Sage Journals: Discover world-class research

Abstract

This study attempts to investigate the asymmetric impacts of oil price and component shocks on categorical economic policy uncertainties (EPUs) in China. A novel multiple thresholds nonlinear autoregressive distributed lagged model (ARDL) was employed to capture the asymmetric impacts of oil price and shocks on EPUs in the short and long run. Additionally, the influence of the Brent oil price on integral EPU was also examined. We find that asymmetric impact is more remarkable in the long run in China. Among the four categorical EPUs, trade policy uncertainty is the most notable in most cases. Based on the results, some implications are provided to policymakers and investors.

Keywords

oil shocks categorical policy uncertainty spillover MTNARDL

Highlights

Asymmetric impact is more remarkable in the long run in China.

EPU declines at a slower rate when oil prices decrease.

Asymmetric impact of oil price shocks on trade policy uncertainty is significant.

Introduction

Crude oil, an important industrial raw material, and its price fluctuations have significant macroeconomic impacts often characterized by asymmetry (Das & Kannadhasan, 2020; Jiang & Cheng, 2021; Loungani, 1986; Mork et al., 1994; Su et al., 2021). Policymakers responsible for mitigating macroeconomic trends to maintain economic stability and growth may have to deal with a trade-off between high inflation and low production stabilization (Sun et al., 2020). Correspondingly, economic policy uncertainties (EPUs) will undoubtedly be affected by such changes in macroeconomic policy (Jiang et al., 2022). Scholars are becoming increasingly interested in the relationship between oil price shock and EPU; accordingly, a growing number of researchers have been evaluating the effect of oil price shocks on EPUs.

Theoretically, growth, investment, and demand all suffer as a result of EPUs impacting oil prices (Antonakakis et al. 2014). Notably, oil price shocks can spread throughout an economy through macroeconomic policies (e.g., monetary and fiscal policies) (Jiang & Cheng, 2021; Pieschacón, 2012). Economic policies are well-known and powerful macroeconomic tools; historically, they have played important roles during oil price shocks—in other words, oil price fluctuation is one of source of EPU (Kara, 2017). Existing studies have demonstrated that different EPUs differently relate to oil price volatility (Jiang et al., 2022; Kang et al., 2017).

The relationship, measurement, and influence of oil price shocks on EPUs have become popular topics of academic interest that have been discussed in numerous papers (Baker et al., 2016; Bloom, 2009). A widely accepted approach is to test the relationship and impact on EPU separately by first disaggregating shocks into supply, demand, and demand-specific shocks depending on the source of the oil price (Antonakakis et al., 2014; Kilian, 2009) decomposed oil prices into three types of shocks: supply shocks, aggregate demand shocks, and precautionary demand shocks, with crude oil production and the global real economic activity index (in terms of dry freight rates) chosen as proxies for oil supply and demand. The effects of different shocks are fully examined and understood by building a structural vector autoregressive (SVAR) model for the three variables. Building on the classification of oil price shocks presented by Kilian and Park (2009), Antonakakis et al. (2014) used a generalized forecast error variance decomposition (Diebold & Yilmaz, 2012) approach to examine the spillover effects between these categorized shocks and EPU. However, according to Ready (2018) and Demirer et al. (2020), this approach does not identify whether fluctuations in precautionary demand are stimulated by concerns about future oil supply or the aggregate demand for oil. Hence, Ready (2018) introduced asset prices into the SVAR model to redefine supply, demand, and risk shocks to overcome these defects. We use this approach to explore the relationship and asymmetric impact between oil price shocks and categorical EPUs (fiscal policy uncertainty, monetary policy uncertainty, trade policy uncertainty, and exchange rate policy uncertainty—henceforth, “policy uncertainty” is abbreviated as “PU”). The reason for this is that we include, in a unified analytical framework, four categories of EPU that are subject to different types of shocks presenting various outcomes: fiscal PU, monetary PU, trade PU, and exchange rate PU. Combining Ready (2018) work on the source of oil price shocks and Huang and Luk (2020) work on categorical EPUs in the same analytical framework allows for a better identification of the influence of oil component shocks on various types of EPUs and their magnitudes.

China is the world’s largest emerging country. After nearly 40 years of opening up to the outside world, it has become an important link in the global industrial chain. While China consumes the world’s largest amount of crude oil, with an external dependence of 69.8%, existing studies on how oil price shocks impact EPUs shows that EPUs differently respond to oil price shocks in the short run and the long run. In one study using a sample of BRIC countries, the impact proved weaker in the short term, but steadily strengthened over long term (Chen, Sun, & Wang, 2019), and the same trend was observed in a study of G7 countries (Sun et al., 2020); notably, this trend has demonstrated asymmetric features (Das & Kannadhasan, 2020; Su et al., 2021; You et al., 2017). Meanwhile, Jiang et al. (Jiang et al., 2022) used a multi-regime-on-scale approach to investigate how fiscal PU (FPU) and monetary PU (MPU) in China responded to global oil price volatility; the study found that China’s FPU and MPU volatility only responded significantly to global oil price volatility in the intermediate to long term in a low oil price volatility regime. Relatively speaking, MPU more strongly reflects the volatility link between China’s EPU and oil prices than FPU. Meanwhile, oil price shocks affect not only monetary and fiscal policies, but also trade (Wei, 2019) and exchange rate policies (Zhu & Chen, 2019) and are thus sources of trade PU and exchange rate PU. On the one hand, because crude oil is a globally traded commodity dominated by the US dollar, crude oil price shocks may affect exchange rate PU; on the other hand, they may also affect trade PU in the context of cost and substitution effects. Wang and Lee (2022) evaluated the associations between categorical policy uncertainty and global oil returns; they found that exchange rate PU has the largest negative average correlation with global oil returns, followed by FPU (negative), MPU (positive), and, finally, trade PU (negative). In contrast to much of the existing literature, Wang and Lee (2022) situate categorical PUs in a non-linear and dynamic analytical framework; however, they exclude the source of oil price shocks and asymmetric impacts from their analysis.

To complement the prior literature, we employ a novel multiple threshold nonlinear autoregressive distributed lag (MTNARDL) model. While we do this, like Pal and Mitra (2019), to analyze the integral oil price, we also assess the asymmetry of component shocks. To clarify how oil price shocks impact China’s EPUs, we classify oil price shocks into supply, demand, and risk shocks (Ready, 2018) and EPUs into fiscal PUs, monetary PUs, trade PUs, and exchange rate PUs. The reason for classifying oil price shocks in this way is that demand and supply shocks stimulate the price of oil, just as they stimulate the price of any other commodity. Not only do different shocks have various time-varying implications on the real oil price, but they also have diverse dynamic effects on economies. In particular, global demand shocks may have direct and indirect repercussions for the economy by impacting the prices of oil and other industrial commodities (Kilian, 2009).

This paper contributes to existing literature in three main ways. First, this study sought to generate a more comprehensive conclusion on the relationship between oil price shock and EPU by investigating the asymmetric impact of the pass-through from oil price and component shocks to China’s categorical PUs; prior literature only highlights fluctuations in oil prices, ignoring disaggregated shocks. More specifically, we add to the literature by introducing Ready’s (2018) innovative reasonable decomposition approach; that is, by analyzing the asymmetric impact of oil price shock on China’s categorical PUs by decomposing the oil price shock into a supply shock, a demand shock, and a risk shock according to the source of the shock to gain new insights on the impact of categorical EUPs, thus expanding the research perspective. Second, this study more precisely evaluates the impact of oil price shock on EPUs by employing multiple threshold models that distinguish between small to large variations in oil price shocks and capture the responses of categorical PUs, In particular, this study split supply shock, demand shock, and risk shock into different quantiles to explore asymmetric transmission across minor to major fluctuations—this approach has a significant advantage over standard nonlinear autoregressive distributed lag (NARDL) models, which simply identify oil price fluctuations. This is, as far as we know, the first study to apply these models to investigate the asymmetric impact of oil price shocks on categorical PUs in China. Finally, our analysis is unique in the present literature since it combines fiscal PU, monetary PU, trade PU, and exchange rate PU with oil price shocks in the same analytical framework; this approach stands as a valuable theoretical support for policymakers seeking to adapt policies to different types of oil price shocks.

The rest of the article is structured as follows. Section 2 provides a brief overview of existing research on oil price changes, disentangling shocks, and the EPU nexus. The methodology and data are provided in Sections 3 and 4. Section 5 presents the empirical results and discussion. Section 6 offers the conclusions.

Literature Review

Since Hamilton (1983) seminal work, early literature on the impact of oil prices has put a spotlight on two main topics: macroeconomic activity and financial markets. In terms of macroeconomic activity, the main foci have been the mechanisms and magnitudes of the effects of oil price shocks on macroeconomic variables such as output, investment, trade, and inflation (Blanchard & Gali, 2007; Cologni & Manera, 2008; Crucini, 2000; Cunado & Gracia, 2005; Du et al., 2010; Hamilton & Herrera, 2004; Jiménez-Rodríguez & Sánchez, 2005; Jones et al., 2004; Cuado & de Gracia, 2003; Peter Ferderer, 1996). The impacts of oil prices on stock markets in financial markets have also been studied (A & B, 2010; Antonakakis et al., 2014; Arouri & Rault, 2012; Filis et al., 2011; Kang & Ratti, 2013a; Kilian & Park, 2009; Lee & Ni, 2002; Maghyereh, 2006; Papapetrou, 2001). When oil price shocks are considered as whole (i.e., without distinguishing between the sources of the shocks), negative shocks outweigh positive ones in terms of their impact. However, as research progresses, there are some extent asymmetries in the impact of the oil price on either macroeconomic or financial markets.

