Chapter 9 Macroeconomic Policy Synergies for Sustained Growth
- Ana Corbacho, and Shanaka Peiris
- Published Date:
- October 2018
Managing boom-and-bust cycles in the presence of global spillovers remains a key policy challenge for Association of Southeast Asian Nations–5 (ASEAN-5) authorities. Recessions that follow a bust can lead to both temporary and permanent output losses. Moreover, these losses are magnified in the presence of financial downturns. At the same time, setting the base for sustained growth not only increases economic opportunities in the longer term but also contributes to stability in the short term by reducing the burden from past debt accumulation and other financial risks.
Countercyclical monetary policy can play a key role in managing boom-and-bust cycles, but on its own its effectiveness can sometimes be diminished. For example, in low inflation–low natural rate environments, the zero lower bound on nominal interest rates can become binding, limiting monetary policy’s ability to deliver on inflation targets. In addition, lack of synchrony between financial and real cycles can make it difficult to manage both cycles with a single instrument (the policy rate). Moreover, setting the basis for sustained growth frequently requires appropriately tailored structural policies that stimulate the main drivers of potential growth over the medium term. In this context, exploiting synergies between monetary, macroprudential, and fiscal policies can bring powerful benefits to managing business cycle fluctuations, preserving financial stability, and sustaining long-term growth.
Guarding both macroeconomic and financial stability usually requires gearing different policy tools toward different objectives: First, macro policies that are effective in managing the business cycle may not be as effective in managing the financial cycle. Second, the business and the financial cycles typically have different lengths and amplitudes, as documented in Chapter 6. Countries may find themselves experiencing a financial boom (bust) during a business cycle bust (boom). As a result, managing the business and financial cycles may require careful calibration of macro-stabilization policies along with macroprudential policies.
Although monetary policy is tasked directly with price and output stability objectives, fiscal policy can also play a key role. Fiscal policy may be an effective tool for managing business cycles and supporting aggregate demand during a recession. Moreover, fiscal stimulus in the form of productive spending on physical and human capital can also support potential growth over the medium to long term, helping overcome some of the possible permanent losses following a recession. Well-designed fiscal policy is important for short-term growth as well as for long-term productivity.
This chapter elaborates on macroeconomic policy synergies for managing economic cycles and sustaining growth. In particular, it analyzes the scope for macroprudential and fiscal policy to aid in overcoming some of the challenges faced by monetary policy in ASEAN-5 economies in the current global environment. The chapter first documents the output costs of boom-and-bust cycles in ASEAN-5 countries. It then presents model-based simulations of the possible synergies between monetary policy and macroprudential policies in smoothing business and financial cycles. It continues with a focus on the interactions between monetary and fiscal policy as stabilization tools. The final section offers concluding remarks.
Recessions and Output Losses in ASEAN-5 Countries
Crises and even recessions can lead to permanent output losses, according to the literature, especially when they accompany financial instability. Using a panel of 190 countries, Cerra and Saxena (2008) document the persistence of large output losses associated with financial (as well as political) crises. In addition, Cerra and Saxena (2017) find that all types of recessions, on average, lead to permanent output losses. These results contrast with the more traditional view that recessions have only a temporary impact on output, based on the view that output will return to its trend once full employment of resources is reestablished.
In line with the results from Cerra and Saxena (2008), novel estimates for ASEAN-5 economies presented in this chapter also suggest that, following a recession, growth has tended to rebound, but not strongly enough to fully restore output to its precrisis trend.1Figure 9.1 shows the impulse response of a “typical” ASEAN growth recovery following a recession. The results show that after about 6 months, growth bounces back, with positive growth for about nine quarters following an economic downturn. However, the response of real GDP growth implies that the growth bounce-back is not directly proportional to the output loss. Figure 9.2 illustrates the effect of the typical recession on real GDP. The estimates imply that recessions have tended, at least over the sample period in question (1996–2016), to have had hysteresis effects on the level of real GDP.2
Figure 9.1.Typical Response of ASEAN Real GDP Growth to a Recession, January 1996–January 2017
Source: IMF staff calculations.
Figure 9.2.Response of ASEAN-Level Real GDP
Source: IMF staff calculations.
These findings reinforce the need for countercyclical fiscal and monetary policies to mitigate economic downturns or prevent them from having long-term adverse effects on economic growth.
Synergies Between Monetary and Macroprudential Policies
The global financial crisis brought into greater focus both the need for macroeconomic policy to include financial stability among its objectives and the debate on whether financial stability should be a mandate of monetary policy. Although the crisis reinforced the importance of well-anchored inflation expectations and long-term price stability, it also challenged the notion that price stability is sufficient for macroeconomic stability. Confronted with a severe financial and economic crisis, central banks and researchers have been called to reconsider the role of financial sector and asset price imbalances in the design and implementation of monetary policy.3
Figure 9.3 illustrates the trade-offs involved when monetary policy is used for financial stability purposes and the circumstances that alleviate or augment them. Traditional monetary policy focuses on targeting inflation and output. In the absence of growing financial risks, this focus allows the authorities to pursue inflation and output targets in the medium term with no financial stability concerns. However, such policy could also lead to growing financial risk and expected losses from a crisis in the medium term, in which case a monetary policy that also targets financial stability by “leaning against the wind” could be beneficial. This result will depend on the effectiveness of the interest rate in lowering financial risks and the expected losses from a crisis in the medium term. This benefit would need to be compared with the cost from short-term deviations from inflation and output targets and the corresponding welfare losses over the forecast period.
Absent other tools, results from mainly closed-economy New Keynesian dynamic stochastic general equilibrium models support the case for leaning against the wind (Curdia and Woodford 2009; Woodford 2012; Ajello and others 2016). However, the implied deviations from more standard inflation-output gap decision rules are quantitatively small in these linear models (Gambacorta and Signoretti 2014; Filardo and Rungcharoenkitkul 2016). Moreover, the case for leaning against the wind is even weaker in small open economies, where the impact of such policy on international capital flows may exacerbate macroeconomic and financial stability concerns (Sahay and others 2014; Menna and Tobal 2017).
Macroprudential policy tools can help alleviate tensions between monetary and financial stability objectives.4 First, divergences between monetary and financial stability mandates stem, to a large extent, from attempts to achieve two different objectives with a single policy instrument, the monetary policy rate. This is particularly relevant given that, as shown in Chapter 6, the real and financial cycles are not always synchronized. Using macroprudential tools to curb systemic risks may ease those tensions. Ghilardi and Peiris (2016), for example, show that monetary and macroprudential policy coordination can enhance the policy effectiveness in attenuating the real and financial cycles. Second, macroprudential policies can also be more effective than interest rate policy in dampening the financial cycle at a lower cost to output. Moreover, financial stability risks come in all shapes and forms. And while macroprudential tools can be customized to address specific risks, movements in the monetary policy rate would have a macroeconomic impact.
