Spillover effects of the US economic policy uncertainty in Latin America

Coronado, Semei; Martínez, José; Venegas-Martínez, Francisco Article Spillover effects of the US economic policy uncertainty in Latin America Estudios de Economía Provided in Cooperation with: Department of Economics, University of Chile Suggested Citation: Coronado, Semei; Martínez, José; Venegas-Martínez, Francisco (2020) : Spillover effects of the US economic policy uncertainty in Latin America, Estudios de Economía, ISSN 0718-5286, Universidad de Chile, Departamento de Economía, Santiago de Chile, Vol. 47, Iss. 2, pp. 273-293 This Version is available at: https://hdl.handle.net/10419/285083 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-sa/4.0/ Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 273Estudios de Economía. Vol.47 - Nº2, Diciembre 2020. Págs. 273-293 Spillover effects of the US economic policy uncertainty in Latin America*1 Efectos indirectos de la incertidumbre de la política económica de EE.UU.en América Latina 2Semei Coronado** 3José N. Martinez*** 4Francisco Venegas-Martínez**** Abstract This paper is aimed at assessing the spillover effects of the US Economic Policy Uncertainty (EPU) in macroeconomic variables of major Latin American Countries (LAC): Mexico, Colombia, Brazil, and Chile. To do that, we estimate a set of two-country Structural Vector Autoregressive (SVAR) models for 1997-2019; each model includes the US and one of the LAC. We use the following variables: EPU indexes, exchange rates, consumer price indexes, industrial production (IP), and interest rates (IR) of the US and the studied LAC. The main finding is that positive shocks in the US EPU index lead to currency depreciation for all four LAC; the largest effect is for Mexico. Other statistically significant results are a brief and small positive impact on Colombia’s IP and a positive impact on Mexico’s IR. The remaining LAC’s estimates are statistically insignificant. For this reason, we applied Rossi and Wang’s (2019) robust Granger causality tests that considers structural breaks. Finally, the estimates before and after the 2008 financial crisis suggest that LAC became slightly more responsive to US EPU shocks after the crisis. Key words: Economic policy uncertainty, structural vector autoregressive, impulse response function, robust Granger-causality tests. JEL Classification: F62, N16. * The authors are grateful for the comments and suggestions of three anonymous referees. Any remaining errors are the sole responsibility of the authors. ** Departamento de Métodos Cuantitativos, Universidad de Guadalajara, México. E-mail: [email protected] *** College of Business Administration and Public Policy, Department of Accounting, Finance, and Economics, California State University Dominguez Hills, USA. E-mail: [email protected]). **** Escuela Superior de Economía, Instituto Politécnico Nacional, México E-mail: fvene[email protected]. Corresponding author. Received: September, 2018. Accepted: October, 2019. Estudios de Economía, Vol.47 - Nº2274 Resumen Esta investigación tiene como objetivo evaluar los efectos indirectos de la incertidumbre de la política económica de EE.UU. en las variables macroeconómicas de los principales países de América Latina: México, Colombia, Brasil y Chile. Para ello, estimamos un conjunto de modelos de vectores autorregresivos estructurales de dos países para 1997-2019; cada modelo incluye a EE.UU. y un país de América Latina. Utilizamos las siguientes variables: índices la incertidumbre de la política económica, tipos de cambio, índices de precios al consumidor, producción industrial y tasas de interés de los Estados Unidos y países de América Latina estudiados. El principal hallazgo es que los choques positivos en el índice incertidumbre de la política económica de EE.UU. conducen a la depreciación de la moneda local en los cuatro países de ALC; el mayor efecto es para México. Otros resultados estadísticamente significativos son un impacto positivo breve y pequeño en la producción industrial