Written by Gerhard Renstler[1]
One of the foundations for developing new macroprudential policies, in addition to traditional macroeconomic stabilization policies, is that financial cycles are different from business cycles. This article identifies the characteristics of the credit and housing cycles, shows how they relate to the GDP cycle, and compares the reliability of real-time estimates.
Financial cycle and macroprudential policy
Systemic instability during the financial crisis led policymakers to develop a macroprudential approach to financial supervision and regulation. An important aspect of these macroprudential policies is to 'tame' the financial cycle (e.g. Bank of England, 2009). When the nature of financial cycles is sufficiently different from that of normal business cycles, monetary and fiscal policies are imperfect tools for dealing with them, and macroprudential policy is considered an independent third stabilizing policy. The evidence for this is strengthened. Moreover, the adjustment of macroprudential policy instruments depends on the characteristics of the financial cycle.
In this regard, credit volume and housing prices are two important financial cycle variables, as historical evidence suggests that many financial crises have been preceded by credit and housing booms (Jorda et al. ., 2014). Empirical studies by central bank researchers have begun to document how central bank behavior differs from business cycles, but the emerging stylized facts have never been digested by the scientific economics community. do not have.
Dreman et al. (2012) report that the housing price and credit cycles are significantly longer than the business cycles. Typically, business cycles are assumed to be 2 to 8 years long, but this study found that financial cycles are 8 to 20 years long. Dreman et al. Therefore, we conclude that business cycles and financial cycles are “separate phenomena.” However, this study specifies cycle length a priori rather than estimating it, so the results are partially open to interpretation. Schuler et al. (2015) use a more flexible approach to determining cycle length, present a composite financial cycle indicator that also includes bond and equity prices, and cover a wider range of euro area countries. They find significant cross-country heterogeneity in the length and amplitude of financial cycles across euro area countries. Claessens et al. (2012) report that although tipping points occur more frequently in GDP than in house prices or credit volumes, large recessions still coincide with a series of financial troughs. Their regression analysis adds that financial turmoil makes recessions longer and more severe. Furthermore, the simultaneous fluctuations between financial and business cycles have received limited attention in the literature so far, and widely used macroeconomic theories do not reflect most of the above facts.
This article summarizes the analysis of Rünstler and Vlekke (2016), which uses a model-based multivariate time series approach to estimate the cyclical components of credit volumes, housing prices, and GDP.[2] This approach goes beyond the aforementioned work based on univariate bandpass filters in three ways. First, it allows us to estimate key characteristics of financial cycles, such as their length and persistence. Second, it allows us to estimate the degree of simultaneous fluctuations between business and financial cycles at different cycle lengths. Finally, because macroprudential policymakers need to evaluate potential actions using real-time indicators of financial cycles (based only on historical data), this study also explores the uncertainty of real-time estimates. ing. The quarterly data used covers the United States and five major European economies (France, Germany, Italy, Spain, and the United Kingdom) from 1973 to 2014.
Figure 1: Personal homeownership and financial cycle characteristics
Characteristics of the credit cycle and housing cycle, and their relationship with the GDP cycle
The more robust methodology confirms two findings of previous research. First, credit volumes and residential real estate prices exhibit significant medium-term cycles, longer than the two to eight years of a typical business cycle. Second, there are important differences between countries. Figure 1 plots the estimated lengths and standard deviations of the credit and house price cycles against the private homeownership rate for the six sample countries. Generally (France, Italy, Spain, UK, USA), the average duration of financial cycles ranges from 13 to 18 years. The standard deviation of the house price cycle (for the same country) ranges from 10% to 20%.
However, the analysis also makes the important point that countries with higher rates of home ownership, particularly Spain and the UK, have larger and longer financial cycles. In contrast, Germany stands out for its very small and short cycles (standard deviation of about 2% and duration of about 7 years) due to its very low home ownership rate.
