The effects of economic policy uncertainty on stock market returns Evidence from Brazil.
| Autor | Gea, Cristiane |
-
Introduction
Economic policy uncertainty (EPU) has interested and challenged policymakers, consumers, investors, economists, and researchers for over three decades (Antonakakis et al., 2013; Vicente and Marins, 2021). According to Baker et al. (2016), economic policy uncertainty refers to the non-zero probability that changes in existing economic policies will affect agents' decisions. Thus, uncertainty may affect economic policy through the following channels: (i) postponement of or a change in decisions (such as those concerning employment, investment, consumption, and savings) of firms and other economic agents; (ii) increased production and financing costs; (iii) intensification of disinvestment and economic contraction through supply and demand channels; (iv) increased risks related to participating in financial markets, through a reduction in the value of protection provided by the government to the markets; (v) changes in inflation, interest rates, and expected risk premiums; (vi) changes in the degree of risk aversion; and (vii) slower economic recovery. (1) In fact, the operation of these channels has been verified in practice in several previous studies. (2)
Given that changes in economic policies affect agents' perception of the economy's behavior, recent studies indicate that uncertainty may be a good predictor of the economy's future performance. Karnizova and Li (2014) show that the economic policy uncertainty measure robustly forecasts recessions in the U.S. over longer forecast horizons, besides improving the predictive performance of financial variables. Fernandez-Villaverde et al. (2015) and Born and Pfeifer (2014) find that an adverse shock in fiscal policy uncertainty leads to a contraction in output, followed by a slow economic recovery. Researchers such as Marcus (1981), Dixit (1989), Rodrik (1991), Aizenman and Marion (1993), Bloom et al. (2007), Bloom (2009), Julio and Yook (2012), Jones and Olson (2013), Wang et al. (2014), and Gulen and Ion (2016) show that uncertainties about government policies have negative impacts on inflation, investment, employment, and economic activity. Colombo (2013) shows that a shock of one standard deviation in a U.S. economic policy uncertainty measure leads, in the short run, to a statistically significant drop in production and prices in the eurozone. (3) Finally, Baker et al. (2016) show that economic policy uncertainty reduces investment and employment, and increases stock price volatility.
In short, the literature on uncertainty surrounding economic policy decisions shows that it affects not only the current state of the economy, but also its future developments. Moreover, Brogaard and Detzel (2015) suggest that measures of economic policy uncertainty are more sensitive to specific shocks--especially those emerging from a "noisy" operation of the political system, which ends up distorting policymaking. These facts imply that measures of economic policy uncertainty can aid in quantifying the aggregate risk of the economy, thus helping explain the equilibrium returns of various financial assets. Furthermore, as they help predict the future behavior of the economy (and therefore its aggregate risk), these measures can also help predict future stock market returns. Indeed, Brogaard and Detzel test both possibilities using the returns observed in the U.S. stock market. Our goal is to adapt their analysis for the Brazilian case.
And why is studying Brazil interesting? First, studies measuring Brazil's economic policy uncertainty are still scarce. In addition, Brazil is a prominent emerging economy, with the most important stock market, in terms of capitalization, liquidity, and participation of foreign investors, in Latin America. Finally, and most importantly, in recent years Brazil has experienced several shocks originating in or propagated by the political system, whose abnormal functioning hindered economic policy decision-making. (4) Indeed, the country can be taken as an extreme case in which economic policy uncertainty has a strong potential to affect the stock market--and an inability to identify this effect in Brazil would be a challenging puzzle.
The work of Phan et al. (2018) is similar to ours. The authors examine whether EPU can predict excess stock returns using data from 16 countries, including Brazil. They show that EPU's ability to forecast stock returns depends on both the country and the sectors studied. In the case of Brazil, EPU was not significant, contrary to our findings. Our analysis differs from Brogaard and Detzel (2015) in the methodology and the use of covariates in the econometric model. Demir and Ersan (2017) include Brazil, but their research centers on the relationship between EPU and cash holdings in the BRIC countries. They discovered that, as local and global instability rises, these countries' businesses tend to keep more cash on hand.