For the ASEAN-5, empirical estimates prepared in this chapter show that macroprudential policies have been effective in taming the financial cycle without significantly affecting the real business cycle. Using a Bayesian vector autoregression, Figure 9.4 (and Annex 9.2) presents the response of credit and output growth to a shock in the macroprudential policy stance in the ASEAN-5 economies. Results suggest that for several countries credit growth is significantly more sensitive to macroprudential policies than is real GDP growth, especially in Malaysia, the Philippines, and Thailand over the medium term.
Figure 9.4.Median Response of Credit and Real GDP Growth to a Tightening in Macroprudential Policy
Source: IMF staff calculations.
Nevertheless, macroprudential tools remain relatively new and untested, and their implementation faces challenges. Macroprudential tools are prone to circumvention and political economy problems (IMF 2012) and may be difficult to adjust depending on institutional settings. Also, financial stability has multiple dimensions and many potential policy indicators and targets. Moreover, bubbles and the imminence of a systemic crisis are difficult to identify in real time, and policy needs to strike a balance between guarding against financial risks and allowing for healthy financial activity (IMF 2014a).5 In sum, the effects of macroprudential policies on financial markets and economic outcomes and the interactions with monetary policy need to be investigated further.
The rest of this section explores how synergies between monetary and macroprudential policies in the ASEAN-5 countries can achieve better macro outcomes. Model-based simulations for the Philippines and Thailand are presented to analyze whether the use of countercyclical macroprudential policies can complement the ability of monetary policy to target inflation while containing financial stability risks in the face of shocks. Then, outcomes for level and volatility of inflation, consumption, investment, and private credit are compared under the different monetary and macroprudential policy reaction functions.
The focus is on excessive leverage in the private sector as the key source of financial instability. Private sector debt has been increasing as a percentage of GDP among ASEAN-5 economies since the global financial crisis (Figure 9.5), and increasing evidence indicates that the buildup of private sector debt is harmful for growth. Schularick and Taylor (2012) show that credit growth is a powerful predictor of financial crises, and Jordà, Schularick, and Taylor (2013) show that more-credit-intensive expansions tend to be followed by deeper recessions (in financial crises or otherwise) and slower recoveries. The composition of private sector debt also matters. For example, Mian, Sufi, and Verner (2017) and IMF (2017) show that increases in household debt can accelerate growth in the short term but can then have significant negative effects over the medium term, thereby producing a boom-bust cycle. They also show that corporate debt has a negative impact on growth in the short term but not in the medium to long term.
Figure 9.5.Asia: Private Sector Debt
Sources: Bank for International Settlements; Dealogic; Haver Analytics; national authorities; and IMF staff calculations.
Note: NFCs = nonfinancial corporations.
Regression analysis for the ASEAN-5 economies indicates that private debt accumulation has a negative impact on growth. Table 9.1 shows that after two and three years of private debt accumulation, growth in ASEAN-5 economies tends to decline by about 0.06 percentage point for every 1 percent of GDP increase in debt. These estimates are statistically significant for the second and third year following the debt buildup. Although the impact is significant, it is smaller than the average impact obtained for the rest of countries in the sample (a pool of 42 countries comprising advanced and emerging markets—see Annex 9.3 for more details).
|Variables||(1) GDP Growth over 3 Years||(2) GDP Growth over 3 Years (f + 1)||(3) GDP Growth over 3 Years (f + 2)||(4) GDP Growth over 3 Years (f + 3)||(5) GDP Growth over 3 Years (f + 4)||(6) GDP Growth over 3 Years (f + 5)|
|Nonfinancial Private Debt|
(t – 1)
|ASEAN-5-Specific Nonfinancial Private Debt|
(t – 1)
|Number of Countries||47||47||47||47||47||47|
|Country Fixed Effects||YES||YES||YES||YES||YES||YES|
Table 9.2 disaggregates the impact of household debt from that of nonfinancial corporate private credit. Within the ASEAN-5, the results are based on data for Thailand and Singapore, the only two countries with household and corporate debt data going back to the Asian financial crisis. In line with Mian, Sufi, and Verner (2017), the results indicate that household debt has a positive impact on growth in the short term but a negative impact in the medium term for the pool of countries as a whole and also for Thailand and Singapore alone. With respect to nonfinancial corporate debt, results for Singapore and Thailand indicate a negative impact on growth in the shorter term (one or two years ahead) that is stronger than the average impact for the other countries in the sample.
|Variables||(1) GDP Growth over 3 Years||(2) GDP Growth over 3 Years|
(f + 1)
|(3) GDP Growth over 3 Years|
(f + 2)
|(4) GDP Growth over 3 Years|
(f + 3)
|(5) GDP Growth over 3 Years|
(f + 4)
|(6) GDP Growth over 3 Years|
(f + 5)
|Household Debt (3-year change) (t – 1)||0.0481|
|Thailand- or Singapore-Specific Household Debt (3-year change)||0.457**|
|Corporate Debt (3-year change) (t – 1)||-0.0857***|
|Thailand- or Singapore-Specific Corporate Debt (3-year change) (t – 1)||-0.295***|
|Number of Countries||47||47||47||47||47||45|
|Country Fixed Effects||YES||YES||YES||YES||YES||YES|
The framework of Anand, Delloro, and Peiris (2014) is extended to incorporate a household borrowing, housing and macroprudential policies. It is based on a New Keynesian dynamic stochastic general equilibrium model for small open economies with price rigidities and financial friction. Figure 9.6 depicts the relationships between agents in this economy.
There are two types of households, patient (with a lower intertemporal discount rate) and impatient, which both derive utility from consumption, leisure, and housing. In equilibrium, the patient households save part of their income, which is invested in domestic bank deposits and foreign bonds. Impatient households end up borrowing to consume and purchase houses.
Entrepreneurs borrow from domestic banks and from abroad to purchase capital. They also hire labor and produce goods that are then sold to retailers who subsequently sell to consumers, capital producers, and foreign markets in a monopolistically competitive environment.
Banks can lend to the government, entrepreneurs, or households. Interest rates are sticky because banks face increasing marginal costs associated with changes in interest rates. At the same time, bank borrowing is subject to macroprudential measures.
Figure 9.6.Structure of the DSGE Model
Source: IMF staff.
Note: DSGE = dynamic stochastic general equilibrium.
Government policies are described by two policy reaction functions: one for the monetary policy interest rate that follows a Taylor rule and one for macroprudential policy measures. The simulations look at two different macroprudential tools. In Thailand, with a high household-debt-to-GDP ratio, the focus is on the impact of a countercyclical loan-to-value ceiling that restricts lending to households to a certain proportion of the value of their houses. In the Philippines, where more concern is placed on rapid bank credit growth and rising corporate debt, the focus is on the impact of a countercyclical capital adequacy ratio (or countercyclical capital buffer, an extension of the Basel III capital conservation buffer), which restricts overall bank borrowing.
