Oil Price Shocks in Major Emerging Economies.

• Author: Azad, Nahiyan Faisal

1. INTRODUCTION

The seven largest emerging market economies (henceforth EM7)--Brazil, China, India, Indonesia, Mexico, Russia, and Turkey--could be double the size of their advanced counterparts, the G7--Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States--by 2040, according to the estimates by Hawksworth et al. (2017). This is a massive shift in world economic power from advanced economies to emerging market economies, especially striking since two decades ago these economies were half the size of their advanced counterparts. Hawksworth et al. (2017) also estimates that by the year 2050, the EM7 economies could increase their share of world gross domestic product to 50% from approximately 35% today. In fact, based on GDP at purchasing power parity, China could be the largest economy in the world, followed by India and Indonesia in fourth place. The EM7 economies will be the primary drivers of world economic growth, growing at an estimated average rate of 3.5% per annum for every year up to 2050, dwarfing the 1.6% annual growth rate of the advanced G7 countries. See Hawksworth et al. (2017) for more details.

With the shift in global economic power to emerging market economies, it is important to examine the vulnerability of these economies to shocks that might have adverse effects on real economic activity in these economies. There is a vast empirical literature that investigates whether positive shocks in the price of oil lead to recessions in advanced countries like the United States--see, for example, Edelstein and Kilian (2009), Elder and Serletis (2010), and Kilian and Vigfusson (2011). Bredin et al. (2011) report that among the G7, uncertainty about oil prices has had an adverse effect on manufacturing activity in Canada, France, the United Kingdom, and the United States. Investigation of similar shocks to members of the EM7 countries has been an area in the empirical literature that has been relatively understudied.

This paper contributes to the empirical literature through its investigation of how oil price shock affect economic activity in the EM7 countries, of whether oil price uncertainty affects real economic activity in the EM7, and whether the relationship between oil prices and the level of economic activity in the EM7 countries is asymmetric. In doing so, we use two classes of empirical models. In particular, we extend the Elder and Serletis (2010) model, incorporating aspects of the Kilian (2009) and Kilian and Park (2009) methodology, to investigate the effects of oil price uncertainty. We also use the Kilian and Vigfusson (2011) methodology to test for symmetry in the impulse responses of real output to positive and negative oil price shocks in the EM7 countries.

In the context of a multivariate GARCH-in-Mean VAR specification, that controls for lagged changes in global crude oil production and world economic activity, we find that oil price uncertainty has a negative and statistically significant effect on real output in India, Indonesia, Mexico, Russia, and Turkey, and a positive and statistically significant effect on real output in Brazil and China. We also find that the responses of real economic activity to oil price shocks in China, India, Indonesia, Mexico, and Turkey are symmetric and those in Brazil and Russia are asymmetric.

The remainder of the paper is organized as follows. In Section 2, we discuss the multivariate GARCH-in-Mean VAR model, incorporating demand and supply side shocks in the world crude oil market, as in Kilian (2009) and Kilian and Park (2009). In Section 3, we discuss the data and their time series properties, and in Section 4 present the empirical results regarding the effects of oil price uncertainty. In section 5, we investigate whether the relationship between real economic activity and the real oil price is nonlinear and asymmetric and in doing so we use the Kilian and Vigfusson (2011) tests of the null hypothesis of symmetric impulse responses. The final section concludes the paper.

(2.) THE MULTIVARIATE GARCH-IN-MEAN VAR

Elder and Serletis (2010) use the Elder (2004) model and investigate the relationship between the price of oil and the level of economic activity, focusing on the role of uncertainty about oil prices. In doing so, they utilize an internally consistent bivariate GARCH-in-Mean structural VAR that accommodates an independent role for the effects of oil price volatility. They find that volatility in oil prices has had a negative and statistically significant effect on several measures of investment, durables consumption, and aggregate output. They also find that accounting for the effects of oil price volatility tends to exacerbate the negative dynamic response of economic activity to a negative oil price shock, while dampening the response to a positive oil price shock.

In this section, we follow Elder and Serletis (2010) and consider an extension of their bivariate structural GARCH-in-Mean VAR model to investigate the relationship between oil price uncertainty and real output, after controlling for global crude oil production and world economic activity, building on the work by Kilian (2009) and Kilian and Park (2009). Existing studies that analyze the relationship between the price of oil and real output in the context of bivariate models suffer from the limitation that oil prices are assumed to be strictly exogenous with respect to the global economy. In this regard, Kilian (2009) provides empirical evidence that fluctuations in global macroeconomic activity have an impact on the price of oil. Therefore, in what follows, we investigate the relationship between oil prices and domestic real output, after we control for economic variables that drive both the price of oil as well as domestic real output.

We assume that the dynamics of the structural system can be summarized by a linear function of the relevant vector of macroeconomic variables, modified to allow the conditional volatility of the real price of oil to affect the conditional mean

[Please download the PDF to view the mathematical expression] (1)

where [z.sub.t] is a column vector in the percentage change in global crude oil production, A In prod,, world economic activity, [wea.sub.t], percentage change in real price of oil, [DELTA]In [0.sub.t], and the growth rates of real output in each of the EM7 countries, [DLETA]In [y.sub.r] In equation (1), dim(B) = dim([[GAMMA].sub.j]) =(4x4) and [Please download the PDF to view the mathematical expression] being the variance-covariance matrix. Also,

[Please download the PDF to view the mathematical expression]

The model is identified by imposing a sufficient number of exclusion restrictions on the B matrix. In this four variable structural VAR case, we estimate n(n -1) / 2 = 6 free parameters in B, subject to a rank condition, such that the diagonal elements of B are assumed to be equal to 1 and B is assumed to be lower triangular. The block-recursive structure of B implies that world crude oil production, world real economic activity, and the real price of oil are predetermined with respect to domestic real output. Global oil production is assumed to be exogenous to the other three variables (it is affected only by a shock to itself but is unaffected by instantaneous feedback from the other variables). Thus, our recursive factorization structure imposes six exclusion restrictions on the B matrix, satisfying a rank condition. The restrictions on the B matrix allow us to differentiate between three structural shocks that affect the real price of oil, namely unanticipated changes in world oil production...