Chapter

Chapter 7. Financial Stability Objectives: Drivers of Gains from Coordinating Monetary and Macroprudential Policies

Author(s):
Luis I. Jacome H., Yan Carriere-Swallow, Hamid Faruqee, and Krishna Srinivasan
Published Date:
October 2016
Share
  • ShareShare
Show Summary Details
Author(s)
Jessica Roldán-Peña, Mauricio Torres-Ferro and Alberto Torres 

This chapter studies the trade-offs that can arise between inflation-targeting and financial stability objectives. We use a simple framework to conduct macroeconomic policy analysis under two strategies that aim to attain both objectives: monetary policy leaning against the wind and policy coordination between monetary and macroprudential instruments. The chapter identifies cases in which the trade-offs between policy objectives create scope for improving macroeconomic performance when monetary and macroprudential policies are coordinated. These improvements are larger when financial shocks are the main driver of macroeconomic fluctuations. The findings further emphasize the importance of assessing the relative effectiveness and interactions of different policies that policymakers have at their disposal to help guide the economy.

The 2008–09 financial crisis shook the global economic landscape in a significant way. The need to take immediate actions to solve time-pressing problems left policymakers—especially central bankers in both advanced and emerging market economies—with the difficult task of restoring or maintaining macroeconomic and financial stability without much guidance on how to do so. Indeed, the existing macroeconomic policy toolkit was generally inadequate to address the challenges that arose, while the effectiveness and externalities of unconventional tools remained largely unknown.1 The aftermath of the crisis inevitably led to a rethinking of the conduct of macroeconomic policy in general, and monetary policy in particular.2 One of the main challenges for central bankers in the years ahead will be to advance their understanding of how their respective financial systems work in order to modify and adapt their macroeconomic policy frameworks.

Without a doubt, two issues that have received the most attention in reexamining the conduct of macroeconomic policy are the recognition that (1) price stability does not necessarily imply financial stability and that (2) a broader perspective in prudential supervision and regulation is needed to prevent the accumulation of financial imbalances. With regard to the latter issue, a new field of macroeconomic policy, namely macroprudential policy, emerged in recognition of the fact that regulatory policy focused on individual institutions was ill-equipped to prevent the buildup of macro-financial and systemic risks. Nevertheless, given the inherent links between the real and financial sectors of the economy and the feedback loops between monetary and macroprudential policies, some basic questions arose. Should monetary policy be responsible for attaining financial stability objectives in addition to price stability objectives? In other words, should monetary policy lean against the wind? Alternatively, should macroprudential policy be tasked with pursuing financial stability objectives? And if so, what should be its relation with monetary policy?3

This chapter examines the trade-offs that arise between the pursuit of price stability objectives through a flexible inflation-targeting regime and financial stability objectives. To do so, we estimate a standard, reduced-form, and small open economy model (commonly used to conduct monetary policy analysis in emerging markets), which we extend to consider a stylized financial sector. Within this framework, we study the performance of the model economy under different policy strategies. Specifically, two cases are examined: (1) a leaning-against-the-wind case in which monetary policy sets the short-term nominal interest rate in order to attain certain predetermined financial stability objectives, in addition to price stability, in the context of a flexible inflation-targeting regime;4 and (2) a coordination case in which both monetary and macroprudential policies jointly set their instruments in order to attain both objectives. Additionally, we also consider as a benchmark the “prior to crisis” case in which monetary policy actions focused only on price stability objectives. Following the literature, we relate traditional inflation-targeting objectives to a particular specification of the monetary policy’s loss function, defined in terms of macroeconomic variables. Likewise, we associate financial stability objectives with a loss function that considers the stabilization of financial variables. To assess each case’s relative effectiveness to attain its policy objectives, we examine the model’s transmission mechanism and policy implications. Therefore, we compare the volatility that each case entails for a set of relevant macroeconomic and financial variables.

The purpose of this work is twofold: first, to provide a simplified framework within which to account for the trade-offs that arise when an inflation-targeting central bank also pursues financial stability objectives; and, second, to identify cases where these trade-offs allow for improvements in macroeconomic performance under coordination of monetary and macroprudential policies.

The results suggest that including financial stability considerations as an additional objective of monetary policy indeed reduces the volatility of financial variables. However, this improvement comes at the expense of increasing the volatility of macroeconomic variables. This cost is present regardless of the type of shocks that affect the economy but tends to increase when fluctuations are driven by financial shocks. These results provide evidence that supports the views of those who do not favor leaning against the wind because of the costly trade-offs involved and advocate for the use of macroprudential policy to address financial stability objectives.5

When identifying the cases where there is an improvement in macroeconomic performance owing to a coordination of monetary and macroprudential policies, our findings emphasize the importance of assessing the relative effectiveness and interactions of the different instruments/policies that policymakers have at their disposal. Our framework features room for gains from coordination, even when macroeconomic shocks are the only type of shocks affecting the economy, either when monetary policy is not an effective tool to address vulnerabilities in the financial sector or when it conflicts with macroprudential policy. In other words, we are referring to cases where one approach responds in a countercyclical way and the other in a procyclical way.

This chapter relates to a growing literature that explores the interaction between monetary and macroprudential policy (Angelini, Neri, and Panetta 2014; Angeloni and Faia 2013; Beau, Clerc, and Mojon 2012; De Paoli and Paustian 2013; Kannan, Rabanal, and Scott 2012; and Lambertini, Mendicino, and Punzi 2013). Our exercise follows closely the approach in Angelini, Neri, and Panetta (2014), who study the benefit of the interaction between capital requirements and monetary policy in a dynamic general equilibrium model featuring a banking sector. They find that an overall improvement in economic stability is attained when monetary and macroprudential policies are coordinated in the presence of financial shocks. Unlike previous studies, which consider a Taylor-type rule to account for monetary policy actions and distinguish between cooperative and noncooperative interactions between monetary and macroprudential policies, we consider optimal policy rules and assume that when monetary and macroprudential policies coexist they do so under a coordination scheme. Although the results of both approaches are similar, our model is much simpler. This allows us to focus on the broad intuition rather than on specific channels driving our results. Hence, our main contribution is to provide a common ground for thinking about the interaction between optimal monetary and macroprudential policies, even in cases when country-specific circumstances call for specific ways to model an economy.