de Colombia y un impacto positivo en la tasa de interés de México. Las estimaciones restantes para América Latina son estadísticamente no significantes. Por esta razón, aplicamos las pruebas robustas de causalidad de Granger propuestas por Rossi y Wang (2019) que consideran cambios estructurales. Por último, las estimaciones antes y después de la crisis financiera de 2008 sugieren que América Latina se tornó un poco más sensible a los choques de incertidumbre de la política económica estadounidense después de la crisis. Palabras clave: Incertidumbre de la política económica, vectores autorregresivo estructurales, función impulso respuesta, pruebas robustas de causalidad de Granger. Clasificación JEL: F62, N16. 1. Introduction The influence of the US economy policy on advanced and emerging countries has been widely studied in the specialized literature. In particular, there are several papers dealing with the impact of the uncertainty of the US economic policy on economic and financial variables on several countries and regions. In this regard, Gupta, Olasehinde-Williams and Wohar (2020) assess the impact of Economic Policy Uncertainty (EPU) shocks on a panel of 50 advanced and emerging market economies These authors find that for advanced economies the exchange rate regime and financial vulnerability account for a large portion of the contraction in activity. In emerging economies the responses do not depend on the exchange rate regime, but the responses become larger when trade openness is high and weakness in the financial system. Also, Kido (2016) analyzes spillover effects of shocks the US economic policy uncertainty on real effective exchange rates of several countries duing the period 2000-2014 by using Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 275 a Dynamic Conditional Correlation Generalized Autoregressive Conditional Heteroskedasticity (DCC-GARCH) model. the other hand, Zouhair, Lanouar, and Ajmi, (2013) examine the volatility spillovers in stock markets, securitized real estate, bond markets, currency markets, and economic policy uncertainty spillovers across 7 countries. The authors’ empirical findings are that spillovers are important and account for, respectively, about 72% and 50% of the dynamics of financial market stress and economic policy uncertainty, respectively, across the 7 economies examined. Also, Liow, Liao, and Huang (2018) find empirical evidence that policy uncertainty spillovers lead financial market stress spillovers in a multi-country context. In other words, changes in international economic policy uncertainty spillovers may be a predictor of changes in international financial market risk spillovers in the short run. Finally, Colombo’s (2013), and Bernal, Gnabo, and Guilmin (2016) assess the impact of economic policy uncertainty on risk spillovers within the Eurozone. Regarding the impact of the US economic policy uncertainty in the country itself, Baker, Bloom, and Davis (2016) find that the US EPU is associated with greater stock price volatility and reduced investment and employment in policysensitive sectors like defense, health care, finance, and infrastructure construction. Also, Bloom (2009) simulate the impact of a large EPU shock in the US and find that it generates a rapid drop, rebound, and overshoot in employment, output, and productivity growth. Hiring and investment rates fall dramatically in the 4 months after the shock because higher uncertainty increases the real-option value to waiting, so firms scale back their plans. Finally, Istiak and Serletis (2020) find that commercial bank leverage rises when geopolitical risk and macroeconomic, policy, and equity uncertainty increase. Moreover, the authors find that the leverage of broker-dealers and shadow banks declines when Chicago risk and macroeconomic, policy, financial, and equity uncertainty increase. Concerning the convergence of uncertainties across the World, Christou, Gozgor, Gupta, Keung, and Lau (2019) analyze the convergence of a newsbased measure of uncertainty across 