The article is organized as follows: Section 2.1 describes the data and methodology that we use, and Section 3 presents the results of our baseline estimations. In Section 4 we show the results of a robustness analysis in which the regressions of Section 2.1 are estimated again using other estimation techniques, and other measures of economic activity and general uncertainty. Finally, our main results are summarized in Section 5.
-
Data and methodology
2.1 Data
The analysis uses data from the period between January 2000 and May 2020.5 Following the literature, the monthly series used to estimate the desired effects are: (6)
Index of economic policy uncertainty (EPU): we use the newspaper-based index developed by Baker et al. (2016), which is comparable to the EPU index for the United States that Brogaard and Detzel (2015) use in their work. (7)
Return of the Ibovespa index (IBOV): the Ibovespa is the leading indicator of the average performance of the Brazilian stock market. It measures the total returns of the shares that comprise it. (8,9)
Excess return of the Ibovespa (EX_IBOV): the portion of the return that exceeds the rate of return of a risk-free investment, which is proxied by the quote of a 30-day DI x Pre swap contract. (10)
Average of squared returns (VAR): we employ the average value of daily squared returns of the Ibovespa index as a proxy of volatility. The averages are calculated within each month.
Term spread (SPREAD): the difference between the interest rates for the 2-year and 30-day maturities. Both are extracted from the term structure of interest rates based on the quotes of DI x Pre swap contracts. (11) According to Bernanke (1990), SPREAD captures the market's perception of the downturn risk of an economy (probability of recession). It also contains information about the stance of monetary policy.
Country risk (EMBI): the Brazilian country risk is measured by a weighted average of the premiums paid for Brazilian external debt securities, compared to those of U.S. Treasury bonds with equivalent maturities (Emerging Markets Bond Index Plus). This indicator is calculated by the J.P. Morgan Chase Bank.
Average low-risk investment indicator (RREL): this is calculated as the difference between the interest rate for 1-month operations (also extracted from the term structure derived from DI x Pre swap contracts) and its moving average in a 12-month window.
Total dividends (DIV) and dividend yield (DIVYIELD): the first series is a measure of total dividends paid by companies whose shares belong to the Ibovespa index. The second series corresponds to the ratio between total dividends distributed in the last 12 months and the market value of the companies that comprise the Ibovespa index. (12)
Global index of economic policy uncertainty (GEPU): developed by Baker et al. (2016), this index is constructed from a gross domestic product weighted average of the national EPU of 20 countries. (13) The purpose of GEPU is to control for the possibility that global economic policy uncertainty can affect the domestic economy (Bhattarai et al., 2020).
Economic activity (CLI): as a measure of economic activity, we use the Brazilian version of the composite leading indicator (CLI) of economic activity calculated by the Organization for Economic Cooperation and Development (OECD). (14)
According to Bernanke (1983), Bloom (2009), and Brogaard and Detzel (2015), economic policy uncertainty has a countercyclical behavior, rising during recessions or other times of economic distress. This stylized fact is also observed in Brazil, as Figure A1 makes clear. This figure suggests that periods of greater economic policy uncertainty and depressed economic activity happen at the same time. More specifically, at least three Brazilian recessions identified by the Economic Cycle Dating Committee (CODACE) are also marked by high values of the economic policy uncertainty index. (15) The figure shows that the index reached its highest values during the global financial crisis of 2008-2009 and the long recession of 2014-2016.
Figure A1 also shows that the Brazilian measure of economic policy uncertainty has remained high since at least 2015. This period coincided (or was preceded) by several political shocks that compromised government action. We highlight the following: (i) the large demonstrations that took place in Brazil throughout 2013, which many believe were triggered by dissatisfaction with government expenses related to the 2014 World Cup; (ii) the attempt to declare invalid the result of the 2014 presidential elections by the defeated candidate Aecio Neves da Cunha and the parties that supported him; (iii) the effects of Operation Car Wash, which accused various politicians of corruption and ended up sending some of them to prison (including former President Luis Inacio Lula da Silva); (iv) the impeachment of President Dilma Rousseff, concluded in August 2016; (v) the scandals of corruption and influence peddling that compromised the Michel Temer administration; (vi) the attempts...
Para continuar a ler
PEÇA SUA AVALIAÇÃOCOPYRIGHT GALE, Cengage Learning. All rights reserved.