The next section outlines the different policy strategies considered to examine the trade-offs between inflation-targeting and financial stability objectives. The chapter then describes the model, presents the main results, provides a sensitivity analysis, and concludes.

Policy Strategies

There are two possible policy objectives: one that is associated with standard inflation targeting, and another that adds a financial stability goal. These goals can be attained by using at most two policy instruments: the short-term nominal interest rate, it, the instrument of monetary policy (henceforth the policy rate); and the coverage ratio, crrt, defined as the ratio of loan-loss reserves to nonperforming loans that banks are required to bear, which serves as the instrument for macroprudential policy.6 We study the following policy strategies throughout the analysis:

  • 1. A benchmark case in which monetary policy is guided by a flexible inflation-targeting regime and as such, it sets the policy rate aiming at stabilizing inflation around its target, the output gap, and changes in the policy rate. Hence, the monetary policy’s optimization problem is to set it to minimize the following loss function:

    where σπ, σx, and σΔi represent the monetary authority’s relative preferences for stabilizing deviations of inflation from its target, (πtπ*); the output gap, xt; and changes in the reference rate, (itit−1), subject to the dynamics of the economy.7

  • 2. A leaning against the wind case considers that monetary policy aims to stabilize both the traditional objectives of inflation-targeting and financial stability objectives by setting the optimal level of its policy rate. The monetary policy’s optimization problem in this case is to set it to minimize the following loss function:

    where σspread, and σcrgap represent the monetary authority’s relative preferences for stabilizing the credit spread, denoted by spread, around its long-term average and the credit-to-GDP gap, crgap, respectively. These variables will be further explained when we describe the model.8

  • 3. The monetary and macroprudential policies coordination case considers that monetary and macroprudential policies aim to attain inflation-targeting and financial stability objectives by setting their respective policy instruments—namely, the policy rate and the coverage ratio. These two instruments are jointly and optimally set in order to minimize the following loss function:

    where σΔcrr represents the relative preference for stabilizing changes in the macroprudential instrument.9

When considering the coexistence of monetary and macroprudential policies, we restrict our analysis to the case where both of them interact in a coordinated manner. This assumption allows for acknowledging an important result found in the literature, namely, the fact that since both policy instruments work less than perfectly, one cannot ignore the limitations of the other and, hence, should account for its effects on the economy (Blanchard, Dell’Ariccia, and Mauro 2013). According to De Paoli and Paustian (2013), this way of interaction between monetary and macroprudential policies is consistent with a first-best solution.

The Model

A growing literature emphasizing the role of macro-financial linkages in macroeconomic models for monetary policy analysis has emerged in recent years.10 Most of these models build on the work developed by Bernanke, Gertler, and Gilchrist (1999) and Iacoviello (2005) that introduce endogenous financial frictions—which arise due to the presence of agency costs (à la Bernanke and Gertler 1986) or the lack of financial-contract enforceability (à la Kiyotaki and Moore 1997)—into general equilibrium frameworks with real and nominal rigidities. In general, these types of models focus on understanding factors affecting the demand for credit that tend to propagate and amplify the transmission of shocks through a “financial accelerator” mechanism, leaving no relevant role for financial intermediaries to play. In response to this, macroeconomic models emphasizing the role of credit supply factors (e.g., the market structure of the banking system, the rate-setting strategy of banks, the role of banks’ balance sheet composition and management, etc.) in the transmission of macroeconomic and financial shocks have also been developed.11 Despite these contributions, no canonical framework is yet available within which to study the relationship between financial frictions, financial intermediation, and macroeconomic activity and its implications both for monetary and macroprudential policies.12

Given this lack of consensus, we use a simple model that accounts for the interaction between a standard macroeconomic setup and some financial variables as the basis for our analysis. Specifically, we follow the approach taken by Sámano (2011), who augments a reduced-form New Keynesian small and open economy model, that is, the macroeconomic block, by appending a macroeconometric financial sector or financial block. This approach allows for the introduction of macro-financial linkages into an otherwise standard macroeconomic model—commonly used to analyze monetary policy in emerging markets—to study the propagation of macroeconomic shocks into the financial sector and vice versa. Admittedly, the nature of this framework has limitations, including: (1) the lack of micro foundations, which makes the model sensitive to the Lucas critique and inadequate for welfare analysis; (2) it is based on a representative agent setup, leaving out specific wedges that arise due to heterogeneity among agents and that characterize models where financial frictions are present; and (3) it is a linearized representation around the steady state of the economy. As a result, the model does not capture potential nonlinear effects of sufficiently large shocks capable of producing financial instability, and thus only features macroeconomic fluctuations under “normal times” (i.e., it is not a model where financial crises occur). Regardless, we consider it useful to provide guidelines about relevant trade-offs between policy objectives.

Macroeconomic Block

The structure of the macroeconomic block is characterized by aggregate supply and demand relationships that include the effect of economic openness (via the impact of the real exchange rate and foreign output, inflation, and interest rates on domestic activity and inflation):

Equation (7.4) is a hybrid Phillips curve that explains core inflation as a linear combination of expected and lagged inflation (with Et representing the time-t rational expectations operator), the output gap and changes in the nominal exchange rate, Δe ≡ Δrer + ππf, and foreign inflation, πf.13Equation (7.5) is a hybrid investment-saving (IS) curve that establishes that the output gap depends on its expected and lagged values, the real ex-ante interest rate, iE[π], (with the policy rate set according to the different cases under analysis, described in the previous section), the real exchange rate, rer, and the credit spread, spread. The latter variable, not present in the standard New Keynesian model, arises with the introduction of financial considerations into the model. It accounts for the distortions on the allocation of expenditures imposed by the presence of financial frictions and/or financial intermediation.14 The spread, for which we provide further details later, is what drives the feedback mechanism between the financial sector and the rest of the economy. Lastly, equation (7.6) models the real exchange rate dynamics imposing uncovered interest rate parity.

