Productivity bias hypothesis: evidence from South Asia

Abstract

Most existing studies have attempted to test the productivity bias hypothesis by making use of the cross-section data. This article utilizes country-level time series data from Penn World Tables to examine the productivity bias hypothesis for five South Asian economies (namely, Bangladesh, India, Nepal, Pakistan and Sri Lanka). We make use of Johansen’s cointegration approach and vector error correction modelling. The empirical analysis presented in this article shows that purchasing power parity theory holds for all countries considered. However, the productivity bias hypothesis appears to hold only in the case of Bangladesh and Nepal.

Keywords:

• purchasing power parity
• exchange rate determination
• productivity bias hypothesis
• South Asia

JEL Classification:

• E44
• F36
• F47

Notes

¹ For an excellent discussion of the related issues, see Ericsson and Irandoust (2004). As far as the broader literature is concerned, the related issues of productivity gains and home bias in consumption are discussed by, among others, Lee (2014).

² For example, see Bergstrand (1992), Asea and Mendoza (1994), Rogoff (1996), Bahmani-Oskooee and Nasir (2004) and Ericsson and Irandoust (2004). In a recent study, Bahmani-Oskooee and Gelan (2006) have attempted to test the productivity bias hypothesis for Chile, Colombia and Costa Rica by making use of unofficial (black market) exchange rates.

³ For Bangladesh and Nepal, respectively, the sample period starts from 1959 and 1960.

⁴ See Bahmani-Oskooee and Gelan (2006) and references therein.

⁵ Within the context of this article, the estimated coefficient is expected to be negative because exchange rates are defined as domestic currency units per US dollar, which means that a negative change represents an appreciation.

⁶ A useful discussion of unit root testing can be found in Greene (2012) and Verbeek (2012).

⁷ In the case of Nepal, the estimated value of the trace statistic is statistically significant only at the 10% level of significance.

⁸ These graphs depict some interesting trends. The trends confirm the validity of our empirical results. In the case of Bangladesh and Nepal, productivity shocks causes an appreciation of the real exchange rate. In case of Pakistan and Sri Lanka, productivity shocks lead to depreciation of the real exchange rate. However, in case of India, productivity shocks initially led to exchange rate appreciation in the first six periods but this effect is reversed thereafter.