باید سیاست های پولی درمقابل سستی تکیه گاه داشته باشند؟: تجزیه و تحلیل بر اساس مدل DSGE با بانکداری

The global financial crisis has reaffirmed the importance of financial factors for macroeconomic fluctuations. Recent work has shown how the conventional pre-crisis prescription that monetary policy should pay no attention to financial variables over and above their effects on inflation may no longer be valid in models that consider frictions in financial intermediation (Cúrdia and Woodford, 2009). This paper analyzes whether Taylor rules augmented with asset prices and credit can improve upon a standard rule both in terms of macroeconomic stabilization and of agents׳ welfare in a DSGE with both a firms׳ balance-sheet channel and a bank-lending channel and in which the spread between lending and policy rates endogenously depends on banks׳ leverage. The main result is that, even in a model in which financial stability does not represent a distinctive policy objective, leaning-against-the-wind policies are desirable in the case of supply side shocks, while strict inflation targeting and a standard rule are less effective. The gains are amplified if the economy is characterized by high private sector indebtedness. Robustness shows that the interaction between financial frictions and debt-deflation effects is potentially very powerful.

Before the global financial crisis the consensus view on the conduct of monetary policy was that the central bank should pay no attention to financial variables over and above their effects on inflation; an aggressive inflation-targeting policy was considered sufficient to guarantee macroeconomic stability. This conclusion emerged from a debate which focused exclusively on how central banks should deal with asset price bubbles and relied on several arguments explaining why monetary policy was, at best, ineffective.1 The theoretical underpinnings of the pre-crisis consensus were the works by Bernanke and Gertler, 2000 and Bernanke and Gertler, 2001, Gilchrist and Leahy (2002) and Iacoviello (2005). A crucial characteristic of these papers is that their results are based on models that consider financial frictions only on the borrowers׳side of credit markets. Credit-supply effects stemming from financial intermediaries׳ behaviour were completely neglected. The financial crisis has instead shown how shifts in credit supply can indeed have a crucial role in macroeconomic fluctuations. Empirical research has pointed out how in many advanced economies loose credit conditions have contributed to amplifying the business cycle prior to the financial crisis, while the tightening of lending standards in the aftermath of the Lehman׳s collapse has contributed to the strong decline in output recorded in 2008–2009 (Adrian and Shin, 2010, Ciccarelli et al., 2010 and Gilchrist et al., 2009). More recently, fears of a credit crunch have resurged in connection with the European sovereign debt crisis (Draghi, 2011). Theoretical analysis has turned its attention to the implications of the credit-supply channel for the conduct of monetary policy (Meh and Moran, 2010 and Gertler and Karadi, 2012); recent work that considers financial frictions on the side of lenders has stressed how “decisions about interest-rate policy should take account of changes in financial conditions” ( Woodford, 2011, p. 39). After the crisis, policymakers have also reconsidered the so-called “Greenspan doctrine”, i.e., the prescription that asset prices should have no role in the conduct of monetary policy over and above that implied by their foreseeable effect on inflation and employment ( Mishkin, 2011). 2 In a formal characterization of these ideas, Cúrdia and Woodford, 2009 and Cúrdia and Woodford, 2010 (henceforth CW) have introduced, in an otherwise standard New Keynesian model, an ad hoc friction in financial intermediation that gives rise to a spread between the loan and the policy rate. In that context, they have shown that—despite standard optimal policy prescriptions continue to hold—spread- or credit-augmented rules are a better approximation to the optimal policy than the standard Taylor rule for a number of different shocks.3 Taking stock of this debate, in this paper we ask if “leaning-against-the-wind” (henceforth LATW)—defined as monetary policy following an “augmented” Taylor rule, which takes into account asset prices or credit—may improve upon a standard rule in the context of a model that combines frictions on both the borrowers׳ and the lenders׳ side; in our model, in particular, loan spreads endogenously depend on banks׳ leverage. Our main contribution to the existing literature is thus to analyze how different instrument rules perform in a model with the simultaneous presence of a borrower balance-sheet and a bank credit-supply channel. In doing so we show that, when credit supply effects are present, responding to financial variables allows the central bank to achieve a better trade-off between inflation and output stabilization and improves both households׳ and entrepreneurs׳ welfare; we thus corroborate CW׳s results using a richer model of the financial sector and analyzing a different range of financial variables that the central bank might want to look at. Our model is a simplified version of Gerali et al. (2010), who estimated a medium-scale model for the euro area. In particular, we share with that paper the main characteristics of the financial sector. A bank lending channel arises due to the presence of a target level for banks׳ leverage; as a consequence, the loan-supply schedule is positively sloped and shifts procyclically with changes in the policy rate and with banks׳ profitability and capital. As pointed out by Woodford (2011), a loan supply curve with these characteristics could be motivated in several ways. For example intermediaries may have costs for originating and servicing loans, with marginal costs increasing with the volume of lending; or leverage could be bounded by regulatory limits or market-based constraints. The balance-sheet channel is modelled along the lines of Iacoviello (2005), assuming that entrepreneurs׳ borrowing capacity is linked to the value of the assets that they can pledge as collateral. On the other hand, the fact that we use a simplified (and calibrated) version of Gerali et al. (2010) implies that our quantitative results should be take with caution; indeed, we aim at obtaining qualitative indications on whether LATW may improve macroeconomic stabilization in a model that provides a sufficiently rich representation of the credit market. Our analysis focuses on aggregate supply shocks, which create a trade-off for a central bank that aims at stabilizing output and inflation since, conditional on the shocks, the two variables tend to move in opposite directions. On the one hand, this choice helps us showing how the presence of