Regarding the influence of oil shocks on EPUs, numerous studies suggest a direct relationship between oil prices and economic policy (Gelb, 1988), advising that oil price volatility significantly affects short-term uncertainties. Hence, changes in the oil price may likewise be one of the main drivers of EPU (Barrero et al., 2017). Through an analysis of data from G7 countries, Hailemariam et al. (2019) observed the time-varying impact of oil prices on EPU (Yang, 2019; Yin, 2016) and argued that oil price shocks cause uncertainty in the economy. Studies have shown that economic activity is severely influenced by oil price. When economic policies are frequently adjusted in order to smooth out economic fluctuations, they become uncertain.

Policymakers make relevant economic policy adjustments according to economic activities in a timely manner. When output is affected, the corresponding fiscal or monetary policies may be adopted. When import and export are affected, the corresponding trade or exchange rate policies may be adopted. On a micro level, oil is the indispensable feed stock for production and oil prices determine the cost of oil and its related products. Producers will adjust their production plans to maximize profits based on oil prices, which will affect production and thus the macroeconomy (Hamilton, 1988; Sorana et al., 2018; Wen et al., 2018). From a macro perspective, higher oil prices stifle output, while lower oil prices do not positively affect output or have the same degree of impact (Mork et al., 1994). Frequent policy adjustments cause economic agents (e.g., consumers, enterprises) to have uncertain expectations and influence their economic behavior. However, the extent to which the oil price affects EPU depends on the source of the shock. For this reason, Kilian (2009) isolated the supply shocks from the real price of crude oil due to political events in the member countries of the Organization of the Petroleum Exporting Countries (OPEC) from other oil supply shocks, crude oil market-specific demand shocks, and aggregate industrial commodity demand shocks. By quantifying the magnitude and timing of these shocks, he revealed. This finding is supported by a number of researchers (Antonakakis et al., 2014; B et al., 2018; Kang & Ratti, 2013a; Ro Ba Ys, 2012). Following Kilian’s analytical approach, Lippi and Nobili (2012).

Macroeconomic variables fluctuate in response to domestic oil supply shocks as well as supply shocks generated abroad. Not all demand shocks are similar and the correlation between US economic activity and oil prices is altered depending on the different kinds of shocks and their sources, which is a qualification of the results in Kilian (2009). Similarly, Kolodzeij and Kaufmann (2014) analyzed the linkage between these costs and oil prices, but challenged Kilian (2009) findings by showing that the conclusions regarding the insignificance of supply shocks and the importance of demand shocks are not reliable. However, the impact of oil price shocks is also not perfectly consistent across economies. Rehman (2018) used Kilian (2009) decomposition of oil prices as the basis for an analysis of EPU in a sample of markets using the SVAR framework. The results highlight that EPUs in India, Spain, and Japan respond to the global oil price, but aggregate demand shocks do not trigger any changes. Only in the context of China and India are oil-specific demand shocks significant in high volatility states. Chen, Sun, and Li (2019) examined the relationship between China’s industrial economic growth, global EPU (GEPU), and oil price shocks using a combination of Granger causality tests and VAR models. They found that oil prices and GEPU had positive and negative effects on China’s industrial economic growth, respectively.

Component oil price shocks can critically impact economic activity, including stock markets and international trade. Arouri et al. (2014) found that an increase in EPU in major net oil importing countries on the Gulf Cooperation Council (GCC) stock market had a persistent negative impact on stock returns, and that this impact interacted with changes in oil prices. According to Kang and Ratti (2013b), oil price shocks and EPU are interrelated and affect stock market returns. Positive oil market-specific demand shocks substantially increase EPU and reduce real stock returns. As for international trade, Novy and Taylor (2020) found that firms selectively cut import orders out of oil price shocks and China’s economy: monetary policy reactions to oil price shocks worried fixed costs and risk in the face of uncertainty, leading to a contraction in international trade activity that was greater than that of domestic trade activity; the 2008 to 2009 trade collapse is excellent evidence of this phenomenon. As international trade is very sensitive to exchange rates, the impact of EPU shocks on exchange rates may be transmitted to trade. Using an empirical analysis of bilateral data for 16 major FDI-inward countries, Choi et al. (2021) found that EPU in host countries significantly reduced FDI inflows and that this effect was more severe in countries with insufficient financial deepening. Although FDI is generally considered to be the most stable form of international capital flow, it is also subject to EPU due to its high fixed costs and investment risks. With respect to China’s real exports, Wei (2019) found that oil supply and EPU shocks posed negative impacts, whereas oil aggregate demand and oil ratio demand shocks posed positive impacts. In the meantime, both oil supply and aggregate oil demand shocks have a positive impact on China’s real imports.

Thinking with such existing literature, this study analyzed the asymmetric impact of oil price shocks from different sources on different types of PU. Existing studies have treated all kinds of EPUs as an integral whole or used only one EPU to analyze the impact of oil prices and shocks (Su et al., 2021; Wang & Lee, 2022; Wen et al., 2019); this approach makes it hard to identify which kind of PU was impacted. From an empirical perspective, researchers have used a conventional quantitative structural model (Wei, 2019), a time-varying parameter structural vector autoregression model (TVP-SVAR) (Wang & Lee, 2022; Wei, 2019), and a wavelet-based BEKK-GARCH model (Chen, Sun, & Li, 2019) to evaluate the impact. To capture the asymmetric and nonlinear relation more precisely, we follow Pal and Mitra (2019) and employ two thresholds and four thresholds NARDL models for China’s categorical economic policy uncertainties.

Methodology

Disentangling the Oil Price

Ready (2018) proposed a novel approach to decompose oil price into supply, demand, and risk shocks; in this study we borrow his method. Based on Ready (2018), three proxy variables are required to construct the SVAR model. He firstly defined “demand shocks” as the portion of returns of global oil producing firms, which is orthogonal to the innovations to the volatility index (VIX) of the Chicago Board Options Exchange (CBOE). The “risk shock” is defined as the innovations to the VIX, which denotes uncertainty in financial markets; therefore, residuals of oil price variations, which are orthogonal to both demand and risk shocks, are expressed as supply shocks. Ready (2018) reasoned that supply shocks are typically caused by specific events; nevertheless, because oil producers have natural hedges against supply difficulties, price increases caused by oil demand would benefit oil producers. More specifically, Ready’s (2018) approach can be divided into the following two steps. Firstly let $ξ_{VIX, t}$ represent unexpected variations in the VIX following an autoregressive moving average model, that is, ARMA (1,1) specification. Regressing $ξ_{VIX, t}$ as the explanatory variable on the production index of oil producers, the demand shocks are defined as the residual values. In the same way, the residuals of the regression of oil price variations, $ξ_{VIX, t}$ , and demand shocks construct the supply shocks. Consider the following matrix:

X_{t} = A Z_{t}

(1)

The supply, demand, and risk shocks are respectively denoted as s_t, d_t, and v_t. Oil price change is denoted as Δpt. $R_{t}^{prod}$ represents the return of oil firms. Therefore,

X_{t} \equiv [\begin{matrix} Δ p_{t} \\ R_{t}^{prod} \\ ξ_{VIX, t} \end{matrix}], Z_{t} \equiv [\begin{matrix} s_{t} \\ d_{t} \\ v_{t} \end{matrix}], A \equiv [\begin{matrix} 1 & 1 & 1 \\ 0 & a_{22} & a_{23} \\ 0 & 0 & a_{33} \end{matrix}]

(2)

Secondly, Ready (2018) imposed the following condition to ensure orthogonality among three shocks:

A^{- 1} \sum_{X} {(A^{- 1})}^{T} = [\begin{matrix} σ_{s}^{2} & 0 & 0 \\ 0 & σ_{d}^{2} & 0 \\ 0 & 0 & σ_{v}^{2} \end{matrix}]

(3)

$\sum_{X}$ denotes the covariance matrix of $X_{t}$ , and $σ_{s}$ , $σ_{d},$ and $σ_{v}$ are the variances of the three kinds of shocks. Essentially, Ready’s (2018) approach is consistent with Kilian’s (2009) work. To effectively define the structural shocks in this SVAR framework, the sum of the three shocks is used as the total change in the oil price instead of normalizing the volatility of the shocks to one.

Asymmetric Approach

The MTNARDL model proposed by Pal and Mitra (2015); Pal and Mitra (2016) was inspired by a seminal nonlinear ARDL suggested by Shin et al. (2014), wherein the regressor(s) can be decomposed into positive and negative partial sum components to investigate the asymmetry. Shin et al. (2014) built their work upon that of Pesaran et al. (2001), which proposed a conventional ARDL model. We present these methods one by one below.

ARDL model

Firstly, consider an ARDL equation:

\begin{matrix} Δ EPU = \sum_{i = 1}^{n 1} φ_{1 i} Δ EP U_{t - i} + \sum_{i = 0}^{n 2} φ_{2 i} Δ Oi l_{t - i} + θ_{1} EP U_{t - 1} \\ + θ_{2} Oi l_{t - 1} + ε_{t} \end{matrix}

(4)

where EPU stands for economic policy uncertainty. We assess the integral EPU and four categorical EPUs one by one. Oil price is presented as Oil and will also be replaced with variety shocks subsequently. The first difference operator is $Δ$ , and $ε_{t}$ means the error term. The lag orders are n1 and n2. Coefficients of $φ_{1}$ and $φ_{2}$ indicate the short term, while $θ_{1}$ and $θ_{2}$ indicate the long term. To examine the potential cointegrating relation, we utilized the bounds test (Pesaran et al., 2001) to test the null hypothesis for H0: $θ_{1} = θ_{2} = 0$ , which indicates no cointegration. One advantage of linear ARDL is that the variables can be used regardless of whether they are of order zero or one. Moreover, they can differentiate short and long-term dynamics, and can be utilized even if the independent variables are endogenous (Pal & Mitra, 2019; Pesaran et al., 2001).

