Household credit cycles and financial crises

Jan Hannes Lang, Peter Welz 11 March 2019

a

A

Excessive credit growth and leverage have been identified as key drivers of the Global Crisis and many other episodes of financial instability (Borio and Lowe 2002, Schularick and Taylor 2012). Recent research has shown that in particular household credit booms take a prominent role in driving financial crises and recessions (Mian and Sufi 2014, Mian et al. 2017).

The analysis of household credit developments is therefore pivotal at institutions in charge of macroprudential policy to address unsustainable credit developments – or at least to mitigate their fallout. However, policymakers face the challenge of disentangling when vigorous credit dynamics are justified by economic developments and when they are “booms gone wrong” (Schularick and Taylor 2009). 

The problem becomes clear when looking at dynamics of household credit-to-GDP ratios, as shown in Figure 1 for 12 EU countries. Oftentimes these ratios trended upwards for a long time, but at some point turned down rapidly, reflecting in many cases financial turmoil. One could argue, for example, that the long phases of credit growth in excess of GDP growth in Portugal and Spain since the mid-1990s were initially justified by economic developments and institutional reforms, but at some point developments became unsustainable. The key question is how to identify the excess – or the deviation of current credit from a sustainable level.

Figure 1 Developments of household credit-to-GDP ratios in EU countries

Sources: Eurostat, BIS, author calculations.

The fundamental determinants of household credit 

In a recent paper (Lang and Welz 2018) we link the sustainable level of household credit to a variety of economic and demographic factors, in order to determine when household credit is excessive. Based on an adjusted version of the overlapping generations model by Eggertsson et al. (2019), we show that the sustainable level of real household credit is driven by four main factors: real potential GDP, the equilibrium real interest rate, the population share of the middle-aged cohort, and the level of institutional quality.

The economic intuition behind these driving forces of household credit is as follows. First, a higher credit stock can be sustained with a higher production/income capacity. Second, lower interest rates reduce the debt service burden, allowing households to finance a higher credit stock. Third, a higher share of the middle-aged population implies a higher share of households that are likely to borrow. Fourth, better institutional quality (e.g. better contract enforcement or effective financial regulation) should help to sustain a higher level of credit without the risk of widespread default or risks to financial stability. 

We use our economic model to measure household credit cycles as deviations of the real household credit stock from the level that is justified by the underlying fundamental economic factors. 

Properties of household credit cycles

Our empirical research for 12 EU countries suggests that household credit cycles are long with an average length of around 20 years.1 Figure 2 illustrates that over the last 35 years most countries experienced two peaks in the household credit cycle: one at the end of the 1980s, and one during the recent global financial crisis. The cycle length for household credit varies between 15 and 25 years across the countries.

Figure 2 Cross-country distribution of household credit cycles over time 

Sources: Lang and Welz (2018).
Notes: The chart shows the mean, median, interquartile range, and 90-10 percentile range of household credit cycles across 12 EU countries (BE, DE, DK, ES, FI, FR, GB, IE, IT, NL, PT, SE).

Household credit cycles are also characterised by large amplitudes in the range of 15% to 25% in most of the countries (Figure 3). In some of the countries that experienced particularly pronounced credit booms, such as Ireland, household credit reached excesses of more than 50% compared to the level justified by economic fundamentals. Hence, substantial boom-bust episodes have occurred over the last 35 years in most EU countries that drove household credit away from the level that is justified by fundamental economic factors.

Figure 3 Amplitudes of household credit cycles in selected EU countries

Sources: Lang and Welz (2018).
Notes: The largest absolute positive and negative deviations of household credit from the sustainable trend level justified by economic fundamentals are computed for the sample 1981q1 to 2014q4.

The link between household credit cycles and financial crises

Household credit excesses tend to build up many years ahead of financial crises and only gradually unwind thereafter (Figure 4). On average, household credit starts to exceed its sustainable trend level around five years prior to the start of a financial crisis. Moreover, household credit excesses usually increase continuously during the run-up to financial crises to reach levels of +20% on average. Once a financial crisis materialises a slow deleveraging process tends to start that takes around five years on average to bring household credit back to a sustainable level. These dynamics indicate the important role that household credit cycles can play for financial crises.

Figure 4 Cross-country distribution of household credit cycles before/after crises

Sources: Lang and Welz (2018)
Notes: The chart shows the mean, median, interquartile range and 90th – 10th percentile range of household credit cycles before and after the onset of financial crises across a sample of 12 EU countries (BE, DE, DK, ES, FI, FR, GB, IE, IT, NL, PT, SE). The dating of crises is based on the new ECB/ESRB financial crises database described in Lo Duca et al. (2017). Purely foreign induced crises are excluded. In total there are 13 systemic financial crises events in the sample.