143 countries in the period of 1996-2018. The authors use a panel data-based unit root test to the ratio of the uncertainty of individual countries relative to that of global uncertainty. These authors find empirical evidence of convergence and hence, the spillover of uncertainty across the economies of the world. Moreover, Gabauer and Gupta (2018) focuss on the transmission mechanism of country-specific and international economic uncertainty spillovers by using a Time-Varying Parameter Vector Autoregressions (TVP-VAR) approach. For Latin American Countries (LAC), the US influence tends to be significant due to both financial and trade ties in the region. In this regard, Alam and Istiak (2020) examine the impact of US policy uncertainty on Mexico by using a Structural Vector Autoregressive (SVAR) model with linear and nonlinear tests. These authors show that an increase in the US economic policy uncertainty leads to a fall in Mexican output (industrial production), price level and interest rate. It is worth mentioning here that there is a lack of research into the possible Estudios de Economía, Vol.47 - Nº2276 effects on uncertainty of the US economic policy on macroeconomic aggregates in broader regions of Latin America. This research examines the spillover effects of US economic policy uncertainty on macroeconomic aggregates in Latin American Countries by estimating a SVAR model during the period 1997-2019. Most of the empirical analysis is mainly based on Impulse Response Function (IRF) estimations from twocountry SVAR1 by using information of the EPU indexes and macroeconomic variables for the US, Mexico, Colombia, Brazil, and Chile. The IRF will be used to quantify the magnitude and persistence of a US economic policy uncertainty shock on LAC for the entire period, 1997-2019, and for specific periods before and after the 2008 financial crisis. The rest of this paper is organized as follows: section 2 describes the nature of data and presents a summary of the descriptive statistics; section 3 depicts the SVAR specification and shows the obtained empirical results; section 4 presents the findings about Granger-causality tests from the EPU index of the US toward the currency depreciation for all four LAC; finally Section 5 concludes. 2. Nature of Data and Summary Statistics Our dataset consists of monthly time series of EPU indexes and macroeconomic data for the US and the following LAC: Mexico, Colombia, Brazil, and Chile. The entire period of study is from January 1997 to December 2019. All series are expressed in growth rates. EPU indexes data for all the countries were collected from the webpage of Economic Policy Uncertainty Index.2 We also include the following macroeconomic fundamentals for all the countries: inflation (CPI), growth rate of industrial production (IP), short interest rate (IR), and the rate of depreciation of the exchange rates (ER). Macroeconomic data comes from different sources; some of them come from the Organization for Economic Cooperation and Development (OECD) webpage, other from the Federal Reserve Economic Data (FRED) database, and other more from National Central Banks of the corresponding LAC. See Appendix A for a full description of the data sources. Figure 1 shows the EPU indexes for the US and each one of the four LAC. The left-hand side presents the EPU indexes in levels and the right-hand side presents the EPU indexes in growth rates. The right-hand graphs suggest the EPU indexes in growth rates are indeed stationary for the US and each of the four LAC. In terms of the levels, Figure 1 also shows that Mexico and the US tend to follow each other closely before the 2008 crisis, but not so much after the crisis. In fact, Mexico’s index tends to be below that of the US and with 1 For measuring uncertainty see Jurado, Ludvigson, and Ng (2015) and Mumtaz and Theodoridis (2018). 