NARDL model

Shin et al. (2014) proposed a nonlinear ARDL to assess the asymmetry of Express Oil as follows:

Oi l_{t} = Oi l_{0} + {Oil}_{t}^{+} + {Oil}_{t}^{-}

(5)

where ${Oil}_{t}^{+}$ and ${Oil}_{t}^{-}$ are partial sums that capture the increase and decrease of the oil price change, respectively:

{Oil}_{t}^{+} = \sum_{i = 1}^{t} Δ {Oil}_{i}^{+} = \sum_{i = 1}^{t} \max (Δ {Oil}_{i}, 0)

(6a)

{Oil}_{t}^{-} = \sum_{i = 1}^{t} Δ {Oil}_{i}^{-} = \sum_{i = 1}^{t} \min (Δ {Oil}_{i}, 0)

(6b)

where $Δ O i l_{i}^{} = O i l_{t} - O i l_{t - 1}$ . Based on equation (7):

\begin{matrix} Δ EPU = \sum_{i = 1}^{n 1} φ_{1 i} Δ EP U_{t - i} + \sum_{i = 0}^{n 2} φ_{2 i} Δ {Oil}_{t - i}^{+} + \sum_{i = 0}^{n 2} φ_{3 i} Δ {Oil}_{t - i}^{-} \\ + θ_{1} EP U_{t - 1} + θ_{2} {Oil}_{t - i}^{+} + θ_{3} {Oil}_{t - i}^{-} + ε_{t} \end{matrix}

(7)

Furthermore, the standard Wald test can be used to test the short and long-term asymmetry. $φ_{2 i} = φ_{3 i}$ indicates that there is no asymmetry in the short term and that, for long-term, the null hypothesis is $θ_{2} = θ_{3}$ . What is more, the cointegration could be examined by the bounds test where the null hypothesis ( $θ_{1} = θ_{2} = θ_{3} = 0$ ) indicates no cointegration.

Multiple threshold NARDL

The MTNARDL proposed by (Pal & Mitra, 2015, 2016) split the regressor into different quantiles to explore the asymmetric transmission of the regressor with minor to major fluctuation. For instance, we pulled into two thresholds at the 30th and 70th quantiles (denoted as $τ_{30}$ and $τ_{70}$ ) to decompose the oil price change into three partial sums:

O i l_{t} = O i l_{0} + O i l_{t}^{} (Φ_{1}) + O i l_{t}^{} (Φ_{2}) + O i l_{t}^{} (Φ_{3})

(8)

In specification (8), $O i l_{t}^{} (Φ_{1})$ , $O i l_{t}^{} (Φ_{2}),$ and $O i l_{t}^{} (Φ_{3})$ comprise the three partial sums series:

O i l_{t}^{} (Φ_{1}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} (Φ_{1}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} I (Δ O i l_{i}^{} \leq τ_{30})

(9a)

O i l_{t}^{} (Φ_{2}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} (Φ_{2}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} I (τ_{30} < Δ O i l_{i}^{} \leq τ_{70})

(9b)

O i l_{t}^{} (Φ_{3}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} (Φ_{3}) = \sum_{i = 1}^{t} Δ O i l_{i}^{} I (Δ O i l_{i}^{} > τ_{70})

(9c)

where $I (.)$ is a dummy variable that equals 1 when the conditions stated within (.) are satisfied and zero otherwise. The two-threshold NARDL model is:

\begin{array}{l} Δ E P U = \sum_{i = 1}^{n 1} φ_{1 i} Δ E P U_{t - i} + \sum_{j = 1}^{3} \sum_{i = 0}^{n 2} φ_{k i} Δ O i l_{t - i}^{} (Φ_{j}) \\ + θ_{1} E P U_{t - 1} + \sum_{j = 1}^{3} θ_{k} O i l_{t - 1} (Φ_{j}) + ε_{t} \end{array}

(10)

where $k = j + 1$ . Rejection of the null hypothesis of H0: $θ_{1} = θ_{2} = θ_{3} = θ_{4} = 0$ could confirm cointegration among the three variables. Besides, with the null of no asymmetry for H0: $φ_{2 i} = φ_{3 i} = φ_{4 i}$ , the Wald test can be employed to investigate short-term asymmetry. Likewise, with respect to long-run asymmetry, H0 is $θ_{2} = θ_{3} = θ_{4}$ . Finally, oil price change could further be decomposed into five partial sums which set four thresholds at the 20th, 40th, 60th, and 80th quantiles:

\begin{array}{l} Δ E P U_{t} = \sum_{i = 1}^{n 1} φ_{1 i} Δ E P U_{t - i} + \sum_{j = 1}^{5} \sum_{i = 0}^{n 2} φ_{k i} Δ O i l_{t - i}^{} (Φ_{j}) \\ + θ_{1} E P U_{t - 1} + \sum_{j = 1}^{5} θ_{k} O i l_{t - 1} (Φ_{j}) + ε_{t} \end{array}

(11)

Eventually, the asymmetry is estimated by standard Wald test both in the long-term short-term. One advantage of MTNARDL is that they are suitable regardless of whether variables are integrated of order one I(1) or zero I(0). Another advantage, according to Pal and Mitra (2019), is that the model can concurrently examine the impact of the regress on the positive and negative outliers of the regressors. However, there are two limitations of the MTNARDL model: one is that when threshold increases, the sample size employed for estimate decreases. For example, the sample is divided into two parts when there is a single threshold, while four thresholds lower the sample size to one-fifth of the original, it’s well known that fewer samples diminish estimate accuracy. In this study we extract more than 250 observations, and we split the series into a maximum of five partial sums. As a result, we still have about 50 observations, which is considered sufficient for the regression. Another drawback is that the model can’t be utilized when any variables are higher than order one, thus we run a unit root test first.

Data

Our data is divided into two parts. The first part is used for oil price decomposition and the second part is used to analyze the asymmetry between oil component shocks and China’s EPUs. The data definitions are shown in Table 1. The categorical EPUs of China provided by Huang and Luk (2020) are monthly data, so we ultimately set all variables to monthly frequencies, spanning the period from January 2000 to March 2021. Table 2 shows the descriptive statistics of the underlying variables and Figure 1 is a visual perspective of the time series for the data.

Table 1.

Data Definition.

Variable	Definition	Source
Part 1. Disentangle oil price
VIX	Logarithm of VIX.	Yahoo Finance
Returns of oil-producing firms	NYSE Arca Oil Index (percentage changes)	Yahoo Finance
Oil price	Crude-light sweet oil futures contract (Percentage changes, New York Mercantile Exchange)	Energy Information Administration (EIA)
Part 2. Asymmetry
Oil price	Brent Spot Price (dollars per barrel)	EIA
Categorical China’s EPUs	Huang and Luk (2020) calculation	https://economicpolicyuncertaintyinchina.weebly.com

Note. In Part 1, the Integrated Oil and Gas Producer Index was selected for returns of the firms by Ready (2018) from DataStream; however, it excludes some large producers, such as Saudi Aramco. Following Ba and Sp (2021), we used the NYSE Arca Oil Index instead. The oil price in Part 1 was only used for oil decomposition; in the main body of the model analysis, we use the entire second part of the oil price.

Table 2.

Descriptive Statistics.

Variable	Min	Max	Mean	SD	Median	Skewness	Kurtosis
Oil price	18.38	132.72	63.692	29.331	60.845	0.444	2.181
Supply shock	−55.638	69.59	0.021	9.952	0.781	0.486	15.356
Demand shock	−25.597	22.289	0.03	5.909	0.108	0.162	5.36
Risk shock	−47.217	78.762	−0.182	19.798	−3.737	0.848	4.516
Total EPU	39.525	238.317	118.035	39.001	122.745	0.099	2.55
Fiscal PU	18.876	450.729	111.853	65.284	99.568	1.136	5.315
Monetary PU	11.435	357.999	106.48	63.35	93.702	0.934	3.531
Trade PU	24.999	1031.009	142.421	136.998	102.152	3.707	20.539
Exchange rate PU	10.741	412.054	106.088	70.963	82.561	1.462	5.465

Figure 1.

Variations of oil price and China’s categorical EPUs.

Empirical Analysis and Discussion

Following Pal and Mitra (2015, 2016, 2019), in ARDL, NARDL, and MTNARDL models, underlying variables should integrated if they are of order one I(1) or zero I(0). Hence, before conducting an asymmetric analysis, we conducted a unit root test by using an ADF test (Fuller, 1979) and a PP test (Phillips & Perron, 1988); we present the results in Table 3. The results indicate that all series are stationary at level value or first difference, namely I(0) or I(1); thus, the conditions for the application of the model are met.

Table 3.

Unit Root Tests.

Variable	Level (I[0])		Difference (I[1])
Variable	ADF	PP	ADF	PP
Oil price	−1.814	−2.217	−10.734***	−10.570***
Supply shock	−14.543***	−14.478***	−24.698***	−33.745***
Demand shock	−18.650***	−18.526***	−32.351***	−48.542***
Risk shock	−15.820***	−15.834***	−26.445***	−37.971***
Total EPU	−5.088***	−4.394***	−22.457***	−26.759***
Fiscal PU	−6.952***	−6.658***	−22.597***	−27.422***
Monetary PU	−4.571***	−3.931***	−19.906***	−22.527***
Trade PU	−8.758***	−9.091***	−28.946***	−34.450***
Exchange rate PU	−6.546***	−6.093***	−19.240***	−23.439***

Note. Significant at 10% (*), 5% (**), and 1% (***) levels.

