قرارداد نرخ تورم، شفافیت بانک مرکزی و مدل عدم قطعیت

Using a New-Keynesian model subject to misspecifications, we examine how the robust monetary policy could be modified by a linear inflation contract when a central bank is opaque about its preference for model robustness. It is shown that a central bank must limit this preference and opacity about it to ensure the dynamic stability of the economy. An optimal inflation contract with a zero penalty rate provides no incentive for a central bank to be opaque. The latter must rebalance the benefit of avoiding very bad outcomes in worst case scenarios and the economic costs due to higher macroeconomic volatility.

A discretionary monetary policy suffers from time-inconsistency problems that arise because central banks optimally respond to the wrong incentive structure that they face. A large principal–agent literature proposes to shape the incentives through the design of an institutional structure within which policy is conducted, providing thus an appropriate perspective for addressing such issues (Persson and Tabellini, 1993; Walsh, 1995; Waller, 1995; Fratianni et al., 1997; Candel-Sánchez and Campoy-Miñarro, 2004; Chortareas and Miller, 2007). Such institutional aspects of the central bank structure and its relationship with the government can be considered as a contract between the government and the central bank. The essential idea of the so-called “contracting” approach is that it is possible for the government, as the principal, to design an optimal incentive scheme for obtaining monetary policy outcomes equivalent to those under credible commitment. Generally, there exists an efficient punishment (transfer) mechanism that sufficiently raises the marginal costs of producing high inflation and hence neutralizes the monetary policymaker's tendency to take such actions. While the initial principal–agent literature assumes full information, subsequent research in the contracting approach is extended to take into account two aspects of imperfect central bank transparency, i.e. the economic transparency and the political transparency.3 Under uncertainty about the central bank's output target, i.e. a lack of economic transparency, the optimal contract induces an equilibrium inflation rate which is certainty-equivalent to that obtained under full economic transparency (Muscatelli, 1999). Political transparency refers to disclosure of information about the central bank preferences. Considering information asymmetries between the central bank and the public about the weight assigned to the arguments of the central bank's objective function, Beetsma and Jensen (1998), and Muscatelli, 1998 and Muscatelli, 1999 have shown that the optimal linear contract implies a persisting inflation bias in equilibrium and hence the trade-off between reducing the inflation bias and stabilization. According to Chortareas and Miller (2003), when the government is uncertain about the central banker's responsiveness to incentive schemes, more complicated incentive schemes are required for them to work effectively. Examining the same transparency issue under common agency, Ciccarone and Marchetti (2008) show that when opacity increases, the average inflation bias falls. Insofar the principal–agent literature ignores the role of model robustness which has been extensively examined by Hansen and Sargent, 2003 and Hansen and Sargent, 2007, Svensson and Woodford (2004), Giannoni and Woodford, 2003a and Giannoni and Woodford, 2003b, Onatski and Stock (2002), Giannoni, 2002 and Giannoni, 2007, and Leitemo and Söderström, 2008a and Leitemo and Söderström, 2008b, among others. It is argued that while a simple model may help the public to understand the monetary policy decisions, it can be criticized for not fitting the data well. One possible response to such criticism is to design more complex models which gain in realism but lose in tractability. The robustness literature suggests that the central bank does not know the true structure of the economy, recognizing that the simple model is a misspecified description of reality. In practice, confronted with such a dilemma, many central banks publish plenty of analysis on their models and policy rules robustness. The aim of this paper is to address in the contracting approach the issues arising from the introduction of model robustness while the central bank does not fully communicate its private information about the accuracy of the model on which is based the monetary policy decisions. We assume that the true model of the economy lies in the neighborhood of the reference model, and we analyze how monetary policy should be designed in order to work reasonably well for all models inside this neighborhood. The robust policymaker is supposed to be unable to formulate a probability distribution over plausible models and hence designs policy for the worst possible outcome within a pre-specified set of models. Its decision problem is solved then using the robust control techniques (Hansen and Sargent, 2007). The main objectives of our study are to investigate whether the introduction of uncertainty about the central bank preferences for model robustness under an inflation contract modifies the implications of the robust control approach for the monetary policy decisions on the one hand, and to examine whether the linear contract is an appropriate response to central bank's opacity about the model robustness on the other hand. An important practical implication of this approach is that the attenuation principle established by Brainard (1967) may not always hold. The concern about worst-case scenarios emphasized by the robust control may likewise lead to amplification rather than attenuation in the response of the optimal monetary policy to shocks in a closed economy (e.g., Giannoni, 2002, Giordani and Söderlind, 2004, Leitemo and Söderström, 2008a and Onatski and Stock, 2002). In effect, the precautionary central bank takes stronger actions in order to prevent particularly costly outcomes. However, Leitemo and Söderström (2008b) show that this result does not carry over to an open economy where the optimal robust policy can be either more aggressive or more cautious than the non-robust policy.4 This study is related to Dai and Spyromitros (2010) who have examined the same issues under flexible inflation targeting rules. They have shown that the disclosure of the knowledge about the true structure of the economy could significantly influence strategic interactions between the central bank, the government and private agents. In particular, monetary policy makers might strategically use their private information in order to gain benefits in terms of output stabilization and goal independence by being opaque about its preference for model robustness. The remainder of the paper is structured as follows. In the next section, we present the model. In the section after, we solve the model to obtain the optimal inflation contract under full transparency about the preference for model robustness, and we examine then the effects of opacity on inflation penalty rate, model robustness as well as the macroeconomic performance. In Section 4, the macroeconomic implications of a serially correlated inflation shock are investigated. The last section concludes.