The set of exogenous variables of the macroeconomic block include noncore inflation (recall that the loss functions in the previous section are defined in terms of the deviations of headline inflation from its target), foreign output, inflation, and interest rates. Noncore inflation is assumed to follow an autoregressive process of order 1, while foreign variables are modeled using a vector autoregressive model of order 2.

Financial Block

A financial block (sketched in Figure 7.1) is appended to the rest of the economy in order to capture, in a stylized fashion, the credit channel of monetary policy. Specifically, we consider a financial sector, characterized by the existence of banks, in charge of intermediating resources among borrowers and lenders. This intermediation is done at a cost, represented by the credit spread. We further capture vulnerabilities in the financial sector by introducing credit dynamics into the model.15

Figure 7.1.Model Mechanism

Source: Authors’ representation.

We assume that there is an intermediation technology that transforms real deposits into real loans subject to costs associated with loan monitoring activities. These costs are related to the presence of some financial friction on the demand side of credit, which gives rise to an external financing premium imposed on borrowers, as in Cúrdia and Woodford (2009). Following Gerali and others (2010), banks devote some of their resources to managing their balance sheets and/or to fulfilling costs associated with regulatory requirements. We also assume that banks enjoy some degree of market power due to the presence of monopolistic competition. This allows banks to charge a markup over the policy rate—that is, the relevant rate at which banks intermediate resources in the interbank market absent other intermediation costs—when setting loan (or active) rates, on the one hand, and a markdown with respect to the policy rate when setting deposit (or passive) rates, on the other.

The active interest rate, itloan, whose dynamics evolve as follows,

is directly affected by monetary policy, with β1(< 1) representing a limited pass-through from policy rates to active rates that arise from market power. It is also affected by a finance premium channel, delinq, to be defined momentarily, and the potential macroprudential policy tool or costs associated with the management of the bank’s balance sheet, crr. For the passive rate we assume a similar structure:

where γ1 determines the pass-through of the policy rate to funding costs that arise from banks’ market power on deposit markets. Thus, the credit spread is given by the following expression:

We capture the buildup of risks in the financial sector through the evolution of the credit-to-GDP gap, which is positively related to economic activity (i.e., credit booms generally start after periods of rapid economic growth), and negatively related to the credit spread. The latter relationship lets us account for the linkage between financial conditions and credit booms:16

The delinquency index, delinq, is assumed to depend on economic activity, the credit-to-GDP gap, and its own lagged value:

The rationale behind this specification is, on one hand, that episodes of economic expansion lead to a reduction in nonperforming loans as debtors’ defaults decline. On the other hand, an accelerated credit expansion, captured by a positive credit-to-GDP gap, may lead to vulnerabilities via a reduction in the quality of new loans as a result of looser lending standards during credit booms. This eventually translates into an increase of the delinquency index.

The coverage ratio rule, crr, has two possible interpretations depending on the policy strategy being considered. In the benchmark case and the leaning against the wind case, where macroprudential policy is not present, it is interpreted as a cost associated with banks’ balance sheet management. In these cases, we simply assume that it follows an autoregressive process of order 1:

The second interpretation of crr is that of a regulatory requirement imposed by the macroprudential authority. This interpretation becomes relevant when we analyze the case in which monetary and macroprudential policies coordinate. In this case, crr is defined as follows:

subject to the rest of the economy.

This ratio should be understood as a dynamic-provisioning-type instrument that reduces the procyclicality of the financial system by forcing banks to build a buffer against losses when the financial cycle is booming and allows for a softer landing when it goes bust. Banks are forced to put aside resources to account for the possible losses incurred as credit quality deteriorates. Hence, the adoption of dynamic provisioning typically pursues three objectives: (1) to allow during good times for the buildup of reserves that would serve as a buffer in bad times; (2) to smooth credit growth over the business cycle; and (3) to shield the real economy from shocks originated in the financial sector.

The financial block is hence represented by equations (7.7) to (7.10), where the delinquency rate and the coverage ratio evolve according to equations (7.11) and (7.12) or (7.13), respectively.

Model Estimation

For illustrative purposes, we estimate the proposed model for the Mexican economy. The data set used in the estimation includes core and noncore inflation, the policy rate, the output gap, the real exchange rate, and the nominal exchange rate. The macroeconomic block includes the U.S. output gap, inflation, and policy rate. For the financial block: the nominal implicit loan rate for new loans, a delinquency index for aggregate credit, the average cost of bank term-deposits, financing to the nonfinancial private sector, and the measure of the coverage ratio. All data are at quarterly frequency.17 We estimate equations (7.4) to (7.12) using the generalized method of moments and ordinary least squares. Our sample ranges from the first quarter of 2001 to the fourth quarter of 2014. All coefficients are summarized in Table 7.1.

Table 7.1Estimation Results
EquationCoefficienti = 1i = 2i = 3i = 4i = 5i = 6Adj. R2J-stat.
(7.4) Phillips curvea10.621**0.019**0.005*0.249**0.700.53
(7.5) Investment-savings equationbi0.508**0.136**0.198**1.823**0.184**;0.046**0.900.99
(7.6) Real exchange rateci0.308**0.674**0.790.66
(7.7) Active rateβi0.625**0.043**0.114*0.530.99
(7.8) Passive rateγi0.778**0.017**0.364**0.950.92
(7.10) Credit gapδi0.875**0.708**0.350.91
(7.11) Delinquency indexηi0.750**0.109**0.052**0.96
Source: Authors’ estimates.Note: Equations (7.4) to (7.8) are estimated by the generalized method of moments. Equations (7.10) and (7.11) are estimated by ordinary least squares. As mentioned previously, it is difficult to pin down a precise estimate for δ3. Therefore, we calibrate it so that the credit spread influences the credit-to-GDP gap by as much as half of its influence on the output gap. ** and * represent the 5 and 10 percent levels of significance, respectively.
Source: Authors’ estimates.Note: Equations (7.4) to (7.8) are estimated by the generalized method of moments. Equations (7.10) and (7.11) are estimated by ordinary least squares. As mentioned previously, it is difficult to pin down a precise estimate for δ3. Therefore, we calibrate it so that the credit spread influences the credit-to-GDP gap by as much as half of its influence on the output gap. ** and * represent the 5 and 10 percent levels of significance, respectively.