The cointegration relation among the variables was first evaluated by the ARDL model. Table 4 displays the results of the linear ARDL. For all four panels, the coefficients shown on the top half are for the short term and those on the bottom are for the long term. The F-statistics raw presents the value of the Wald test which was compared with the asymptotic upper bound critical value estimated by Pesaran et al. (2001). Besides, the −1 in brackets indicates the lagged one-period value. Due to the significance of the F-statistics, cointegrated relationships were confirmed not only between oil price and categorical EPUs, but also between most component shocks and categorical EPUs. However, the asymmetric effect cannot be evaluated using the ARDL model. To achieve our target, we built the NARDL model to split the independent variables—that is, oil price and variety shocks—into two partial sums to differentiate the responses of EPUs to increases (positive variation) and decreases (negative variation) in the oil price and to shocks. The results of the NARDL are displayed in Table 5.

Table 4.

ARDL Result.

	Total	Fiscal	Monetary	Trade	Exr
Panel A: Oil price
Short term
Δ(EPU(−1))	−0.288***	−0.424***	−0.127**	−0.205***	−0.0352
Δ(Oil)	−0.428*	−0.495	−0.804**	−1.174**	−0.947*
Δ(Oil(−1))	−0.16	0.334	0.48	−0.461	0.318
Long term
EPU(−1)	−0.117***	−0.228***	−0.219***	−0.316***	−0.309***
Oil(−1)	0.203***	0.434***	0.360***	0.528***	0.484***
F-statistic	7.46***	9.87***	12.90***	17.16***	18.42***
R²	.178	.343	.172	.255	.18
Panel B: Supply shock
Short term
Δ(EPU(−1))	−0.328***	−0.481***	−0.216***	−0.304***	−0.152**
Δ(S)	−0.15	−0.445	−0.143	−0.106	−0.246
Δ(S(−1))	−0.233*	0.263	−0.255	−0.315	0.0294
Long term
EPU(−1)	−0.00918	−0.113***	−0.0285*	−0.0544**	−0.0748***
S(−1)	−0.0593	−0.6	0.349	−0.446	−0.415
F-statistic	0.39	4.78***	2.03	2.94*	4.29**
R²	.131	.316	.072	.149	.068
Panel C: Demand shock
Short term
Δ(EPU(−1))	−0.323***	−0.489***	−0.203***	−0.287***	−0.150**
Δ(D)	−0.00663	0.721	−0.489	−0.392	−0.181
Δ(D(−1))	0.114	0.843	−0.284	0.706	0.0218
Long term
EPU(−1)	−0.00943	−0.113***	−0.0296*	−0.0544**	−0.0736***
D(−1)	0.00438	−0.579	0.202	−0.0844	0.263
F-statistic	0.38	4.62**	1.59	2.62*	4.04**
R²	.114	.317	.074	.156	.068
Panel D: Risk shock
Short term
Δ(EPU(−1))	−0.353***	−0.493***	−0.235***	−0.343***	−0.173***
Δ(R)	0.243***	−0.565	0.359***	0.507***	0.671***
Δ(R(−1))	0.0415	−0.222	0.202**	−0.0465	0.158
Long term
EPU(−1)	−0.00854	−0.110***	−0.0284*	−0.0544**	−0.0668***
R(−1)	0.334***	0.199	0.304*	0.758***	0.507*
F-statistic	4.41**	4.58**	2.71*	6.38***	5.10***
R²	.172	.33	.126	.173	.131

Note. Significant at 10% (*), 5% (**), and 1% (***) levels. The F-statistic was calculated by the Wald test (with H0: θ1 = θ2 = 0) and was compared with the upper bound critical value.

Table 5.

NARDL Results.

	Total	Fiscal	Monetary	Trade	Exr
Panel A: Oil price
Short term
Δ(EPU(−1))	−0.284***	−0.371***	−0.139**	−0.191***	−0.0376
Δ(Oil⁻)	−1.023**	0.374	−1.834***	−2.566***	−2.795***
Δ(Oil⁻(−1))	−0.17	0.893	0.729	−0.928	1.005
Δ(Oil⁺)	0.425	−2.396	0.726	0.976	1.511
Δ(Oil⁺(−1))	0.278	0.202	0.709	1.215	0.449
Long term
EPU(−1)	−0.128***	−0.335***	−0.198***	−0.348***	−0.280***
Oil⁻(−1)	0.112**	−0.0611	0.301***	0.331***	0.303**
Oil⁺(−1)	0.134**	0.132	0.313***	0.379***	0.332***
F-statistic	4.56***	10.29***	7.00***	12.63***	10.56***
Wald-short	2.01	0.49	3.44**	3.73**	2.93**
Wald-long	4.02**	15.70***	1.07	6.56**	2.5
R²	.179	.373	.167	.269	.177
Panel B: Supply shock
Short term
Δ(EPU(−1))	−0.278***	−0.354***	−0.184***	−0.201***	−0.0617
Δ(S−)	−0.694***	−1.367	−0.764**	−1.615***	−0.87
Δ(S−(−1))	−0.167	0.568	0.0629	0.177	0.584
Δ(S+)	−0.0814	−0.000112	−0.0669	0.531	−0.271
Δ(S+(−1))	−0.0682	0.109	−0.201	−0.424	0.154
Long term
EPU(−1)	−0.154***	−0.367***	−0.107***	−0.292***	−0.263***
S−(−1)	−0.479**	−0.731	−0.147	−0.840**	−1.188***
S+(−1)	−0.462**	−0.66	−0.135	−0.812**	−1.165***
F-statistic	4.92***	11.30***	3.31**	10.53***	10.37***
Wald-short	2.25*	0.66	1.03	3.54**	1.67
Wald-long	7.54***	10.83***	3.31*	8.64***	5.83**
R²	.184	.379	.099	.255	.148
Panel C: Demand shock
Short term
Δ(EPU(−1))	−0.257***	−0.363***	−0.153**	−0.183***	−0.0569
Δ(D−)	−0.625*	0.635	−1.109**	−0.783	−1.134
Δ(D−(−1))	0.004	−0.296	−0.0826	−0.206	−0.0939
Δ(D+)	0.913***	1.132	0.366	0.571	1.337*
Δ(D+(−1))	−0.33	1.253	−0.900*	0.869	−0.654
Long term
EPU(−1)	−0.155***	−0.367***	−0.118***	−0.281***	−0.255***
D−(−1)	0.595*	0.262	0.627	0.883	1.205*
D+(−1)	0.616**	0.366	0.638	0.919	1.228*
F-statistic	5.57***	11.28***	4.32***	10.65***	10.50***
Wald-short	4.22***	0.35	1.29	1.66	2.06
Wald-long	6.36**	10.74***	1.36	6.71**	2.79*
R²	.187	.379	.119	.242	.158
Panel D: Risk shock
Short term
Δ(EPU(−1))	−0.322***	−0.371***	−0.204***	−0.250***	−0.102
Δ(R−)	0.0671	−1.323**	0.259	0.690***	0.500*
Δ(R−(−1))	0.0407	0.266	0.0593	−0.763***	−0.205
Δ(R+)	0.406***	0.229	0.472***	0.396*	0.797***
Δ(R+(−1))	0.0472	−0.704	0.353**	0.634***	0.563**
Long term
EPU(−1)	−0.0860***	−0.363***	−0.0976***	−0.242***	−0.207***
R−(−1)	0.312**	0.0838	0.304	0.862***	0.449
R+(−1)	0.316**	0.118	0.307	0.871***	0.454
F-statistic	5.20***	11.01***	4.48***	31.61***	9.38***
Wald-short	3.61**	1.35	1.76	9.41***	4.59***
Wald-long	3.31*	13.02***	0.87	5.12**	2.04
R²	.204	.398	.154	.274	.19

Note. Significant at 10% (*), 5% (**), and 1% (***) levels. The F-statistic is calculated by the Wald test (with H0: θ1 = θ2 = 0) and is compared with the upper bound critical value.

The F-statistics in Table 5 re-verify the cointegration among underlying variables. This is consistent with the result of the ARDL. Wald-short provides the F-statistics of the Wald test which used to evaluate asymmetry in the short term (with H0: φ_2i = φ_3i), so does Wald-long which test the long-term asymmetry (with H0: θ₂ = θ₃). Panel A shows that, generally, in both the short term and the long term, there is only an asymmetric influence between oil price and trade PU. More explicitly, in the short term, there is an asymmetry between oil price and monetary PU, trade PU, and exchange rate PU, while in the long term, an asymmetry exists between oil price and total EPU, fiscal PU, and trade PU. These asymmetric results in the short term are similar to those reported by Pal and Mitra (2016, 2019).

The Wald-short only shows the asymmetric impacts of short-run supply shocks on trade PU, but the asymmetric impact of long-run supply shocks can be observed in all categorical EPUs. According to our findings, demand shocks have considerable asymmetric impacts on fiscal PU, trade PU, and exchange rate PU in the long term, but no asymmetry exists in the short term. From the perspectives of both supply and demand shocks, the asymmetric impact is much more significant in the long term than in the short term; this suggests the possibility of a sticky of economic policy, with most significant changes occurring in both the short and long term due to trade PU. Further, the asymmetric effect of oil price and three component shocks on trade PU in the short and long term is always significant. This may be because China is the world’s second-largest trading nation and international trade is very sensitive to such shocks. Regarding risk shocks, we follow Ready (2018) and Li and Guo (2022) and display unexpected variations in market discount rates.

