Asymmetric oil price transmission to the purchasing power of the U.S. dollar: A multiple threshold NARDL modelling approach

Abstract

This paper studies, whether the responses of the purchasing power of the U.S. dollar asymmetric to crude oil price fluctuations? A multiple threshold nonlinear autoregressive distributed lag model (MTNARDL) is developed to address this question. Working on the monthly data from January 1990 to June 2019, this study shows a short-run asymmetric transmission of oil price fluctuations into the purchasing power of the U.S. dollar. The results of the MTNARDL model apprehend, more minutely, the magnitude of variations in purchasing power in response to minor to major change in the oil price. It finds corroborative evidence for more rapid response of the purchasing power to the upward oil price shocks than to the downward movement in the oil price. This implies that the purchasing power of the U.S. dollar experiences sharp reduction with the rise in the oil price, nonetheless, the advantage of a drop in the oil price is not entirely transmitted to the purchasing power as its correction takes much longer period than expected.

Introduction

Effect of crude oil price fluctuations is experienced worldwide. As documented in several studies (e.g., Hamilton, 1983, 2011; Karras, 1993; Dogrul and Soytas, 2010), large fluctuation in crude oil prices since 1970s has been associated with economic downturn. Even though there holds some deliberation as to whether crude oil price shocks are primarily responsible for recession (e.g., Kilian, 2014), it is widely accepted that crude oil price variability, at least partially, passed-through into the purchasing power (Edelstein and Kilian, 2009).

Purchasing power represents the worth of any currency stated in terms of the volume of services and/or goods that a single unit of currency could buy. As suggested by Edelstein and Kilian (2009), mechanisms by which energy price changes directly affects the purchasing power of money are as follows: First, increasing energy prices would leave a lesser fund with consumers after settling their energy bills, known as discretionary income effect. Second, consumers facing a fluctuating energy price might be uncertain regarding its future path and would postpone the purchase of consumer durables (Bernanke, 1983; Pindyck, 1991), termed as uncertainty effect. Third, consumption might experience a slowdown while savings might increase in anticipation of future unemployment, a negative outcome of higher energy price on output (Gokmenoglu et al., 2015), referred as precautionary savings effect. Fourth, usage of durables that were complementary in consumption of energy (e.g., automobiles, motor cycles) would be postponed or even might be foregone to save higher operating cost (Hamilton, 1988), identified as operating cost effect. As, in the short-run, those energy-complementary durables face a relatively elastic demand in response to the change in gasoline price, a marginal increase in the price of gasoline would lead to a substantial drop in the demand of automobiles. Furthermore, as the dollar value of automobiles is considerably high as compared to the price of gasoline, even a marginal rise in the energy price would result in a major fall in the revenue of the firms who manufacture energy-complementary durable goods. Such major changes in the energy-intensive durable industry is argued to carry a major inter- as well as intra-industry reallocation of capital and labor. Such reallocation effect is expected to carry a significant influence on slow economic growth and unemployment (Bresnahan and Ramey, 1993; Davis and Haltiwanger, 2001; Lee and Ni, 2002) which is likely to further limit the purchasing power.

Primarily the uncertainty effect and the reallocation effect are more likely to yield asymmetric response in the purchasing power (Hamilton, 2011). Both these effects are believed to be amplifying the impact of rise in oil price while dampening the effect of oil price fall on the purchasing power. In response to the rise in energy price, a consumer, expecting further increase in the energy price, would postpone the decision to purchase high-ticket energy-complementary durable goods, thus pushing the economy to reallocate the capital and labor. However, in the event of downward correction of crude oil price, consumers still remain uncertain about direction of the energy price movement, at least in the short-run, and they do not take immediate purchase decision of energy-intensive durables. Consumers are likely to watch the energy price movement for a relatively longer period before becoming optimistic about the energy price and then decide to purchase costly energy-complementary durable goods. This involves longer time period for the economy to recover and dampens the effect of energy price fall on the purchasing power.

The asymmetric response of macroeconomic indicators to rise and fall of energy price still remains a debatable topic in the literature. Hamilton (2003, 2011) argues that transmission of the oil price variations to macroeconomic indicators including inflation and output are asymmetric in nature. An asymmetric price transmission indicates that inflation and output responded rapidly to the upsurge in the oil price than to the decrease. Similarly, Hammoudeh and Reboredo (2018) find support for an asymmetric effect of oil price fluctuations on inflation risk premium. In a more recent investigation, in the context of inflation targeting nations, López-Villavicencio and Pourroy (2019) report that oil price transmission is higher during oil price decrease compared to oil price increase. However, there are dissenting voices, such as Kilian and Vigfusson, 2011, Kilian and Vigfusson, 2013 that argue for a more symmetric transmission of crude oil price shocks to the macroeconomic aggregators. We add to this thin yet growing body of literature by exploring the following research question: whether the responses of the purchasing power of the U.S. dollar asymmetric to crude oil price fluctuations? Understanding how sensitive is purchasing power to crude oil price variability is important as most monetary authorities aim to increase the purchasing power of money by controlling inflation. Knowledge of the extent to which positive crude oil price variations would have larger impact than negative crude oil price variations will help monetary authorities to craft appropriate policies to adjust such fluctuations.

On the methodological front, we differ from earlier contributions by employing a multiple threshold nonlinear autoregressive distributed lag model (MTNARDL). MTNARDL extends the seminal work of Shin et al. (2011, 2014) by offering a generalization of the two-way classification nonlinear ARDL model (NARDL). The original NARDL model decomposes series into two partial sum series for evaluating effect of positive and negative changes of independent variables. We find that, by splitting the underlying independent variable series into multiple thresholds, nonlinear analysis can be extended beyond asymmetric analysis of positive and negative changes of independent variables. MTNARDL allows splitting of the independent variable in quantiles at pre-determined thresholds. Therefore, MTNARDL is expected to offer greater precision in capturing asymmetry in oil price transmission to purchasing power compared to the conventional two-way nonlinear ARDL approach as adopted by Lacheheb and Sirag (2019).

We add to the literature on asymmetric price transmission by providing corroborative evidence for more rapid response of the purchasing power of the U.S. dollar to the upward crude oil price shocks than to the downward movement in the price of crude oil, in the short-run. This indicates, in the short-run, the purchasing power of the U.S. dollar drops at a faster rate in response to the increase in oil price changes, while, recovers its strength at a much slower rate with the decrease in oil price changes. The results of the MTNARDL model apprehend, more minutely, the magnitude of variations in purchasing power of the U.S. dollar in response to minor to major fluctuation in the oil price.

Our results would be useful for the policy makers for setting policy rates as well as deciding on the stimulus for the economy. For investors, our results would be helpful as the interest of short-term investors differ from that of the long-term investors. While short-term investors would be inclined to understand the short-run behavior of an asset, a long-term investor would take interest in assets over a longer time horizon (Devpura et al., 2018; Pal and Mitra, 2019). As our results unearth a short-run asymmetric transmission of oil price to the purchasing power of the U.S. dollar, short-term investors would be benefitted while building a portfolio of crude oil with other assets and more specifically with assets from energy-complementary durable industry (e.g., automobile).

Section 2 briefly reviews the relevant literature. Section 3 deals with data followed by an outlay of the econometric specifications in Section 4. Section 5 highlights the results, discusses the outcome of analysis and validates the asymmetric transmission of oil price variations into the purchasing power of the U.S. dollar. The final section concludes.