Do exchange rate changes have symmetric or asymmetric effects on the demand for money in Turkey?

ABSTRACT

As a result of the research conducted by Nobel Laureate Robert Mundell (1963), most studies estimating the demand for money today do include the exchange rate in their specification to account for currency substitution. Previous studies that did this for the Turkish demand for money assumed that exchange rate changes do have symmetric effects on the demand for money in Turkey. In this article, we question this assumption. By using the nonlinear ARDL approach, we show that indeed exchange rate changes do have short-run and long-run asymmetric effects on the M1 demand for money. Introducing nonlinearity also yields a stable money demand.

KEYWORDS:

• Money demand
• exchange rate
• asymmetry
• nonlinear ARDL
• Turkey

JEL CLASSIFICATION:

• F31
• E40

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ One assumption is that the money market is at equilibrium and, therefore, the quantity of the money supplied is the same as the quantity of money demanded.

² The study period was 1949:03 – 1987:04.

³ Mutluer and Barlas (2002, p. 17).

⁴ For more on this, see Laidler (1993) and Bahmani-Oskooee and Bohl (2000).

⁵ For details on normalization, see Bahmani-Oskooee and Tanku (2008).

⁶ For the application of this approach in some other areas, see Halicioglu (2007, 2013), Narayan et al. (2007), Tang (2007), Mohammadi, Cak, and Cak (2008), Wong and Tang (2008), De Vita and Kyaw (2008), Payne (2008), Bahmani-Oskooee and Gelan (2009), Dell’Anno and Halicioglu (2010), Chen and Chen (2012), Wong (2013), Hajilee and Al-Nasser (2014), and Tayebi and Yazdani (2014).

⁷ Note that since the money market is assumed to be at equilibrium, the quantity of money supplied is equal to quantity demanded.

⁸ This critical value comes from Pesaran, Shin, and Smith (2001, Table CI-Case III, page 300) and it is for a model that includes three exogenous variables. When we move to the nonlinear model, where there are four exogenous variables, the comparable critical value is 4.01.

⁹ The critical value is, again, for the case of three exogenous variables and comes from Banerjee, Dolado, and Mestre (1998, Table 1, p. 276). When there are four variables, the comparable value is −4.03.

¹⁰ For a graphical representation of the CUSUM and CUSUMSQ tests in this context, see Bahmani-Oskooee and Fariditavana (2015).

¹¹ Note that CUSUMSQ is indicated by CUSMSQ and instability is indicated by NS for not stable.

¹² These are supported by insignificant Wald statistics, both in the short run and long run.