Results

Trade-offs between inflation-targeting and financial stability objectives arise when growing financial risks warrant a higher interest rate than the one necessary to tame inflation. In contrast, no trade-offs arise when higher interest rates due to an expansion of economic activity are also sufficient to contain financial stability risks. This section examines the trade-offs between policy objectives that result from the policy arrangements described in the previous section using our estimated model for the Mexican economy. To do so, we examine the model’s transmission mechanism and policy implications under each case. In particular, we assess the cases’ relative effectiveness in attaining their policy objectives (inflation targeting in the benchmark case and inflation targeting and financial stability in the leaning-against-the-wind and coordination cases) by comparing the volatility that each of them entails for a set of relevant macroeconomic and financial variables.18 We conduct this analysis for two economic environments: one in which the economy is only affected by macroeconomic shocks, and another in which the economy is only affected by financial shocks.19

The general procedure used consists in feeding the model with 1,000 draws of shocks and simulating it 10,000 times to obtain the model’s invariant distribution. The variances of the relevant variables and the values of the loss functions are then compared across cases and economic environments. To derive optimal policy rules for each case, we need to choose specific weights for the “preference parameters” in the loss functions (i.e., to assign values for σi, for all i in equations (7.1), (7.2), and (7.3)).20 Unlike micro-founded models that assign welfare-based weights, our choice is arbitrary. In particular, we set all weights equal to 1 so that, in principle, the results are not distorted by a given preference for stabilizing one particular variable over another.

Macroeconomic Shocks

Table 7.2 shows our results when the economy is affected only by macroeconomic shocks. A monetary policy response that leans against the wind is effective in reducing fluctuations in financial variables, and even in output, relative to the case where monetary policy has traditional objectives (labeled “Benchmark” in the table). Intuitively, in the former case, monetary policy recognizes the effect that movements in the policy rate entail for the financial sector and internalizes this when setting the rate. As a result, the volatility of the policy rate is also reduced. However, this improvement does not come without a cost. A less reactive interest rate also entails somewhat larger inflation fluctuations. In particular, for the estimated version of our model and the specific “preference parameters” that we consider, the differences between the model dynamics in the two cases are relatively small: output and financial variables are between 8 and 6 percent less volatile, while inflation volatility increases by only 3 percent.

Table 7.2Simulation Results: Macro Shocks
BenchmarkLeans against the WindMonetary and Macroprudential CoordinationRelative Gains
(a)(b)(c)(a)/(b)(b)/(c)
Lπ0.2010.2070.2070.9720.999
Lx4.7704.3914.3131.0861.018
LΔi54.30050.07950.0071.0841.001
Lspread2.7132.5782.0971.0531.229
Lcrgap20.65719.35517.9481.0671.078
LΔcrr0.015
Source: Authors’ estimates.
Source: Authors’ estimates.

When comparing the model dynamics where monetary and macroprudential policies coordinate (labeled “Monetary and Macroprudential Coordination”) vis-à-vis those where monetary policy leans against the wind, we observe a further reduction in the volatility of output and of financial variables but no relative gains in terms of inflation volatility. Smets (2014) points out that incorporating financial considerations into the central bank’s objectives may lead to an inflationary bias problem, since monetary policy is kept looser than necessary for price stability. As will be discussed when we perform our sensitivity analysis, this lack of improvement can be related to the relative ability of both the monetary and the macroprudential policy instruments to affect the financial sector’s spread in a significant way.

Figures 7.2 and 7.3 depict the model’s response to a positive demand shock and a positive cost-push shock, respectively. In the first case, inflationary pressures call for a restrictive monetary policy reaction. The policy rate increases to raise the real interest rate and moderate the increased economic activity. In the second case, inflation expectations increase mechanically due to the increase in inflation, which calls for a mild increase in the policy rate. Although these responses are the same in all cases, notice that the increase in the policy rate in the benchmark case slightly exceeds that in the other two. This confirms the intuition described above. When internalizing the effect of movements in the policy rate on the stabilization of the financial sector, monetary policy’s responses are toned down. This policy reaction comes along with a tepid increase in the coverage ratio—the macroprudential instrument—in the coordination case. Notice, however, that the difference in the response of all variables across cases is negligible, as expected from the magnitudes of the changes in volatilities previously described.

Figure 7.2.Impulse Response to a One Standard Deviation Demand Shock

(Percentage points from steady state)

Source: Authors’ estimates.

Figure 7.3.Impulse Response to a One Standard Deviation Push Shock

(Percentage points from steady state)

Source: Authors’ estimates.

Financial Shocks

Table 7.3 shows our results when the economy is only affected by financial shocks. In particular, we examine two shocks: (1) a shock to credit spreads due to changes in the loan rate, which can be interpreted as a reduction in the risk premium, and (2) a shock to the credit-to-GDP gap, which will allow us to analyze the policy response to an exogenous buildup of financial imbalances.

Table 7.3Simulation Results: Financial Shocks
BenchmarkLeans against the WindMonetary and Macroprudential CoordinationRelative Gains
(a)(b)(c)(a)/(b)(b)/(c)
Lπ0.0010.0020.0020.7131.108
Lx0.4080.4480.3990.9101.122
LΔi10.29010.2449.7451.0041.051
Lspread2.2022.1301.6991.0341.254
Lcrgap5.2235.5914.1810.9341.337
LΔcrr0.011
Source: Authors’ estimates.
Source: Authors’ estimates.