Simulations for Thailand: Adding a Countercyclical Loan-to-Value Ratio
For Thailand, the model is calibrated to assess the impact of a temporary positive shock on the demand for housing. The starting conditions mirror the current juncture: inflation is below target and there is a small negative output gap, but household debt and house prices are relatively high. In this context, lowering the monetary policy rate to improve the inflation and growth outlook could fuel further imbalances in the housing sector. In turn, further accumulation of household debt could produce a negative feedback loop on growth.
The response of the economy to this shock is examined under two variants of the Taylor rule and two variants of the macroprudential policy rule. The two variants of the Taylor rule are as follows:
1. Standard Taylor rule—focused on inflation and output gaps
in which i is the policy interest rate, inflation gap is the difference between actual inflation and the target, and output gap is the difference between actual and potential output.
2. Modified Taylor rule—focused on inflation, output, and credit gaps
in which credit gap is the difference between the actual stock of household credit and the steady-state level.
The two variants for the macroprudential policy measure are as follows:
1. Constant macroprudential policy measures: The loan-to-value ceiling applied to household credit and the minimum capital adequacy ratio applied to bank credit are kept constant.
2. Countercyclical loan-to-value ratios: The loan-to-value ceiling applied to household credit decreases as the stock of household loans increases relative to the steady-state value.
These variants in policy functions yield four possible scenarios:
1. Standard Taylor rule + constant macroprudential measures
2. Modified Taylor rule + constant macroprudential measure
3. Standard Taylor rule + countercyclical macroprudential measure
4. Modified Taylor rule + countercyclical macroprudential measure
In the simulations the authorities can use monetary policy, macroprudential policy, or both to respond to the impact of the housing demand shock. The policy reaction functions follow the four possible scenarios described above. A first set of results compares outcomes when monetary policy is the only instrument available for tackling both macroeconomic and financial imbalances. Macroprudential policy is passive with a constant loan-to-value ratio that does not react to signals of growing financial imbalance in the housing sector.
In scenario 1 (dark blue line, Figure 9.7), with monetary policy following a standard Taylor rule and a constant loan-to-value ratio, a positive shock to the demand for housing leads to an increase in house prices and in the level for housing loans in the economy. Inflation increases and output falls slightly on impact, but then they converge to the steady state on a cyclical path. The authorities respond with an initial increase in the nominal and real policy rates, then reduce them in subsequent quarters.
In scenario 2 (red line, Figure 9.7), where the authorities are concerned about a buildup in household loans, but use interest rates as the single instrument (modified Taylor rule and constant loan-to-value scenario), the increase in household loans is significantly mitigated, while output and inflation drop. Interestingly, the paths of the nominal interest rates are below those under scenario 1. The threat of an increase in the real rate in response to higher housing loans reduces the equilibrium demand for housing loans and goods. This actually preempts the need for nominal and real rates to be raised above those in scenario 1, except on impact, when the real rate is now higher.
Figure 9.7.Response to Positive Shock on Demand for Housing
Source: IMF staff simulations. Note: LTV = loan to value.
These two scenarios illustrate the trade-off faced by the authorities when trying to target both household debt and inflation using the policy interest rate as the only instrument. When the cycles for household debt and inflation are not synchronized, achieving the inflation target requires letting go of the household debt target. Conversely, moderating the household debt increase requires letting go of the inflation target. Moreover, incorporating household debt concerns in the Taylor rule leads to even lower nominal rates over the medium term, thereby increasing the likelihood of hitting the zero lower bound given the current low-interest-rate environment.
A second set of results looks at the benefits of introducing a countercyclical loan-to-value ratio into the authorities’ toolbox.
In scenario 3 (green line, Figure 9.7), with a standard Taylor rule and a countercyclical loan-to-value ratio, the loan-to-value ceiling is tightened in response to an increase in household debt, and the impact of the housing demand shock on household debt is mitigated. In turn, interest rate policy remains focused on inflation and output, delivering on macroeconomic stability. Some downward adjustment in nominal rates takes place, but much less than in scenario 2.
In scenario 4 (light blue line, Figure 9.7), with a modified Taylor rule and a countercyclical loan-to-value ratio, the imbalance in the housing sector is addressed, but inflation and output are still lower than in scenario 3.
Considering all scenarios together, results suggest that the separation principle holds. Better outcomes in growth, inflation, and financial stability can be achieved with monetary policy focused on its traditional targets of inflation and output and macroprudential policy targeting the specific source of financial instability. Asking monetary policy to do too much (that is, to also target financial stability) comes at the cost of suboptimal inflation and growth. Moreover, in a low inflation–low interest rate environment, it may increase the risk of hitting the zero lower bound.
Critical in these results is that both interest rate and macroprudential policies are effective in achieving their respective targets. In this model, macroprudential policy does not suffer from leakages, and interest rate policy transmission operates smoothly. Also, countercyclical loan-to-value ratios and interest rates under the modified Taylor rule are assumed to react immediately and strongly to dynamics in household loans. Implementation lags or weaker transmission in either macroprudential or monetary policy could potentially affect the conclusion on the size of the gains from the separation principle or even its superiority.6 For example, alternative simulations7 in which the interest rate transmission mechanism is diminished by a flatter Phillips curve (one in which the impact of the output gap on inflation is much lower) yield smaller trade-offs between scenarios 1 and 2.8 As a result, it is still optimal to use macroprudential policy to contain housing credit growth while focusing the policy rate only on inflation and output gaps, but the gains are smaller. Other simulations also show that, in cases in which the ability to adjust macroprudential tools is very limited or the macroprudential tools applied are too blunt and affect credit well beyond vulnerable sectors, the separation principle is sometimes an inferior option.
An Application to the Philippines: A Countercyclical Capital Buffer
In the Philippines, the model is first calibrated to assess the impact of a negative interest rate shock. Historic low real interest rates since the global financial crisis have fueled rapid bank credit growth, raising financial stability risks. Household debt is relatively low compared with corporate debt, and bank credit is concentrated in lending to businesses, particularly large conglomerates (Figure 9.8). External borrowing by businesses has also risen since the global financial crisis and, as a result, the corporate sector in the Philippines is also exposed to global financing conditions. Hence, a key policy challenge is managing rising bank credit and corporate leverage while sustaining robust growth momentum.
Figure 9.8.Private Credit
The simulations focus on the synergies between monetary policy and a countercyclical macroprudential tool, in particular, a countercyclical capital buffer, in response to a decline in interest rates. Such a buffer could potentially help mitigate a credit boom in a very-low-interest-rate environment such as that of the Philippines over the past decade or so. Exploiting synergies between monetary and macroprudential policies may also help insulate the economy from global financial volatility that would raise the external financing premium.
