Journal Issue

Research Summaries: Measures of Financial Integration

International Monetary Fund. Research Dept.
Published Date:
March 2009
  • ShareShare
Show Summary Details

Understanding the costs and benefits of financial integration has been a topic of intensive research during the past two decades. Central to any empirical research investigating this topic is the measurement of financial integration. A large number of such measures have been suggested, including de jure measures, reflecting the extent of legal restrictions on cross-border financial flows, and de facto measures, reflecting a country’s actual degree of financial integration. This article summarizes research on measures of financial integration, an area to which IMF research has contributed substantially.

Global financial integration, as measured by the magnitude of cross-border financial asset holdings, has grown exponentially in recent years. While it can benefit economies through improved access to capital and better risk diversification, it may also facilitate the transmission of adverse shocks across countries. Better understanding the relative costs and benefits of financial globalization is important for policy analysis: should policymakers impose restrictions on cross-border capital flows or should they undertake policies to attract more flows? Central to empirical research investigating this and related questions is the measurement of financial integration.

Over the past several years, an increasing number of such measures have been made available, including de jure measures, aiming to reflect the extent to which countries impose legal restrictions on cross-border financial flows, and outcome-based de facto measures, aiming to capture a country’s actual degree of financial integration. For the purpose of policy analysis, de jure measures, which are under policymakers’ direct control, are more relevant, while in other applications, de facto measures may be more appropriate; in still other situations, both may be necessary, for example, if the research question centers on the effectiveness of capital controls in stemming de facto outcomes.

Most de jure measures rely on information contained in the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER). Until 1995, the AREAER summarized a country’s openness to capital flows using a binary dummy variable. Since 1995, the AREAER has provided additional information on capital account restrictions in several subcategories. The structural break in the AREAER’s format confronts researchers with a trade-off between sample coverage and detail.

The binary AREAER indicator provides the largest sample coverage, with a (unbalanced) panel starting in 1966 and covering 184 countries. Grilli and Milesi-Ferretti (1995) were among the first to use this indicator. However, the binary index only crudely approximates a country’s degree of capital account openness and provides no information on the composition of capital controls.

The AREAER reported three additional binary variables: on current account openness; on export proceeds’ surrender requirements; and on multiple exchange rate practices. Mody and Murshid (2005) extend all four variables until 1999 and interpret their sum as a financial integration index. (Grilli and Milesi-Ferretti, 1995, also use all four variables, but separately.) Chinn and Ito (2008) further extend these four variables for 182 countries up to 2006 and aggregate them taking a principal components approach. While these measures provide more finely gradated information, they arguably capture information that extends beyond a narrow definition of capital controls.

Other authors have chosen a more focused approach. Bekaert, Harvey, and Lundblad (2005) date equity liberalization episodes for 42 countries during 1960–2006; Edison and Warnock (2003) focus on equity restrictions in 31 countries during 1989–2006 at a monthly frequency, by measuring the fraction of a country’s market capitalization that is open to foreign investment.

An alternative index by Quinn (1997) has recently been used in IMF research on structural reforms (IMF, 2008). An updated version covers 94 countries during 1950–2005; it captures the intensity of controls by ranking different control instruments by their (assumed) economic importance, which involves a certain degree of judgment, and it also distinguishes between residents and nonresidents.

Some authors have utilized the greater richness of the post-1995 AREAER structure to capture more dimensions of capital account restrictiveness, including by asset categories, residency status, and the direction of flows. Tamirisa (1999) codes the various subcategories in the new AREAER structure for 40 countries in 1996; Johnston and Tamirisa (1998) analyze the dataset and its various subcomponents in more detail. Miniane (2004) follows a similar approach, but extends some of the post-1995 structure backward, covering 1983–2000 for 34 countries, at the cost of a more limited country coverage and less detail, including the inability to distinguish between inflow and outflow restrictions.

“Should policymakers impose restrictions on cross-border capital flows or should they undertake policies to attract more flows?”

More recently, and broadly following Tamirisa’s (1999) approach, Schindler (forthcoming) constructs a dataset containing information for a subset of the categories contained in the new AREAER structure, covering 91 countries during 1995–2005. In this dataset, indices are coded at the level of individual types of transactions, allowing for different data aggregations, including by asset category, residency status, and inflows versus outflows. (Dis)aggregations of this nature are likely to be important. As Henry (2007, p. 889) notes, existing evidence suggests that opening equity markets to foreign investors may avoid some of the problems associated with the liberalization of debt flows, and so, “[a]t a minimum, the distinction between debt and equity is critical.” The resulting, more finely gradated indices also allow for more meaningful comparisons across countries and over time.

The de jure measures discussed so far share some drawbacks: they do not reflect the extent to which legal controls are enforced in practice; even the more disaggregated indices may not capture subtle, but possibly important differences between countries’ capital control regimes; and they do not necessarily reflect a country’s actual degree of financial integration, which is presumably the key issue of interest. For example, Dell’Ariccia and others (2008) document that even countries with relatively closed capital accounts became substantially more financially integrated over the past decades.

De facto indicators avoid these issues by focusing directly on outcomes. Lane and Milesi-Ferretti (2007) construct a database of external stocks of assets and liabilities by using official estimates from countries’ international investment position and then generating estimates for stock positions in earlier years based on capital flows data and capital gain/loss calculations. Their database is the most comprehensive and widely used de facto measure of financial integration, covering 145 countries during 1970–2004.

Overall, a wide array of measures exists from which researchers can choose those that best fit their research question. For example, the inflow/outflow distinction in Schindler (forthcoming) allows Prati, Schindler, and Valenzuela (forthcoming) to identify the differential effects of capital account liberalizations on different subsets of firms; Binici, Hutchison, and Schindler (forthcoming) use both de jure (Schindler, forthcoming) and de facto (Lane and Milesi-Ferretti, 2007) measures and find significant differences in the effectiveness of capital controls between equity and debt flows and inflows and outflows; and Kose and others (2006) argue for the use of de facto measures. For further discussion of financial integration measures and related issues, researchers are referred to Edison and others (2004), Kose and others (2006), Miniane (2004), and Schindler (forthcoming).


Other Resources Citing This Publication