آزمون عدم تقارن در اولویت های بانک مرکزی در یک اقتصاد کوچک باز : مطالعه برای برزیل

In this paper, we extend Surico's (2007a) model to an open economy and test if the Brazilian Central Bank's loss function is asymmetric with regard to positive and negative deviations of the output gap and of the inflation rate from its target. Furthermore, we use tests for structural breaks to investigate changes in the conduct of policy and monetary authority's preferences. The results revealed that the Central Bank reacted more strongly to deviations of inflation from the target and the output gap after 2003. With regard to the monetary authority's preferences, estimates indicate only an asymmetric preference over an above-target inflation rate until mid-2003. After this period, the evidence obtained by estimating the reaction function with inflation expectations showed that the Central Bank has been more averse to positive deviations of inflation from its target.

Since the early 1990s, the economic literature on monetary policy actions estimated by reaction functions has increasingly expanded. Taylor (1993) rule is probably the most widely known reaction function specification in this literature. According to this rule, the monetary authority responds to deviations of output and inflation from the targets through nominal interest rate movements, regarded as a policy instrument. Another specification that has received considerable attention is the forward-looking reaction function proposed by Clarida et al., 1998 and Clarida et al., 2000. In this type of policy rule, the policymaker adjusts the current interest rate based on the expected future inflation and output gap rates. These two types of interest rate rules share a common characteristic: they are linear functions related to economic variables. This can be explained by the fact that both specifications are theoretically supported by the linear-quadratic paradigm, where the monetary authority's loss function is assumed to be quadratic while the equations for the economic structure are linear. At recent times, however, two theoretical approaches have challenged the linear-quadratic framework underlying the linear reaction function. The first approach disproves the assumption that the economic structure is linear. Orphanides and Wieland (1999) derive optimal policy rules for the case where the monetary authority makes use of a quadratic loss function and a zone-linear Phillips curve that allows for nonlinearities in the short-term tradeoff between inflation and output. Nobay and Peel (2000) analyze an optimal discretionary monetary policy under a nonlinear Phillips curve and find out that the monetary authority can no longer eliminate the inflationary bias by defining an output target that is the same as the natural rate. Dolado et al. (2005) demonstrate that the central bank's optimal reaction function for an economy with a nonlinear Phillips curve is the forward-looking interest rate rule augmented to include the interaction between expected inflation and output gap. The second theoretical approach considers that policymakers can have asymmetric preferences in terms of their goals. According to Cukierman (2000), politicians and the public at large are usually more averse to negative than to positive gap differences from the potential output. The vice chairman of the Federal Reserve, Alan Blinder, states that “in most situations the CB will take far more political heat when it tightens pre-emptively to avoid higher inflation than when it eases pre-emptively to avoid higher unemployment” Blinder (1998, pp. 19–20). Since in democracies independent central banks are not totally insensitive to political organs, this type of asymmetry may be present in the policymaker's loss function. In addition, in periods in which the monetary authority is more concerned with lending credibility to its disinflation policy, the loss ascribed to positive deviations of the inflation rate from the inflation target is likely greater than that from negative deviations with the same magnitude. The consequences of introducing asymmetric preferences into the monetary authority's loss function have been investigated by several authors. Cukierman (2000) demonstrates that, when the policymaker is unsure about economic conditions and is more sensitive to negative output gaps, there is an inflationary bias even in the case where the effective output target is the potential output of the economy. This finding has been supported by empirical evidence gathered by Cukierman and Gerlach (2003) for a group of 22 OECD countries. Gerlach (2000) and Surico (2007a) show that the Federal Reserve was more concerned about negative than about positive output gaps in the pre-1980 period. Bec et al. (2002) verify that the business cycle phase, measured by output gap, has mattered in the conduct of monetary policy for Central Banks of Germany, the United States, and France. Cukierman and Muscatelli (2003) give evidence of nonlinearities with respect to inflation and to output gap in reaction functions estimated for Germany, the United Kingdom, and the USA. Dolado et al. (2004) observe that Federal Reserve preferences about inflation were asymmetric during the Volcker-Greespam's term. Bearing in mind what was presented earlier, the present paper aims to estimate the nonlinear reaction function for the Central Bank of Brazil so as to test asymmetries in its goals regarding the deviations of output gap and inflation from the targets. To achieve that, we extend the theoretical model proposed by Surico (2007a) for an open economy and we obtain the optimal monetary policy for the monetary authority considering that the loss function is potentially asymmetric. Given that the presence of asymmetries in the goals leads to nonlinear responses of the interest rates to inflation and to output gap, we check whether the policymaker's preferences are symmetric by testing the null hypothesis of a linear reaction function. Moreover, we estimate the Central Bank asymmetric coefficients and test whether they are statistically significant. Several research studies in the Brazilian literature seek to estimate monetary policy reaction functions.1 Although a smaller number of these studies assess reaction function nonlinearities, only Aragón and Portugal, 2010 try to check whether there exist asymmetric preferences in the Central Bank of Brazil. The present paper shows some progress in relation to that of Aragón and Portugal (2010) in two points. First, we consider asymmetries in the policymaker's preferences and derive the optimal monetary rule based on a macroeconomic model for a small open economy. Second, we follow Bai and Perron, 1998 and Bai and Perron, 2003 and run structural break tests for nonlinear reaction function parameters. The analysis of structural breaks allows investigating possible changes in monetary policy conduct, as well as in the Central Bank asymmetric preference coefficients. The major findings of this paper can be summarized as follows. First, the structural break tests strongly reject the hypothesis of stable reaction function parameters. For two out of the three monetary rule specifications analyzed, the date estimated for the break indicates that the change in the conduct of the Brazilian monetary policy occurred in the third quarter of 2003. Second, the estimates for the reaction function coefficients reveal that the Central Bank of Brazil has had stronger reactions to deviations of (current and expected) inflation from its target and to output gap in the more recent subperiod. Third, the evidence indicates, in general, that there is not an asymmetric response of the Selic rate to the output gap, suggesting that the loss function of the Central Bank of Brazil is quadratic relative to this state variable. Finally, the coefficient estimates of asymmetric preference concerning inflation show differences in the analyzed monetary rules. For the specification that regards the inflation target as constant, results indicate an increase in asymmetry in favor of above-target inflation after 2003. When we consider the deviation of inflation from a variable target, the asymmetry coefficient of inflation was not significant. Conversely, for the policy rule with deviation of the expected inflation from a variable target, results show that the Central Bank was more averse to a below-target inflation until mid-2003 and that it was averse to an above-target inflation thereafter. Aside from this introduction, the paper is organized into another four sections. Section 2 introduces the basic theoretical model used in the study, as well as the extension to this model and the derivation of the optimal reaction function for the interest rate. Section 3 presents the reduced-form interest rate rule to be estimated in order to check for the existence of monetary authority's asymmetric goals. Section 4 describes and analyzes the results obtained from the estimations. Section 5 concludes.

