Do Trade Openness and Output Gap Affect Inflation? Empirical Evidence from BRICS Nations

1. Introduction

For a long time, when inflation had not seemed to be as much of an issue for modern economies, the sudden emergence of the coronavirus pandemic has put it back on the desks of policymakers. Inflation has regained its glory, not only in developed countries but also in middle-income and low-income countries as well. Recent internal and external economic shocks, such as hiked oil prices, the outbreak of coronavirus pandemic, disrupted supply chains and upsurges in government spending among others, have disrupted the whole functioning of economies. Such external shocks hit those economies especially hard when a high share of the imported goods basket is used to fulfil domestic gaps. This includes both developed and developing nations, where these recent factors have highly contributed to driving up the prices that are hard to combat. Such inflationary shocks shake the developing economies with a much higher magnitude than developed countries necessitating actions to control inflation. In this context, the ‘good-luck’ hypothesis states that high economic integration helps to reduce the inflationary pressures by making external shocks more conducive to economic growth. Rapid trade integration among the lower-cost countries depresses the import prices and enhances price competition that results in the availability of goods at lower prices. Consistent with this approach, the ‘globalization’ hypothesis enunciates that more openness to the global markets alters the sensitivity of inflation to its chief driving factors, such as the output gap. In the short run, the domestic output gap acts as one of the chief economic drivers of inflation. However, with the internationalization of the domestic markets, the foreign output gap also arises as one of the crucial factors affecting inflation. This foreign measure of slack generates synchronicity of price shocks across the economies and acts as a common driving force for inflation, such as the Russia–Ukraine war. The Ukraine war has worldwide influenced the oil import prices which increased inflation in all the countries, particularly in developing countries such as Brazil, Russia, India, China and South Africa (BRICS). This invokes the question: Whether opting for the trade liberalization policies is really a good choice to combat inflation? And what roles do the domestic and foreign output gaps play in this process?

Many empirical studies conducted, in this context, report that openness to trade suppresses inflationary pressures (Bowdler & Malik, 2017; Haq et al., 2016; Jafari et al., 2012; Jedidia et al., 2019; Kim et al., 2012; Lin, 2010; Lin et al., 2017;). On the other hand, some studies mention that internationalization of trade intensifies inflation to a higher degree (Ajaz et al., 2016; Lotfalipour et al., 2013; Mukhtar et al., 2019; Zakaria, 2011) or has a negligible impact on the price (Ghosh, 2014). As a result, ambiguities still persist about the question that whether trade openness is an effective measure to control inflation or not.

Most of the previously conducted studies exclusively focused on analysing the impact of trade openness on inflation and do not take into account the impact of the output gap on the trade openness and inflation nexus. There is only a dearth of studies that emphasize examining the effect of the output gap on the relationship between trade openness and the output gap. Some studies contend that the effect of the domestic output gap is more prominent than the foreign output gap (Yolcu Karadam, 2020). Whereas, the evidence also found a strong positive relationship between the foreign output gap and inflation levels (Ihrig et al., 2010; Milani, 2010; Nawatmi et al., 2020; Wynne & Kersting, 2007). This suggests that, in the presence of globalization, both domestic and foreign output gaps have a significant impact on domestic inflation and should be revisited to control any inflationary bias.

Since the 1990s, high economic integration in many developing economies, including BRICS, has resulted in dynamic changes in the trading patterns of the respective countries. With an average share of approximately 17% of the total world’s trade, BRICS is among the few highest exporting and importing countries that represent more than half of the world’s total population with a total GDP of more than $16.16 trillion. The pattern of trade openness (defined as total trade as a percentage of GDP) is also changing in BRICS nations, from 20.98% to 28.98% in Brazil, 33.52% to 37.56% in China, 24.81% to 43.59% in India, 69.39% to 51.58% in Russia and 42.25% to 54.51% in South Africa, from 1999 to 2018 (as shown in Figure 1). Among these economies, China accounts for the largest country of all the five with the GDP of $14.34 trillion (World Bank, n.d. matched but without year). Moreover, the changes in consumer price index (CPI) values show a dramatic increase in prices from 34.43 to 114.67 in Brazil, 70.50 to 104.97 in China, 34.52 to 109.91 in India, 15.14 to 112.52 in Russia and 43.42 to 114.99 in South Africa, from 1999 to 2018 (as shown in Figure 2). This shows that both trade openness and inflation are trending upward but with fluctuating output gaps. The overall picture of the output gap shows that China has the highest whereas South Africa has the lowest domestic output gap throughout (as shown in Figure 3). However, the situation is opposite for the foreign output gap, with China at the lowest and South Africa and Brazil at the highest (as shown in Figure 4). This also highlights a boost in inflation and motivates a more thorough analysis of the effects of trade openness and output gap on inflation more thoroughly in the BRICS countries. In addition, the uneven weightage of exports and imports weights in the BRICS trading basket highlights the need to measure the impact of exports and imports separately. The lack of studies in this context for BRICS motivates further evidence gathering, which will help decision-makers structure better policy frameworks.

