Policy Uncertainty and the Demand for Money in Korea: An Asymmetry Analysis

ABSTRACT

Previous studies included money supply volatility as well as output volatility as measures of uncertainty in estimating the demand for money. However, a more comprehensive measure of uncertainty is now constructed for many countries and is known as policy uncertainty. When we included this new measure in the formulation of the demand for money in Korea and relied upon a nonlinear specification of the money demand which allows us to assess the asymmetric effects of changes in the policy uncertainty measure, we found asymmetric long-run effects of policy uncertainty on the demand for cash in Korea. Our conjecture is that increased uncertainty induces Koreans to hold less cash in favor of safer assets and decreased uncertainty has opposite effects, though at different rate.

KEYWORDS:

• Policy uncertainty
• asymmetry
• money demand
• Korea

JEL CLASSIFICATION:

• F40

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Bahmani-Oskooee and Baek (2017) have reviewed these studies and there is no need to repeat them here.

2 See Economic Uncertainty Policy Group:  http://www.policyuncertainty.com/korea_monthly.html

3 In relation to Friedman’s volatility hypothesis, the volatility of money supply, perhaps, reflected in too frequently announcement of money and interest rate policy is reflected in the news, and eventually in the measure of policy uncertainty.

4 The above specification closely follows Bahmani-Oskooee, Bahmani, Kones, and Kutan (2015) and Bahmani-Oskooee, Kutan, & Kones (2016).

5 Note that as the Appendix reveals, by way of construction a decline in the nominal effective exchange rate, EX, reflects depreciation of won.

6 Note that the theoretical foundation of the demand for money originates from the Quantity Theory of Money. Including the exchange rate in money demand has its origins in the writings of Mundell (1963) who mentioned it without much elaboration or derivation. Others (e.g. Arango & Nadiri, 1981) tried to justify Mundell’s conjecture through an empirical approach. Additional examples include Bahmani-Oskooee (1991, 1996, 2001), Bahmani-Oskooee and Malixi (1991), Bahmani-Oskooee and Shabsigh (1996), Bahmani-Oskooee and Bohl (2000), Bahmani-Oskooee and Chomsisengphet (2002), Bahmani-Oskooee and Rehman (2005), Bahmani-Oskooee and Economidou (2005), Bahmani-Oskooee and Karacal (2006), Bahmani-Oskooee and Wang (2007), Bahmani-Oskooee and Gelan (2009), Bahmani-Oskooee and Xi (2011, 2014)), Bahmani-Oskooee et al. (2012, 2013), Bahmani-Oskooee and Kones (2014), Bahmani-Oskooee and Bahmani (2015), Bahmani-Oskooee, Halicioglu, and Bahmani (2017), and Bahmani-Oskooee and Baek (2017).

7 Note that another advantage of this approach is that the short-run dynamic adjustment process allows all feedback effects among variables to take place in estimating the long-run coefficient estimates, resolving the endogeniety issues. Pesaran et al. (2001, p. 299) alert us to this issue by writing ‘our approach is quiet general in the sense that we can use a flexible choice for the dynamic lag structure in … as well as allowing for short-run feedbacks.’

8 See Shin et al. (2014, p. 291).

9 For another application of this approach to money demand in Australia, see Bahmani-Oskooee and Maki Nayeri (2017).

10 For some other application of these methods in recent literature, see Gogas and Pragidis (2015), Durmaz (2015), Al-Shayeb and Hatemi-J (2016), Lima, Foffano Vasconcelos, Simão, and de Mendonça (2016), Bahmani-Oskooee and Fariditavana (2016), Nusair (2017), Aftab, Shah Syed, and Katper (2017), Arize, Malindretos, and Igwe (2017), and Gregoriou (2017).

11 Note that we also included a dummy variable in all models to account for 1997 Asian Financial Crisis. Since our study period is dominated by post 1997 period, the dummy was insignificant.

12 We also suspected that different lag length may contribute to the difference. To that end, we estimated the linear model by imposing the same number of lags as in the nonlinear model. Since ΔPOS took six lags but ΔNEG took five lags, we estimated the linear model once by imposing six lags on the policy uncertainty measure and once by imposing five lags. Although the size of long-run estimates was slightly different, there was no change in their significance.

Additional information

Notes on contributors

Mohsen Bahmani-Oskooee

Mohsen Bahmani-Oskooee is a University Distinguished Professor and Patricia and Harvey Wilmeth Professor of Economics at the University of Wisconsin-Milwaukee. He is also the director of the Center for Research on International Economics at the same institution. He has published extensively in international finance and open economy macroeconomics including papers in the Journal of Political Economy, the Review of Economics and Statistics, Journal of Development Economics, Economic Development and Cultural Change, Southern Economic Journal, etc. He currently serves as the editor of the Journal of Economic Studies.

Majid Maki Nayeri

Majid Maki Nayeri is a Ph.D. student in economics at the University of Wisconsin-Milwaukee. He has published articles in Australian Economic Papers, Quarterly Review of Economics and Finance, and International Economic Journal.