In the above-mentioned analysis, the NARDL model does not precisely identify all small to large variations in independent variables. Therefore, to examine the asymmetric impact more precisely, we employed the MTNARDL model with the 30th and 70th quantile thresholds and thus split the variables into three partial sums. The results of the two-threshold NARDL model based on equation (10) are shown in Table 6. The significant coefficients of $Δ oil (Φ 1)$ in Panel A for monetary PU, trade PU, and exchange rate PU indicate that in the short term the decreases of oil price had negative influences on these three EPUs, but positive effects in the long run. Nevertheless, short-run influences are more remarkable than long-run ones because when an oil price decrease lowers production cost, there is an increase in economic activity, such as consumption, investment, and trade (Beckmann et al., 2020). With this increase in economic activity, the economy runs smoothly and the degree of economic policy uncertainty naturally becomes lower. On the other hand, the difference between $Δ oil (Φ 1) (- 1)$ and $Δ oil (Φ 3) (- 1)$ shows that monetary PU, trade PU, and exchange rate PU are only positively impacted by oil price change in the long term. Supply shock (coefficients of $S (Φ 1) (- 1)$ and $S (Φ 3) (- 1))$ has a negative impact on trade PU and exchange rate PU in the long run, while only exchange rate policy uncertainty is affected by demand shock in the long run. Monetary PU, trade PU, and exchange rate PU are significantly positively impacted by risk shocks in the short-term, while in the long term only trade PU is affected. The MTNARDL approach supports the idea that oil price, supply shock, and demand shock are much more remarkable in the long term than the short run in most cases, but risk shocks hold positive impacts on monetary PU, trade PU, and exchange rate PU in the short term. Notably, these findings contrast with those of other studies that claim that supply shocks are relatively unimportant (Antonakakis et al., 2014; B et al., 2018; Kang & Ratti, 2013a; Kilian & Park, 2009). Here, we emphasize the short-term impact of risk shocks that deserve attention.

Table 6.

Results of Two-Threshold MTANARL Model.

	Total	Fiscal	Monetary	Trade	Exr
Panel A: Oil price
Short term
Δ(EPU(−1))	−0.266***	−0.341***	−0.140**	−0.181***	−0.03
Δ(Oil( $Φ 1))$	−1.108***	1.084	−1.991***	−2.714***	−2.616***
Δ(Oil( $Φ$ 1)(−1))	−0.0671	1.439	0.861	−0.816	0.946
Δ(Oil( $Φ$ 2))	2.840*	1.357	3.253	4.377	1.487
Δ(Oil( $Φ$ 2)(−1))	−1.61	−2.11	−0.518	1.237	1.545
Δ(Oil( $Φ$ 3))	0.299	−4.515*	0.685	0.915	1.164
Δ(Oil( $Φ$ 3)(−1))	0.0558	−1.351	0.589	0.98	0.271
Long term
EPU(−1)	−0.178***	−0.404***	−0.206***	−0.363***	−0.294***
Oil( $Φ$ 1)(−1)	0.154***	−0.051	0.314***	0.358***	0.331**
Oil( $Φ$ 2)(−1)	0.484***	1.851***	0.433**	0.539*	0.622**
Oil( $Φ$ 3)(−1)	0.127**	−0.154	0.304***	0.377***	0.309**
F-statistic	4.45***	9.77***	5.31***	9.49***	8.02***
Wald-short	1.74	1.1	2.24*	1.95*	1.37
Wald-long	4.73***	10.72***	0.54	2.99*	1.46
R²	.212	.395	.177	.275	.181
Panel B: Supply shock
Short term
Δ(EPU(−1))	−0.264***	−0.355***	−0.175***	−0.175***	−0.0542
Δ(S( $Φ$ 1))	−0.673***	−1.548	−0.696**	−1.536***	−0.765
Δ(S( $Φ$ 1)(−1))	−0.0998	0.625	0.155	0.373	0.628
Δ(S( $Φ$ 2))	−0.394	−1.48	−0.82	−1.282	−1.666
Δ(S( $Φ$ 2) (−1))	−0.206	1.957	−0.153	−2.015	0.21
Δ(S( $Φ$ 3))	−0.0962	0.119	−0.0406	0.610*	−0.215
Δ(S( $Φ$ 3)(−1))	−0.0799	−0.0224	−0.27	−0.47	0.0872
Long term
EPU(−1)	−0.182***	−0.363***	−0.123***	−0.340***	−0.275***
S( $Φ$ 1) (−1)	−0.530**	−0.721	−0.134	−0.818**	−1.136**
S( $Φ$ 2) (−1)	−0.19	−0.685	0.386	0.116	−0.588
S( $Φ$ 3) (−1)	−0.479**	−0.64	−0.075	−0.701*	−1.064**
F-statistic	4.27***	8.22***	3.13**	9.63***	8.18***
Wald-short	1.25	0.55	0.62	2.74**	1
Wald-long	5.21***	5.45***	2.94*	6.99***	3.56**
R²	.194	.381	.112	.28	.157
Panel C: Demand shock
Short term
Δ(EPU(−1))	−0.257***	−0.361***	−0.152**	−0.183***	−0.0567
Δ(D( $Φ$ 1))	−0.506	0.633	−1.032**	−0.894	−0.927
Δ(D( $Φ$ 1)(−1))	−0.0366	−0.137	−0.177	−0.14	−0.201
Δ(D( $Φ$ 2))	0.0982	0.598	−0.34	1.97	−1.472
Δ(D( $Φ$ 2)(−1))	−0.532	−5.453	−0.989	0.726	0.115
Δ(D( $Φ$ 3))	0.839***	1.169	0.273	0.512	1.248*
Δ(D( $Φ$ 3)(−1))	−0.315	1.366	−0.753	0.767	−0.544
Long term
EPU(−1)	−0.154***	−0.368***	−0.119***	−0.276***	−0.255***
D( $Φ$ 1)(−1)	0.659**	0.244	0.583	0.805	1.231*
D( $Φ$ 2)(−1)	0.564	0.991	0.86	1.179	1.365
D( $Φ$ 3)(−1)	0.685**	0.348	0.59	0.84	1.255*
F-statistic	4.02***	8.49***	3.26**	7.68***	7.72***
Wald-short	2.35**	0.47	0.71	1.15	1.28
Wald-long	3.22**	5.47***	0.73	3.48**	1.47
R²	.185	.382	.118	.245	.159
Panel D: Risk shock
Short term
Δ(EPU(−1))	−0.314***	−0.362***	−0.193***	−0.239***	−0.0942
Δ(R( $Φ$ 1))	0.0904	−1.237**	0.332**	0.706***	0.628**
Δ(R( $Φ$ 1)(−1))	0.0846	0.259	0.0747	−0.611***	−0.174
Δ(R( $Φ$ 2))	0.27	1.295	0.0427	−0.00985	0.148
Δ(R( $Φ$ 2)(−1))	−0.346	0.464	0.269	0.0848	0.389
Δ(R( $Φ$ 3))	0.355***	−0.0098	0.397***	0.377*	0.671***
Δ(R( $Φ$ 3)(−1))	0.0306	−0.768	0.320*	0.430*	0.493*
EPU(−1)	−0.105***	−0.392***	−0.108***	−0.252***	−0.216***
R( $Φ$ 1)(−1)	0.293**	0.0793	0.295	0.856***	0.442
R( $Φ$ 2)(−1)	0.401***	0.4	0.420**	1.037***	0.630**
R( $Φ$ 3)(−1)	0.295**	0.11	0.295	0.863***	0.444
F-statistic	4.40***	8.88***	3.77***	9.80***	7.18***
Wald-short	2.28**	1.25	0.77	4.32***	2.16*
Wald-long	3.71**	8.28***	2.07	4.69***	2.96*
R ²	.218	.407	.163	.272	.198

Note. Significant at 10% (*), 5% (**), and 1% (***) levels. The F-statistic is calculated by the Wald test (with H0: θ1 = θ2 = 0) and is compared with the upper bound critical value.

With quantiles of 20th, 40th, 60th, and 80th, we decomposed the oil price series into five partial sums to ensure the results were legitimate and to discover any probable unequal link between oil price and categorical EPU. Table 7 shows the outcomes of equation (11). Similarly, the preceding conclusions have been tested and are correct. In Tables 4 and 5, some of the coefficients of the three kinds of shocks on fiscal policy uncertainty are not remarkable. It is very clear that there are no asymmetric impacts in the short term, but that there is a very strong asymmetric impact in the long term.

Table 7.

Results of Four-Threshold MTNARDL Model.

