Market Structure, Excess Returns in the Foreign Exchange Market and Deviations from Uncovered Interest Parity

Abstract

This paper contributes to the uncovered interest parity literature by analyzing the role of financial market concentration in determining deviations from the uncovered interest parity condition. The theoretical section of the paper demonstrates that countries with concentrated financial markets will increase their welfare by discouraging financial flows through their manipulation of domestic interest rates. The empirical results support this finding and indicate that the correlation between financial flows and excess returns changes when the concentration ratio is above 0.68. This article suggests that the recent increases in banking sector concentrations around the world are likely to limit international financial flows and this will have welfare implications in countries with concentrated and competitive financial markets.

Keywords:

• Financial market integration
• uncovered interest parity
• emerging markets
• bank concentration

Notes

¹In order to show the robustness of my results to various estimation techniques, the results will be presented for pooled regression, within and between regressions in this and the following sections.

²While within-estimations are the least square dummy variable estimations here, between-effect estimates are based on cross-sectional group means.

³Tables 5, 6 and 11 are replicated by using alternative measures of market power to test the sensitivity of results and they are found to be robust. Details of sensitivity analysis are available in the working paper version and also available upon request.

⁴Since pooled regressions overlook cross-sectional heterogeneity, I will rely on the threshold found from the fixed and between effect estimations in the forward premium bias analysis below.

⁵Table 7 shows how the countries in my sample did in terms of their concentration ratios in the time period we analyzed. For each country in the sample, I have 24 observations. This table shows how many quarters their concentration values were higher than the threshold value, 0.68. The table suggests that 1/3 of our sample was below the threshold.

⁶It is harder to apply this paper's approach to developed countries since most of the developed countries have low concentration ratios in my time period. For instance, among the developed countries analyzed in Bansal (1997) only Germany passes the threshold level in my time period for a very limited number of observations and, unsurprisingly, the results did not hold for this country.