Figure 9.9 presents the response to a negative shock to the domestic interest rate under different macroprudential policy reaction functions. Scenario 1 (blue line) plots the results from a standard Taylor rule for the policy rate (similar to that specified for Thailand) with a constant capital ratio, while scenario 2 (red line) plots the results from a standard Taylor rule with a countercyclical capital buffer. The use of a countercyclical buffer yields better results; that is, it tempers the rise in bank credit and real estate prices by mandating that banks hold more capital, thereby discouraging a reduction in the lending rate. Consumption does not rise as much as in scenario 1, so there is a welfare cost, but consumption is still higher than it was before the shock. In summary, the countercyclical buffer in response to a negative interest rate shock mitigates a generalized credit and asset price boom while allowing the economy to benefit from the lower borrowing costs.
Figure 9.9.Response to a Negative Domestic Policy Rate Shock
Source: Author simulations.
The use of countercyclical capital buffers to mitigate the procyclicality of the financial system can help reduce systemic risks and macroeconomic volatility, but may not be optimal in response to all types of shocks (see IMF 2009). A second simulation for the Philippines looks at the response to a productivity shock that is welfare enhancing. In this case, the use of a countercyclical capital buffer would result in lower investment and potential growth compared with a policy framework that relies only on interest rate policy. Under a technology or productivity shock, reducing the procyclicality of the financial system would result in a real cost (Figure 9.10, red line). The policy trade-offs would depend partly on whether the economy is more frequently and severely affected by productivity shocks as opposed to financial shocks. The conventional view that financial shocks tend to dominate over productivity shocks in emerging markets provides a rationale for considering the use of countercyclical capital buffers, a macroprudential tool still absent from the toolkit in ASEAN-5 countries.
Figure 9.10.Response to a Positive Technology Shock
Source: Author simulations.
Synergies Between Monetary and Fiscal Policy
As discussed in the previous section, low interest rates since the global financial crisis have spawned challenges for monetary policy and financial stability in the ASEAN-5 economies. Historically low global interest rates have spilled over to domestic interest rates (see Chapter 4). In some countries, the resulting easy financing conditions have amplified domestic financial cycles, resulting in elevated household debt, corporate debt, or both, even when inflation pressure has remained subdued (Chapter 7). Diverging business and financial cycles have put to the test the calibration of monetary policy and possible synergies with macroprudential policies.
Yet the same environment that has put monetary policy to the test has likely made fiscal policy all the more powerful as a tool for stabilizing the business cycle. In countries with economic slack and low inflation, fiscal stimulus can stimulate aggregate demand, reducing the burden on countercyclical monetary policy and any potential trade-offs with financial stability. In contrast, in countries with robust growth and inflation, fiscal stimulus and the spur to domestic demand may need to be compensated for by monetary policy tightening. In this case, fiscal and monetary policy may work at cross-purposes rather than in sync.
The use of fiscal policy to stabilize the cycle, however, is not free of challenges. Changes in fiscal policy frequently require legislative changes and are often opposed by well-organized groups whose interests would be harmed. Even if such changes are approved, implementation is often cumbersome and subject to delays. These hurdles help explain why fiscal policy in most emerging markets has been found to be procyclical over the past few decades (Ilzetzki and Vegh 2008). Coordination with monetary authorities can also be challenging, given the importance of a central bank that is independent and that is regarded as such.
This section looks at the conditions under which exploiting synergies between monetary and fiscal policies may improve macroeconomic outcomes. In particular, it analyzes the impact of fiscal stimulus from public investment in infrastructure under different reactions of monetary policy as well as different government financing strategies. Although development needs vary across ASEAN-5 countries, many face common challenges from infrastructure gaps. There is scope to increase the quality of infrastructure to catch up with regional leaders such as Singapore (Figure 9.11). Better infrastructure could also help increase total factor productivity where there is currently a significant gap with respect to advanced economies (Figure 9.12).
Figure 9.11.Indicator of Quality of Infrastructure
Sources: IMF, World Economic Outlook database; World Economic Forum; and IMF staff calculations.
Note: Figure uses International Organization for Standardization country codes. EMEUR = emerging Europe; LAC = Latin America and the Caribbean.
Figure 9.12.Total Factor Productivity
Sources: Penn World Table 9.0; and IMF staff calculations.
Note: All TFP is weighted by PPP. EMDEs include Bangladesh, Bhutan, Cambodia, Fiji, Lao P.D.R., Myanmar, Nepal, Sri Lanka, and Vietnam. EMDEs = emerging market and developing economies; PPP = purchasing power parity; TFP = total factor productivity.
The simulations are based on the APDMOD, a module of the IMF’s Flexible System of Global Models.9 This is a semistructural model of the global economy, with individual blocks for 16 Asian countries and 8 additional regions to represent the rest of the world. The model has a rich fiscal sector with seven possible instruments, including spending on consumption, infrastructure, or transfers; lump sum taxation on households; and distortionary taxation on consumption, labor, and capital. Only some households hold debt as a source of wealth, which allows them to smooth consumption in the face of shocks or policy changes. Other households cannot save effectively and live off only their current income. These non-Ricardian properties allow fiscal policy to have a powerful role in the long term, not just the short term. Monetary policy is assumed to follow an inflation-targeting regime. With its rich fiscal sector, this model is well suited to simulate the impact of different fiscal policies under different financing scenarios.10
This first set of results illustrates an expansion of 1 percent of GDP in infrastructure investment every year over a period of five years for all ASEAN-5 countries. All of the infrastructure push is implemented through traditional public investment financed by lower government transfers to households in a budget-neutral manner. In turn, monetary policy follows a standard Taylor rule, with the interest rate path set to close inflation and output gaps over the forecast horizon.
Figure 9.13.Impact of Infrastructure Investment on ASEAN-5 Countries
Source: IMF staff calculations.
The simulations show that the total 5 percentage point increase in infrastructure investment leads to an increase in real GDP within the range of 0.8–1.2 percentage points in ASEAN-5 countries over five years (Figure 9.13, panel 1), with growth increasing by 0.16–0.24 percentage point a year, on average. Infrastructure investment raises growth in the short term, but it also has a lasting impact on growth through higher productivity. In these baseline simulations, the impact of additional infrastructure spending on growth is partly dampened by two factors: (1) the reaction of monetary policy, with an increase in real interest rates of up to 0.3–0.5 basis point at their maximum (Figure 9.13, panel 2); and (2) budget-neutral financing through lower transfers. Both factors crowd out private domestic demand, and the resulting impact of fiscal stimulus on inflation is relatively small.
