,
12 tweets,
4 min read
Read on Twitter
Many private-sector economists use the concept of "credit-impulse" as if it was totally obvious that this is a reasonable and valid indicator. But this concept does not belong to the language of economics! I would like to make a few points. 1/ 
It was coined in 2008 by Michael Biggs, then an economist at Deutsche Bank. He found that a suitably transformed measure of credit flows had a strong contemporaneous correlation with GDP growth. He also wrote an (unpublished) paper on the issue. 2/ dnb.nl/binaries/Worki…
Since then, sell-side analysts have used it extensively, usually in plots that suggest it is a powerful predictor (driver?) of (future?) growth. But it has not picked up among mainstream economists. Why? 3/
First (how many times do we need to say this?) CORRELATION IS NOT CAUSATION. The habit of plotting two time series together and the name "impulse" suggests a particular direction of causality: from credit to growth. That's not necessarily the case! 4/
I could easily come up with the "Investment Impulse", or the "PMI impulse" or the "Consumer Confidence Impulse". In Macroeconomics EVERYTHING is correlated. Pick two random macro aggregates, chances are they will be correlated over the business cycle!
It reminds me of the old debate of Money vs Income. Which one causes which? Chris Sims wrote extensively about it, and went on to receive a Nobel prize for it. See here for a nice summary: economistsview.typepad.com/economistsview… 6/
We have the tools to (try to) address these issues. Running a VAR with additional covariates and Granger-causality tests are the bare minimum that an honest, applied economist should be expected to perform. Time to end with DISHONEST "casual econometrics"! 7/
All of this is not to say that credit is unimportant for output fluctuations. We know that it is. But it becomes crucial to identify EXOGENOUS movements in credit that reflect changes in the risk bearing capacity of the financial system, rather than responses to credit demand. 8/
The substantial body of research since 2008 suggests to focus instead on PRICE measures, i.e. the External Finance Premium as a measure of binding financial constraints. See, e.g. the paper by Gertler and Gilchrist. 9/ pubs.aeaweb.org/doi/pdfplus/10…
This is the object that MODERN MACRO theory (which includes financial frictions) tells us we should go out and measure. Not some arbitrary transformation of an arbitrary credit flow. G&G explain why this is better than looking at quantities. 10/
A FAR BETTER measure, which has been shown to have substantial predictive power for GDP growth, is the Excess Bond Premium developed by Gilchrist and Zakrajšek. It is conveniently updated by the Fed Board, so there is NO EXCUSE for not using it! federalreserve.gov/econresdata/no…
In general, this is another manifestation of the BAD PRACTICE of the Financial-Entertainment Complex of plotting correlated series (with loose theoretical grounding) to "suggest" causation in a hand-wavy way. BAD! \End