financial frictions may amplify financial cycle fluctuations—with an effect on macroeconomic stability—even during “normal times”. On the other hand, this choice limits the scope of the paper, as it does not allow the model to account for periods of financial distress. In particular, the model cannot explain the genesis of the Great Recession for which demand effects have likely been far more important than supply side effects. Our main results support the view that LATW is indeed desirable when the economy is driven by supply side shocks, while both strict inflation targeting and a standard Taylor rule are less effective. Consider first strict inflation targeting or, equivalently, a standard rule with a very aggressive response to inflation. In this case, the strong response to inflation reduces inflation volatility less than it increases that of output, due to the impact of policy rates on credit-market developments. Following a positive technology shock, for example, this type of policy calls for a reduction of the policy rate, which tends to counteract the fall in consumer prices. In the presence of a broad credit channel, however, the easing of policy has a very strong expansionary effect on output, due to its impact on asset prices and loan supply, which in turn sustains a boom in consumption and investment by borrowing agents. In one word, this type of rules implies that monetary policy is “too loose” in the face of a positive supply shock, generating a procyclical behaviour by financial sector variables which, in turn, amplifies volatility in the real economy. Consider now a standard rule with a non-negligible response to output or with a response to financial variables. In this case, the response of monetary policy is less accommodative, possibly becoming countercyclical, and partly counteracts the amplification effects stemming from the presence of financial frictions. Simulations show that in this case, a rule that entails LATW delivers superior results in terms of macroeconomic stabilization as compared to a standard rule, suggesting that financial variables are better indicators than output of the procyclical effects stemming from financial frictions. As a criterion for evaluating the relative performance of the different rules, we use two different approaches. First, we assume that the central bank׳s objective is the minimization of a standard loss function in the (weighted) variance of inflation and output. Second, we consider welfare-based optimal simple rules in the spirit of Schmitt-Grohé and Uribe (2007). In both cases, LATW turns out to be superior to a standard Taylor rule. In particular, with the loss-function approach, LATW reduces the value of the central bank׳s loss by as high as 20–30%, depending on the shock considered and the central bank׳s preferences assumed. Gains tend to be larger for greater weight assigned to output stabilization by the central bank. Moreover, we find that gains from LATW are amplified in an economy characterized by a high degree of private sector indebtedness. With the welfare-based approach, we find that LATW improves welfare of both households and entrepreneurs. Considering a technology shock, gains in terms of consumption equivalents under the augmented rule are 36% for households and almost 100% for entrepreneurs. Gains are smaller but still substantial when a cost-push shock is considered. As for the specific variables that the central bank should look at, in general a rule responding to asset prices performs better than a rule responding to credit; this reflects the fact that in our model LATW does a better job in stabilizing inflation, because it reduces the fluctuations of investment, the return to capital and thus marginal cost. We run a number of robustness checks on the calibration used in the baseline. We find that qualitatively our results are robust to changes in the parameters determining bank׳s target leverage and the loan rate elasticity to bank leverage; price stickiness, investment adjustment costs and labor supply elasticity; monetary policy and shock persistence. Quantitatively, differences with respect to the baseline calibration, in terms of the gains attainable by LATW, are modest. We also investigate how our results are affected by the assumption that debt contracts are indexed in nominal terms. This hypothesis removes the so-called “Fisher׳s debt-deflation” effect, which is typically at work in models with a balance-sheet channel (Iacoviello, 2005). We find that our results continue to hold qualitatively, but they are significantly reduced quantitatively, suggesting that the interactions between financial frictions and debt-deflation effects can potentially be substantial. Since the onset of the financial crisis, other contributions have reassessed the case for LATW also in models that do not have a credit-supply channel, focusing on shocks on expected future economic conditions (Lambertini et al., 2013 and Christiano et al., 2010).4 Compared with these studies, our results are more general. First, we show that LATW is also desirable when the economy is hit by current (rather than expected) supply shocks; the reason is that credit supply conditions in our model are affected by the current state of the business cycle and that expansions of economic activity are associated with a boom in bank leverage and lending. Second, we evaluate optimal rules considering a wider range of variables and possible coefficients for the central bank׳s response to inflation, output and financial variables and a variety of central bank׳s relative preferences for inflation versus output stabilization. It is important to stress that our results indicate that LATW is desirable even when the central bank is concerned only with macroeconomic objectives; indeed, by using a linearized model where all variables eventually return to their steady state level, we rule out any consideration regarding financial (in)stability. Nonetheless, it can be argued that LATW would likely bring about even more social benefit if the central bank (or any other public authority) was also concerned about financial stability; as a simple hint in this direction it is worth noticing that, in our model, LATW has the effect of reducing the variability of many financial variables besides that of output and inflation. The remainder of the paper is organized as follows. Section 2 discusses the dynamic stochastic general equilibrium (DSGE) model used in the simulations; Section 3 analyzes the main financial channels at work; Section 4 describes the simulations of a technology shock and a cost-push shock; Section 5 examines whether a policy of leaning-against-the-wind is more effective in economies with highly leveraged borrowers; Section 6 studies the role of nominal debt; Section 7 provides some robustness checks; Section 8 summarizes the main conclusions.