2 Available in http://www.policyuncertainty.com/ Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 277 lower volatility after the 2008 crisis. The rest of LAC tend to follow the US more closely. FIGURE 1 EPU INDEXES TIME SERIES PLOT IN LEVELS (LEFT) AND GROWTH RATES (RIGHT). THE EPU INDEX FOR THE US PLOTTED AGAINST MEXICO IS IDENTIFIED WITH HIGHER PEAKS (A,B), COLOMBIA IS IDENTIFIED WITH THE LIGHT LINE (C,D), BRAZIL IS IDENTIFIED WITH HIGHER PEAKS (E,F), AND CHILE IS IDENTIFIED WITH THE LIGHT LINE (G,H). Source:Authors’ own elaboration based on the Bayesian Estimation, Analysis and Regression Toolbox (BEAR) from the European Central Bank. Table 1 shows the descriptive statistics and the unit root test for all series under study. Results from the Residual Augmented Least Squares (RALS) test from Im, Lee, and Tieslau (2014) does not require a specific density function for the error term are reported in the last column of Table 1. For the RALS test, the null hypothesis is that there exists a unit root. The number of lags was selected using the sequential analysis proposed by Im et al. (2014). The null hypothesis Estudios de Economía, Vol.47 - Nº2278 TABLE 1 TIME SERIES DESCRIPTIVE STATISTICS (N = 275) Country Variable Mean Median Min Max Variance SD Skewness Kurtosis RALS US CPI –0.041 –0.003 –11.054 6.914 0.920 0.959 –5.167 79.819 –11.239 IP 0.001 0.001 –0.035 0.024 0.000 0.007 –1.174 8.351 –5.319 IR 0.003 0.000 –0.454 0.800 0.014 0.117 0.997 13.045 –4.158 EPU 0.043 –0.018 –0.601 1.935 0.103 0.321 2.207 11.930 –6.306 Mexico (MEX) CPI –0.005 –0.010 –0.249 0.404 0.006 0.075 0.738 6.472 –6.153 IP 0.002 0.002 –0.053 0.049 0.000 0.012 –0.281 6.380 –7.879 IR –0.003 0.000 –0.225 0.484 0.004 0.062 1.488 17.532 –6.887 ER 0.004 0.000 –0.084 0.188 0.001 0.025 1.588 12.957 –13.175 EPU 0.119 –0.002 –0.742 9.015 0.534 0.731 6.980 81.660 –19.543 Colombia (COL) CPI –0.004 –0.004 –0.201 0.233 0.004 0.067 –0.097 3.507 –4.369 IP 0.002 0.004 –0.098 0.158 0.001 0.031 0.599 7.692 –5.671 IR –0.005 –0.004 –0.218 0.154 0.002 0.043 –0.932 7.848 –4.859 ER 0.005 0.001 –0.073 0.120 0.001 0.030 0.838 4.776 –12.322 EPU 0.049 0.011 –0.548 1.869 0.116 0.340 1.716 8.631 –8.647 Brazil (BRA) CPI 0.001 –0.003 –0.277 0.538 0.009 0.094 1.465 9.017 –4.076 IP 0.001 0.002 –0.120 0.141 0.000 0.020 –0.511 18.475 –22.638 IR 0.000 –0.007 –0.310 0.970 0.014 0.116 3.498 26.979 –4.265 ER 0.006 0.004 –0.107 0.274 0.002 0.043 1.996 12.556 –4.601 EPU 0.132 0.041 –0.736 3.155 0.349 0.591 1.728 7.700 –8.412 Chile (CHI) CPI 0.296 –0.007 –3.160 83.725 25.621 5.062 16.369 270.336 –160.178 IP 0.002 0.002 –0.184 0.223 0.001 0.026 0.960 31.599 –4.162 IR 0.004 0.000 –0.503 1.880 0.031 0.175 5.251 54.317 –5.224 ER 0.002 0.000 –0.073 0.176 0.001 0.026 1.179 9.911 –5.002 EPU 0.060 0.003 –0.572 1.921 0.131 0.362 1.112 5.340 –5.322 Source: Authors’ own elaboration Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 279 is rejected for all variables using a 99% confidence level. The results suggest the series in growth rates can be considered stationary. 3. SVAR Specification and Empirical Results In this section, we report the empirical results to attempt to quantify the impact of an EPU shock from the US on the macroeconomic indicators of the LAC under study. Following Colombo (2013), we estimate a set of two-country structural VAR models. Each VAR model includes the US and one of the LAC from the country sample. The IRF are estimated and account for the magnitude of the impact, if any, and its dynamics.3 Consistent with the main objective of this paper, countries are ordered assuming that the variables in the US are “more” exogenous than the variables from LAC. Within each economy and consistent with the literature, we order the macroeconomic variables from “more” to “less” exogenous. Macroeconomic variables for each country are ordered as follows: Mexico: yt = [EPUUSA CPIMEX IPMEX IRMEX ERMEX EPUMEX]', Colombia: yt = [EPUUSA CPICOL IPCOL IRCOL ERCOL EPUCOL]', Brazil: yt = [EPUUSA CPIBRA IPBRA IRBRA ERBRA EPUBRA]', Chile: yt = [EPUUSA CPICHI IPCHI IRCHI ERCHI EPUCHI]'. In this research IRF identification is obtained via Cholesky decomposition of the variance-covariance matrix of the reduced-form VAR estimates. SVAR