OPEN IN VIEWER

Figure 1.

Source: The authors.

OPEN IN VIEWER

Figure 2.

Source: The authors.

OPEN IN VIEWER

Figure 3.

Source: The authors.

OPEN IN VIEWER

Figure 4.

Source: The authors.

In the above context, the study contributes to the literature in three ways: (i) It investigates the impact of trade openness on inflation by incorporating all the trade aspects that is, export, import and overall trade openness separately for the BRICS nations. (ii) It analyses the impact of the output gap on the relationship between trade openness and inflation, which to the best of the author’s knowledge has been examined by very few studies, especially for the BRICS countries. (iii) Since most studies assume cross-sectional independence among countries, which is extremely unrealistic in this highly globalizing era, this study attempts to examine the impact of trade openness and output gap on inflation in the presence of cross-sectional dependence state-of-the-art technique, providing a more robust analysis.

Against the above backdrop, the study answers three research questions: (i) Whether trade openness increases or decreases the inflation in the BRICS countries?, (ii) Does the output gap (domestic and global) affect the price levels in these economies and if yes, then how? (iii) Do the trade openness and output gap act as causal factors affecting inflation or vice versa?

The rest of the paper is structured in the following way: the next section of the paper shows the empirical review of the present literature assessing the impact of trade openness on inflation. The third section reflects the various data sources and variable formulation techniques for the variables used in the study. The details relating to the econometric methodology utilized for the empirical analysis have been discussed in the fourth section. The results of the empirical analysis for the respective link have been mentioned in the fifth section of the study followed by the last section containing the conclusion and policy implications.

3. Data Description

To examine the effect of trade openness and output gap on inflation, the study employs a quarterly dataset for BRICS from 1999Q1 to 2018Q4. The motivation for the time of this study is entirely based on the quarterly data availability. While data for many variables are extensively available, for a few indicators such as money supply, the quarterly data are only available since 1999Q1. This section of the paper deals goes into detail about the data sources, definitions and variable construction methods. All the variables in the study have been de-seasonalized and are used in the natural log form.

To capture the effect of trade openness on inflation, the study utilizes the measure of CPI inflation or headline inflation over core inflation as a proxy for inflation. When compared to core inflation, the headline inflation includes foreign goods in its basket of goods. Therefore, the CPI inflation is considered as the best measure of inflation in the study. The data are derived from Federal Reserve Economic data with 2015 as the base year. Further, compared to the other estimates of trade openness such as export and import intensity indices (Sawhney & Kiran, 2019) and following the literature, the study uses the trade-to-GDP ratio as a measure of trade openness. It is defined as the ratio of the total trade of goods and services to GDP. It measures the extent of the economy’s dependence on foreign market goods and services, including exports and imports. Since the exports and imports impact differently to inflation, therefore, this study examines their impact separately on inflation, using export openness and import openness. Here, export openness can be described export of goods and services to the total GDP, whereas import openness refers to the import of goods and services to the total GDP of an economy. The available literature shows studies use only the trade-to-GDP ratio as an openness indicator (Bowdler & Malik, 2005; Farvaque & Sarfaraz, 2009; Hanif & Batool, 2006; Jin, 2006; Kurihara, 2013; Lotfalipour et al., 2013; Yiheyis, 2013; Zakaria, 2011), only import openness as an indicator (Bowdler & Nunziata, 2006; Jin, 2006; Lin, 2010; Romer, 1993; Temple, 2002; Terra, 1998), as well as both export openness and import openness as a proxy to openness (Alfaro, 2005). As a result, this study uses all three proxies to represent the economy’s openness. The variable is constructed following the methodology as per definition using the Direction of Trade Statistics dataset by IMF.

To analyse the output–inflation trade-off and capture the effect of globalization through the Phillips curve, a separate variable, the output gap has been constructed. The output gap is the difference between the potential and the actual output of the economy in terms of potential output. Here, the potential output can be expressed in terms of the decelerating or non-accelerating price range, which is not provided by various simple techniques. Furthermore, the required datasets and the size of econometric models used to measure it are impractical. Thus, following in the footsteps of Jašová et al. (2020), the study directly estimates the domestic output gap by applying the univariate Hodrick-Prescott filter to the quarterly estimates of real GDP. The method estimates the long-run series by discounting any short-run price fluctuations using the smoothing parameter λ. For the quarterly data, the λ that is smoothing parameter is set at 1,600. For this purpose, the study uses real GDP data series from the Global Economic Monitor data by World Bank. It defines the goods and services produced in an economy during a particular period by using the prices in the base year. While, the output affects prices through marginal costs of the firms that is by pushing up the prices of imported raw material, machinery and consumer goods, its effect depends upon the domestic output as well as foreign output (Borio & Filardo, 2007). A larger foreign output gap raises demand for domestically produced goods, causing domestic wage levels to rise. This raises the firm’s production costs, resulting in high inflation (Engel, 2011). Therefore, the study includes both domestic and foreign output gap for the analysis. For this purpose, the trade-weighted foreign output gap is constructed as