In this paper, we have examined the accountability issue in a framework of delegation, under a linear Walsh's inflation contract, where the government and private agents are subject to uncertainty about the preference for model robustness of the central bank, while the latter faces model uncertainty. Furthermore, we have studied the implications of these two kinds of uncertainty for macroeconomic performance. We have shown that a central bank with a very strong preference for model robustness and too opaque about this preference leads to an unstable dynamic process of adjustment of the expected inflation rate. This will induce higher welfare costs in dynamic terms because the economy will stay longer in disequilibrium or is indeterminate. By limiting to the case of dynamic stability, we have eliminated the possibility that the equilibrium solutions of inflation, the output gap and model misspecifications could tend from plus to minus infinity at the critical values of model robustness and central bank opacity in the event of arbitrarily small inflation shocks. If the inflation shock is serially uncorrelated, under full transparency and in the absence of robust control, it is not necessary to impose any inflation penalty under the contract scheme. This result is explained by the fact that we do not introduce an inflation bias in the loss function of the central bank. If the central bank applies robust control techniques in monetary policy decision, an increase in the preference for model robustness does not lead the government to impose an inflation penalty rate different from zero. In effect, a linear contract does not influence the choice of the central bank about the model misspecification. However, under such a contract, while opacity has no effect on the level of inflation and the output gap, it increases the volatility of these variables and hence the social welfare loss. Therefore, the central bank has no incentive to be opaque under a linear inflation contract. In the case of serially correlated inflation shocks, our study suggests that the optimal penalty rate is always set at zero under full transparency or opacity. A higher degree of persistence of inflation shocks implies a stronger response of inflation, the output gap and model misspecifications to inflation shocks and reinforces the effect of opacity on the sensitivity of these variables to inflation shocks. Under the assumption that the central bank does not choose a too high degree of opacity compromising the dynamic stability, opacity and persistence are positively related to the macroeconomic volatility (and hence the social welfare loss) and their effects on the latter are mutually reinforced. They therefore reinforce the central bank's concern about worst-case scenarios emphasized by the robust-control approach, leading hence to further amplification in the response of the optimal monetary policy to inflation shocks. Stronger action by the precautionary central bank may be justified to prevent particularly costly outcomes. Nevertheless, the central bank must rebalance the benefit of avoiding very bad outcomes in worst case scenarios and the economic costs due to a higher volatility in the inflation rate and the output gap, since high uncertainty about the central bank's preferences for model robustness and the persistent inflation shocks increase the economic costs. In general, a case for central bank transparency in this trade-off could be made, particularly in the presence of persistent inflation shocks.