Why the Effects of Oil Price Shocks on China's Economy are Changing.

AuthorWang, Shouyang

    The effects of oil price shocks on the macroeconomy have attracted much attention. Many oil-importing countries see a correlation between increases in oil prices and subsequent economic downturns (e.g., Hamilton, 1983; Dhawan and Jeske, 2008, 2010; Gronwald, 2012; Abhyankar et al., 2013; Morana, 2014; Huntington, 2017; Guesmi et al., 2018). However, some recent studies reveal these effects have changed with time (Lescaroux, 2011). For example, the effects of oil price shocks in America have decreased since the 1980s. Reasons for the changes can be summarized in four aspects. First, oil price shocks have changed (e.g., Blanchard and Gali, 2007; Kilian, 2009; Chen, 2009b; Hamilton, 2009; Katayama, 2013), such as a shift from oil supply shocks to oil demand shocks since 2003 (Kilian, 2009). Second, the features of oil consumption have changed, such as deregulation in oil consumption sectors, improved energy efficiency (Katayama, 2013), a smaller share of oil used for production (e.g., Blanchard and Gali, 2007; Miguel, 2009; Bachmeier and Cha, 2011), changes in both the demand and supply elasticity of oil (Baumeister and Peersman, 2013; Ven and Fouquet, 2017), and the declining importance of oil consumption sectors (Edelstein and Kilian, 2007). Third, the transmission of oil price shocks has changed through factors such as a more flexible labor market (Blanchard and Gali, 2007), declining effects of oil price shocks on expectations and learning (Milani, 2009), and a decreasing exchange rate pass-through (De Gregorio et al., 2007). Fourth, the role of the monetary policy has changed. Conclusions regarding this aspect are inconsistent. Some researchers indicate that the stronger commitment of central banks to maintaining a stable rate of inflation contributes to a reduction in the impacts of oil price shocks on both inflation and the output (e.g., Blanchard and Gali, 2007; Chen, 2009a; Miguel, 2009). Other researchers-such as Bernanke et al. (1997, 2004), Hooker (2002), Leduc and Sill (2004), Herrera and Pesavento (2009), and Nakov and Pescatori (2010)-show that the reaction of the monetary policy to oil price shocks (instead of oil price shocks themselves) causes the dynamics of the macroeconomy. In addition, Barsky and Kilian (2001) argue that aggregate fluctuations can be partly attributed to an exogenous monetary policy that happens simultaneously rather than oil price shocks.

    In China, economic growth is highly oil intensive and is accompanied by low efficiency (Tang et al., 2010). Existing studies show that a correlation exists between oil price shocks and macroeconomic dynamics in China (Lin and Mou, 2008; Mu and Ye, 2011; Ju et al., 2014; Ji et al., 2015; Wei and Guo, 2016; Zhang and Broadstock, 2017; Ji and Zhang, 2018; Ji et al., 2018), but conclusions are inconsistent. Du et al. (2010) indicate that oil price shocks positively affect economic growth and inflation in China, while Zhang and Xu (2012) find that energy prices do not have a significant impact on energy consumption and economic growth because energy prices are regulated. Some researchers have revealed that the effects of oil price shocks in China are time varying. For example, Kim et al. (2017) find that oil price shocks have negative effects on interest rates in the earlier sample period (1992:4-2001:10) but positive effects in the subsequent period (2001:11-2014:5). Moreover, oil price shocks have played an increasingly important role in driving the volatility the interest rates. Gong and Lin (2018) indicate that the effects of oil price shocks on China's economy not only evolve with time but also change directions from 1995 to 2015. Cross and Nguyen (2018) conclude that the responses of both real GDP growth and inflation to energy price shocks have declined from 1994 to 2016. However, few studies explore the problem synthetically.

    How do macroeconomic factors impact the relationship between oil price shocks and the macroeconomy? The answers for China may differ from the conclusions for some developed oil-importing countries because of the following two considerations. First, China is a net crude oil-importing country, and its dependency on imported crude oil has gradually increased over the last twenty years. This situation is different from some developed countries for which oil import dependency has decreased (United States (1)) or remained relatively stable (Japan and most countries in the European Union) over the same period. Second, China has undergone a series of economic reforms and structural changes (Huang et al., 2019) and has been a typical transition economy since the 1970s (Campos and Coricelli, 2002). Stylized facts and many deep parameters in the structural model of China have changed over time. That situation is very different from the situation in some developed countries that have relatively stable parameters (such as the U.S.). Therefore, our paper explores how the effects of oil price shocks on China's macroeconomy change over time and discusses possible reasons for the changes.

