Lesson 8 of 8
Credit growth predicts crises
Fourteen countries, 1870 to 2008, and the strongest result in the field: credit booms predict crises, and predict how deep the recession will be.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- The finding: credit growth is a powerful predictor of financial crises, measured across fourteen advanced countries from 1870 to 2008 with a purpose-built dataset.
- The second finding is about severity: more credit-intensive expansions are followed by deeper recessions and slower recoveries, and crisis recessions cost more output than ordinary ones.
- It is not a vindication of any one theory. Several accounts in this course predict the same relationship, and none of them is separated from the others by it.
This is the last lesson of the course and the strongest evidence in the track. It is mainstream work, published in mainstream outlets, built on a dataset its authors assembled because the standard ones were too short.
The two findings, and their samples
Credit booms gone bust
- 14 advanced countries, 1870 to 2008.
- Money, credit and macroeconomic aggregates, assembled by the authors.
- Credit growth is a powerful predictor of financial crises.
- Financial sector leverage rose sharply after the mid twentieth century.
When credit bites back
- The same 14 countries, over 200 recession episodes.
- Local projections conditioned on macroeconomic controls and their lags.
- Crisis recessions are more costly in lost output than normal ones.
- More credit-intensive expansions are followed by deeper recessions and slower recoveries.
Long samples were the enabling condition. Neither result is visible in a few decades of quarterly data.
What was measured
The first study builds a long-run annual dataset for fourteen advanced economies from 1870 to 2008 and uses it to study rare events. Its headline result is that credit growth predicts financial crises, which the authors summarise by calling such crises credit booms gone wrong. They also report that leverage in the financial sector rose strongly in the second half of the twentieth century, visible as a decoupling of money and credit aggregates, and that post-1945 policy responses to crises were more aggressive without the output costs of crises becoming smaller.
The second study takes the same countries and the universe of more than two hundred recession episodes, and asks a different question: not whether a crisis happens, but what the recession looks like afterwards. It reports two facts. Recessions associated with financial crises are more costly in lost output than normal recessions. And for both kinds of recession, the more credit-intensive the preceding expansion, the deeper the recession and the slower the recovery.
Why this is the strongest result here
Three reasons, and they are the three the earlier courses kept asking for.
The sample is long enough to contain many crises, which is the only way to say anything about rare events. The denominator is present: these are studies of all recessions in the sample, not a selection of the interesting ones. And the second paper conditions on macroeconomic controls, so the relationship is not simply a reflection of the business cycle it sits inside.
It is also independent of this track in the strict sense that matters. No author is defending a school, the data are published, and the results have been extended by other researchers using the same database.
Why it is not a vindication
Because too many accounts predict it.
The Austrian account expects credit expansion to be followed by trouble. Minsky's expects a long expansion to shift financing postures toward fragility, with the same consequence. The monetarist reading expects monetary disturbance to produce contraction. The BIS financial cycle work treats credit and property prices as the measure of the thing itself, which makes the relationship close to definitional.
A finding consistent with four accounts does not choose between them. What it does is rule out the position that credit conditions are a detail, which was the mainstream position for several decades and which the survey in the previous course describes as treating finance as a veil.
That is a real result and it should be stated at its true size: the question of whether credit matters is settled, and the question of why is not.
Check yourself
What is the headline finding of the first study?
- That recessions are becoming shorter
- That credit growth is a powerful predictor of financial crises
- That monetary policy has become less effective
Show the answer
Answer: B.
It is measured across fourteen advanced countries from 1870 to 2008, using a dataset the authors assembled for the purpose.What does the second study add?
- That crises are unpredictable
- That more credit-intensive expansions are followed by deeper recessions and slower recoveries
- That financial crises have no output cost
Show the answer
Answer: B.
It studies more than two hundred recessions in the same countries and conditions on macroeconomic controls, and also finds crisis recessions costlier than normal ones.Why is this not a vindication of the account in this track?
- Because the data stop in 2008
- Because several rival accounts predict the same relationship and the result does not separate them
- Because the authors reject the theory
Show the answer
Answer: B.
Austrian, Minskyan, monetarist and financial cycle accounts all imply trouble after credit booms. Separating them needs evidence about the mechanism, not about the correlation.Sources
- Schularick and Taylor, 'Credit Booms Gone Bust: Monetary Policy, Leverage Cycles and Financial Crises, 1870 to 2008', NBER Working Paper 15512, November 2009, revised July 2011, abstract and introduction: the fourteen-country dataset, credit growth as a powerful predictor of financial crises, the rise in financial sector leverage after 1945 seen as a decoupling of money and credit, and the finding that more aggressive post-war policy responses did not reduce the output costs of crises.
- Jorda, Schularick and Taylor, 'When Credit Bites Back: Leverage, Business Cycles, and Crises', Federal Reserve Bank of San Francisco Working Paper 2011-27, October 2012, abstract: the universe of over two hundred recessions in fourteen advanced countries between 1870 and 2008, the local projection method with macroeconomic controls, and the two findings about the cost of crisis recessions and the effect of credit-intensive expansions.
- Borio, 'The Financial Cycle and Macroeconomics: What Have We Learnt?', BIS Working Paper 395, December 2012, introduction, for the characterisation of the earlier mainstream position that treated finance as a veil.
What this leaves out
- This is a heterodox tradition. The mainstream account of the same phenomenon differs, and nothing here is a forecast or a view on any security.
- These are studies of advanced economies over a long period. Nothing in them is a statement about current credit conditions anywhere, and nothing here is a forecast.
- The result establishes a relationship between credit and crises. It does not establish a mechanism, and this course does not use it to choose between the accounts in the lessons before it.