In this paper, we estimated the nonlinear reaction function for the Central Bank of Brazil in order to check for asymmetries in its goals related to output gap and deviations of inflation from its target. To achieve that, we extended the theoretical model of Surico (2007a) for an open economy and we obtained an optimal monetary policy rule for the monetary authority considering that its loss function is potentially asymmetric. Given that the presence of asymmetries yields nonlinear responses of the interest rate to inflation and to the output gap, we investigated whether the policymaker's preferences are symmetric by testing the null hypothesis of linearity of the reaction function. In addition, we estimated the Central Bank asymmetric preference coefficients and tested whether they were statistically significant. From an empirical standpoint, our paper followed Bai and Perron, 1998 and Bai and Perron, 2003 and we ran structural break tests for the nonlinear reaction function parameters. The analysis of structural changes allowed checking for possible changes in the monetary policy conduct, as well as in the Central Bank asymmetric preference coefficients. The structural break tests strongly rejected the null hypothesis of no break. For two out of three reaction function specifications, the estimated period for the change in the monetary rule was the third quarter of 2003. In general, the estimates of reaction function coefficients revealed that the Central Bank of Brazil reacted more strongly to deviations of (current and expected) inflation from the inflation target and to the output gap after 2003. The estimated values for the coefficient of asymmetric preference on the stabilization of inflation, α, indicated distinct changes in the Brazilian monetary authority's loss function. For the specification of the reaction function with a constant inflation target, the estimates showed an increase in asymmetry in favor of an above-inflation inflation after the date of the structural change. When we considered the deviation of inflation from a variable target, the asymmetry coefficient for the inflation was not significant in any of the analyzed subperiods. However, in the specification with the Djt variable, commonly used in the Brazilian literature on monetary rule, the results demonstrated that the Central Bank was more averse to a below-target inflation up to mid-2003, but more averse to an above-target inflation thereafter. Finally, the results show that the estimates for parameter γ were not statistically different from zero in any of the specifications. This indicates that the Central Bank did not have an asymmetric preference for an output above or below the potential output.