    To reveal how the economic responses to oil price shocks have changed over time, a time-varying parameter vector autoregressive (TVP-VAR) model for China's economy is established and estimated. The results present significant time-varying effects of the macroeconomic responses to oil price shocks. Changes of effects can be attributed to several factors. Including more factors in the VAR model will increase the model dimension and require a longer data sample. These requirements cannot be supported by the available data. In addition, some factors are difficult to quantify into time series data (such as price stickiness or monopoly power), and are therefore difficult to include in the VAR model. Considering these two reasons, a more comprehensive structural model is developed to investigate how macroeconomic factors impact the economy's vulnerability and resilience to oil price shocks in China. The model is a New Keynesian Dynamic Stochastic General Equilibrium (DSGE) model with sticky prices and adjustment costs. It is extended in two aspects to better fit China's macroeconomy. First, the Cobb-Douglas production function is augmented by adding oil as an input element that provide channels for analyzing both the nominal and the real effects of oil price shocks. Second, the transportation sector is incorporated in the model, which is crucial for the conduction of oil price shocks in China.

    The main findings of this paper can be summarized as follows. First, oil price shocks have significant effects on the macroeconomic dynamics in China. The effects are decreasing but fluctuating from 1997 to 2018. Second, the responses of the real output are much greater and last longer than those of inflation. Third, among all the factors, decreasing oil intensity and monopoly power play crucial roles in increasing the economy's resilience to oil price shocks, while increasing capital intensity in production is an important factor that amplifies the responses. Changes in price stickiness have only weak effects on the responses to oil price shocks. In addition, the monetary policy with less inflation targeting increases interest rates and inflation but has insignificant impacts on the other variables. Oil inflation targeting in the monetary policy and the deregulation of refined oil prices matter only in the very short term (within half a year).

    The remainder of this paper is organized as follows: Section 2 provides empirical evidence for the changing effects of oil price shocks in China. Section 3 describes the structure of the DSGE model. Section 4 presents the calibrated parameters. The results are provided in section 5. A discussion and the conclusions are provided in section 6.


    To examine whether the oil price-macroeconomy relationship in China has changed, a TVP-VAR model specification proposed by Primiceri (2005) is used:

    [y.sub.t] = [c.sub.t] + [B.sub.L,T][Y.sub.T-1] +... + [B.sub.k,t][Y.sub.t-k] +[u.sub.t] (1)

    where [y.sub.t] is a 4*1 vector of variables; [c.sub.t] is a 4*1 vector of the timevarying coefficients of intercept terms; [B.sub.k,t] is a 4*4 matrix of the time-varying coefficients; [u.sub.t] is a 4*1 vector of heteroskedastic unobservable shocks with variance covariance matrix [[OMEGA].sub.t], which can be transformed as [A.sub.t] [[OMEGA].sub.t] [A'.sub.t]; [A.sub.t] is a time-varying lower triangular matrix, and [[OMEGA].sub.t] is a time-varying diagonal matrix. All time-varying parameters are modeled as random walks.

    Considering the small sample size of Chinese macro variables, monthly data are used to estimate the model. Variables used in this model include the output, inflation oil price shocks, and external demand. The output is represented by the real growth rate of industrial added value (IAV) due to the lack of monthly GDP data. Inflation is represented by the growth rate of the consumer price index (CPI), which is consistent with other studies, such as Kilian (2009), Wei and Guo (2016), Gong and Lin (2018).

    The net oil price increase (NOPI) proposed by Hamilton (2003) is used to represent oil price shocks. It is defined as the oil price deviation from the mean of the past three years. (2) Oil price (3) is the West Texas Intermediate (WTI) crude oil spot price deflated by the U.S. CPI. Although Kilian (2016) points out that Brent is a better benchmark for the world crude oil price between 2010 and 2015, WTI is recognized as more representative of the global oil price and leads Brent during most of the data sample (Lu et al, 2014). A robustness check using Brent instead of WTI in the TVP-VAR model is provided in Appendix A.

    External demand is included in the model since it has been a major driving force of the rapid economic growth in China over the past two decades. Moreover, it is an important transmission channel of global oil price shocks to domestic...

To continue reading

Request your trial