A monetary policy that leans against the wind implies larger volatility for both inflation and output relative to one with only traditional objectives. The reason why output volatility increases when the economy faces financial shocks is that, in this environment, the policy rate aims at stabilizing the financial sector not only through its effect on economic activity but also through its effect on the credit channel (i.e., the credit spread). Inflation volatility is affected in a twofold manner. First, when monetary policy has an additional objective, the room for maneuver to temper the volatility of more variables declines, and this in turn affects inflation expectations. Second, larger output volatility necessarily translates into more inflation volatility. Additionally, this increase in output volatility translates into larger fluctuations in the credit-to-GDP gap, which leaves the monetary policy instrument at odds when trying to stabilize both objectives. In particular, for the case that we analyze, inflation and output volatility increase by 38 and 9 percent, respectively, while the credit-to-GDP gap increases by nearly 7 percent. The credit spread is less volatile by about 3 percent.

To further understand these responses, we examine the model dynamics when the economy is affected by a reduction in the risk premium that decreases the loan rate, which fuels economic activity through lower credit spreads (Figure 7.4). In the benchmark case, the responses of variables are straightforward. A decrease in credit spreads increases output and inflation. The improvement in economic activity initially translates into a lower delinquency index (i.e., a decrease in nonperforming loans), but the higher credit-to-GDP gap eventually offsets this effect, leading to a higher delinquency index. Monetary policy response is countercyclical in order to moderate economic activity and inflation. When monetary policy leans against the wind, the policy rate is set so as to contain the initial decrease in the credit spread through the credit channel. Since in our baseline estimation monetary policy has a negative net effect on credit spreads, the policy rate is set in a procyclical manner, which implies a reduction in the rate. This response amplifies the mechanism described in the benchmark case and leads to even higher levels of the credit-to-GDP gap and the delinquency index.

Figure 7.4Impulse Response to a One Standard Deviation Spread Shock

(Percentage points from steady state)

Source: Authors’ estimates.

Figure 7.5 illustrates the model’s response to an increase in the credit-to-GDP gap. In the benchmark case, this shock implies higher levels for the delinquency index, which leads to an increase in credit spreads due to an upward adjustment in the loan rate. The policy reaction is merely to tone down output and inflation. On the other hand, a monetary policy that leans against the wind counteracts the effect on the credit-to-GDP gap through lower economic activity, which implies a higher level for the policy interest rate and thus a decrease in inflation.

Figure 7.5Impulse Response to a One Standard Deviation Credit-to-GDP Gap Shock

(Percentage points from steady state)

Source: Authors’ estimates.

After examining monetary policy responses to financial shocks, one can easily understand why macroeconomic variables become more volatile when monetary policy has to lean against the wind. In our model, the policy rate goes above and beyond what would be required in the absence of financial stability objectives, in order to affect credit and the credit spreads. These dynamics suggest that the increase in the volatility of macroeconomic variables in our simulation exercise depends on the relative strength of the credit channel, that is, the effect of the policy rate on credit spreads, and on the feedback of output fluctuations into the financial sector. We will return to this point in the next section.

Table 7.3 also shows the results of the model’s dynamics when both monetary and macroprudential policies coordinate. Coordination of the two instruments leads to a subtle improvement in the effectiveness of monetary and macroprudential policies to attain price and financial stability relative to the case in which monetary policy leans against the wind. These results are illustrated in Figures 7.4 and 7.5 as well, where it can be seen that the introduction of a macroprudential policy tool (crr) entails more favorable dynamics for financial variables under the shocks described above and hence less distortions in traditional objectives.

To sum up, this chapter has presented several main findings. First, adding a financial stability objective to monetary policy reduces the volatility of some financial variables. However, this improvement with respect to financial stability may come at the expense of increasing the volatility of macroeconomic variables. Second, this cost is present regardless of the type of shocks that affect the economy, but tends to increase when fluctuations are driven by financial shocks. This is probably what critics of monetary policy leaning against the wind had in mind prior to the 2008–09 financial crisis when they argued that the costs of such a strategy would likely be high (see Mishkin 2011 and references therein). Third, regardless of the type of shocks hitting the economy, we observe modest gains from coordination between monetary and macroprudential policies to achieve macroeconomic stability relative to having monetary policy leaning against the wind in our estimated model for the Mexican economy. Moreover, as can be seen in Tables 7.2 and 7.3, despite the observed reductions in volatility of most variables, inflation volatility remains high relative to the benchmark case.21 The next section will shed light on the properties of the model driving this last result.

Sensitivity Analysis

The results presented above are particular to the estimated model for the Mexican economy and hence cannot be generalized (i.e., it may be the case that country-specific aspects of the transmission of shocks in the Mexican economy are the main drivers of our results). In what follows, we assess the robustness of our findings. We mainly focus on identifying the reasons behind the lack of significant improvement in macroeconomic performance under coordination of monetary and macroprudential policies relative to monetary policy leaning against the wind. Specifically, we explore the importance of the ability of monetary and macroprudential instruments to affect financial variables through changes in credit spreads.

Ability of Monetary and Macroprudential Instruments to Affect Financial Variables

There is no doubt that the effectiveness of any policy strategy to attain its objective(s) should rest on the ability of the corresponding instrument to affect the economy in the desired way. In our model, this ability is given by the magnitude of the pass-through of the policy rate to loan and deposit rates—and thus its overall effect on credit spreads—and by the effect of the coverage ratio on the loan rate.

Let us first analyze the sensibility of our results to monetary policy’s effectiveness to affect credit spreads. To do so, recall equations (7.7) and (7.8) that account for the effect of the policy rate on credit spreads. By substituting (7.7) and (7.8) into (7.9) we obtain the following:

with β1 denoting the pass-through of the policy rate to the loan rate, γ1 to the deposit rate, and (β1 – γ1) its “net effect.”