Gap m G = ∑ ​ k ∈ K ( w m . n i Gap n D + w m . o i Gap o D + w m . p i Gap p D + w m . q i Gap q D ) 5 ,

where w m . n i defines the trade weights, G a p n D explains the domestic gap for the trading country n. The weights for the trade-weighted global output gap are assigned by using the following formula:

w m . n = i m p o r t s m . n + exp o r t s m . n T o t a l t r a d e m

Looking at the significance of the foreign output gap effect, the expression uses the exports and imports done by each country within BRICS. The positive output gap indicates that actual output outweighs the potential output resulting in higher inflation due to increased demand, and vice versa for the negative output gap. However, the zero-output gap reflects price stability.

To measure the effectiveness of the monetary policy, the study uses the M3 money supply indicator for the sample countries in terms of national currency taken from the Federal Reserve of Economic Data. The data have been standardized using the Purchasing power parity rate by the OECD. The study assumes that the money supply posits a direct proportional relationship with inflation. According to the quantity theory of money and the real balance effect, the holding amount of real cash balance depends on the predicted inflation rate that affects people’s purchasing power. According to Fisher, assuming the full employment and speed of money flow, any change in the volume of money is reflected in the price level (Sepehrivand & Azizi, 2016). This suggests that the expected relationship between inflation and money supply is positive.

To measure the effect of fiscal policy on inflation, the study uses government final consumption expenditure by the World Development Indicators. Government spending impacts inflation in three ways: by the change in the money supply, growth of output and tax base (Varvarigos, 2010). Increased government spending necessitates an increase in public revenue, which is typically generated through high taxation or a monetary base. This results in high prices and reflects a positive relationship between government spending and inflation (Basu, 2001).

To examine the impact of the exchange rate on inflation, the real effective exchange rate (REER) was used. According to proponents of the cost-push theory, exchange rate depreciation leads to more expensive imports, resulting in imported inflation (Ajaz et al., 2016; Jedidia et al., 2019; Mukhtar, 2010). In this light, the expected direction of effect is assumed to be positive and extracted from the Bruegel database.

4. Methodology

This section lays out the framework for investigating the relationship between trade openness, output gap and inflation. The main estimating model related to inflation that is, measured by the consumer price inflation, for economy ‘i’ in the time ‘t’ can be expressed as

CPI i , t = α + β 1 TO i , t + β 2 RGDP i , t + β 3 REER i , t + β 4 GEXP i , t + β 5 MS i , t + μ i , t ,

(1)

C P I i , t = α + β 1 I O i , t + β 2 R G D P i , t + β 3 R E E R i , t + β 4 G E X P i , t + β 5 M S i , t + μ i , t ,

(2)

C P I i , t = α + β 1 X O i , t + β 2 R G D P i , t + β 3 R E E R i , t + β 4 G E X P i , t + β 5 M S i , t + μ i , t ,

(3)

where C P I i , t is the consumer price inflation, TO _i,t denotes openness to trade, RGDP _i,t shows the real gross domestic product, REER _i,t denotes real effective exchange rate, GEXP _i,t stands for government final consumption expenditure, MS _i,t is the money supply and μ _i,t represents error term. Since the sample countries are a mix of export-oriented and import-oriented countries, the study examines the impact of export openness (XO _i,t ) and import openness (IO _i,t ) separately, as mentioned in (2) and (3). To measure the impact of trade openness on inflation with the output gap, the following equations have been examined for all the openness measures:

C P I i , t = α + β 1 T O i , t + β 2 D O i , t + β 3 F O i , t + β 4 R G D P i , t + β 5 R E E R i , t + β 6 G E X P i , t + β 7 M S i , t + μ i , t

(4)

C P I i , t = α + β 1 I O i , t + β 2 D O i , t + β 3 F O i , t + β 1 R G D P i , t + β 1 R E E R i , t + β 1 G E X P i , t + β 1 M S i , t + μ i , t

(5)

C P I i , t = α + β 1 X O i , t + β 2 D O i , t + β 3 F O i , t + β 1 R G D P i , t + β 1 R E E R i , t + β 1 G E X P i , t + β 1 M S i , t + μ i , t