The financial crisis has reaffirmed the importance of financial factors for explaining macroeconomic fluctuations. Empirical research has pointed out how credit boom-bust cycles may affect the business cycle; theoretical analysis has turned its attention to the implications of this channel on the conduct of monetary policy. The pre-crisis consensus on the conduct of monetary policy, namely that the central bank should pay no attention to financial variables over and above their effects on inflation, has been reconsidered. In this paper we contribute to the existing literature by analyzing how different instrument rules perform in a model with the simultaneous presence of a balance-sheet and a credit-supply channel. We show that when credit supply conditions matter for the real economy, responding to financial variables allows the central bank to reach a better trade-off between inflation and output stabilization and to improve agents׳ welfare in the economy. In particular, assuming that the central bank׳s objective is the minimization of a standard loss function in the (weighted) variance of inflation and output we find that LATW may bring about gains as high as 20–30%, depending on the shock considered and the central bank׳s preferences assumed. If we assume that the central bank objective is agents׳ welfare maximization, then LATW brings about gains (in terms of consumption equivalents) in the range of 2–36% for households and 20–99% for entrepreneurs depending on the shock considered. These gains are further amplified if the economy is characterized by a high level of the loan-to-value ratio, which mimics an economy with more indebted borrowers (which in turn could reflect institutional factors or a different degree of development in the financial industry). Our results are robust to changes in a number of key model parameters. Qualitatively, they are also not affected when we remove the assumption that debts are indexed to current inflation, reintroducing the so-called “Fisher׳s debt-deflation” effect, which is typically at work in models with financial frictions. Quantitatively, however, gains from LATW are substantially smaller when we consider this alternative model. Our study corroborates previous results (Cúrdia and Woodford, 2009, Cúrdia and Woodford, 2010, Lambertini et al., 2013 and Christiano et al., 2010) using a richer model of the financial sector and analyzing a different range of financial variables that the central bank might want to look at. From a policy perspective our results, obtained in the context of a linearized model that rules out financial (in)stability and default, highlight the potential gains from leaning against the wind from a strict macroeconomic stabilization perspective; gains are likely to be larger if financial stability was also to be the concern of policymakers. Some caution is obviously required when interpreting the results of this paper. First, admittedly, the model studies the conduct of monetary policy in normal times, not during a period of financial stress: many important aspects that are typical ingredients of boom-bust cycles in asset markets such as sudden shifts in borrowers׳ credit risk perception and the possibility of banks׳ default are not modelled. This could hide the information content of some financial variables such as credit that the literature indicates as a good indicator to detect the development of a financial crisis. Moreover, it would be particularly interesting to analyze the effect of the augmented rules in a model where households—rather than firms—are constrained borrowers, with the constraint tied to the value of housing collateral; this exercise could show whether the presence of a credit-supply channel might challenge the result of irrelevance of asset prices found by Iacoviello (2005). Second, the typical solution techniques used for the DSGE models, based on log-linearization, do not allow for the non-linear dynamics that typically characterize boom-bust episodes. However, despite the lack of these features, the importance for monetary policy to lean against the wind is fully recognized by simply giving a non-negligible role to financial flows and credit intermediation. Third, our results should not be interpreted as providing precise quantitative prescriptions of the optimal values to be assigned to asset prices in an operational rule. The fact that they are obtained from numerically optimized rules, calculated on a finite and discrete grid of possible parameters, and based on a stylized model, suggests that their indications are mainly of a qualitative nature. In addition, our analysis considers technology and cost-push shocks one at a time. This simplification is crucial for understanding how the transmission mechanism of the various shocks works and for studying the trade-offs that each of them entails for the conduct of monetary policy; however, this approach further restricts the ability of our analysis to give quantitative prescriptions, because it implicitly assumes that the central bank can perfectly disentangle the source of business cycle fluctuations. Finally, our study, focusing on the interaction between asset-price developments and monetary policy, underscores the importance of co-operation between the central bank and the macroprudential authorities (Borio, 2006 and Angelini et al., 2011). All these issues are important avenues for future research