	Total	Fiscal	Monetary	Trade	Exr
Panel A: Oil price
Short term
Δ(EPU(−1))	−0.242***	−0.353***	−0.134**	−0.141**	0.000544
Δ(Oil( $Φ 1))$	−1.244***	1.634	−1.910***	−3.028***	−2.547**
Δ(Oil( $Φ$ 1)(−1)	−0.281	1.092	0.718	−1.738*	0.953
Δ(Oil( $Φ$ 2))	−1.653	9.366	−1.841	−7.159*	−6.043
Δ(Oil( $Φ 2) (- 1))$	−4.370**	−17.07*	−0.861	−9.357**	1.982
Δ(Oil( $Φ 3))$	3.436	−24.54	3.166	10.09	3.021
Δ(Oil( $Φ 3) (- 1)$ )	1.085	13.21	2.554	7.401	2.963
Δ(Oil( $Φ$ 4))	1.494	−9.834	0.811	3.426	−0.743
Δ(Oil( $Φ$ 4)(−1))	0.602	−3.362	0.485	5.453**	0.18
Δ(Oil( $Φ$ 5))	0.336	−6.320**	0.563	1.373	1.87
Δ(Oil( $Φ 5) (- 1))$	0.59	0.0535	0.684	1.498	0.414
Long term
EPU(−1)	−0.219***	−0.406***	−0.217***	−0.457***	−0.331***
Oil( $Φ$ 1))(−1)	0.150**	−0.3	0.332***	0.296**	0.294*
Oil( $Φ$ 2)(−1))	0.00261	4.125	0.497	3.347***	−1.122
Oil( $Φ$ 3)(−1))	0.209	8.054**	0.376	4.127**	0.176
Oil( $Φ 4$ )(−1))	0.288	0.272	0.602	0.344	−0.336
Oil( $Φ$ 5)(−1))	0.104	−0.316	0.269**	0.514***	0.292*
F-statistic	3.54***	6.86***	3.75***	8.54***	5.99***
Wald-short	1.22	1.29	0.92	1.77*	0.99
Wald-long	2.51*	6.12***	0.32	2.96**	1.23
R²	.229	.415	.178	.31	.198
Panel B: Supply shock
Short term
Δ(EPU(−1))	−0.193***	−0.335***	−0.143**	−0.119*	−0.00187
Δ(S( $Φ$ 1))	−0.628***	−0.871	−0.31	−1.091**	−0.259
Δ(S( $Φ$ 1)(−1))	−0.126	0.451	0.00768	0.313	0.718
Δ(S( $Φ$ 2))	0.85	1.568	0.908	1.131	1.521
Δ(S( $Φ$ 2)(−1))	0.288	−0.977	0.294	0.663	1.16
Δ(S( $Φ$ 3))	−0.974	2.322	0.98	0.499	−0.0848
Δ(S( $Φ$ 3)(−1))	−0.211	−4.429	−0.834	−0.879	−1.644
Δ(S( $Φ$ 4))	−0.0565	0.296	−0.711	−0.527	−1.569
Δ(S( $Φ$ 4)(−1))	0.137	2.238	1.206	−0.0211	1.127
Δ(S( $Φ$ 5))	−0.102	−0.205	0.00334	0.636*	−0.148
Δ(S( $Φ$ 5)(−1))	−0.114	0.19	−0.447	−0.716*	−0.193
Long term
EPU(−1)	−0.345***	−0.414***	−0.203***	−0.449***	−0.382***
S( $Φ$ 1)(−1)	−0.459**	−0.515	0.291	−0.392	−0.679
S( $Φ$ 2)(−1)	−0.221	0.429	0.322	0.18	−0.536
S( $Φ$ 3)(−1)	1.046**	−1.622	2.079***	2.225**	2.325**
S( $Φ$ 4)(−1)	0.0379	0.723	0.669	0.412	−0.258
S( $Φ$ 5)(−1)	−0.376*	−0.485	0.324	−0.197	−0.529
F-statistic	6.11***	6.04***	4.06***	7.72***	7.72***
Wald-short	1.5	0.29	0.8	1.45	0.69
Wald-long	6.96***	3.29**	3.84***	5.13***	4.54***
R²	.274	.393	.171	.309	.209
Panel C: Demand shock
Short term
Δ(EPU(−1))	−0.238***	−0.335***	−0.134**	−0.152**	−0.0109
Δ(D( $Φ$ 1))	−0.648*	2.151	−1.213**	−0.767	−1.375
Δ(D( $Φ$ 1)(−1))	0.0212	0.231	0.0525	−0.119	0.119
Δ(D( $Φ$ 2))	−0.819	7.23	−0.585	2.564	−2.158
Δ(D( $Φ$ 2)(−1))	−0.509	−10.75**	−0.256	−2.489	−1.487
Δ(D( $Φ$ 3))	1.313	−9.684	1.748	−3.159	−1.822
Δ(D( $Φ$ 3)(−1))	−0.548	−1.21	−4.152	1.664	−5.29
Δ(D( $Φ$ 4))	0.397	−9.583**	−0.185	−1.972	−0.375
Δ(D( $Φ$ 4)(−1))	−0.97	−0.743	−1.943	0.706	−1.485
Δ(D( $Φ$ 5))	1.016***	0.417	0.457	0.472	1.577**
Δ(D( $Φ$ 5)(−1))	−0.216	1.814	−0.854	0.998	−0.291
Long term
EPU(−1)	−0.196***	−0.404***	−0.150***	−0.331***	−0.348***
D( $Φ$ 1)(−1)	0.526	0.461	0.493	0.725	0.529
D( $Φ$ 2)(−1)	0.727**	1.196	0.829	0.913	1.377*
D( $Φ$ 3)(−1)	−0.915	11.02*	−1.162	−3.779	−5.347**
D( $Φ$ 4)(−1)	0.697*	0.41	0.769	0.704	1.142
D( $Φ$ 5)(−1)	0.615*	0.613	0.582	0.992	0.82
F-statistic	3.09***	6.00***	2.57**	5.54***	7.02***
Wald-short	1.77*	1.63	0.72	1.33	1.22
Wald-long	2.63**	3.88***	1.16	3.24**	3.61***
R²	.205	.422	.136	.276	.206
Panel D: Risk shock
Short term
Δ(EPU(−1))	−0.221***	−0.347***	−0.157**	−0.172***	−0.0315
Δ(R( $Φ$ 1))	0.122	−1.422**	0.373**	0.835***	0.631**
Δ(R( $Φ$ 1)(−1))	−0.0531	0.261	−0.0254	−0.833***	−0.222
Δ(R( $Φ$ 2)	−0.153	−0.63	0.205	0.241	−0.369
Δ(R( $Φ$ 2)(−1))	−0.349	0.269	−0.00734	−0.105	−0.0576
Δ(R( $Φ$ 3))	0.426	2.016	0.387	1.1	−0.335
Δ(R( $Φ$ 3)(−1))	0.097	6.644*	1.243	1.646	0.198
Δ(R( $Φ$ 4))	0.378*	0.719	0.428	0.563	0.955*
Δ(R( $Φ$ 4)(−1))	−0.0509	0.024	0.306	0.364	0.37
Δ(R( $Φ$ 5))	0.367***	−0.037	0.376**	0.237	0.670***
Δ(R( $Φ$ 5)(−1))	0.0582	−0.799	0.265	0.292	0.335
Long term
EPU(−1)	−0.316***	−0.428***	−0.206***	−0.413***	−0.359***
R( $Φ$ 1)(−1)	0.420***	−0.116	0.464**	1.145***	0.623**
R( $Φ$ 2)(−1)	0.275**	0.175	0.317	0.908***	0.331
R( $Φ$ 3)(−1)	0.761***	1.678	0.419	1.101**	0.786
R( $Φ$ 4)(−1)	0.371***	−0.00075	0.475**	1.176***	0.441
R( $Φ$ 5)(−1)	0.394***	−0.059	0.411**	1.063***	0.595**
F-statistic	7.37***	6.38***	4.46***	10.53***	7.85***
Wald-short	1.71*	1.18	0.62	3.52***	1.45
Wald-long	7.12***	3.83***	2.80**	5.43***	3.34**
R²	.292	.426	.204	.346	.254

Note. Significant at 10% (*), 5% (**), and 1% (***) levels. The F-statistic is calculated by the Wald test (with H0: θ1 = θ2 = 0) and is compared with the upper bound critical value.

Regarding MPU, demand shocks bring significant negative impacts. Notably, referring to $Δ oil (D Φ 1)$ in Tables 6 and 7, the lowest threshold supply shocks decrease in the short term, but referring to $Δ oil (R Φ 1)$ and $Δ oil (R Φ 3)$ in Table 6 and $Δ oil (R Φ 1)$ and $Δ oil (R Φ 5)$ in Table 7, both the lowest and highest threshold risk shocks increase in the short term.

Regarding trade policy uncertainty, the impact of the lowest threshold supply shock is only reduced in the short run (refer to $Δ oil (S Φ 1)$ in Tables 6 and 7), and it intensifies from the highest one (refer to $Δ oil (S Φ 3)$ and $Δ oil (S Φ 5)$ in Tables 6 and 7). Likewise, demand shock has almost no impact on it. In terms of risk shocks, it is evident from the corresponding coefficients of $Δ oil (R Φ 1) (- 1)$ , $Δ oil (R Φ 3) (- 1), Δ oil (R Φ 5) (- 1)$ in Tables 6 and 7 that in the long term both the lowest and highest threshold shocks have significant positive effects on trade PU.

Regarding exchange rate PU, the coefficients of $Δ oil (D Φ 1) (- 1)$ and $Δ oil (D Φ 3) (- 1)$ in the long term presented in Table 6 display negative asymmetry effects caused by demand shock in the lowest and highest thresholds; the asymmetric effect of demand shock on exchange rate PU only exists in the long term in the two-threshold MTANARL model. This result is consistent with that of Beckmann et al. (2020). Moreover, risk shocks have an impact in the two- and four-threshold MTANARL model in the short run (refer to $Δ oil (R Φ 1)$ and $Δ oil (R Φ 3)$ in Table 6 and $Δ oil (R Φ 1)$ and $Δ oil (R Φ 5)$ in Table 7), and in the four-threshold MTANARL model in the long run (refer to $Δ oil (R Φ 1) (- 1)$ , $Δ oil (R Φ 5) (- 1)$ in Table 7).