(6)

where, DO _i,t means domestic output gap, and FO _i,t represents foreign output gap for economy ‘i’ in the time ‘t’. To begin with the empirical analysis of the models, the study examines the phenomenon of cross-sectional dependence between the countries. In today’s increasingly globalized economies, cross-sectional dependence is a common occurrence, and ignoring it may result in inefficient and inaccurate statistics. Therefore, the diagnostic to examine the presence of cross-sectional dependence among panels is critical to the analysis and further useful in determining the subsequent steps to be taken for the analysis. To detect the issue, the study employs the Breusch and Pagan LM test (1980), Pesaran CD test (2004) and Pesaran scaled LM test (2004). The diagnostics test the null hypothesis that there is no cross-sectional dependence in the panels. For a panel with small cross-sectional units, the Breusch and Pagan LM test (1980) comes out as an appropriate choice. On the other hand, the Pesaran scaled LM test (2004) is the best choice for the larger panels with larger time dimensions. However, the results for the tests deteriorate in the case of only the cross-sectional units and the time dimension is large. To overcome the issue, the Pesaran CD test (2004) is conducted. The test is effective for the panels specifying both smaller units in cross-section and the time dimension. In addition, the study also investigates the homogeneity of the panels using the delta test by Pesaran and Yamagata (2008). The test examines whether the coefficients of the slope are identical or homogeneous across the individual units.

Since the first-generation unit root fails to provide efficient results in the presence of cross-sectional dependence and heterogeneity, the study employs second-generation unit root tests to check the stationarity of the variables. One of the second-generation tests used for the analysis is the cross-section ADF (CADF) by Pesaran (2003). The test assumes the null hypothesis of non-stationary data. To check the stationarity among the heterogeneous panels, CADF uses ADF or DF regression by augmenting it with cross-section averages of the first-differenced and lagged values of the individual test series which eliminates the cross-section dependence and provides efficient results for the models showing linear trends and the presence of serial correlation among residuals. Assuming a similar null hypothesis, the study also employs the cross-sectionally augmented IPS test (CIPS) by Pesaran (2007) as another method used to examine stationarity and to check the reliability of results.

After confirming the stationarity of data, the study employs the panel cointegration technique by Westerlund (2007) to examine if there exists a long-run relationship among the panels. Over other tests of cointegration such as the Pedroni and Kao cointegration tests, the Westerlund cointegration test allows the presence of cross-sectional dependence and heterogeneity among the panels, which makes the Westerlund cointegration the best approach for this study. This error-correction-based test assumes the null hypothesis of no cointegration. The rejection of the null hypothesis indicates, and that there is a long-run relationship among the variables and that all the panels or only some panels are cointegrated.

Further, to estimate the long-run coefficients for the variables in the model, the study applies the modern method of the dynamic common correlated effect model (DCCE model) by Chudik and Pesaran (2015), which checks for the cross-sectional effects and has proven to be a better approach than GMM, fixed effect and random effects method. The DCCE model is developed and based on the technique of mean group designed by Pesaran and Smith (1995), pooled mean group designed by Pesaran et al. (1996), and CCE designed by Pesaran (2006). While the analysis for many countries and variables in the panel data models at the same time causes the presence of some commonly unobserved factors, the DCCE technique considers the multifactorial error structure and filters the individual regressors by the means of the cross-section averages. Thus, this method helps to remove such unobserved common factors among the panels which results in more reliable outcomes.

In addition, the study employs the Dumitrescu and Hurlin panel causality test to examine the relationship between the variables (2012). While the classic Granger causality approach relies on the homogeneity and no interconnection of the panel units, the Dumitrescu–Hurlin test is more flexible. Dumitrescu–Hurlin causality test allows the heterogeneity and interconnection between the individual units and helps to identify the Granger non-causality from endogenous variable to exogenous variable as well. As a result, the Dumitrescu–Hurlin causality test is better for the investigation at hand.

This section of the paper presents the findings on the relationship between trade openness, output gap and inflation. The significant delta test values obtained by applying the Pesaran and Yamagata (2008) test of homogeneity in Table 1 conclude that the null hypothesis of homogeneous panels in the model is rejected. This lends credence to the notion that panels are not homogeneous.

Since greater trade integrations increase the likelihood of cross-sectional dependence among the countries, the study used the Breusch-Pagan LM test, Pesaran scaled LM test and Pesaran test to examine the cross-sectional dependence in the panels. The empirical evidence in Table 2 from all three tests indicates the presence of cross-sectional dependence in the residuals. As a result, the null hypothesis of no cross-section dependence is rejected. This affirms the use of the second-generation unit root approach to test stationarity.

OPEN IN VIEWER

Table 2.

Cross-section Dependence Test Results

Tests	Statistics	P Value	Results
Breusch-Pagan LM	250.0480	0.0000*	There is cross dependence in residuals
Pesaran Scaled LM	52.5583	0.0000*
Pesaran CD	6.3517	0.0000*

Source: The authors.

Notes: The table exhibits the estimates for the presence of cross-country dependence among the BRICS countries. The asterisk signifies the significance level of the t-statistic value: *P < 0.01.

