Thursday, 13 December 2018

Our financial cycles

Two economists at the University of Auckland, Caitlin Davies and Prasanna Gai, have come up with a really useful bit of practical macroeconomics. They've devised a Financial Cycle measure for New Zealand - an indicator of the overall tightness or looseness of financial conditions. Their paper, 'The New Zealand financial cycle 1968–2017', is available here in the online version of New Zealand Economic Papers.

Financial conditions indices (FCIs) are an established thing overseas. They were always relevant - changes in financial conditions have played a lead part in many business cycles, and even when not the lead have been important channels for propagating non-financial shocks - but have naturally become of greater interest since the GFC. As Davies and Gai say (p1), "Recent [academic] work ... suggests that strong credit growth and house price booms are good predictors of crisis and significantly shape macroeconomic outturns".

In the States, for example, there are a herd of them. The chart below shows five FCIs - three produced by various regional Federal Reserve Banks (Chicago, Kansas City, St Louis), two by the private sector (Bloomberg, Goldman Sachs) - plus a market-derived measure, the VIX, which is the volatility investors expect from holding the S&P500 index and which can be backed out of the prices for S&P500 options. They've all been normalised to be comparable, as described here by the St Louis Fed. Higher values for these indices mean tighter conditions.

You can see, for example, how financial conditions (ex the VIX) had been tightening ahead of the GFC, and then hit all-time highs for financial distress and unavailability of credit through the GFC itself. And if you subscribe to Austrian or Minsky style theories of business cycles, you'd argue that the pronounced period of unusually easy monetary conditions you can see in 2004-2006 sowed the seeds for the over-exuberant risk-taking that fuelled the eventual GFC bust.

Highly useful and informative things, these FCIs. And now we've got one of our own. To get it, Davies and Gai went the principal components route - take a bunch of finance and credit indicators, and see if there's a common influencing component in the background - and found that yes, there was. It combines six variables into an overall index: real credit, credit to GDP, credit to the M3 measure of money supply, real house prices, real share prices, and housebuilding to GDP.

Here's what the results look like, in raw form: for this New Zealand index, you read it the other way round to the US ones, in that higher values show easier financial conditions. The authors say that "Our measure of the New Zealand financial cycle appears to be broadly consistent with the main economic developments during the period", and I agree. You can see, for example, the surge in credit availability in the mid 1980s following banking deregulation, and the subsequent bust after the 1987 sharemarket crash. You can also see our experience of the GFC.

The authors usefully superimposed their financial index results on the business cycles identified by Viv Hall and John McDermott. The FCI for this comparison has been expressed in smoothed cyclical terms showing whether it is rising or falling (there's econometrics behind this we don't need to explore here), but same diff. Here's how they compare.

"Of the six contractionary episodes during the past fifty years, five occur less than three years after a peak in the financial cycle", the authors say (p8), and while I wouldn't immediately leap to cause and effect (and they don't either), it's a suggestive pattern.

The authors modestly say (p14) that their work "should be regarded as a tentative first-step in constructing a set of stylised facts on the financial cycle in New Zealand", but it's more than that. We had a rather large gap in documenting our recent macro history, and they've filled it. They've also created something that could easily be kept up to date, and serve as a real-time indicator of trouble brewing. As they mention, it's of obvious relevance to macroprudential policy: you might want to keep a weather eye out for where the FCI is before, for example, tightening or loosening LVRs. Indeed, you'd wonder why the RBNZ hadn't developed an FCI of their own by now.

And you can see extensions to it. This FCI is based on whatever quarterly series were available all the way back to 1968, and for a paper looking at the grand sweep of history, that's fine and unavoidable. But you can easily imagine an FCI using data that became available only more recently:I think it's highly likely, for example, that moves in corporate credit spreads, unavailable back to 1968 but available for more recent years, would feature strongly. And I think it's plausible that you could get to a monthly FCI: the Americans certainly have, and the Fed of Chicago has even gone as far as producing a weekly one (conditions are currently on the easy side of normal).

In any event this is a great start: I hope there's someone out there - the RB? a bank? the University itself? - who'll take up the baton and turn this into an ongoing up-to-date macro indicator.

No comments:

Post a Comment