible nature of FDI (Tong and Wei 2011). Nevertheless, we are also aware that there is another related concept in the international economics literature (perhaps more on risk-sharing and economic growth), usually called total equity flows (e.g. Alfaro, Kalemli-Ozcan, and Volosovych 2005, 2008), which is the sum of flows of FDI and flows of portfolio equity.

3 The underlying rationale that is a more volatile form of capital will be more likely to fly out of the country in crisis. For instance, Tong and Wei (2011) do not find a connection between a country’s exposure to aggregated capital flows and the extent of the liquidity crunch experienced by its manufacturing firms, which masks an important compositional effect, as a different but consistent pattern emerges when they disaggregate capital flows into three types (FDI, foreign portfolio flows and foreign loans). Other empirical evidence also suggests that aggregating different capital flows may not be appropriate when one wishes to understand the relationship between capital flows and local stock market returns (e.g., Milesi-Ferretti and Tille 2011; Yan, Phylaktis, and Fuertes 2016).

4 It is in line with most of the literature in this area (Froot, O’Connell, and Seasholes 2001; Bekaert, Harvey, and Lumsdaine 2002b; Griffin, Federico, and Stulz 2004; Richards 2005; Froot and Ramadorai 2008; Jinjarak, Wongswan, and Zheng 2011; Yan, Phylaktis, and Fuertes 2016).

5 Most of the liberalizations in EMs are in the 1980s and earlier 1990s. For official liberalization dates and the chronology of important financial, economic, and political events in EMs, c.f.: Bekaert and Harvey (2000), Bekaert, Harvey, and Lumsdaine (2002b).

6 Bekaert and Harvey (1995) provide a detailed comparison between the MSCI indices and the IFCG indices. Share counts used in IFCI are reduced to reflect any limits or restrictions on investments by foreign investors or entities. For instance, if a company has a market capitalization of 600 million U.S. dollars and the national law restricts foreign ownership to 50% of any company, the IFCG uses the full 600 million U.S. dollars as the market capitalization while the S&P/IFCI uses 300 million U.S. dollars. IFCI data were not available for most of past research – such data did not exist in the 1990s. A few applications of IFCI data include Bekaert, Harvey, and Lumsdaine (2002a), Bae, Chan, and Ng (2004), Boyer, Kumagai, and Yuan (2006).

7 The only exception to our knowledge is Aizenman and Pasricha (2013), which focuses on capital outflows.

8 We omit the covariance decomposition method in Froot, O’Connell, and Seasholes (2001, equation 5, page 180) for brevity.

9 The sample size of 12 markets is large enough to provide results that are potentially fairly general, yet is small enough to allow more attention to market-specific analysis and presenting results market-by-market in an intelligible way than might be impossible in datasets with a larger number of markets. Our sample markets have been studied in earlier literature. For example, Choe, Kho, and Stulz (1999; Choe, Kho, and Stulz 2005) have studied South Korea at the individual stock level; Richards (2005) have a look at Indonesia, South Korea, Philippines, Thailand and Taiwan China. Fuertes, Phylaktis, and Yan (2016) and Yan, Phylaktis, and Fuertes (2016) include all these market as a subsample of their studies.

10 Note that other changes (such as changes in institutions and other drivers of capital flows as well as the new implementation of capital controls) may also have affected this pattern. We thank a referee for pointing it out.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (NNSFC Grant No.71703045), Newton Fund, Durham University Business School, Higher Education Funding Council for England.