Table 7.4 illustrates the losses in terms of volatility of our baseline model relative to others with different values for this “net effect” when the economy is affected by financial shocks. First, notice that an economy is better off with a net effect equal to 0, when the loan and deposit rates are equally affected by movements in the policy rate. In such a case, changes in credit spreads are not directly obtained through movements in loan and/or deposit rates but rather through the effect that the policy rate has on other macroeconomic variables such as the output gap. For positive values of the net effect—which can be associated with a countercyclical effect of monetary policy on the financial sector—losses begin to increase, but not by as much as for negative values of the net effect, where losses can more than double.

Table 7.4Simulation Results: Monetary Policy, Net Effect
Negative Net EffectPositive Net EffectZero Net Effect
Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b)/(c)Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b)/(c)Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b)/(c)
Lπ0.0210.0161.3700.0150.0111.3170.0010.0011.086
Lx1.3411.1151.2020.9400.8161.1520.3910.3491.122
LΔi11.37610.1721.1186.8616.5701.0449.4029.0241.042
Lspread15.6839.8501.59210.4097.6081.3680.7350.7141.029
icrgap46.50727.1261.71415.2149.5651.5911.3601.1981.136
LΔcrr0.1710.0820.001
Source: Authors’ estimates.
Source: Authors’ estimates.

We now focus on the case where the impact of macroprudential effectiveness on credit spreads is changed—that is, when we vary β3 in equation (7.7). Table 7.5 illustrates the losses in terms of volatility of our baseline model relative to others with different values for this parameter, when the economy is affected by financial shocks. The larger the impact of macroprudential policy on credit spreads, the greater the gains of coordination versus leaning against the wind. Results are straightforward, as in this framework the transmission mechanism of macroprudential policy is quite simple. Figure 7.6 summarizes the results presented above and shows that our framework features room for gains from coordination between monetary and macroprudential policies, either when monetary policy is not an effective tool for the financial sector or when it enters into clear conflict with macroprudential policy. This result emphasizes the importance of assessing the relative effectiveness of the different tools that policymakers have at their disposal to achieve policy objectives, as well as the interactions among these tools. The room for improvement is larger when the shocks affecting the economy come from the financial sector.

Table 7.5Simulation Results: Macroprudential Policy, β3
β3 = 1β3 = 0.5β3 = 0.8
Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b/c)Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b/c)Leans against the Wind (b)Monetary and Macroprudential Coordination (c)Relative Gains (b/c)
Lπ0.0020.0021.0990.0020.0021.2080.0020.0021.287
Lx0.4690.4191.1190.4690.4101.1440.4690.4041.162
LΔi10.3579.8751.04910.3579.6551.07310.3579.5261.087
Lspread2.6472.2251.1902.6471.4971.7682.6471.1672.268
Lcrgap6.9495.5151.2606.9493.7481.8546.9493.0222.300
LΔcrr0.0100.0290.034
Source: Authors’ estimates.
Source: Authors’ estimates.

Figure 7.6.Gains from Policy Instruments, LLAW/LCoord (Financial Shocks)

Source: Authors’ estimates.

Conclusions

Despite the simplicity of the framework used in this chapter, our findings highlight some important aspects regarding how the interlinkages between the real and financial sectors determine the appropriate policy responses to shocks. In particular, our results suggest that the trade-offs to attain price and financial stability objectives increase in the presence of financial shocks.

The decision whether to use monetary policy as the main tool to counter the financial cycle by leaning against the wind versus using coordinated monetary and macroprudential policies to do so is ultimately a decision between having one versus two instruments to attain two different but inherently interconnected policy objectives. The fact that under a given situation (i.e., under the presence of financial shocks) there is a trade-off between inflation-targeting and financial stability objectives opens up the possibility to improve macroeconomic performance with coordinated monetary and macroprudential policies. The main result of this chapter is that the magnitude of these benefits depends on a wide range of determinants such as the effectiveness of policy instruments to affect credit spreads in the economy.

This result highlights the importance of further research on at least two fronts. The first is the price-setting behavior of banks. In particular, understanding the main determinants of profit margins in the banking system is of vital importance. The second is how the characteristics of the financial system shape the amplification and persistence of financial shocks to the economy. Within this task lies the need to understand the effect of having a particular degree of financial penetration in the economy, and detecting and assessing the importance of the financial frictions that account for the transmission of financial shocks.

A deeper understanding of these topics will lead to developing larger, more complex, and more wide-ranging country-specific macroeconomic models that allow for capturing the main features of the economies which policymakers attempt to influence.

References

    AndrésJ. and O.Arce.2012. “Banking Competition, Housing Prices and Macroeconomic Stability.Economic Journal122(565): 134672.

    AngeliniP.S.Neri and F.Panetta.2014. “The Interaction between Capital Requirements and Monetary Policy.Journal of Money Credit and Banking46(6): 1073112.

    AngeloniI. and E.Faia.2013. “Capital Regulation and Monetary Policy with Fragile Banks.Journal of Monetary Economics60(3): 31124.

    BeauD.L.Clerc and B.Mojon.2012. “Macro-prudential Policy and the Conduct of Monetary Policy.Working Paper 390Banque de France, Paris.

    BernankeB. and M.Gertler.1986. “Agency Costs, Collateral, and Business Fluctuations.NBER Working Paper 2015National Bureau of Economic ResearchCambridge, Massachusetts.

    BernankeB. and M.Gertler and S.Gilchrist.1999. “The Financial Accelerator in a Quantitative Business Cycle Framework.” In Handbook of Macroeconomics Volume 1 edited by J. B.Taylor and M.Woodford.Amsterdam: North-Holland.

    BlanchardO.G.Dell’Ariccia and P.Mauro.2010. “Rethinking Macroeconomic Policy.IMF Staff Position Note 2010/03International Monetary FundWashington, DC.