models have a lag length of 1 according to Bayesian Information Criterion (BIC). Figure 2 shows the IRF estimates from a shock of one standard deviation in the US EPU index on macroeconomic variables for each LAC obtained from a two-country SVAR model.4 In particular, Figure 2 shows that a positive shock 3 For robustness, we applied two main specifications for VAR models: a Bayesian VAR model with stochastic volatility (BVARSV) and time-varying parameters Bayesian VAR (TV-BVAR). It has been documented extensively in the literature that stochastic volatility is important in fitting the dynamics of macro/finance variables as the ones analyzed in this research. The authors estimated the models by using Bayesian Estimation, Analysis and Regression Toolbox (BEAR) from the European Central Bank. For the structural VAR model specification, we follow Dieppe, Legrand, and Van Roye (2018). For a full description of the SVAR, BVARSV and TV-BVAR models and estimation procedures see Cogley and Sargent (2005), and Dieppe et al. (2018). 4 We carry out the estimation exercise using BVARSV and TV-BVAR models, but all estimates were statistically insignificant. The results of BVARSV and TV-BVAR are available upon request. We also replied the same exercise with the variables of the US as in Colombo (2013) to LAC. Estudios de Economía, Vol.47 - Nº2280 to the US EPU index results in statistically significant currency depreciation for each LAC, and the magnitude of the impact is the largest for Mexico, followed by Colombia, Brazil, and Chile. This suggests the impact of an EPU index shock on LAC currencies is inversely related to the geographical distance from the US to the specific LAC. FIGURE 2 TWO-COUNTRY SVAR MODELS (US EPU SHOCK TO LAC) Source:Authors’ own elaboration based on the sample and using the Bayesian Estimation, Analysis and Regression Toolbox (BEAR) from the European Central Bank. In terms of the EPU index from each LAC, the impacts from a shock to the EPU index from the US are positive and statistically significant for each country, and the effects are larger for Mexico. The estimates suggest that economic policy uncertainty from the US has a direct and strong impact on economic Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 287 5. Conclusions We have studied how a shock of economic policy uncertainty from the US might spillover to macroeconomic conditions of the LAC under study. We use EPU from the United States to attempt to account for uncertainty spillovers. We estimate a set of two-country structural VAR models. The SVAR model had a better performance to analyze the uncertainty spillovers than other models. Each model has included the US EPU index and one LAC with their corresponding macroeconomic variables from the country sample. The impulse response functions (IRF) are computed accounting for the magnitude of the impact, if any, and its shot-run dynamics. Considering the entire period, structural VAR model estimates showed that a shock of one standard deviation in the EPU index of the US tend to lead to currency depreciation and positive impacts on EPU indexes for all four LAC, and these estimates tend to be larger for Mexico. Estimates from before and after the 2008 financial crisis suggest that LAC economies became slightly more responsive to EPU index shocks from the US after the 2008 financial crisis. Finally, robust Granger causality tests that consider structural breaks were applied for the entire period and for the periods before and after the 2008 financial crisis. The estimates before and after the 2008 financial crisis suggest that LAC became slightly more responsive to US EPU shocks after the crisis. Estudios de Economía, Vol.47 - Nº2288 Appendix A Variables Source EE.UU. Consumer Price Index (Inflation) OECD Data https://data.oecd.org Industrial Production (Industrial) OECD Data https://data.oecd.org Exchange Interest (Interest Rate)* https://fred.stlouisfed.org/series/IR3TIB01USM156N Economic Policy Uncertainty (Economy Policy, EPU) http://www.policyuncertainty.com