However, a legacy of the global financial crisis has been subdued inflation pressure on a global scale, which has affected some ASEAN-5 countries in a significant way (see Chapter 7). In countries facing persistently low inflation, central banks have plenty of room for inflation to rise before breaching their targets. These central banks could follow a policy of monetary accommodation in the short term, which would lead to real interest rates noticeably lower than those that would prevail if monetary policy worked to offset the inflationary impact of the fiscal stimulus. Such a strategy could be justified to prevent a low-growth, low-inflation trap and mitigate the risks of hitting the zero lower bound. The combination of fiscal and monetary stimulus would allow inflation to converge to target from above (that is, some temporary overshooting),11 and the impact on real activity would be larger. Moreover, the longer the period of monetary accommodation, the greater the gain from fiscal stimulus because the private sector would expect inflation to be more responsive, further reducing the real interest rate.
A second set of results illustrates the payoff from exploiting synergies between fiscal and monetary policy in a low-inflation, low-interest-rate environment. Figure 9.14 shows simulations calibrated for Singapore and Thailand. Consider first the baseline scenario with the infrastructure push financed through lower general transfers and monetary policy working to offset the inflationary impact (Figure 9.14, blue line). If the central bank instead chose to accommodate the stimulus over the five years, there would be a significant increase in the fiscal multiplier. Over a five-year horizon, the impact on real GDP more than doubles to 2.4–2.6 percentage points with monetary accommodation, while prices increase by about 1.5 percentage points in cumulative terms (Figure 9.14, red line).
Figure 9.14.Impact of Infrastructure Investment under Alternative Scenarios
Source: IMF staff calculations.
Yet fiscal policy can still do more. Low interest rates also imply cheap financing costs for the government. If the government were to use debt financing rather than a budget-neutral strategy, the size of the multiplier would increase further. Monetary accommodation with debt financing (Figure 9.14, light blue line) leads to an additional 0.5 percentage point increase in real GDP and a 0.4 percentage point increase in prices over a five-year period. Real GDP and growth increase for two reasons: first, there is no cut in transfers to households, and second, the real interest rate is lower. Both factors provide greater support to private domestic demand, amplifying the initial impact of fiscal stimulus on growth and inflation.
Through their positive impact on nominal GDP, over the short and medium term joint public investment and monetary accommodation policies can also help preserve fiscal space. Financing the scaled-up investment with debt and without monetary accommodation (Figure 9.14, green line) leads to a significant increase in the debt-to-GDP ratio (about 5 percentage points in Singapore and 4 percentage points in Thailand) as the fiscal primary deficit increases. In contrast, in Thailand, allowing for monetary accommodation significantly reduces the debt buildup to less than 1 percent of GDP because the growth-adjusted real interest rate that applies to its old debt stock drops significantly. In Singapore, the impact of lower real interest rates on debt accumulation is negligible given that its initial stock of net debt is close to zero. A case can be made for debt financing in countries with low inflation and low interest rates, as long as the medium-term debt profile remains sustainable and the sovereign risk premium is contained.
For countries where monetary accommodation and debt financing are not appropriate, increasing the efficiency of public investment would be a way to raise the multiplier. Figure 9.15 shows the impact of high-efficiency investment in Indonesia, a country with more limited fiscal space and inflation currently within the target range. A budget-neutral scale-up in investment, with no monetary accommodation, would allow for nearly 0.8 percentage point in additional real GDP over five years at the current average investment efficiency level. In contrast, new investment with efficiency levels comparable to those of Singapore (a country at the investment efficiency frontier within the region) would raise the impact on real GDP by another 0.2 percentage point.
Figure 9.15.Indonesia: Impact of High-Efficiency Infrastructure Investment on Real GDP
Source: IMF staff calculations.
Exploiting synergies between monetary, macroprudential, and fiscal policies can be beneficial for guarding the economy against fluctuations along the real and financial cycles. Recessions have been seen to have permanent impacts on GDP, including in the ASEAN-5, and both household and corporate debt accumulation could have a significantly negative impact on growth a few years down the road. At the same time, the business and financial cycles have different amplitudes and durations in ASEAN-5 countries (just as in many other countries); therefore, a single policy instrument is insufficient for smoothing out both cycles.
In this context, macroprudential policies can play an important role in complementing countercyclical monetary policy. Model simulations show that a strategy based on countercyclical loan-to-value ratios and a Taylor rule focused on price and output stability yields better results than a lean-against-the-wind monetary policy rule when there is a shock to the demand for housing. This result is robust to a relative flattening in the Philips curve. Countercyclical capital buffers could also help mitigate the procyclicality of the financial system and reduce systemic risks with minimal real costs in response to a wide array of shocks.
Fiscal policy can complement monetary policy in smoothing out the cycle while supporting medium- and long-term growth. Infrastructure investment can lead to significant increases in real GDP. When coupled with monetary accommodation, the investment multiplier doubles. When financed through debt and coupled with monetary accommodation, the investment multiplier can even triple. These policy options are particularly attractive for countries with persistently low inflation. The additional growth also allows these countries to protect their fiscal space even in scenarios in which the investment scale-up is financed with debt. For countries with more limited fiscal space and high inflation, a focus on high-efficiency investment is likely to be the best option for achieving a higher multiplier effect.
This annex examines whether recessions (no matter how long or deep) affect long-term trend real GDP. Following DeLong and Summers (1988) and Beaudry and Koop (1989), the current depth of a recession (denoted CDRt) is defined as the gap between the current level of output and the economy’s historical maximum level, that is:
The values taken by the CDR variable over the period 1996 through 2017 are measured using real quarterly GDP data for Indonesia, Malaysia, the Philippines, and Thailand. As an example, Annex Figure 9.1.1 shows the CDRt indicator for Thailand. The impacts of the Asian financial crisis and the global financial crisis episodes stand out in the figure and are likely to be important drivers of the results for all ASEAN-5 countries.
Annex Figure 9.1.1.Thailand: Depth of Recession
Source: IMF staff calculations.
Using Pesaran’s panel mean group estimator, the response of real GDP can be drawn as follows:
Figure 9.1 presents the impulse response function of the “typical” ASEAN country following an economic recession. The model shows that, after about 15 months, growth bounces back, with positive growth for about 9 months following an economic downturn. However, the response of real GDP growth implies that the growth bounce-back is not directly proportional to the output loss, as would be predicted by the natural rate theory (which essentially states that output fluctuates around a fixed trend and, therefore, booms help predict busts).
This can be illustrated by looking at the response in Figure 9.2, which illustrates the level effect of the typical recession on real GDP. The estimates imply that recessions have tended, at least over the sample period in question, to have had hysteresis effects on the level of real GDP.
These findings reinforce the need for countercyclical fiscal and monetary policies to prevent economic downturns from having long-term adverse effects on economic growth.
