The whole point of seasonally adjusting key economic indicators is to smooth distortions, so we can see through the quirks of the calendar and get at the underlying trend. But what if those seasonal adjustments were twisting the information like a fun-house mirror?
Thanks to the Great Recession (the gift that keeps on giving), that may be exactly what has happened over the past couple of months. Seasonal adjustments may be dressing up economic data to look better than they actually are – sending the markets false signals of well-being, and setting us up for disappointment.
Andrew Tilton, an economist with Goldman Sachs, says the financial crisis – when many indicators plunged to their lowest levels since the Great Depression in the winter of 2008-09, then rebounded sharply in the summer of 2009 – is distorting seasonal-adjustment formulas for some of the market's most closely watched economic measures, including U.S. employment, housing starts and the Institute for Supply Management's manufacturing index.
Seasonal adjustments compare the average value of an economic measure for a given month against the average for all other months over the past few years (anywhere from three to seven years is typical), then tailor the data to account for routine seasonal variations. These adjustments typically provide the biggest boost to economic data in the winter months, when many forms of economic activity are routinely slower because of cold weather, while providing the biggest drag in the robust spring and summer. Those peak and valley periods are precisely where the 2008-09 extremes landed – and that has exaggerated the average seasonal swings used in calculating seasonal adjustments.
Now, economists aren't oblivious to this danger. Seasonal-adjustment formulas do, typically, attempt to identify outlier economic years and dampen their distorting effect on the broader seasonal calculations. But the 2008-09 events were so far out of the norm that they have confounded these attempts.
The evidence is in the data. Mr. Tilton found that the gap between seasonally adjusted figures and unadjusted figures for many of the most important U.S. economic indicators is significantly wider, in both the peak adjustment season and the trough, than it was prior to the financial crisis. He estimates that this widening accounted for nearly one-third of the upside economic surprises from December, 2010, through February, 2011 – a string of apparently good economic news that helped fuel a stock market rally.
It may be happening again. The S&P 500 is up nearly 15 per cent since late November, at the same time that Citigroup's U.S. Economic Surprise Index spiked to near-record levels.
And this winter, there's an added wrinkle to the seasonal distortions: The weather.
Normally, seasonal adjustment corrects for slowdowns in economic activity that stem from inclement weather – which winter, particularly in the industrial heartland of North America, usually has in spades. But this winter has so far been unusually mild, creating conditions for unseasonably brisk economic activity. The statistics-based seasonal adjustments can't capture this; they are upwardly adjusting data to account for economically stifling weather that simply isn't there.
Toss in the fact that U.S. Thanksgiving – a critical date for retail sales and the labour market – was unusually early last year, and you have a quagmire of seasonal anomalies mucking up the recent economic numbers.
Mr. Tilton insists that the bulk of the improvement in the economic data – both last winter and now – reflects genuine improvements in economic activity. Nevertheless, the magnitude of the improvements are being exaggerated on a seasonally adjusted basis. And we'll pay the price in a few months – when downward seasonal adjustments will underestimate economic performance.
In stock markets, where values are linked inextricably to growth expectations, that's a dangerous condition. Far from smoothing the economy's contribution to market sentiment, it would appear seasonal adjustments are going to add to the volatility.
David Parkinson