Brazil Consumer Price Index (Inflation) OECD Data https://data.oecd.org Industrial Production (Industrial) OECD Data https://data.oecd.org Exchange Interest (Interest Rate)* Banco Central do Brasil https://www.bcb.gov.br/ Exchange Rate OECD Data https://data.oecd.org Economic Policy Uncertainty (Economy Policy, EPU) http://www.policyuncertainty.com Chile Consumer Price Index (Inflation) OECD Data https://data.oecd.org Industrial Production (Industrial) OECD Data https://data.oecd.org Exchange Interest (Interest Rate)* Banco de Chile http://www.bancochile.cl Exchange Rate OECD Data https://data.oecd.org Economic Policy Uncertainty (Economy Policy, EPU) http://www.policyuncertainty.com Colombia Consumer Price Index (Inflation) OECD Data https://data.oecd.org Industrial Production (Industrial) OECD Data https://data.oecd.org Exchange Interest (Interest Rate)* Banco de Colombia https://www.grupobancolombia.com/ Exchange Rate OECD Data https://data.oecd.org Economic Policy Uncertainty (Economy Policy, EPU) http://www.policyuncertainty.com Mexico Consumer Price Index (Inflation) OECD Data https://data.oecd.org Industrial Production (Industrial) OECD Data https://data.oecd.org Exchange Interest (Interest Rate)* Banco de México http://www.banxico.org.mx Exchange Rate OECD Data https://data.oecd.org Economic Policy Uncertainty (Economy Policy, EPU) http://www.policyuncertainty.com * 3 months. Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 289 Appendix B FIGURE B.1 WALD STATISTICS TESTING OF STRUCTURAL BREAKS IN GRANGER CAUSALITY FOR PERIOD 1997M2-2019M12 Note: (a) EPUUSA → IRMEX, (b) EPUUSA → ERMEX (c) EPUUSA → ERCOL, (d) EPUUSA → IRBRA, and (e) EPUUSA → ERCHI against the alternative of break in Granger causality at time on axis. In all figures, a solid line represents the sequence of the Wald Statistic over time. The dashed line < 5% critical value, and the dotted line < 10% critical value. Source: Prepared by the authors based on the sample and using the STATA Software. Estudios de Economía, Vol.47 - Nº2290 FIGURE B.2 WALD STATISTICS TESTING OF STRUCTURAL BREAKS IN GRANGER CAUSALITY FOR PERIOD 1997M2-2008M12 Note: (a) EPUUSA → IRMEX, (b) EPUUSA → ERMEX (c) EPUUSA → ERCOL, (d) EPUUSA → IRBRA, and (e) EPUUSA → ERCHI against the alternative of break in Granger causality at time on axis. In all figures, a solid line represents the sequence of the Wald Statistic over time. The dashed line < 5% critical value, and the dotted line < 10% critical value. Source: Prepared by the authors based on the sample and using the STATA Software Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 291 FIGURE B.3 WALD STATISTICS TESTING OF STRUCTURAL BREAKS IN GRANGER CAUSALITY FOR PERIOD 2008M1-2019M12 Note: (a) EPUUSA → IRMEX, (b) EPUUSA → ERMEX (c) EPUUSA → ERCOL, (d) EPUUSA → IRBRA, and (e) EPUUSA → ERCHI against the alternative of break in Granger causality at time on axis. In all figures, a solid line represents the sequence of the Wald Statistic over time. The dashed line < 5% critical value, and the dotted line < 10% critical value. Source: Prepared by the authors based on the sample and using the STATA Software. Estudios de Economía, Vol.47 - Nº2292 TABLE B.1 GRANGER CAUSALITY TESTS IN THE DIRECT MULTISTEP VAR-LP FORECASTING MODEL PANEL A (1997m2-2019m12) h=0h=3 Variables ExpW Statistics (p-value) MeanW Statistics (p-value) Nyblom Statistics (p-value) SupLR Statistics (p-value) ExpW Statistics (p-value) MeanW Statistics (p-value) Nyblom Statistics (p-value) SupLR Statistics (p-value) EPUUSA → IRMEX 14.874 (0.000) 9.147 (0.041) 3.778 (0.016) 40.074 (0.000) 61.166 (0.000) 51.598 (0.000) 4.473 (0.000) 132.815 (0.000) EPUUSA → ERMEX 56.892 (0.000) 31.619 (0.000) 1.768 (1.149) 124.082 (0.000) 272.664 (0.000) 138.050 (0.000) 3.730 (0.017) 555.820 (0.000) EPUUSA → ERCOL 14.553 (0.000) 7.059 (0.110) 0.766 (0.492) 37.913 (0.000) 46.490 (0.000) 26.184 (0.000) 4.413 (0.000) 102.447 (0.000) EPUUSA → IRBRA 41.946 (0.000) 30.919 (0.000) 1.643 (0.178) 94.046 (0.000) 36.660 (0.000) 31.394 (0.000) 2.243 (0.086) 