Second, credit and house price cycles are poorly correlated with standard business cycles of 2 to 8 years, but highly correlated with medium-term GDP cycles of more than 8 years. In multivariate estimation, the GDP cycle appears as a mixture of standard business cycles and medium-term cycles. Previous literature has already documented the existence of such medium-term GDP cycles ( Comin and Gertler, 2006 ), but so far they have not been linked to financial cycles. We find that the house price cycle is largely independent of the standard business cycle, but closely linked to the medium-term GDP cycle. Credit cycles exhibit some correlation with standard business cycles.[3] These patterns are shown in Figure 2. The main peaks and troughs of the GDP cycle are consistent with those of the house price cycle, but additional cyclical fluctuations occur.
Figure 2: GDP and financial cycle estimates (percent deviation from trend)
Third, the uncertainty in real-time estimates of the house price cycle is about the same order of magnitude as the business cycle. Estimates for longer cycles have higher uncertainty, but the house price cycles are also larger, so the estimates have lower uncertainty. Figure 3 shows the results of a simulation run that evaluates the uncertainty ratio of the real-time estimates from the two types of filters. The uncertainty ratio is defined as the standard error of the estimate relative to the standard deviation of the original data. Similar to business cycle research (Rünstler, 2002; Basistha and Startz, 2008), multivariate model-based filters also provide more accurate estimates of the house price cycle than univariate bandpass filters. The uncertainty ratio is approximately 1.5 for the univariate filter, but remains close to 1 for the model-based filter. A value of 1 means that periodic locations near peaks and valleys are most often correctly identified.
Figure 3: Uncertainty in real-time estimates of business and housing price cycles
conclusion
Applying a multivariate time series approach to credit, house prices, and GDP allows us to analyze how financial cycles match GDP cycles across different frequencies. Our results for the United States and five large European economies show that (i) there are important differences in the length and size of financial cycles across countries, but in most countries financial cycles are on average longer and larger than GDP cycles; suggests. (ii) The correlation between the financial cycle and the GDP cycle is limited at normal business cycle frequencies of 2 to 8 years, but increases at lower frequencies. The study also found that the accuracy of real-time estimates of credit and housing price cycles is nearly as accurate as that of business cycles.
One interpretation of these results is that the significant differences between both types of cycles (and the accuracy of estimates of financial cycles) justify macroprudential stabilization policies that are distinct from monetary and fiscal policies. is. However, the lower frequency dependence suggests that it is also important to consider the link between macroprudential policies and traditional stabilization policies. It should be emphasized that the above analysis does not address the causal relationship between financial and GDP cycles, the nature of the underlying shocks, or the specific transmission channels. Future research that addresses these three points will further improve our understanding of why and how financial cycles differ from business cycles and what the implications for policy are.
References
Bank of England (2009), “The role of macroprudential policy”; discussion paper,November.
Basistha, A. and Startz, R. (2008), “Measuring NAIRU with reduced uncertainty: a multi-indicator common cycle approach”; Economics and Statistics ReviewVol. 90, No. 4, pp. 805-811.
Claessens, S., Kose, M.A. and Terrones, M. (2012), “How do business and financial cycles interact?” International Economics JournalVol. 87(1), pp. 178-190.
Comin, D. and Gertler, M. (2006), “Intermediate Cycles”; American Economic ReviewVol. 96(3), pp. 523-556.
Drehmann, M., Borio, C. and Tstasaronis, K. (2012), “Characterizing financial cycles: Don’t lose sight of the medium term!” BIS working paperNo 380, Bank for International Settlements, June.
Jarocinski, M. and Lenza, M. (2016), “Inflation-forecasting measures of the euro area output gap”, Working paper serieshenceforth, ECB, Frankfurt am Main.
Jorda, O., Schularick, M. and Taylor, A. (2014), The Great Mortgage: Housing Finance, Crisis, and Business Cycles. NBER working paper, No. 20501, September.
Rünstler, G. (2002), “Information content of real-time output gap estimation: Application to the euro area”, Working paper seriesNo 182, ECB, Frankfurt am Main, September.
Rünstler, G. and Vlekke, M. (2016), “Business, housing and the credit cycle”. Working paper seriesNo 1915, ECB, Frankfurt am Main, June.
Schüler, Y., Hiebert, P., and Peltonen, T. (2015), “Characterizing financial cycles: multivariate and time-varying approaches”, Working paper seriesNo 1846, ECB, Frankfurt am Main, September.