Regarding the NARDL and two and four thresholds MTNARDL models, Table 8 sums up the results of the influences of various shocks on categorical EPUs. To assess the strength of the asymmetric influence in the short and long runs, we collected all Wald short and Wald long values from Tables 5 to 7. The null of no asymmetry in Panel A solely applies to short-run fiscal PU. The rest have weak asymmetric effects. While this implies a weak asymmetric transmission among monetary PU, trade PU, exchange rate PU, and oil price change, the asymmetric impact between oil price and fiscal PU and trade PU are strong in the long run.

Table 8.

Test Statistic to Ascertain Asymmetry Strength in the Short and Long Runs.

Categorical EPUs		MTNARDL	MTNARDL	Strength of asymmetry	NARDL	MTNARDL	MTNARDL	Strength of asymmetry
	NARDL	2 thresholds	4 thresholds			2 thresholds	4 thresholds
	Short term					Long term
Panel A: Oil price
Total	2.01	1.74	1.22	Nil	4.02**	4.73***	2.51*	Strong
Fiscal	0.49	1.1	1.29	Nil	15.70***	10.72***	6.12***	Very strong
Monetary	3.44**	2.24*	0.92	Weak	1.07	0.54	0.32	Nil
Trade	3.73**	1.95*	1.77*	Weak	6.56**	2.99*	2.96**	Strong
Exr	2.93**	1.37	0.99	Weak	2.5	1.46	1.23	Nil
Panel B: Supply shock
Total	2.25*	1.25	1.5	Nil	7.54***	5.21***	6.96***	Very strong
Fiscal	0.66	0.55	0.29	Nil	10.83***	5.45***	3.29**	Very strong
Monetary	1.03	0.62	0.8	Nil	3.31*	2.94*	3.84***	Strong
Trade	3.54**	2.74**	1.45	Strong	8.64***	6.99***	5.13***	Very strong
Exr	1.67	1	0.69	Nil	5.83**	3.56**	4.54***	Very strong
Panel C: Demand shock
Total	4.22***	2.35**	1.77*	Strong	6.36**	3.22**	2.63**	Strong
Fiscal	0.35	0.47	1.63	Nil	10.74***	5.47***	3.88***	Very strong
Monetary	1.29	0.71	0.72	Nil	1.36	0.73	1.16	Nil
Trade	1.66	1.15	1.33	Nil	6.71**	3.48**	3.24**	Strong
Exr	2.06	1.28	1.22	Nil	2.79*	1.47	3.61***	Strong
Panel D: Risk shock
Total	3.61**	2.28**	1.71*	Strong	3.31*	3.71**	7.12***	Strong
Fiscal	1.35	1.25	1.18	Nil	13.02***	8.28***	3.83***	Very strong
Monetary	1.76	0.77	0.62	Nil	0.87	2.07	2.80**	Nil
Trade	9.41***	4.32***	3.52***	Very Strong	5.12**	4.69***	5.43***	Very strong
Exr	4.59***	2.16*	1.45	Weak	2.04	2.96*	3.34**	Weak

Note. Significant at 10% (*), 5% (**), and 1% (***) levels. The F-statistic is calculated by the Wald test (with H0: θ1 = θ2 = 0) and compared with the upper bound critical value.

Panels B and C of Table 8 show almost identical results. In particular, the strength of asymmetry is nearly absent in the short term for all models (NARDL, MTNARDL with two and four thresholds), but asymmetry is very strong in the long term. This indicates that the supply and demand shocks have the same effect on categorical EPU. (i.e., null in the short term, but very strong in long term). These results show that the effects of both supply shock and demand shock are almost equal; demand shock does not lead the impact (Antonakakis et al., 2014; B et al., 2018; Kang & Ratti, 2013b; Ro Ba Ys, 2012). In Panel D (risk shock), asymmetry is very strong for trade PU, but weak for exchange rate PU in the short term; however, in the long term it is very strong for fiscal and monetary PU.

Conclusion and Policy Implications

We examined the asymmetric responses of categorical EPUs to oil price and shocks. We followed the approach of Ready (2018) and Li and Guo (2022) and decomposed the oil price change into supply, demand, and risk shocks. Consequently, we set up an empirical structure to catch uneven pass-through utilizing the MTNARDL model provided by Pal and Mitra (2015, 2016). In contrast with traditional NARDL, which disentangles the independent variable to a positive and negative fractional total, we decomposed the oil price change and variety shocks into three and five partial sums. On the basis of the analysis, we yielded careful assessments and more point by point and moment results. More completely, we feature the consequences of the ARDL and NARDL models to uncover the nexus among underlying variables.

The empirical results are as follows: First, the strength of asymmetry between oil price and monetary PU, trade PU, and exchange rate PU is weak in the short term. More specifically, through an analysis of the NARDL model, oil price proved very significant in these three kinds of EPUs in the short term. However, the asymmetric influence only exists given trade PU in the MTNARD model (two and four thresholds). In the long run, the strength of asymmetry is very strong in fiscal PU and trade PU. This implies that when oil prices fall the EPU declines at a slower rate, but increases at a faster pace when they rise. Second, with respect to supply shocks, there are strong asymmetries for all categorical EPUs in the long run, wherein only trade PU has strong asymmetry in the short run; for demand shocks, strong asymmetry exists in the long run for all categorical EPUs except monetary PU, and no asymmetry exists in the short term for all categorical EPUs. Third, regarding the risk shock, very strong asymmetry exists in trade PU in the short and long term, while it only exists in fiscal uncertainty in the long term. Most notably, the asymmetry between oil price shocks and trade PU is significant in most cases. These findings are highly consistent with the economic reality, which is inextricably linked to China’s important role in global industrial chains and reflects the depth of its participation in the global division of labor.

Our empirical analysis indicated that asymmetric impact is more remarkable in the long run in China. It is unprecise to expect that oil price and component shocks have indistinguishable and remarkable impacts on straight out EPU because every EPU holds various sources and channels of transmission. Accordingly, oil price shocks may impact EPUs in a different manner. Future studies may expand on the current one by applying this framework to different countries or industries. Moreover, the usage of MTARDL could be extended.

Our empirical results have three key policy implications. Firstly, Chinese enterprises should pay more attention to the asymmetric influence supply shocks have on trade PU both in the short run and long run; in particular, they should be attentive to the fact that this impact intensifies as the oil price declines. Export enterprises should adjust their quotations based on the production cost to ensure the international competitiveness of their products in global markets. Import enterprises should keep an eye on the price of their import products in order to guarantee business profits and consumer welfare. Secondly, both Chinese enterprises and policymakers ought to identify the source of oil price shocks; long-run supply and demand shocks have the greatest impact for all categorical EPUs. Production and investment activities of enterprises need to be flexibly adapted and some policy tools should be followed up over time. Thirdly, because the asymmetric relationship between oil price shock and trade PU is significant in most cases, enterprises and policymakers need to prioritize trade issues.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Yingce Yang

Junjie Guo

References

Antonakakis

Chatziantoniou

Filis

(2014). Dynamic spillovers of oil price shocks and economic policy uncertainty. Energy Economics, 44, 433–447.

Arouri

M. E. H.

Rault

(2012). Oil prices and stock markets in GCC countries: Empirical evidence from panel analysis. International Journal of Finance & Economics, 17(3), 242–253.

Arouri

M. E. H.

Rault

Teulon

(2014). Economic policy uncertainty, oil price shocks and GCC stock markets. Economics Bulletin, 34(3), 1822–1834.

Baker

S. R.

Bloom

Davis

S. J.

(2016). Measuring economic policy uncertainty*. The Quarterly Journal of Economics, 131(4), 1593–1636. http://doi.org/10.1093/qje/qjw024

Barrero

J. M.

Bloom

Wright

(2017). Short and long run uncertainty (No. w23676). National Bureau of Economic Research.

Beckmann

Czudaj

R. L.

Arora

(2020). The relationship between oil prices and exchange rates: Revisiting theory and evidence. Energy Economics, 88, 104772. https://doi.org/10.1016/j.eneco.2020.104772

Blanchard

O. J.

Gali

(2007). The macroeconomic effects of oil shocks: Why are the 2000s so different from the 1970s? (NBER Working Papers) In Gali

Gertler

(Eds.), International Dimensions of Monetary Policy 2009. University of Chicago Press..

Bloom

(2009). The impact of uncertainty shocks. Econometrica, 77(3), 623–685. https://doi.org/10.3982/ECTA6248

Chen

Sun

(2019). How does economic policy uncertainty react to oil price shocks? A multi-scale perspective. Applied Economics Letters, 27, 188–193.

10.

Chen

Sun

Wang

(2019). Dynamic spillover effect between oil prices and economic policy uncertainty in BRIC countries: A wavelet-based approach. Emerging Markets Finance and Trade, 55(12), 2703–2717.

11.

Choi

Furceri

Yoon

(2021). Policy uncertainty and foreign direct investment. Review of International Economics, 29(2), 195–227. https://doi.org/10.1111/roie.12495

12.

Cologni

Manera

(2008). Oil prices, inflation and interest rates in a structural cointegrated VAR model for the G-7 countries. Energy Economics, 30(3), 856–888.

13.

Crucini

(2000). Oil prices and the terms of trade. Journal of International Economics, 50(1), 185–213.

14.

Cuado

de Gracia

F. P.

(2003). Do oil price shocks matter? Evidence for some European countries. Energy Economics, 25(2), 137–154.

15.

