In my experience, though the really big economic/financial crises shake the seismograph, they don't arrive like earthquakes out of the blue - rather, there's a steady and enduring stream of disappointment lasting usually years before the cashflow and balance sheet consequences become unendurable. In the late 1990s, SE Asia's terms of trade were quietly falling apart for years, eroding cashflows and current accounts, before the consequences became clear in 1997/98. In the US, new home sales peaked in mid-2005 and spent the next three and a half years grinding relentlessly lower before finally bad debts (or rather, the attempt to 'insure' them by the CDS market) brought the Reaper's sickle to the banking system.
It's not a rule, but it is a tendency. And so to today's topic: the danger implicit in a perceived collapse in consumer confidence. This, after all, is what is now dominating the headlines - just this morning we had the Nielsen global survey which concluded that we are at our most miserable since early 2009. Last week we had the Uni of Michigan US consumer confidence shocker also serving up the same conclusion. But because the really serious problems grind slow but grind small, I tend to view short-term movements in consumer confidence readings with considerable caution. Are they cause, or are they effect, or are they, indeed, anything truly measurable in the first place?
First, let's look at what is happening to measured global consumer confidence. To create this global index I've used subindexes in the US, Eurozone and NE Asia (including China), and as usual weighted them according to 5yr GDP averages.
The first thing that strikes me is that although global confidence has certainly taken a knock, we're not really back at early 2009 levels. The Uni of Michigan may or may not be accurate in that story, but it's not one which yet has purchase in the Eurozone (105.1 in June 2011 plays 70.6 in March 2009), or in NE Asia (97.2 in June plays a low of 81.7 in Jan 2009).
Now let's look at the relationship between confidence and demand momentum. Using the great statistical method of smoothing to a 6m purl and then eyeballing, it seems that usually there's a pretty good fit between the consumer confidence measured, and direct measurements of domestic demand.
But one needs to go considerably beyond eyeballing the smoothed data to explore any seeming relationship. What we are looking for is evidence that a change in consumer confidence has a useful correlation between a change in domestic demand momentum - with no smoothing or averaging allowed. I've looked at the change occurring over three months (ie, the changes in both measures between, say, December and September) between 2000 and now. This shows there's a correlation coefficient of 0.24 between the changes in these two over 134 observations - sailing past the 1% significance test. But the coefficient gets even stronger if you view consumer confidence as being a lagging indicator: it rises to 0.256 on a one-month lag, and 0.349 on a 2 month lag (this is its peak).
If you treat confidence as a leading indicator, however, the coefficient rapidly dribbles away to insignificance.
The moral? Those surveys we're watching - the Nielsen, the Conference Board, the University of Michigan - they're more likely to be lagging indicators than leading. Perhaps we're miserable because we're not spending, rather than not spending because we're miserable.
But if that's true (and it's not a nice thought) it suggests there's a more powerful reason that mere transient misery why we're not spending, if we're not spending. And so on we go, down the analytical trail.