Situations where household credit cycles are positive and increasing should alert policy makers that financial imbalances might be building up. Figure 5 shows that during the four years that precede systemic financial crises both the level and the two-year change of household credit cycles usually display positive values. This is in contrast to the vast majority of tranquil periods, where either the level of household credit cycles is negative or cycles are decreasing. Early warning signalling models suggest that vulnerabilities tend to be present whenever household credit cycles exceed -1.1% and at the same time increased by more than 2.9 percentage points over the last two years. Based on past data for the 12 EU countries covered in our study, these joint signalling thresholds correspond to 8.7% of missed pre-crisis periods and 14.4% of tranquil periods that are incorrectly classified as vulnerable pre-crisis states.

Figure 5 Signalling information of household credit cycles for financial crises

Sources: Calculations based on results in Lang and Welz (2018).
Notes: The chart shows realisations of the level and the 2-year change of household credit cycles for 12 EU countries (BE, DE, DK, ES, FI, FR, GB, IE, IT, NL, PT, SE) since 1980q1. The 2-year percentage point change is expressed as an annual average. See notes to Figure 5 regarding the dating of financial crises.

In addition, the magnitude of household credit excesses before financial crises contains information about the magnitude of subsequent real GDP declines. Figure 6 shows that, on average across the 13 systemic financial crises in our sample, higher household credit excesses at the start of a systemic financial crisis tended to be associated with larger declines in real GDP following the onset of a crisis. While this is not indicating any causality, it shows that household credit cycles could potentially be useful for informing the calibration of macroprudential instruments. 

Figure 6 Relationship between household credit cycles and the severity of crises

Sources: Calculations based on results in Lang and Welz (2018).
Notes: The chart shows the peak level of household credit cycles at the start of past financial crises plotted against the maximum drop in real GDP from peak to trough that materialised during the same crises. See notes to chart 5 regarding the dating of financial crises.

Implications for macroprudential policy 

Our method that links household credit developments to underlying economic fundamentals can help in identifying when credit is excessive and providing a risk narrative. All else equal, household credit cycles are pushed up by higher credit growth and driven down by changes in the factors that raise the sustainable level of household credit, which include the level of institutional quality, real potential GDP, the middle-aged population share and the equilibrium real interest rate. 

For example, Figure 7 shows how these factors shaped the evolution of the Swedish household credit cycle. During the run-up to the Swedish financial crisis at the beginning of the 1990s, high household credit growth and increases in the equilibrium real interest rate were the main factors pushing the household credit cycle to a peak level of around 30%. Subsequent decreases in the household credit cycle were initially driven by deleveraging, but subsequently also by reductions in the equilibrium real interest rate, improvements in institutional quality, and prolonged increases in real potential GDP. Ahead of the Global Crisis Sweden did not display household credit excesses. However, the level of household credit has continuously exceeded the level justified by economic fundamentals over the last couple of years, which seems to be in line with increased macroprudential action regarding household debt by Swedish authorities.

Figure 7 Evolution and driving factors of the Swedish household credit cycle 

Sources: Lang and Welz (2018).
Notes: The coloured bars show the contributions of fundamental driving factors to changes in the Swedish household credit cycle.

To summarise, our findings suggest that household credit cycles should be carefully monitored by policymakers to ensure financial stability. In particular, positive and increasing household credit cycles should raise warning signs about the potential build-up of vulnerabilities. Moreover, the larger household credit excesses are, the more vigilant macroprudential policy should be to contain the potential fallout from the unravelling of imbalances. However, given that measurement of household credit cycles is always subject to uncertainty, additional analysis and considerations will always remain essential for robust policy decisions.    

Authors’ note: The views expressed in this column are those of the authors and do not necessarily reflect those of the ECB.

References

Borio, C and P Lowe (2002), “Asset prices, financial and monetary stability: exploring the nexus”, BIS Working Papers 114.

Eggertsson, G B, N R Mehrotra and J A Robbins (2019), "A Model of Secular Stagnation: Theory and Quantitative Evaluation", American Economic Journal: Macroeconomics, American Economic Association 11(1): 1-48.

Lang, J H and P Welz (2018), "Semi-structural credit gap estimation", ECB Working Paper No 2194.

Mian, A and A Sufi (2014), House of Debt: How They (and You) Caused the Great Recession and How We Can Prevent It From Happening Again, University of Chicago Press.

Mian, A, A Sufi and E Verner (2017), "Household Debt and Business Cycles Worldwide", The Quarterly Journal of Economics 132(4): 1755-1817.

Schularick, M. and A. Taylor (2009), “Credit booms go wrong”, VoxEU.org, 8 December.

Schularick, M and A Taylor (2012), "Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870-2008", American Economic Review 102(2): 1029-61.

Endnotes

[1] The 12 EU countries are: Belgium, Germany, Denmark, Spain, Finland, France, Ireland, Italy, The Netherlands, Portugal, Sweden, and the United Kingdom.

a

A

Topics:  Macroeconomic policy

Tags:  household credit, financial crises, credit-to-GDP ratios

Principal Financial Stability Expert, Macroprudential Policy Division, ECB

Senior Financial Stability Expert, ECB

Events

CEPR Policy Research