For this purpose, CADF and CIPS test have been employed. For the cross-dependent panels, both the tests assume the null hypothesis of non-stationary data. The results indicate the combination of stationarity order from both the tests, that is, at the level or first difference (as shown in Table 3). None of the variables is found stationary at the second difference. Few variables such as CPI, TO, DO, FO, REER and MS are stationary at first difference whereas others are stationary at level.

OPEN IN VIEWER

Table 3.

Unit Root Test Results

Variables		CADF Test		CIPS Test
		Constant	Constant and Trend	Constant	Constant and Trend
CPI	Level	−2.20	−2.25	−2.49	−2.37
TO		−2.70	−3.06*	−3.36	−3.64 *
XO		−3.39*	−3.95*	−3.79*	−4.27*
IO		−3.14*	−3.18*	−4.44*	−4.44*
DO		−2.74*	−2.77	−4.58	−4.64*
FO		−2.76*	−2.95	−3.41	−3.61*
RGDP		−1.62	−2.21*	−1.28	−2.30
REER		−2.76*	−2.73	−2.54	−2.28*
GEXP		−3.24*	−4.62*	−2.28	−2.15
MS		−3.42*	−2.21	−3.52	−2.14
CPI	First difference	−5.08*	−5.24*	−5.48	−5.65*
TO		−6.07*	−6.42*	−6.19	−6.42*
XO		−6.06*	−6.23*	−6.19*	−6.42 *
IO		−6.19*	−6.42*	−6.19*	−6.42 *
DO		−6.19*	−6.41*	−6.19	−6.42 *
FO		−6.03*	−6.19*	−6.19	−6.42*
RGDP		−4.54*	−5.15*	−5.47*	−6.03*
REER		−5.40*	−5.42*	−6.12	−6.30*
GEXP		−4.66*	−4.79*	−2.22*	−4.66*
MS		−4.46*	−4.53*	−6.19	−6.42*

Source: The authors.

Notes: The table presents the results for checking the presence of stationarity in the variables. Each column shows the calculated t-statistic values for the respective variables. The asterisk signifies the significance level of the t-statistic value: *P < 0.05.

After confirming the stationarity, the study employs Westerlund’s (2007) cointegration approach to examine the long-run relationship among the variables. This approach examines the presence of a long-run relationship assuming the null hypothesis of no cointegration in the panel. The results show that the statistical values for all models with dependent variables trade openness, import openness and export openness are significant (as shown in Table 4). This implies the rejection of the null hypothesis for all the three dependent variables of the model and confirms the long-run relationship among the variables.

OPEN IN VIEWER

Table 4.

Westerlund Test of Cointegration Results

Model with Varying Openness Indicator	H1: Some Panels Are Cointegrated		H1: All Panels Are Cointegrated
	Statistics	P Value	Statistics	P Value
Trade openness	2.9137	0.0018**	3.3639	0.0004**
Export openness	2.6373	0.0042**	2.6121	0.0045**
Import openness	1.9866	0.0235*	1.7861	0.0370*

Source: The authors.

Notes: The table exhibits the estimates for testing the cointegration in the presence of cross-country dependence as given by Westerlund (2007). It shows the outcomes for the existence of long-run relationship among the variables taking into account all three different measures of openness as varying independent variables in the models. The asterisk signifies the significance level of t-statistic value: *P < 0.05, **P < 0.01.

Further, Table 5 presents the estimated results for the long-run coefficients using the DCCE test. The empirical evidence demonstrates the existence of a negative significant relationship between trade openness and inflation. The findings are in line with the outcomes of Sahu and Sharma (2018) against the positive relationship obtained by Chhabra and Alam (2020). A percentage change in trade openness causes a 0.64% negative change in inflation, resulting in lower CPI. This implies that as global trade integration increases, domestic prices will fall. In the case of import openness, the outcomes are consistent with both theoretical and empirical literature (Bhatti et al., 2021; Bianchi & Civelli, 2015; Satti et al., 2013) confirming a positive relationship between import openness and inflation. Increased imports lead to increased trade, which raises the steepness of the Phillips curve, resulting in higher prices and a rise in quantity demanded. This increases the prices of imported goods such as crude oil, petroleum products, etc. causing inflation at both the consumer end as well as the producer end. In contrast to import openness, export openness posits a negative relationship with the prices and indicates that a percentage increase in exports causes a 2.29% decrease in inflation. This implies that the global economic integration flattens the Phillips curve for the BRICS countries in terms of export openness.

OPEN IN VIEWER

Table 5.