    BlanchardO.G.Dell’Ariccia and P.Mauro.2013. “Rethinking Macro Policy II; Getting Granular.IMF Staff Discussion Note 13/003International Monetary FundWashington, DC.

    BorioC.2011. “Rediscovering the Macroeconomic Roots of Financial Stability Policy: Journey, Challenges and a Way Forward.BIS Working Paper 354Bank for International SettlementsBasel.

    BorioC.2014. “Monetary Policy and Financial Stability: What Role in Prevention and Recovery?BIS Working Paper 440Bank for International SettlementsBasel.

    BorioC. and P.Lowe.2002. “Asset Prices, Financial and Monetary Stability: Exploring the Nexus.BIS Working Paper 114Bank for International SettlementsBasel.

    BorioC.W.English and A.Filardo.2003. “A Tale of Two Perspectives: Old or New Challenges for Monetary Policy?BIS Working Paper 127Bank for International SettlementsBasel.

    CecchettiS.H.GenbergJ.Lipsky and S.Wadhwani.2000. “Asset Prices and Central Bank Policy.Geneva Reports on the World Economy No. 2 (July). Centre for Economic Policy ResearchLondon.

    CúrdiaV. and M.Woodford.2009. “Credit Spreads and Monetary Policy.NBER Working Paper 15289National Bureau of Economic ResearchCambridge, Massachusetts.

    CúrdiaV. and M.Woodford.2011. “The Central-Bank Balance Sheet as an Instrument of Monetary Policy.Journal of Monetary Economics58(1): 5479.

    De PaoliB. and M.Paustian.2013. “Coordinating Monetary and Macro-prudential Policies.Staff Report 653Federal Reserve Bank of New York.

    Dell’AricciaG.D.IganL.Laeven and H.Tong.2014. “Policies for Macro-financial Stability: Dealing with Credit Booms and Busts.” In Financial Crises: Causes Consequences and Policy Responses edited by S.ClaessensA.KoseL.Laeven and F.Valencia.Washington, DC: International Monetary Fund.

    DrehmannM. and M.Juselius.2013. “Evaluating Early Warning Indicators of Banking Crises: Satisfying Policy Requirements.BIS Working Paper 421Bank for International SettlementsBasel.

    GalatiG. and R.Moessner.2013. “Macroprudential Policy: A Literature Review.Journal of Economic Surveys27(5): 84678.

    GeraliA.S.NeriL.Sessa and F. M.Signoretti.2010. “Credit and Banking in a DSGE Model of the Euro Area.Journal of Money Credit and Banking42(s1): 10741.

    GertlerM. and P.Karadi.2011. “A Model of Unconventional Monetary Policy.Journal of Monetary Economics58(1): 1734.

    IacovielloM.2005. “House Prices, Borrowing Constraints, and Monetary Policy in the Business Cycle.American Economic Review95(3): 73964.

    KannanP.P.Rabanal and A. M.Scott.2012. “Monetary and Macroprudential Policy Rules in a Model with House Price Booms.B.E. Journal of Macroeconomics12(1): 144.

    KiyotakiN. and J.Moore.1997. “Credit Cycles.Journal of Political Economy105(2): 21148.

    LambertiniL.C.Mendicino and M. T.Punzi.2013. “Leaning against Boom–Bust Cycles in Credit and Housing Prices.Journal of Economic Dynamics and Control37(8): 150022.

    MehC. A. and K.Moran.2010. “The Role of Bank Capital in the Propagation of Shocks.Journal of Economic Dynamics and Control34(3): 55576.

    MishkinF. S.2011. “Monetary Policy Strategy: Lessons from the Crisis.NBER Working PaperNational Bureau of Economic ResearchCambridge, Massachusetts.

    Roldán-PeñaJ.D.Sámano and A.Torres.2014. “Monetary and Macro Prudential Policies: Interaction and Complementarity.Banco de MéxicoMexico City. Unpublished.

    SámanoD.2011. “In the Quest for Macroprudential Policy Tools.Working Paper 2011-17Banco de MéxicoMexico City.

    SmetsF.2014. “Financial Stability and Monetary Policy: How Closely Interlinked?International Journal of Central Banking10(2): 263300.

    SöderlindP.1999. “Solution and Estimation of RE Macro Models with Optimal Policy.European Economic Review43(4–6): 81323.

    SvenssonL. E. O.2014. “Inflation Targeting and Leaning against the Wind.International Journal of Central Banking10(2): 10314.

    VlcekJ. and R.Scott.2012. “Macrofinancial Modeling at Central Banks; Recent Developments and Future Directions.IMF Working Paper 12/21International Monetary FundWashington, DC.

    WilliamsJ. C.2014. “Financial Stability and Monetary Policy: Happy Marriage or Untenable Union?FRBSF Economic Letter 2014-17 (June9).

    WoodfordM.2003. Interest and Prices: Foundations of a Theory of Monetary Policy.Princeton, New Jersey: Princeton University Press.

    WoodfordM.2012. “Inflation-Targeting and Financial Stability.NBER Working Paper 17967National Bureau of Economic ResearchCambridge, Massachusetts.

This chapter is related to previous work done by Roldán-Peña, Sámano, and Torres (2014). We are grateful for Ana Aguilar’s support throughout the completion of this project. We thank Giovanni Dell’Ariccia and participants of the Midpoint Seminar at the IMF and the 2015 LACEA Meetings for their comments. We also thank María Diego and Carlos Zarazúa for helpful suggestions. Finally, we thank Cid Rodriguez for providing excellent econometric assistance. The views presented here are exclusively of the authors and do not necessarily reflect those of Banco de México or Secretaría de Hacienda y Crédito Público.

Some notable exceptions, however, can be found in emerging market economies that experienced financial crises during the 1980s and 1990s.

See Blanchard, Dell’Ariccia, and Mauro (2010, 2013), Borio (2011), and Mishkin (2011).

