In his 2010 book, Fault Lines, Raghuram Rajan argued that rising inequality in the past three decades led to political pressure for redistribution that eventually came in the form of subsidised housing finance. Political pressure was exerted so that low-income households who otherwise would not have qualified received improved access to mortgage finance. The resulting lending boom created a massive run-up in housing prices and enabled consumption to stay above stagnating incomes. The boom reversed in 2007, leading to the banking crisis of 2008. Along these lines, Kumhof and Rancière (2011) explore the links between inequality, leverage and crises within the context of a DSGE model. They motivate their model with examples from the US in the 1920s and the more familiar events leading up to the subprime crisis.
There is reason to pause before accepting the generality of this new view. Income inequality plays no significant role in the large literature on financial instability and credit booms. Standard variables in the international finance literature are current account deficits, pegged exchange rates and the business cycle. Borio and White (2003) also elaborate an influential view on credit booms. Periods of expected low and stable inflation, strong economic growth, and liberalised finance can give rise to complacency among borrowers, lenders, and regulators. Endogenous market forces that might normally ‘rein in’ these imbalances seem to be absent. Massive buildups in credit lead to financial instability in this case.
In recent research (Bordo and Meissner forthcoming) we present some empirical evidence on whether rising income concentration has any explanatory power in accounting for credit booms and financial crises after holding these other factors constant. We focus on a much larger sample than the two unique periods in US economic history that motivate Rajan et al. Our data cover a panel of 14 mainly advanced countries from 1920 to 2008 covering a wide number of boom-bust episodes and financial crises. We proceed in three steps. First, we offer cross-country regressions relating changes in income inequality to credit growth. Next, we relate credit growth to systemic banking crises. Finally, we present further historical evidence.
Inequality and credit growth
After controlling for a number of the variables cited above, rising income concentration, measured by changes in the income share of the top 1% of tax units, plays no significant role in explaining credit growth. Instead, the two key determinants of credit booms are the upswing of the business cycle or economic expansion and low interest rates. This is very much consistent with a broader literature on credit cycles.
While inequality often ticks upwards in the expansionary phase of the business cycle, this factor does not appear to be a significant determinant of credit growth once we condition on other macroeconomic aggregates. Figure 1 presents an added-variable plot linking credit growth with changes in the income share of the top 1% after controlling for several other variables including growth in GDP and interest rates. Figure 2 presents a similar plot relating credit growth to changes in real GDP. The relationship is clearly much tighter in Figure 2 than in Figure 1.
Figure 1. Change in loans versus changes in top 1% income shares, 14 countries, 1972–2008
Notes: The y-axis graphs the residuals from a panel OLS regression of the annual change of the log of real loans on country fixed effects, year fixed effects, and lagged changes of the following: the log of real GDP, short-term interest rates, the log of the real money supply, the log of the ratio of investment to GDP, and the ratio of the current account to GDP. The x-axis plots residuals from a similar regression using changes in the top 1% income share as the dependent variable.
Figure 2. Change in loans versus the change in GDP, 14 countries, 1972–2008
Notes: The y-axis plots residuals from a panel OLS regression of the annual change of the log of real loans (credit) on country fixed effects, year fixed effects, and lagged changes in the share of income of the top 1% of tax units, short-term interest rates, log of the real money supply, log of the ratio of investment to GDP, and the ratio of the current account to GDP. The x-axis plots residuals from a similar regression using contemporaneous changes in the log of real GDP and contemporaneous explanatory variables.
Credit and crises
On the other hand, we do agree with Rajan et al that financial instability and banking crises often follow above-average growth in credit. Our evidence, which reproduces that found in Schularick and Taylor (forthcoming), finds a fairly strong relationship between growth in real credit and the probability of a banking crisis. This is consistent with many different models of financial instability, but we take no stand on this in our paper. What can be said is that inequality is not significantly related to systemic banking crises in our large sample. Since income concentration is not a good predictor of credit growth, it is hard to see how it can be related to crises by the channels proposed in the work cited above.
History, credit, and crises
Historical evidence from several major credit booms finds scant support for the inequality/crisis hypothesis. In the 1920s in the US, consumer and mortgage debt did indeed rise as the top 1% share in total income climbed from 15% in 1922 to 18.42% in 1929 (Piketty and Saez 2003). In that period, consumer credit was closely related to the rise of new, big-ticket consumer durables in the 1920s such as automobiles, washing machines, and radios. The rise of consumer credit arguably came from supply-side innovations rather than from increased household demand to maintain consumption in the face of stagnant incomes as in Kumhof and Rancière (Olney 1989). The housing boom that ended in 1926, well before the Depression started, reflected a significant amount of postwar pent-up demand, higher quality housing, and a favourable interest-rate environment (White 2009). It appears to have had little to do with the subsequent Depression and series of banking crises that would begin in mid-1929.
Time series evidence from other countries is not consistent with the inequality link either. In Japan in the 1980s, credit growth rose in advance of top income shares. On the other hand, top income shares started rising in 1995 in Japan while credit growth languished. In Sweden in the 1980s, sharp rises in top incomes followed rather than led credit growth. Again, in Sweden, top income shares continued to grow in the aftermath of the banking and real estate bust in 1991, but credit fell. In Australia, credit growth was unrelated to top income shares in the 1970s. Top income shares followed rather than led credit growth in the housing boom of the late 1980s in Australia.
Conclusions and lessons
If income inequality drove the credit boom that preceded the subprime crisis in the US, the event was an outlier by historical standards. Comparative evidence from the last century shows little relationship between rising inequality and credit booms. Even in the US, a more plausible interpretation of events in the first decade of the 21st century is that interest rates were at historical lows. That situation coupled with financial innovation allowing low-income workers to buy houses at unrealistic prices given forecasts of permanent incomes and the likely reversion in interest rates.
If there is a policy lesson in all of this, it might be related to the fact that market-determined rates of leverage can lead to a systemic financial crisis and ensuing negative spillovers. But an increase in the supply of credit that generates a financial crisis has very different policy implications from those that might be prescribed by an increase in the demand for credit that allegedly arose to maintain consumption in an increasingly unequal society. In the former case, financial regulations and reforms to limit leverage and systemic risk that have been discussed and enacted since 2008 are potentially more appropriate remedies to achieve financial stability. While inequality is arguably problematic for many other reasons, we remain sceptical of claims that it engenders financial crises.
Bordo, MD and CM Meissner (forthcoming), “Does Inequality Lead to a Financial Crisis?”, Journal of International Money and Finance.
Borio, C and WW White (2003), “Whither monetary and financial stability? The implications of evolving policy regimes”, in Monetary Policy and Uncertainty: Adapting to a Changing Economy: A Symposium sponsored by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, 28-30 August, pp. 131–211.
Kumhof, M and R Rancière (2011), “Inequality, leverage, and crises”,VoxEU.org, 4 February.
Olney, M (1989), “Credit as a Production-Smoothing Device: The Case of Automobiles, 1913-1938”, Journal of Economic History, 49(2):377–91.
Piketty, T and E Saez (2003), “Income Inequality in the United States, 1913-1998”, Quarterly Journal of Economics, CXVIII(1):1–39.
Rajan, R (2010), Fault Lines, Princeton University Press.
Schularick, M and AM Taylor (forthcoming), “Credit Booms Gone Bust Monetary Policy, Leverage Cycles and Financial Crises, 1870–2008”, American Economic Review.
White, Eugene N (2009), “Lessons from the Great American Real Estate Boom and Bust of the 1920s”, NBER Working Paper No. 15573.