DCCE Mean Group Test Results

Variables	Long-run Coefficients
	(1)	(2)	(3)	(4)	(5)	(6)
RGDP	0.2363 (1.18)	0.3037 (1.26)	0.3727 (1.12)	0.3150 (1.12)	0.2851 (1.13)	0.5636 (1.01)
REER	−0.2763** (−2.22)	−0.2386 (−1.57)	−0.1964 (−1.10)	−0.3823*** (−2.59)	−0.3540** (−2.29)	−0.3197* (−1.95)
GEXP	−0.0284 (−0.50)	−0.0336 (−0.59)	−0.0513 (−0.60)	−0.0306 (−0.68)	−0.0184 (−0.66)	−0.0738 (−0.89)
MS	0.0968*** (2.77)	0.0908*** (2.84)	0.0901*** (2.59)	0.1049*** (2.19)	0.1051** (2.42)	0.0983* (1.72)
TO	−0.0984 (−0.40)	–	–	−0.6442* (−1.72)	–	–
IO	–	0.6235 (0.80)	–	–	1.3941 (1.37)	–
XO	–	–	−1.0620 (−1.06)	–	–	−2.2945 (−1.21)
DO	–	–	–	−0.0018 (−0.07)	0.0060 (0.24)	−0.0076 (−0.23)
FO	–	–	–	0.0562** (2.85)	0.0569*** (2.58)	0.0855* (1.91)
R 2	0.50	0.50	0.41	0.43	0.42	0.41

Source: The authors.

Notes: The table indicates the results for the long-run coefficients estimated by the DCCE model for the mean group. The inflation (measured by the CPI) is used as a dependent variable for each column model. The initial three columns estimate the openness effect on price levels without including the effect of the output gap effect. However, the later columns represent the estimates accounting for the domestic and foreign factors as well. The Hodrick-Prescott filter has been used for the construction of the output gap variable. The parenthesis reports the t-statistic values calculated for each column. The asterisk signifies the significance level of t-statistic value: *P < 0.10, **P < 0.05 and ***P < 0.01.

Furthermore, the statistical results for BRICS countries show a positive relationship between the foreign output gap and inflation, with the effect becoming stronger as export openness increases. In contrast, the pattern of the domestic output gap effect differs from that of the foreign output gap. The findings indicate that domestic forces are effective in combating inflation and contribute to the flattening of the Phillips curve. However, the prices in these countries are relatively more sensitive to global factors than to domestic factors. This indicates that the underpowered effect of the domestic output gap is not sufficient to reduce inflation, which eventually results in a steeper Phillips curve. The results are consistent with the sub-sample outcomes of Jašová et al. (2020) and highlight that with the possible increase in trade openness, the prices become more vulnerable to global factors than local or domestic factors.

Similar to the effect of global factors, the results highlight that an increase in money supply creates inflationary pressure in BRICS. The empirical finding of Ashra (2002), Jedidia et al. (2019) and Van (2019) supports the result of this study. The results also support the theoretical arguments by monetarists and demonstrate that 1% increase in money supply results in approximately a 0.10% hike in prices. The reason behind this is that the adoption of expansionary monetary policies causes depreciation of the exchange rate which results in bloated import prices. These price increments are further reflected in the domestic prices in proportion to the economy’s sensitivity to global factors and degree of trade openness. As a result, it can be said that the money supply is a significant contributor to rising inflationary pressures.

In contrast, the results show that an expansionary fiscal policy is conducive to the reduction in prices, where a 1% increase in government final consumption expenditures reduces inflation by approximately 0.03%. Consistent with the results of Batool et al. (2022), the results show that the dominance of fiscal policy is vital to attaining price stability. In this instance, the effect of government spending is channelized via interest rates. Increased government spending raises interest rates, which harms private investment and aggregate demand. This generates a crowding-out effect in the economy and brings down the price levels by targeting the government spending on lower-income households (Nguyen, 2019; Romer, 2012). Comparing the effect of monetary and fiscal policy, the results show that the prices are relatively more responsive to the monetary policy tools. This demonstrates that, when used prudently, monetary policy can be a more effective tool to combat inflation in BRICS countries.

The findings also show that the real GDP posits a positive relationship with inflation, where, a 1% increase in economic growth causes approximately a 0.30% increase in prices. The findings are consistent with those of Cai (2008) and Sahu and Sharma (2018) for India. Furthermore, the findings highlight a negative relationship between the REER and inflation, where, a 1% decrease in the REER leads to approximately a 0.35% increase in inflation. The possible rationale behind this can be the heavily weighted import products to the total trading baskets. The heavy reliance of countries on imported products such as crude oil and petroleum products may cause imported inflation in the economy resulting in increased domestic prices in BRICS countries.