83.777 (0.000) EPUUSA → IRCHI 32.476 (0.000) 27.285 (0.000) 1.687 (0.169) 73.459 (0.000) 47.288 (0.000) 25.564 (0.000) 2.914 (0.039) 104.874 (0.000) PANEL B (1997m2-2007m12) h=0h=3 EPUUSA → IRMEX 221.281 (0.000) 59.746 (0.000) 1.745 (0.155) 450.901 (0.000) 51.124 (0.000) 50.425 (0.000) 5.076 (0.000) 111.162 (0.000) EPUUSA → ERMEX 24.294 (0.000) 14.294 (0.000) 1.807 (0.142) 56.802 (0.000) 134.009 (0.000) 91.971 (0.000) 4.755 (0.000) 276.990 (0.000) EPUUSA → IPBRA 491.281 (0.000) 131.192 (0.000) 1.078 (0.000) 991.568 (0.000) 134.631 (0.000) 64.155 (0.000) 2.772 (0.046) 278.238 (0.000) EPUUSA → IRBRA 165.113 (0.000) 27.790 (0.000) 1.112 (0.331) 339.232 (0.000) 76.951 (0.000) 63.126 (0.000) 1.464 (0.216) 162.831 (0.000) EPUUSA → EPUBRA 146.483 (0.000) 139.878 (0.000) 4.492 (0.000) 301.972 (0.000) 16.549 (0.000) 9.830 (0.029) 0.737 (0.511) 41.795 (0.000) PANEL C (2008m1-2019m12) h=0h=3 EPUUSA → IRMEX 23.615 (0.000) 14.585 (0.000) 1.449 (0.220) 55.459 (0.000) 268.339 (0.000) 182.245 (0.000) 4.031 (0.011) 545.432 (0.000) EPUUSA → ERCOL 92.248 (0.000) 37.701 (0.000) 1.493 (0.210) 193.368 (0.000) 39.135 (0.000) 14.24 (0.000) 7.409 (0.000) 85.831 (0.000) EPUUSA → ERBRA 74.998 (0.000) 72.625 (0.000) 1.820 (0.140) 159.092 (0.000) 30.847 (0.000) 22.997 (0.000) 2.222 (0.088) 70.572 (0.000) EPUUSA → ERCHI 18.963 (0.000) 16.440 (0.000) 3.614 (0.019) 46.321 (0.000) 83.473 (0.000) 42.691 (0.000) 3.789 (0.000) 176.117 (0.000) Note: ExpW, MeanW, Nyblom and SupLR are statistics for the null hypothesis that the VAR series xt does not Granger cause series yt assuming heteroskedastic and serially correlated idionsyncratic shocks. The bolded values are not statistically significant. The rest are significant at the < 5% level. Source: Prepared by the authors based on the sample and using the STATA Software. Spillover effects… / S. Coronado, J. N. Martinez, F. Venegas-Martínez 293 Bibliography Alam, M. R., & Istiak, K. (2019). Impact of US policy uncertainty on Mexico: Evidence from linear and nonlinear tests. Quarterly Review of Economics and Finance. https://doi.org/10.1016/j.qref.2019.12.002 Cogley, T., & Sargent, T. J. (2005). Drifts and volatilities: Monetary policies and outcomes in the post WWII US. Review of Economic Dynamics, 8(2 SPEC. ISS.), 262-302. https://doi.org/10.1016/j.red.2004.10.009 Colombo, V. (2013). Economic policy uncertainty in the US: Does it matter for the Euro area? Economics Letters, 121 (1), 39-42. https://doi.org/10.1016/j. econlet.2013.06.024 Dieppe, A., Legrand, R., & Van Roye, B. (2018). The Bayesian Estimation, Analysis and Regression (BEAR) Toolbox Technical guide. Granger, C. W. (2001). Forecasting white noise. In Essays in Econometrics; Collected Papers of Clive W. J. Granger. Vol. 32 (pp. 457-471). Cambridge University Press. Im, K. S., Lee, J., & Tieslau, M. A. (2014). More Powerful Unit Root Tests with Non-normal Errors. In Festschrift in Honor of Peter Schmidt (pp.315- 342). https://doi.org/10.1007/978-1-4899-8008-3_10 Jurado, K., Ludvigson, S. C., & Ng, S. (2015). Measuring uncertainty. American Economic Review, 105(3), 1177-1216. https://doi.org/10.1257/aer.20131193 Mumtaz, H., & Theodoridis, K. (2018). The Changing Transmission of Uncertainty Shocks in the U.S. Journal of Business and Economic Statistics, 36 (2), 239-252. https://doi.org/10.1080/07350015.2016.1147357 Rossi, B. (2005). Optimal tests for nested model selection with underlying parmeter instability. Econometric Theory, 21 (5), 962–990. https://doi. org/10.1017/S0266466605050486 Rossi, B., & Wang, Y. (2019). Vector autoregressive-based Granger causality test in the presence of instabilities. Stata Journal, 19(4), 883–899. https:// doi.org/10.1177/1536867X19893631 Yang, M.-H. (2015). Revisit Export and GDP Nexus in China and Taiwan: A Rolling Window Granger Causality Test. In Theoretical and Applied Economics. Zeileis, A., Leisch, F., Kleiber, C., & Hornik, K. (2005). Monitoring structural change in dynamic econometric models. Journal of Applied Econometrics, 20 (1), 99-121. https://doi.org/10.1002/jae.776