Investigating Capital Mobility

1. Literature review

There are two main approaches to investigate the capital mobility. The first approach comprises a structural model based on the direct comparison of the rates of return on physical capital across countries (Obstfeld, 1985). This approach involves controlling for a number variables including inflation, expected inflation, currency fluctuations, expected appreciation or depreciation, political risk, tax rates, transaction costs, capital controls, investor preferences and additional uncertainty (Herzog, 2008). The second approach originated by Feldstein and Horioka (FH) requires testing the relationship between domestic saving and investment rates. Both lines of research have advantages and drawbacks. The first estimation process is guided by the structural model. Nonetheless, the model is often grounded in economic theory; thus the empirical models are more only as good as the underlying theory. Besides, the select of appropriate variables is difficult. FH’s approach is more novel, but whether a simple saving-investment relation can truly represent the degree of capital mobility is still under controversy.

Feldstein and Horioka (1980) were the first researchers who analyzed the connection between domestic saving and investment rates to measure capital mobility. They started from the fact that in the world of perfect capital mobility, domestic savings would seek for the highest returns on the world capital market irrespective to domestic demand for investment; thus there should be no systematic relationship between domestic saving and domestic investment. They identify a cross-section association between national saving and investment for a sample of 16 OECD countries from 1960 to 1974. Their findings point out that the systematic relationship does exist which is inconsistent with the view that there is sufficiently high international capital mobility. This interpretation has posed an uncomfortable puzzle which generated a large debate on saving-investment relationship. The main criticism of FH model is the potential endogeneity problem. The changes in domestic investment may affect changes in domestic savings. Moreover, the estimated saving-investment correlation may be indicative of a common effect which affects both variables simultaneously (Obstfeld, 1986, 1995). Despite the criticisms on theoretical and econometric grounds, the basic intuition behind the FH approach that it is possible to investigate the existence of international capital mobility from saving and investment association remains tempting. FH puzzle has sparked off a growing body of literature which is divided into two main threads of research.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

The first thread use the saving-investment relationship to form a better understanding of capital mobility through alternative regressions and controlling relevant variables. The FH result has been mainly replicated using cross section regressions (Artis & Bayoumi, 1992; Dooley, Frankel, & Mathieson, 1987; Feldstein, 1983; Feldstein & Bachetta, 1991; Murphy, 1984; Obstfeld, 1995; Penati & Dooley, 1984; Tesar, 1991). Those researches are criticized due to the use of cross section regression might lead to the significant loss of data information.

On the other hand, Obstfeld (1986) states that it is more adequate to apply time series instead. He argues that time series regressions provide a set of empirical regularities which might inspire more powerful tests. The measures of correlation are unlikely to be affected by some common factors that may limit current account imbalances over longer period. In fact, Obstfeld found significantly lower saving-investment correlation and even sharp decrease after the ending of the fixed exchange rate period 1973. He concluded that the international capital mobility is high and increasing.

Miller (1998) argues that Obstfeld (1998) applied the simple time series regression without adjusting for cyclical factors and the regression might be spurious, while co-integration techniques may convey more significant information. He was the first to employ Engle-Granger’s two step approach to test for saving-investment cointegration. He believes the increase of capital mobility will sever the long-run saving-investment link. His findings applied for United States show that saving and investment rates are non-stationary and cointegrated during Bretton Woods period, but capital is more mobile during the flexible exchange rate period.

Taylor (1996) augments the literature by investigating the saving-investment relationship for capital mobility measurement using both time-series and cross-section analysis. The study applied annual data for 12 countries over the period 1850 – 1992. Taylor contrasts the results of saving-investment correlations in cross section and time series. The conclusion is that both approaches offer helpful and complementary insights into the evolution of the world capital market. The results present a nuanced picture of capital market evolution. In cross section, the results show a marked tightening of the saving-investment correlation in periods of global economic crisis, symptomatic of diminishing capital mobility in such periods. In time series, the results show considerable heterogeneity in individual country; nevertheless, capital tends to be more mobile overtime. The puzzle remained unsolved.

