Inflation targeting and the forward bias puzzle in emerging countries

Highlights

• In emerging countries, inflation targeting leads to the appearance of the forward bias puzzle.

• This reflects the performance of inflation targeting in lowering the level and volatility of inflation.

• This result is robust to various specifications and estimation techniques.

Abstract

Based on quarterly data on 31 emerging countries (among which 16 are inflation targeting countries) from 1990Q1 to 2014Q3, we obtain a strong support for the conjecture that the implementation of inflation targeting weakens the Fisherian relation between expected depreciation and the interest rate differential (uncovered interest parity condition) and thus is conducive to the appearance of the forward bias puzzle in emerging countries. We show that this reflects the performance of inflation targeting regimes in lowering the level and volatility of inflation. Our finding holds when controlling for country-specific effects, time-specific effects, global disinflation, exchange rate management, crises, and using different econometric techniques.

Introduction

After being initiated by New Zealand in 1990, inflation targeting (hereafter IT)¹ has been adopted by a large number of industrial and emerging countries as a strategy to conduct monetary policy. Several studies have found evidence that adopting an IT strategy has lead to overall economic performance (Bernanke and Mishkin, 1997, Svensson, 1997, Mishkin and Schmidt-Hebbel, 2007). More specifically, the adoption of IT by some emerging countries has helped to reduce the level and volatility of inflation in these countries (Gonçalves and Salles, 2008, Lin and Ye, 2009, De Mendonça and Guimaraes e Souza, 2012).

The contribution of this paper is to examine the impact of IT on the uncovered interest parity (hereafter UIP) condition in emerging countries through the reducing effect of IT on inflation uncertainty or a greater weight on the inflation in monetary policy rule during IT. The forward bias puzzle (hereafter FBP) referring to the failure of UIP to hold is well documented (Fama, 1984, Bansal and Dahlquist, 2000, Chinn and Meredith, 2004, Ito and Chinn, 2007, Frankel and Poonawala, 2010). Its corresponds to a slope coefficient less than 1 in the so-called Fama regression (regression of the expected currency depreciation and interest rate differential). The forward premium puzzle (hereafter FPP) (i.e. a negative slope coefficient in the Fama regression) which is an extreme form of the FBP occurs in developed economies but not in emerging countries. This feature is explained by the high level of inflation uncertainty in the latter ones (Bansal and Dahlquist, 2000, Frankel and Poonawala, 2010). More specifically, as noticed by Bansal and Dahlquist (2000), the absence of FPP in emerging countries is consistent with the reasoning developed in models incorporating non-Fisherian effects (Lucas, 1990, Fuerst, 1992). In these models, in presence of high inflation uncertainty or high expected inflation the Fisherian effects are overwhelming. However, in low inflation environments, the non-Fisherian fundamentals may be predominant. As emerging economies display large inflation uncertainty and/or high expected inflation the Fisherian relation between expected depreciation and the interest rate differential holds. As a result, the FPP is absent. On the contrary, as developed countries are generally low inflation economies, the non-Fisherian effects are important, leading to the FPP. Farhi and Gabaix (2016) theoretically rationalized the empirically evidence of Bansal and Dahlquist (2000) by developing a disaster-based tractable framework for the analysis of exchange rates. They show that the UIP tends to hold in countries with very variable inflation (typically countries with high average inflation) since these countries have a high risk of depreciating in bad states. Using a standard open macroeconomic model, Park and Park (2017) also theoretically rationalized the finding of Bansal and Dahlquist (2000) showing that the UIP is more likely to hold in low-inflation countries since their monetary rules likely put higher weight on inflation. Therefore, by lowering inflation or putting higher weight on inflation in policy rule, the adoption of IT in some emerging countries may lead to the appearance of the FBP in these countries, in contrast with the non-adopting IT emerging countries. We test this conjecture in this paper.

Compared with previous empirical studies on the influence of monetary policy on the UIP condition (Park and Park, 2017, Kim et al., 2017), we contribute to the literature on the FBP by investigating the link between the adoption of IT and FBP in emerging countries. Contrary to Park and Park (2017) which examine the FBP for inflation targeting and non-targeting countries, before and after the global financial crisis, we investigate the FBP in emerging countries by comparing inflation targeters before and after IT adoption and inflation targeters with non-targeters. Given its focus on inflation targeting in emerging countries, our study also differs from Kim et al. (2017) who, using data on 14 US-trading partners (advanced countries), find that UIP fails to hold during the Volker disinflation episode and tends to hold during the post-Volker era.

To test our conjecture, we conduct an empirical investigation using quarterly data on 31 emerging countries (among which 16 are IT countries) from 1990Q1 to 2014Q3. Our results are consistent with the conjecture that IT modifies the pattern of interest rates determination which reverberates on the behavior of exchange rates. Specifically, we find that the implementation of IT in emerging countries expands the FBP. We show that this reflects the fact that IT helps creating a low and stable inflation environment. This finding holds after controlling for country-specific effects, time-specific effects, global disinflation, exchange rate management, currency crises and global financial crisis. It still holds when we use alternative econometric techniques.

The rest of the paper is organized as follows. Section 2 provides the theoretical background on the link between inflation targeting and the FBP. Section 3 describes the data used in the empirical estimation. Section 4 presents the empirical analysis. Finally, Section 5 contains concluding comments.