حکومت فدرال رزرو سیاست های پولی و تورم در ایالات متحده : در مورد اولویت های عدم تقارن

This paper investigates the empirical relevance of a new framework for monetary policy analysis in which the decision makers are allowed, but not required, to weight differently positive and negative deviations of inflation and output from the target values. The estimates of the central bank's Euler equation indicate that the preferences of the Fed had been asymmetric only before 1979, with the interest rate response to output contractions being larger than the response to output expansions of the same magnitude. We show that this asymmetry on output implied an average inflation bias around 1.5%1.5%. While the implicit inflation target also declined, the asymmetric preferences induced inflation bias appears to account for a sizable fraction of the historical decline in the inflation mean.

A popular method of monetary model building is to regard policy interventions as the solution of an optimal control problem in which the central bank minimizes some quadratic criterion subject to a linear structure of the economy. The quadratic characteristic of the objective and the linear feature of the constraints give rise to a linear first-order condition, usually referred to as a targeting rule (see Svensson, 1999), which describes the optimal response of the central bank to the developments in the economy. While the quadratic specification implies that the monetary authorities weight evenly positive and negative deviations of inflation and output from the target values, such modeling choice had been questioned by several practitioners at the policy committees of various central banks on the ground that it has little justification beyond analytical tractability.1 Blinder (1997, p. 6) argues that ‘academic macroeconomists tend to use quadratic loss functions for reason of mathematical convenience, without thinking much about their substantive implications. The assumption is not innocuous, […] practical central bankers and academics would benefit from more serious thinking about the functional form of the loss function’. Describing his experience as Fed Vice-Chairman, Blinder (1998, pp. 19–20) pushes the argument even further and claims ‘in most situations the central bank 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’, suggesting that political pressures can induce asymmetric central bank interventions. Similar concerns emerge also at other central banks like the ECB and on the occasion of an interest rate cut of 50 basis point Duisenberg (2001) states ‘the maintenance of price stability remains our first priority. […] today’s action could be taken “without prejudice to price stability”, and it thereby supported the other goals of EMU, such as economic growth’. On the theoretical side, a number of recent studies explore some novel mechanisms through which the costs of the business cycle can be asymmetric. Persson and Tabellini (1999) combine retrospective voting with imperfect information about the incumbent's talent to show that career concerned politicians can make reappointment more likely by endowing the central bank with an asymmetric objective that requires a larger monetary policy response in periods of poor economic performance. Galí et al. (2003a) construct a theoretical measure of welfare gap based on price and wage markups, and find that the costs of output fluctuations for the U.S. had historically been large and asymmetric. Erosa and Ventura (2002) introduce transaction costs and heterogeneity in portfolio holdings in an otherwise neo-classical model and show that these frictions can make the costs of inflation variation asymmetric. Lastly, the psychology of choice reveals that people tend to place a greater weight on the prospect of losses than on the prospect of gains in decision making under uncertainty (see Kahneman and Tversky, 1979), also suggesting that policy makers, who aggregate over individual welfare, may be loss-averse. On the empirical side, only a few studies, developed independently, estimate asymmetric reaction functions. Cukierman and Muscatelli (2003) and Martin and Milas (2004) show some international evidence supporting the notion of nonlinear interest rate rules. Ruge-Murcia (2003) and Cukierman and Gerlach (2003) adopt an inflation rate reaction function that is nonlinear in either inflation or the output gap, and they favor the hypothesis of an asymmetric objective for some OECD economies. Dolado et al. (2004) estimate an interest rate rule that is asymmetric in inflation only, and find evidence of nonlinearity after 1983 for the U.S. Despite the increasing number of empirical works, the inflation bias associated with asymmetric preferences had not been quantified yet. In particular, no study had assessed, to our knowledge, the average contribution of asymmetric preferences to inflation during the 1960s and 1970s. This paper attempts to fill the gap. The specification of a potentially asymmetric loss function generates the testable prediction that the monetary authorities respond nonlinearly to the inflation and the output gaps. This prediction is used to identify the degree of asymmetry with respect to both objectives. A main result of the paper is that the Fed's monetary policy can be characterized by a nonlinear interest rate rule only before 1979 and with respect to the output gap. According to the model, these estimates imply an average inflation bias around 1.5%1.5% before 1979 but a value not statistically different from zero over the last two decades. The fall in the bias is found to account for a larger fraction of the historical decline in average inflation relative to a reduction in the Fed's implicit inflation target. Asymmetric preferences seem thus to provide a new, additional explanation for the great inflation of the 1960s and 1970s, which may also be relevant for other countries. The paper proceeds as follows. Section 2 presents the model and derives the interest rate rule as the first-order condition of the central bank optimization problem. Section 3 reports the results of the hypothesis testing for symmetric preferences and the estimates of the asymmetric preference parameters. The following part maps the estimates of the nonlinear Euler equation into a measure of inflation bias. Section 5 concludes.

The contribution of this paper is twofold. At the theoretical level, it derives the analytical solution of the central bank optimization problem when policy preferences are asymmetric in both inflation and output gaps, and the monetary transmission mechanism is New-Keynesian. The specification of the policy objectives is general enough to nest the quadratic form as a special case and translates into a potentially nonlinear targeting rule. This feature forms the basis of our hypothesis testing for the presence of asymmetric preferences as it allows us to reversely engineer potential evidence of nonlinearities in the reaction function into evidence of asymmetries in the policy objective. At the empirical level, this paper shows that U.S. monetary policy can be effectively characterized by a nonlinear policy rule during the pre-Volcker regime only, with the interest rate response to the output gap being the dominant type of nonlinearity. In particular, the Fed attached a larger weight to output contractions than to output expansions of the same magnitude such as to induce an average inflation bias of about 1.5%1.5%. The latter can account for a sizable fraction of the decline in the inflation mean moving from the pre- to the post-1979 period. This paper focuses on the notion of average inflation bias and hence offers a new explanation for the fact that the inflation mean fell across monetary policy regimes. An interesting avenue for future research is to investigate further the properties of the implicit inflation target by modeling also its time variation within regimes. The random walk assumption put forward by Smets and Wouters (2003a) and Ireland (2005) is a promising step in this direction.