To consider the impact of macroprudential policies on the financial and real cycles, a Bayesian panel vector autoregression model is estimated for ASEAN-5 countries. The specification uses accounts for cross-sectional heterogeneity and can be expressed in basic terms as follows:
in which c = 1, . . . ,5 denotes the country; l = 1,…,L denotes the number of lags; yct is a vector of n endogenous variables; wtis a vector of W exogenous variables that are common across the ASEAN-5 countries; and zctincludes country-specific constant terms. Individual country results can be drawn from the panel by extracting the individual country coefficients assumed to be drawn from a normal distribution with a common mean β and a variance Λc, which may be country-specific:
The final assumption pertains to Λc, the variance of βc around the common mean. The model sets these parameters with respect to the uncertainty about these variances and how heterogeneous these countries are. In the vector autoregression for country c, the coefficient of the variable k (in which k = 1, . . . ,K run over lags of endogenous variables and common exogenous controls) in equation n has a variance equal to
Since some of the country coefficients (βc) are large and some small, each country coefficient’s variance is scaled by a factor that adjusts for the size of the country coefficient λ.
The model contains an index of macroprudential policy, real GDP growth (as a measure of the real cycle), credit growth (as a measure of the financial cycle), and the stock price to capture asset price dynamics. The data run from 2000 to 2012 at a monthly frequency. The ASEAN panel results show the following, as illustrated in Figure 9.4:
The response functions show that a tightening of macroprudential policy leads to a statistically significant decline in credit growth (red lines). The impact on the real cycle of a tightening in macroprudential policy is much smaller and less persistent, with the response becoming progressively more statistically insignificant at conventional levels over the medium to long term.
A joint cumulative equality test on the null hypothesis that the responses of real and financial cycles to a tightening in macroprudential policies are the same can be rejected at the 10 percent significance level, based on a p-value of 0.06 from a chi-squared distribution.
One could construe these findings as being consistent with the idea that real and financial cycles operate at different frequencies and, therefore, their responses to different policy instruments (such as macroprudential policy) are likely to differ. These estimates for the ASEAN-5 imply that the economic adjustment to a tightening in macroprudential policy falls principally on the financial—as opposed to the real—cycle.
This annex examines the impact of a buildup in private sector debt on current and future GDP growth with a focus on Singapore and Thailand (the only two ASEAN-5 countries with sufficiently long household and corporate debt series). The methodology follows Mian, Sufi, and Verner (2017) and uses the Bank for International Settlements database on nonfinancial credit, which allows for an irregular panel of 47 countries spanning 1990 to 2016 for the longest series.
The regression analysis intends to explain the cumulative growth in country i’s real GDP over three years (∆3
yi,t + k + k)12 as a function of the buildup over three years in private sector debt13
Using fixed effects, a panel regression is run for the entire pool of countries, allowing for one set of coefficients for ASEAN-5 and another set of coefficients for the rest of the countries in the pool.
The first set of regressions (Table 9.1) looks at the impact of private sector debt (household plus corporate debt) on growth. Accumulations of private sector debt in ASEAN-5 countries between years t − 4 and t − 1 have an initially positive but statistically insignificant impact on growth accumulated over the three years to t, but this impact turns negative and significant in years t + 1 and t + 2.
The second set of regressions (Table 9.2) looks at the impact of household and corporate debt on growth. Accumulation of household debt in Singapore and Thailand (the only ASEAN-5 countries with sufficiently long series for household and corporate debt) has a positive and significant impact on growth two and three years ahead, whereas the impact turns negative and significant four and five years ahead. This result is similar to the one obtained for the other countries in the pool, although the positive effect in the first two years is significantly higher in Singapore and Thailand, and the negative impact starts one year earlier. The accumulation of household debt produces a boom-bust pattern in growth in Singapore and Thailand.
In contrast, the accumulation of corporate debt in Singapore and Thailand has a negative and significant impact starting just one year ahead and through the medium term. The impact of corporate debt on growth in Singapore and Thailand is significantly stronger (in statistical and economic terms) than in the rest of the countries in the pool.
The Model Economy
The open economy dynamic stochastic general equilibrium (DSGE) model of Anand, Delloro, and Peiris (2014) is extended to incorporate household borrowing, housing, and macroprudential policies. The sectoral breakdown and model structure is as follow:
The economy is populated by two groups of households (patient P and impatient I ) as well as entrepreneurs (E ). Each group of households is assumed to be composed of a unit mass of identical and infinitely lived households indexed by j. The key difference between the two groups is the degree of impatience: the discount factor of patient household (βP) is higher than that of impatient households (βI). The heterogeneity in discount factors determines the direction of financial flows in equilibrium: patient households purchase a positive amount of deposits and do not borrow, while impatient households (as well as entrepreneurs) borrow a positive amount of loans.
(1.1) Patient Households
The representative patient household j maximizes expected lifetime utility
wherein resources are spent on consumption
Combining equations (4) and (5) gives the uncovered interest parity condition.
(1.2) Impatient Households
The representative impatient household j maximizes expected lifetime utility
subject to its budget constraint
wherein resources are spent on consumption
The first-order conditions of the optimization problem are given by
The aggregated consumption bundle
in which ϑ is the elasticity of substitution between domestic and foreign goods and γ refers to the home bias. Each group of households minimizes consumption expenditure Pt Ct = PH,t CH,t + PF,t CF,t. The demand function for domestic and imported consumption goods, as well as the consumer price index are given by the following expressions:
(1.4) Deposit and Loan Demand
Patient households determine how much to deposit in retail banks by maximizing the interest payments from deposits
subject to the deposit and deposit rate aggregator functions, respectively given by
The first-order condition gives the ith retail bank’s deposit demand function:
Similarly, impatient households decide how many loans to take out by minimizing interest payments from loans:
subject to analogous loan and lending rate aggregator functions. Similarly to the deposit demand of patient households, loan demand by impatient households will take a functional form of
Entrepreneurs purchase capital Kt+1 for the next time period at price qt. They finance capital acquisition partly through their net worth nt and partly through borrowing. Total borrowing
A proportion of total borrowing
Entrepreneurs decide how many domestic loans to take out by minimizing interest payments from those loans:
subject to analogous loan and lending rate aggregator functions. Similarly to impatient households, loan demand by entrepreneurs will take a functional form of
We assume that there exists an agency problem between foreign banks and entrepreneurs, which makes foreign external finance more expensive than internal funds. The entrepreneur’s marginal external financing cost 𝔼 t ft+1 is given by
in which Γt is a shock to the cost of borrowing. We specify the external finance premium as
At the end of the period, entrepreneurs lease their undepreciated capital to capital goods producers. The expected marginal real return on capital yields the expected gross return
The optimal loan contract condition between banks and entrepreneurs is given by
which states that the marginal return of capital should equal its marginal cost. The net worth of an individual entrepreneur Vt is given by
We assume that a proportion ν of entrepreneurs survive until the next period. A fraction 1 – ν of entrepreneurs exits the economy and is similarly replaced by new entrepreneurs. We further assume that the new entrepreneurs receive an exogenous transfer H from the exiting entrepreneurs. The transfer of resources is necessary to ensure that all entrepreneurs have sufficient funds to borrow and settle their loans. Aggregate entrepreneurial net worth evolves according to
Entrepreneurs exiting the economy consume and transfer some funds to new entrepreneurs. Thus, the consumption of entrepreneurs, denoted by
(3) Capital Producers
Capital producers combine the existing capital stock leased from entrepreneurs to transform gross investment It into new capital. We assume that the production of new capital entails quadratic adjustment costs. Capital accumulation in the economy is given by a linear technology:
in which ςI,t is a shock to the marginal efficiency of investment. Gross investment consists of domestic and foreign final goods, denoted respectively as IH,t and IF,t. We further assume that it has the same aggregation function as the consumption bundle.