Further, the study also uses a causality test developed by Dumitrescu and Hurlin (2012) to investigate the causal relationship among the variables. The test assumes the null hypothesis for the absence of any causal relationship between the variables. For the overall sample, the evidence in Table 6 shows the existence of a unidirectional relationship from trade openness, import openness, domestic output gap and foreign output gap to inflation. This exhibits that the rising trend towards openness and high weightage of foreign goods has triggered domestic prices throughout the time. The possible explanation for the unidirectional relationship can be attributed to the increasing firm competitiveness in the global market at the expense of small and local domestic firms. When entering the global market, most domestic firms fail to survive competition against the global firms that provide goods at relatively cheaper prices. Moreover, the overall results highlight the existence of a feedback effect between inflation and other variables, namely, export openness, real GDP, REER, government final consumption expenditure and money supply. The rationale behind this is that high inflation increases the cost of inputs and machinery which eventually shoots up the production cost and hence increases the prices of export goods. This reduces the global competitiveness of the domestic economies leading to higher current account deficits that cause currency depreciation. Such currency depreciation can also be the result of expansionary government policies aimed at boosting economic growth. This increases the aggregate demand for both foreign and domestically produced goods, where a greater proclivity of demand towards the foreign goods increases the deficits, resulting in currency depreciation.

OPEN IN VIEWER

Table 6.

Dumitrescu and Hurlin (2012) Granger Non-causality Test Results

Direction of Causality	Overall	Brazil	China	India	Russia	South Africa
	z-bar value	Coefficient Values
CPI → TO	0.0721	0.0048 (1.03)	-0.256 (-1.16)	-0.0043 (-0.62)	0.0048 (0.72)	0.0113 (0.82)
TO → CPI	7.3373***	-0.0342 (-0.97)	0.0421*** (3.90)	0.0811*** (3.44)	0.0092 (0.38)	0.0032 (0.19)
CPI → IO	0.1169	0.0035 (1.06)	0.0003 (0.02)	-0.0001 (-0.03)	0.0101* (1.92)	0.0071 (0.75)
IO → CPI	8.1648***	-0.0098 (0.856)	0.0703*** (3.81)	0.1219*** (3.77)	0.0572 (1.16)	0.0201 (0.82)
CPI → XO	2.4786***	0.0156*** (2.81)	0.0281 (1.12)	0.0001 (0.02)	0.0081 (1.35)	0.0117 (0.179)
XO → CPI	7.2574***	-0.1215* (-1.83)	0.0601*** (3.56)	0.1895*** (3.37)	0.0195 (0.56)	-0.0169 (-0.54)
CPI → DO	-1.2974	0.0584 (-0.48)	0.1355 (0.46)	0.0190 (0.20)	-0.0388 (-0.62)	-0.0099 (-0.17)
DO → CPI	7.1196***	0.0025 (1.43)	0.0038*** (2.60)	-0.0017 (-0.44)	0.0034 (1.11)	0.0125*** (4.15)
CPI → FO	-1.0459	0.269 (0.42)	0.0084 (0.09)	0.141 (0.79)	0.0324 (0.86)	0.0258 (0.37)
FO → CPI	0.0511**	-0.0034 (-1.27)	0.0055** (3.4)	-0.0013* (2.12)	-0.0026 (-0.68)	0.0023 (0.69)
CPI → RGDP	35.2057***	-0.0131 (-0.66)	-0.4864*** (-10.73)	-0.0157 (-0.30)	-0.0023 (-0.10)	-0.0164 (-0.77)
RGDP → CPI	10.2046***	0.0045 (0.62)	0.0107*** (4.47)	0.0271*** (2.42)	0.0036 (0.33)	0.0387*** (3.31)
CPI → REER	1.8785*	0.0088 (0.44)	0.0904*** (3.10)	0.0065 (0.90)	0.0014 (0.10)	-0.0127 (0.566)
REER → CPI	13.0403***	-0.0138** (-2.05)	-0.0534*** (-4.74)	-0.0212 (-0.58)	-0.0080 (-0.72)	-0.0343*** (-4.33)
CPI → GEXP	18.0920***	-0.0174 (-0.30)	-1.4331*** (-7.46)	-0.0952* (-1.77)	-0.0818* (-1.83)	0.0010 (0.02)
GEXP → CPI	13.1543***	-0.0025* (-1.75)	0.0139*** (5.34)	0.0151*** (3.41)	-0.0016 (-0.85)	-0.0031 (-1.64)
CPI → MS	11.6255***	-0.0483 (-0.51)	-0.5771*** (-3.08)	-0.4181*** (-5.09)	-0.0694 (-0.28)	-0.2044*** (-2.46)
MS → CPI	10.9307***	0.0036 (1.19)	0.0084*** (4.01)	0.0184*** (3.95)	0.0082* (1.71)	0.0044* (1.88)

Source: The authors.

Notes: The table indicates the estimates for the presence of causality among the variables. The results show the coefficient values for all the variables. The parenthesis reports the t-statistic values calculated for each country separately. The asterisk signifies the significance level of t-statistic value: *P < 0.10, **P < 0.05 and ***P < 0.01.