There is other research taking different angles. Kim et al. (2010) applied framework extended from FH model using panel approach with controlled common shocks’ effect and reducing endogeneity problem by instrument variable. By evaluating the role of domestic saving, regional saving and global saving in financing domestic investment, their study introduces new perspective for inference of capital mobility level from saving-investment relationship. If the foreign savings contribute considerably to the funding of domestic investment, then there is existent of capital mobility. They focus their efforts toward 11 emerging Asian countries and the results suggest that there is an increase in capital mobility, yet the capital is obviously more mobile within the region than with the rest of the world.

The second thread proposes alternative theoretical explanations for the high correlation between saving and investment rates. This line of research argues that the quantity approach of Feldstein and Horioka is useful but incomplete, because the evidence of positive saving-investment relationship is also consistent with other modern open economy macroeconomics models[1] (Eslamloueyan and Jafari, 2010). These models allows domestic saving and investment to diverge from their long run equilibrium. Put it in other word, temporary current account imbalances are allowed (Blanchard and Fisher, 1989 and Jansen, 1996). According to these theories, short run saving-investment dynamics can be different from their long run behavior. There are many recently empirical works support this approach which focus on jointly modeling the short-run and long-run relationship to infer capital mobility.

Coackley, Kulasi & Smith (1996), Coackley and Kulasi (1997) attempt to solve the FH puzzle for OECD countries by developing a theoretical model in which saving-investment behavior is related to the current account by means of a solvency constraint. Their results state that under solvency constraint, the current balance is stationary and therefore saving and investment cointegrated with a unit coefficient irrespective of the degree of capital mobility. They strongly conclude that saving- and investment cointegration is uninformative about capital mobility and FH cross-section regression captures the long-run solvency regression rather than the measure of capital mobility.

Nevertheless, Jansen (1996), (1997), (1998) argue that the short and long run saving-investment relationship could be used jointly to detect the degree of international capital mobility. Jansen states that it might not be appropriate to rely only on cointegration test for capital mobility inference. Instead, error correction model which can capture both short and long run relationship is more adequate representation. Capital mobility can manifest itself through three possible outcomes in error correction framework: (i) a non-cointegration of saving and investment rates; (ii) a non-stationary current account; (iii) a small or zero short run saving-investment relationship. Using time series approach, Jansen’s researches support the theory that solvency constraints binding in the long-run and suggest that there is an increase in capital mobility within the OECD area.

Following the pioneer work of Jansen (1996), Pelgrin & Schich (2004) replicated the model for 20 OECD countries with recent annual data. The study applied panel error correction method to detect the existence of international capital mobility. They pointed out that with the use of a panel of time-series and cross-section data, increased power and more precise estimates are obtained. Additionally, it is less preferable to focus on individual country than measure capital mobility for a large number of countries jointly, and hence take into consideration relations between them. Their results show that there is a saving-investment cointegration and the current account follows a stationary process which imply solvency constraint is binding. The existence of capital mobility allows for national saving and investment to deviate temporarily from each other and it is proved to increase over time.

The extensive empirical literature on the FH puzzle varies significantly depend on approaches, methodologies employed, along with the data set and sample periods covered. However, the recently studies have proved the superior advantages using panel approach and error correction model to investigate the degree of capital mobility from the short and long run saving-investment behavior (Pelgrin & Schich, 2004; Kollias et al., 2006; Eng & Habibullah, 2006). The tight correlation between saving and investment rates is explained by the current account balance in the long run, while this method still allows for capital to mobile in the short-run which presents conventional wisdom that there is a high international capital mobility and global capital markets are rapidly integrated.

[1] There are many researches proposed different explanations for the strong saving-investment correlation such as large country bias (Harberger, 1980 and Murphy, 1984), information constraints (Obstfel, 1986), government policies such as constant current account targeted by governments or solvency constraints (Bayoumi,1990; Coackley et al., 1996; Coackley and Kulasi, 1997).