Minimizing the capital producers’ investment expenditure Pt It = PH,t IH,t + PF,t IF,t gives the demand function for domestic and imported investment goods, respectively,
Capital-producing firms seek to maximize expected profits:
The first-order condition gives the capital supply equation:
(4) Wholesale Sector
The wholesale sector in the economy is assumed to be a perfectly competitive market. It is composed of the economy’s entrepreneurs, who combine labor provided by each group of households and capital purchased from capital-producing firms, in order to produce wholesale goods Yt through a constant return to scale (CRS) Cobb-Douglas production function.
in which θt is a shock to total factor productivity. Entrepreneurs determine how much labor and capital to employ by maximizing profits subject to the production function
in which mct is the real marginal cost of production.
(5) Retail Sector
The retail sector of the economy is assumed to be monopolistically competitive and is composed of a continuum of retailers with a unit mass. Retailers purchase wholesale goods, and differentiate them at no cost, to produce domestic goods
with a corresponding demand function facing each retailer
For simplicity, we assume that the aggregate export demand function is given by
in which variables with asterisks indicate their exogenous counterpart. We also assume that the law of one price holds in the export market, so that Px,t = et PH,t.
To incorporate nominal rigidity in the model, we assume that in each period, only a fraction 1 − α of firms can change their prices. All other firms can only index their prices to the previous price set. Retailers seek to maximize expected profits
in which λt + s is the stochastic discount factor derived from patient household utility maximization. Profit maximization yields the New Keynesian Phillips Curve
We assume that the price of imported goods is set in the similar way.
(6) Banking Sector
The banking sector operates in a monopolistically competitive environment in which it sets the deposit and lending rates, correspondingly. It is divided into a wholesale and retail branch. The retail branch consists of deposit and loan banks. We incorporate nominal rigidities in interest rate setting by assuming that deposit and loan banks face quadratic adjustment costs when setting their respective rates.
(6.1) Retail Branch
Each deposit bank collects deposits Dt(i) from patient households and passes them on to the wholesale branch, which pays them a wholesale deposit rate
subject to deposit demand of patient households given in equation (A9.4.13). In a symmetric equilibrium, the first-order condition gives the optimal retail deposit rate
Each loan bank obtains wholesale loans
subject to the loan demand of impatient households and entrepreneurs given in equations (A9.4.14) and (A9.4.16). Similarly, in symmetric equilibria, the optimal retail loan rates for impatient households and entrepreneurs are
(6.2) Wholesale Branch
The wholesale branch takes the deposits from the deposit bank. We assume that the wholesale branch meets the cash reserve ratio (CRR) and the statutory liquidity ratio (SLR) imposed by the central bank. The latter can be thought of as an exogenously determined share of deposits in government securities. The central bank varies these requirements to control credit supply by changing the availability of resources with which the banks can make loans. Let
The wholesale branch combines bank capital Zt(i) with the remaining deposit
We assume that there exists an exogenously given capital-to-assets (leverage) ratio κt for banks. The bank pays a quadratic cost whenever the capital-to-asset ratio moves away from κt. This modeling choice gives bank capital a key role in providing the conditions of credit supply.
Bank capital is accumulated each period out of retained earnings according to
in which 1 − ωb summarizes the dividend policy of the bank, δb measures the resources used in managing bank capital and conducting overall banking activity, and
The problem for the wholesale branch is to choose loans and deposits so as to maximize profits subject to the balance sheet identity:
The solution yields an optimality condition that links the spread between wholesale loan and deposit rates to the degree of the bank’s leverage position.
We assume that banks can invest excess liquidity in the special deposit account (SDA) facility of the central bank, from which they are remunerated at rate
which links the wholesale loan rate to the central bank policy rate and Treasury bill rate, as well as to the leverage of the banking sector. Overall, profits of banks are the sum of earnings from the wholesale and retail branches. After deleting intragroup transactions, bank profit is given by
(7) Public Sector
Government spending and the government security rate are assumed to be determined exogenously. The central bank sets the policy rate using a Taylor-type rule
while it sets the cash reserve ratio and statutory liquidity ratio according to
The central bank exercises macroprudential regulation on the banking sector by setting the capital adequacy ratio requirement using the following rule:
This macroprudential policy rule is analogous to the Basel III countercyclical capital buffer—the capital requirement of banks is increased when economic conditions are good and relaxed during downturns. Similarly, we assume that the loan-to-value ratio for impatient households is determined by the following rule:
(8) Market Clearing Conditions
Households, exiting entrepreneurs, capital producers, government, and the rest of the world buy final goods from retailers. The economy-wide resource constraint is given by
Note that the aggregated housing stock is fixed:
in which the left side of the equation is the current account, and the right side is the capital account. In order to close the small open economy model, we specify a foreign debt elastic risk premium whereby holders of foreign debt are assumed to face an interest rate that is increasing the country’s net foreign debt:
in which χ is the degree of capital mobility.
(9) Specification of the Stochastic Processes
The model includes 13 structural shocks: shocks to technology (θt), investment efficiency (ςI,t), and housing demand
Ajello, A.,T.Laubach,D.Lopez-Salido, andT.Nakata. 2016. “Financial Stability and Optimal Interest-Rate Policy.” FRB Working Paper 2016–067, Federal Reserve Board, Washington, DC.
Anand, R.,V.Delloro, andS. J.Peiris. 2014. “A Credit and Banking Model for Emerging Markets and an Application to the Philippines.” Proceedings of the2014 BSP International Research Conference, Issue 2. Manila: Bangko Sentral ng Pilipinas.
Andrle, M.,P.Blagrave,P.Espaillat,K.Honjo,B.Hunt,M.Kortelainen,R.Lalonde, and others. 2015. “The Flexible System of Global Models—FSGM.” IMF Working Paper 15/64, International Monetary Fund, Washington, DC.
Beaudry, P., andG.Koop. 1989. “Do Recessions Permanently Change Output?” Journal of Monetary Economics 31:149–63.