Looking at the country-specific causal linkages, in the case of Brazil, the results in Table 6 show evidence of a one-way causal relationship from the REER and government final consumption expenditure to CPI. On the other hand, export openness and inflation posit a two-way causal relationship. This highlights that adopting an expansionary fiscal policy increases the harmful effects of inflation which is detrimental to the economic growth of Brazil. For China, the results are similar to the findings for overall samples. The evidence shows one-way causal effect from every explanatory variable to CPI, whereas the two-way causal link can be observed between inflation and export openness, real GDP, REER, government final consumption expenditure and money supply. Similarly, for India, one-way causal linkage can be evidenced from every explanatory variable to CPI, except the domestic output gap. This shows that compared to the domestic factors, the foreign factors play a more significant role in determining domestic prices for India. Since the external factors intensify prices, there is a possibility that to some extent, the pace of economic growth is dependent on the trading activities which makes a large consumer share leans towards foreign-based firms in the domestic market. The results also show the existence of a two-way relationship between inflation and money supply, and government expenditure. This implies that the prices are highly sensitive to both fiscal and monetary policy, hence, requires a more cautious policy approach to attain higher growth with controlled inflation. Concerning the causal linkages for Russia, the results show a unidirectional relationship from inflation to import openness and government final consumption expenditure, and from money supply to inflation. This underlines that the high domestic prices with increased money supply results in chasing more goods than are produced domestically, which attracts the demand for cheaper imported products that eventually increases inflation. Similarly, for South Africa, the money supply and inflation show the presence of a significant causal relationship but bidirectional. On the other hand, the evidence highlights a significant unidirectional causal link from the domestic output gap, real GDP and REER to inflation. This clarifies that domestic factors are more dominant than foreign factors in determining the domestic price levels.

Without unequivocal empirical evidence about the direction of effects, the debate over the impact of trade openness and output gap on inflation is far from over. Multiple previous studies in the literature report conflicting results. Moreover, they assume cross-sectional independence among countries, which is extremely unrealistic in this highly globalizing era. To address this issue and contribute empirically, this paper undertakes a cross-country analysis to examine the impact of trade openness and output gap on inflation in BRICS countries using the quarterly dataset from 1999 to 2018. The study employs a state-of-the-art DCCE model, taking into account the existing cross-sectional dependence, serial correlation and heterogeneity among the panels, to estimate the long-run relationship among variables. Further, Dumitrescu and Hurlin’s (2012) causality test has been used to investigate the causal relationship between the variables. The study utilizes CPI as an independent variable and a proxy for inflation. Trade-to-GDP, export-to-GDP and import-to-GDP have been used as independent variables and taken as proxies for openness indicators. In addition, the model incorporates the foreign output gap and domestic output gap as control variables. Other control variables such as money supply, REER and government final consumption expenditure have also been incorporated into the analysis.

The results show that higher trade openness is a panacea for the rising domestic inflation in the BRICS countries. The results are consistent with the ‘new growth’ theory suggesting higher economic integration as a tool to combat inflation. Analysing the openness measures separately reveals that import openness and export openness creates an insignificant effect on prices. It is observed that higher import openness creates an unfavourable impact on inflation, whereas, a higher inclination towards exports creates favourable impact on inflation. In BRICS nations, the price-reducing impact of export openness outperforms the unfavourable impact of imports resulting in a flattened Phillips curve. Notably, the results contradict the conventional wisdom which states that only domestic measures determine domestic inflation in the economy. The findings show that the prices in these countries are relatively more responsive to global factors than domestic factors. This indicates that, contrary to the previous belief, the under-powering effect of the domestic output gap is not sufficient to reduce inflation. The effect of the global factors on domestic prices expands more with the increasing export activities. This makes export-inclined countries such as Brazil, China and India more vulnerable to foreign demand shocks. Furthermore, the findings support the monetarist’s arguments by finding that expansionary monetary policy contributes to rising inflationary pressures. In contrast, an expansionary fiscal policy promotes price reduction. Since the prices are more responsive to the monetary policy tools than fiscal policy, the results demonstrate monetary policy are more effective when implemented cautiously.

Therefore, the study highlights that both domestic and foreign output gaps are crucial to domestic prices. Therefore, policymakers should pay close attention to the fluctuating price levels of foreign commodities in the global market. By anticipating the higher import prices, the central banks and authorities should cushion the price levels in the domestic market with the help of effective monetary and fiscal policy. This will help to reduce the susceptibility of domestic prices to international markets. It is also suggested to emphasize offsetting of other economic factors such as oil and petroleum products, which weigh heavily in import good baskets of these developing countries. Since export openness comes out as an effective tool for price reductions, policymakers should encourage export-based firms by providing subsidies and tax concessions while keeping the monetary policy in check. Also, in recent years, the rapidly rising geopolitical and geoeconomic tensions around the globe have resulted in the gradual erosion of bilateral and multilateral trade agreements causing disturbances in trading circles. This endangers the global economic cooperation among major economic powers like BRICS. To counter this, it is suggested that policy structures are framed so that coordinated actions with trading and neighbouring countries are promoted. This will help in sustaining the achieved economic growth in the long run and help nations realize maximum trade benefits in the form of lower prices.