The question on whether monetary policy should lean against the wind is not new. The possible dangers associated with asset-price bubbles had already been pointed out before the crisis by economists such as Cecchetti and others (2000), Borio and Lowe (2002) and Borio, English, and Filardo (2003). They argued that central banks should at times lean against the financial wind by raising interest rates to stop bubbles from getting out of hand. In their view, raising interest rates to slow a bubble’s growth would produce better outcomes because it would either prevent the bubble or result in a less severe bursting of it, with far less damage to the economy. A contrasting dominant view argued that monetary policy should not try to lean against asset-price bubbles, but rather just clean up after they burst, since monetary policy was, at best, ineffective in dealing with such bubbles (Mishkin 2011).

Woodford (2012) argues that it is possible to incorporate financial stability considerations into a model in a way that it represents a natural extension of “flexible inflation targeting.” We associate the leaning-against-the wind case examined in this chapter with his definition of flexible inflation targeting.

See Williams (2014). For Borio (2014), attaining monetary and financial stability simultaneously calls for monetary policy that leans more deliberately against booms and eases less aggressively during busts, within a context where all macroeconomic policies play a mutually supportive role. In this respect, Svensson (2014) argues that monetary and macroprudential instruments and policies vary greatly from country to country, such that each economy must be scrutinized before judging whether there is a case for leaning against the wind.

More details about the coverage ratio and the rationale for its role as a macroprudential policy instrument are provided in the next section of this chapter.

Woodford (2003) shows that a loss function determined by inflation deviations from its target and the output gap can be justified as a quadratic approximation to the goal of maximizing the representative agent’s utility in a simple version of a New Keynesian model. As estimated central bank reaction functions also incorporate some degree of partial-adjustment dynamics of the interest rate itself, the last term in equation (7.1) ensures that movements in the policy rate are not excessively volatile.

In a micro-founded model, σπ and σx are functions of deep parameters of the economy and reflect the degree by which distortions affect economic welfare. A framework characterized by the use of a semistructural model, like ours, is not able to trace down the effects of deep parameters on economic welfare. Instead we assume that these weights characterize policymakers’ preferences, which ultimately should be associated with the deep parameters of the economy. Angelini, Neri, and Panetta (2014), who analyze the interaction between monetary and macroprudential policies, use loss functions like the ones described in this work.

Cúrdia and Woodford (2009) show that a loss function with the usual inflation and output gap stabilization goals and other terms that represent the welfare consequences implied by financial frictions and/or the loss in resources incurred by the financial intermediary sector, can be justified as a quadratic approximation to the goal of maximizing the average expected utility of households.

Angelini, Neri, and Panetta (2014) propose a loss function in the same spirit as the one proposed here when studying the interaction between monetary and macroprudential policies in the case of cooperation. The last term in equation (7.3) represents policymakers’ concern with the variability of the macroprudential policy instrument.

Vlcek and Scott (2012) provide an extensive survey of models featuring financial frictions and/or financial intermediation in use by central banks.

In this respect, Galati and Moessner (2013, 854) point out that “[w]hile the literature on monetary policy has provided a common conceptual framework over the past two or three decades, research on macroprudential policy is still in its infancy and appears far from being able to provide an analytical underpinning for policy frameworks. . . . [This may be due to, among other reasons, the fact that] we lack a thorough understanding and established models of the interaction between the financial system and the macroeconomy.”

One can further consider that firms’ costs, and thus their price-setting behavior, are affected by credit market conditions. We leave this consideration for future work.

Alternatively, credit volumes or lending standards could have been used instead of the credit spread. However, according to Sámano (2011), preliminary evidence for the case of Mexico points out that credit volumes do not Granger-cause the output gap. Furthermore, for the case of Mexico, lending standards are available for a very short period, making it highly imprecise to perform statistical inference with the latter variable.

According to Dell’Ariccia and others (2014), who analyze the characteristics of financial booms that end up in busts or crises, credit growth can be a powerful predictor of financial crises.

Drehmann and Juselius (2013) find that the credit-to-GDP gap is a good early warning indicator of banking crises.

Inflation is calculated as the quarter-over-quarter percent change of the corresponding index. The nominal interest rate is the short-term interbank funding rate. The output gap is estimated with a tail-corrected Hodrick-Prescott filter over real GDP. For the real exchange rate, we consider a bilateral index between Mexico and the United States constructed by Banco de México; the change in nominal exchange rate is thus the quarter-over-quarter percent change in the bilateral nominal rate. The lending spread corresponds to the difference between the implicit rate and the average cost of bank term deposits. The delinquency index is an adjusted index constructed by Banco de México, consisting of the sum of overdue loans and loans written off in the prior 12 months divided by total loans plus loans written off in the last 12 months. For the credit-to-GDP gap we filter the ratio of total financing from commercial banks to the nonfinancial private sector over GDP, using a Hodrick-Prescott filter with a smoothing parameter of 400,000. Finally, the variable capturing the coverage ratio rule (crr) is the ratio of loan-loss reserves to nonperforming loans for the Mexican banking system. Notice that to account for the crr we consider a realized measure of the ratio of loan-loss reserves to nonperforming loans as opposed to a legal actual requirement (the Mexican banking regulation does not include a dynamic provisioning requirement; instead it considers an expected loss approach consistent with the Basel III framework). The source of all data is Banco de México, Comisión Nacional Bancaria y de Valores, and Instituto Nacional de Estadística y Geografía.

For all cases presented below, each of the terms in the loss function is weighed out by the inverse of the variance of the corresponding historical series from 2001 to 2014.

Evidently, the overall effectiveness of each particular policy strategy is a combination of its effectiveness in both environments and depends on the relative importance of macroeconomic versus financial shocks in the economy. However, separating the studied effects by type of shock makes it easier to understand the sources of the results.

Optimal policy rules are calculated following Söderlind (1999).

According to these findings, and given that the size of gains from coordination for other variables is relatively small, the risk that financial stability considerations undermine the credibility of the central bank’s price stability mandate might eventually generate further welfare losses.

    Other Resources Citing This Publication