Cunado

Gracia

(2005). Oil prices, economic activity and inflation: Evidence for some Asian countries. The Quarterly Review of Economics and Finance, 45(1), 65–83.

16.

Das

Kannadhasan

(2020). The asymmetric oil price and policy uncertainty shock exposure of emerging market sectoral equity returns: A quantile regression approach. International Review of Economics & Finance, 69, 563–581.

17.

Demirer

Ferrer

Shahzad

S. J. H.

(2020). Oil price shocks, global financial markets and their connectedness. Energy Economics, 88, 104771.

18.

Diebold

F. X.

Yilmaz

(2012). Better to give than to receive: Predictive directional measurement of volatility spillovers. International Journal of Forecasting, 28(1), 57–66.

19.

Chu

(2010). The relationship between oil price shocks and China’s macro-economy: An empirical analysis. Energy Policy, 38(8), 4142–4151.

20.

Filis

Degiannakis

Floros

(2011). Dynamic correlation between stock market and oil prices: The case of oil-importing and oil-exporting countries. International Review of Financial Analysis, 20(3), 152–164.

21.

Fuller

D. W. A.

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statal Association, 79(366), 355–367.

22.

Gelb

(1988). Oil windfalls: Blessing or curse. Oxford University Press.

23.

Hailemariam

Smyth

Zhang

(2019). Oil prices and economic policy uncertainty: Evidence from a nonparametric panel data model. Energy Economics, 83, 40–51.

24.

Hamilton

J. D.

(1983). Oil and the macroeconomy since World War II. Journal of Political Economy, 91(2), 228–248.

25.

Hamilton

J. D.

(1988). A Neoclassical Model of Unemployment and the Business Cycle. Journal of Political Economy, 96, 593–617. http://doi.org/10.1086/261553

26.

Hamilton

J. D.

Herrera

A. M.

(2004). Comment: Oil shocks and aggregate macroeconomic behavior: the role of monetary policy. Journal of Money, Credit and Banking, 36(2), 265–286. https://www.jstor.org/stable/3839020

27.

Huang

Luk

(2020). Measuring economic policy uncertainty in China. China Economic Review, 59, 101367.

28.

Jiang

Cheng

(2021). How the fiscal and monetary policy uncertainty of China respond to global oil price volatility: A multi-regime-on-scale approach. Resources Policy, 72, 102121.

29.

Jiang

Cheng

Cao

Wang

(2022). The asymmetric and multi-scale volatility correlation between global oil price and economic policy uncertainty of China. Environmental Science and Pollution Research, 29(8), 11255–11266.

30.

Jiménez-Rodríguez

Sánchez

(2005). Oil price shocks and real GDP growth: empirical evidence for some OECD countries. Applied Economics, 37(2), 201–228.

31.

Jones

D. W.

Leiby

P. N.

Paik

I. K.

(2004). Oil price shocks and the macroeconomy: What has been learned since 1996. The Energy Journal, 25(2), 1–32.

32.

Kang

Ratti

R. A.

(2013a). Structural oil price shocks and policy uncertainty. Economic Modelling, 35, 314–319.

33.

Kang

Ratti

R. A.

(2013b). Oil shocks, policy uncertainty and stock market return. Journal of International Financial Markets, Institutions and Money, 26, 305–318.

34.

Kang

Ratti

R. A.

Vespignani

J. L.

(2017). Oil price shocks and policy uncertainty: New evidence on the effects of US and non-US oil production. Energy Economics, 66, 536–546.

35.

Kara

(2017). Does US monetary policy respond to oil and food prices? Journal of International Money and Finance, 72, 118–126. https://doi.org/10.1016/j.jimonfin.2016.12.004

36.

Kilian

(2009). Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review, 99(3), 1053-1069.

37.

Kilian

Park

(2009). The impact of oil price shocks on the US stock market. International Economic Review, 50(4), 1267–1287.

38.

Kolodzeij

Kaufmann

R. K.

(2014). Oil demand shocks reconsidered: A cointegrated vector autoregression. Energy Economics, 41, 33–40.

39.

Lee

(2002). On the dynamic effects of oil price shocks: A study using industry level data. Journal of Monetary Economics, 49(4), 823–852.

40.

Guo

(2022). The asymmetric impacts of oil price and shocks on inflation in BRICS: A multiple threshold nonlinear ARDL model. Applied Economics, 54(12), 1377–1395.

41.

Lippi

Nobili

(2012). Oil and the macroeconomy: A quantitative structural analysis. Journal of the European Economic Association, 10(5), 1059–1083.

42.

Loungani

(1986). Oil price shocks and the dispersion hypothesis. The Review of Economics and Statistics, 68(3), 536–539. https://doi.org/10.2307/1926035

43.

Maghyereh

(2006). Oil price shocks and emerging stock markets: A generalized VAR approach. In Global stock markets and portfolio management (pp. 55–68). Springer.

44.

Mork

K. A.

Olsen

Mysen

H. T.

(1994). Macroeconomic responses to oil price increases and decreases in seven OECD countries. The Energy Journal, 15(4), 19–35.

45.

Novy

Taylor

A. M.

(2020). Trade and uncertainty. Review of Economics and Statistics, 102(4), 749–765.

46.

Pal

Mitra

S. K.

(2015). Asymmetric impact of crude price on oil product pricing in the United States: An application of multiple threshold nonlinear autoregressive distributed lag model. Economic Modelling, 51, 436–443.

47.

Pal

Mitra

S. K.

(2016). Asymmetric oil product pricing in India: Evidence from a multiple threshold nonlinear ARDL model. Economic Modelling, 59, 314–328.

48.

Pal

Mitra

S. K.

(2019). Asymmetric oil price transmission to the purchasing power of the US dollar: A multiple threshold NARDL modelling approach. Resources Policy, 64, 101508.

49.

Papapetrou

(2001). Oil price shocks, stock market, economic activity and employment in Greece. Energy Economics, 23(5), 511–532.

50.

Pesaran

M. H.

Shin

Smith

R. J.

(2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Economics, 16(3), 289–326.

51.

Peter Ferderer

. (1996). Oil price volatility and the macroeconomy. Journal of Macroeconomics, 18(1), 1–26. https://doi.org/10.1016/S0164-0704(96)80001-2

52.

Phillips

P. C.

Perron

(1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346.

53.

Pieschacón

(2012). The value of fiscal discipline for oil-exporting countries. Journal of Monetary Economics, 59(3), 250–268. https://doi.org/10.1016/j.jmoneco.2012.03.001

54.

Ready

R. C.

(2018). Oil prices and the stock market. Review of Finance, 22(1), 155–176.

55.

Rehman

M. U.

(2018). Do oil shocks predict economic policy uncertainty? Physica A: Statistical Mechanics and its Applications, 498, 123–136.

56.

Ro Ba Ys

I. V.

(2012). Macroeconomic uncertainty and the impact of oil shocks (Cesifo Working Paper). European Central Bank (ECB).

57.

Shin

Nwoodnimmo

(2014). Modelling asymmetric cointegration and dynamic multipliers in a nonlinear ARDL framework. In Sickles

Horrace

(Eds.), Festschrift in honor of Peter Schmidt: Econometric methods and applications (pp. 281–314). Social Science Electronic Publishing.

58.

Sorana

V. T.

Oana-Ramona

Iulia

Andrei

Fenghua

(2018). Addressing oil price changes through business profitability in oil and gas industry in the United Kingdom. PLoS ONE, 13(6), e199100.

59.

Huang

Qin

Umar

(2021). Does crude oil price stimulate economic policy uncertainty in BRICS? Pacific-Basin Finance Journal, 66, 101519.

60.

Sun

Chen

Wang

(2020). Multi-scale interactions between economic policy uncertainty and oil prices in time-frequency domains. The North American Journal of Economics and Finance, 51, 100854. https://doi.org/10.1016/j.najef.2018.10.002

61.

Wang

Lee

(2022). The dynamic correlation between China’s policy uncertainty and the crude oil market: A time-varying analysis. Emerging Markets Finance and Trade, 58(3), 692–709.

62.

Wei

(2019). Oil price shocks, economic policy uncertainty and China’s trade: A quantitative structural analysis. The North American Journal of Economics and Finance, 48, 20–31.

63.

Wen

Min

Zhang

Yang

(2019). Crude oil price shocks, monetary policy, and China’s economy. International Journal of Finance & Economics, 24(2), 812–827. https://doi.org/10.1002/ijfe.1692

64.

Wen

Yang

(2018). Exploring the rebound effect from the perspective of household: An analysis of China’s provincial level. Energy Economics, 75, 345–356.

65.

Yang

(2019). Connectedness of economic policy uncertainty and oil price shocks in a time domain perspective. Energy Economics, 80, 219–233.

66.

Yin

(2016). Does oil price respond to macroeconomic uncertainty? New evidence. Empirical Economics, 51(3), 921–938.

67.

You

Guo

Zhu

Tang

(2017). Oil price shocks, economic policy uncertainty and industry stock returns in China: Asymmetric effects with quantile regression. Energy Economics, 68, 1–18.

68.

Zhu

Chen

(2019). Asymmetric effects of oil prices and exchange rates on China’s industrial prices. Energy Economics, 84, 104551.

The Asymmetric Impact of the Oil Price and Disaggregate Shocks on Economic Policy Uncertainty: Evidence From China

Abstract

Keywords

Highlights

Introduction

Literature Review

Methodology

Disentangling the Oil Price

Asymmetric Approach

ARDL model

NARDL model

Multiple threshold NARDL

Data

Empirical Analysis and Discussion

Conclusion and Policy Implications

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iDs

References