DeLong, J. B., andL. H.Summers. 1988. “How Does Macroeconomic Policy Affect Output?” Brookings Papers on Economic Activity2 (Fall): 433–80.
Cerra, V., andS. C.Saxena. 2008. “Growth Dynamics: The Myth of Economic Recovery.” American Economic Review98 (1): 439–57.
Cerra, V., andS. C.Saxena. 2017. “Booms, Crises, and Recoveries: A New Paradigm of the Business Cycle and Its Policy Implications.” IMF Working Paper 17/250, International Monetary Fund, Washington, DC.
Corbacho, A.,D.Muir,M.Nozaki, andE.Zoli. Forthcoming. “The New Mediocre in Asia: What Can Fiscal Policy Do?” International Monetary Fund, Washington, DC.
Curdia, V., andM.Woodford. 2009. “Credit Spreads and Monetary Policy.” Journal of Money, Credit and Banking42 (S1): 3–35.
Filardo, A., andP.Rungcharoenkitkul. 2016. “The Quantitative Case for Leaning against the Wind.” BIS Working Paper 594, Bank for International Settlements, Basel.
Gambacorta, L., andF. M.Signoretti. 2014. “Should Monetary Policy Lean against the Wind?” Journal of Economic Dynamics and Control43 (June): 146–74.
Ghilardi, M. F., andS. J.Peiris. 2016. “Capital Flows, Financial Intermediation and Macroprudential Policies.” Open Economies Review27 (4): 721–46.
Ilzetzki, E., andC.Vegh. 2008. “Procyclical Fiscal Policy in Developing Countries: Truth or Fiction.” NBER Working Paper 14191, National Bureau of Economic Research, Cambridge, MA.
International Monetary Fund (IMF). 2009. Global Financial Stability Report: Navigating the Financial Challenges Ahead. Washington, DC, October.
International Monetary Fund (IMF). 2012. “The Interaction of Monetary and Macroprudential Policies.” Policy Papers, Washington, DC.
International Monetary Fund (IMF). 2014a. “Monetary Policy in the New Normal.” IMF Staff Discussion Note, Washington, DC.
International Monetary Fund (IMF). 2014b. “Staff Guidance Note on Macroprudential Policy.” Washington, DC.
International Monetary Fund (IMF). 2015. “Monetary Policy and Financial Stability.” IMF Policy Paper, Washington, DC. www.imf.org/external/np/pp/eng/2015/082815a.pdf.
International Monetary Fund (IMF). 2017. “Household Debt and Financial Stability.” Chapter 2 in Global Financial Stability Report. Washington, DC, October.
Jordà, Ò.,M.Schularick, andA.Taylor. 2013. “When Credit Bites Back.” Journal of Money, Credit and Banking45 (s2): 3–28.
Menna, L., andM.Tobal. 2017.“Financial and Price Stability: The Role of the Interest Rate.” Unpublished, Banco de Mexico.
Mian, A.,A.Sufi, andE.Verner. 2017. “Household Debt and Business Cycles Worldwide.” Quarterly Journal of Economics132 (4): 1755–817.
Olsen, Ø.. 2015. “Integrating Financial Stability and Monetary Policy Analysis.” Speech given at the Systemic Risk Centre, London School of Economics, London, April 27.
Sahay, R.,V. B.Arora,A.Arvanatis,H.Faruqee,P.N’Diaye, andT.Mancini-Griffoli. 2014. “Emerging Market Volatility: Lessons from The Taper Tantrum.” IMF Staff Discussion Note, Washington, DC.
Schularick, M., andA.Taylor. 2012. “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008.” American Economic Review102 (2): 1029–1061.
Smets, F.2014. “Financial Stability and Monetary Policy: How Closely Interlinked?” International Journal of Central Banking10 (2): 263–300.
Stein, J. C.2014. “Incorporating Financial Stability Considerations into a Monetary Policy Framework.” Speech given at the International Research Forum on Monetary Policy, Washington, DC, March 21.
Svensson, L. E.O.2014. “Inflation Targeting and ‘Leaning against the Wind’.” International Journal of Central Banking10 (2): 103–14.
Svensson, L. E.O.2016. “Cost-Benefit Analysis of Leaning against the Wind: Are Costs Larger Also with Less Effective Macroprudential Policy?” NBER Working Paper 21902, National Bureau of Economic Research, Cambridge, MA.
Woodford, M.2012. “Inflation Targeting and Financial Stability.” NBER Working Paper 17967, National Bureau of Economic Research, Cambridge, MA.
Annex 9.1 presents vector autoregression estimates for ASEAN-5 countries for the period 1996–2017, using Pesaran’s panel mean group estimator to generate the impulse response for a typical ASEAN country when hit by a recession.
The Asian financial crisis of the late 1990s and the 2008–09 global financial crisis were certainly big shocks during the sample period, and the results likely reflect the impact of these two events to a large extent.
There have been voices strongly advocating in favor (for example, Olsen 2015) and against (for example, Svensson 2014) such a role for monetary policy, while others portray a more balanced view (for example, Yellen 2014). Detailed discussions can be found in Smets (2014), Stein (2014), and Svensson (2014), among others.
Macroprudential policy has been defined as “the use of primarily prudential tools to limit systemic risk, that is the risk of disruptions to the provision of financial services that is caused by an impairment of all or parts of the financial system, and can cause serious negative consequences for the real economy” (IMF 2014b). It includes a range of instruments, such as measures to address sector-specific risks (for example, loan-to-value and debt-to-income ratios), countercyclical capital requirements, dynamic provisions, reserve requirements, liquidity tools, and measures to affect foreign-currency-based or residency-based financial transactions.
Difficulties in identifying financial risks hinder the use of monetary policy to attenuate the financial cycle as well.
For example, in Thailand, nonbank financial institutions play a significant role in household credit, and the reach of macroprudential instruments is more limited in this sector. In addition, the supervision of banks and nonbanks is carried out by different institutions, which could present coordination challenges when implementing a countercyclical macroprudential policy.
Available on request from the authors.
In this case, the increase in household credit in response to the positive housing demand shock would be smaller under scenario 1, but the drop in inflation and output when leaning against the wind (scenario 2) would also be smaller.
Its results, however, are not directly comparable with those obtained in the previous section, where the model used had a more sophisticated financial sector (but a very simple fiscal sector) and where policies simulated included a monetary and macroprudential policy component but no fiscal component.
Similar results can be obtained under a scenario in which monetary policy, rather than a standard linear Taylor rule, follows a policy reaction function that minimizes a quadratic loss function of inflation and output gaps.
Cumulative growth in GDP as of t + k is calculated as the percent change in growth between year t + k − 3 and t + k.
Changes in private sector debt are calculated as the change in the ratio of debt to GDP over years t − 4 to t − 1.