DRI-364 for week of 9-23-12: Revisiting the DRI and Related Economic Indices

An Access Advertising EconBrief:

Revisiting the DRI and Related Economic Indices

Periodically this space has revisited Access Advertising’s Driver Recruiting Index (DRI). This index tries to estimate the real-time ex ante demand for commercial drivers. It samples the classified ads placed to recruit drivers in 32 geographically dispersed major-metropolitan newspapers throughout the U.S. Every now and then, we place the DRI alongside other trucking, transport and freight indices and gauge its movements in relation to indices of overall income and employment.

The DRI in the 2nd and 3rd Quarters

Like the overall economy, the DRI got off to a reasonably fast start after a promising end to 2011. But it soon became clear that, like the U.S. economy, the DRI was going to fall short of expectations (let alone hopes) in 2012. The Index pierced the 200 barrier for the only time so far this year on March 25, flirting with a raw score of 500 (an achievement it exceeded 14 times in 2011).

After that, it has been all downhill. Year-over-year comparisons have occasionally fallen over 20% short of 2011; 10% shortfalls have been commonplace. Driver demand has never really caught fire, peaking in late spring and skipping both the normal summer high and the fall upturn.

Second-quarter GDP growth was also disappointing, ending up at a tepid 1.7%, even below the moderate 2% in first quarter. Although unemployment has now declined fractionally to 8.1%, this reflects only the decline in the size of the labor force. The number of new jobs created has continued to languish.

The only good news has been a Wall Street rally engineered largely in response to the latest quantitative easing by Ben Bernanke’s Federal Reserve. This is probably responsible for a recent upsurge in consumer confidence in the economy.

The DRI as Economic Indicator

If the DRI’s disappointing performance has mirrored that of general economic indicators, this very congruence speaks well for the DRI’s performance as an economic indicator. Trucking handles some two-thirds of all freight by volume and about 80% by value; many production inputs and final goods travel the nation’s roads and highways. We expect changes in trucking activity and driving demand to track overall trends in production, income and employment. (Whether the DRI is or should be a leading, coincident or lagging indicator is a complicated question that will be broached later.)

The DRI continues to display other desirable properties as well. One of those is (relative) stability. The very ubiquity and prominence of trucking strongly suggests that we should not expect to find the Index fluctuating widely from week to week. While trucking firms ordinarily experience high rates of turnover, this is a long-term phenomenon, responsive to demographic factors (average age, cultural shifts) and cyclical variables (wage changes, politico-regulatory changes). A particular trucking firm may well experience a sudden, substantial need for drivers, but this will usually represent a geographic shift of demand that is offsetting in the aggregate. Significant increases or decreases, when they occur, are usually persistent, cyclical changes in trend, not random fluctuations.

For a concrete illustration, compare the DRI with TransCore’s DAT North American Freight Index, which compiles data from the company’s load board network in the U.S. and Canada. Neither the DRI nor the NAFI is seasonally adjusted, but the contrast in variability is stark. Earlier in 2012, NADI racked up this record of month-over-month fluctuations: March 2012 – up 40%, April 2012 – up 3.5%, May 2012 – up 1.7%, June 2012 – down 2%, July 2012 – down 20%, August – up 1.1%. Monthly fluctuations of 20% in the DRI would be observed only at a seasonal or cyclical peak or turning point, and a 40% monthly change is unheard of.

The DRI and Comparable Economic Indicators

Given the importance of trucking in the production chain, it is surprising that the DRI is one of few time-tested, reliable trucking indices – and the only one to track driver demand. A rundown of the others reaffirms our understanding of what makes for a good index and confirms the DRI’s recent congruence with its brethren.

The Cass Freight Index compiles data from the expenditures of 350 of the largest freight shippers. Early in the year, the Index flexed its muscles with a 2.5% increase in February 2012. Then, along with the economy at large, it began to lose vitality. Increases diminished to 2.1% in March, 1.9% in April, 1.8% in May and 1.3% in June. In July, the Index turned negative with a 0.1% decrease, followed by a 1.1% fall in August. Sponsors and analysts cited a buildup of inventories, a fall in international trade and recent declines in manufacturing output as recorded by the Institute of Supply Management’s index, which had fallen for three straight months.

The American Trucking Associations are the Establishment of the trucking industry and Chief Economist Bob Costello is the voice of starched, high-collared authority. The ATA Truck Tonnage Index is perhaps the most widely cited trucking index, not only by private-sector analysts but even by the federal government. Although the ATA’s respect for authority imparts a big-government bias to its occasional obiter dicta, its economic data are accorded the utmost respect.

The TTI’s recent numbers paint the same by-now-familiar picture of a stalling, sluggish trucking sector. March saw a 0.2% increase, April improved to 1.1% but May backtracked to 0.9%. June went back up to 1.1% but July was unchanged. Typically, Costello’s public comments wrapped the language of an economist inside the rhetoric of a politician. He depicted “…an economy that has lost some steam but hasn’t stalled.” The ATA’s corporate commentary was more telling, noting that “…the index… has been moving mostly sideways in 2012.”

The Truck Tonnage Index is an important component in the federal government’s Transportation Services Index (TSI), which is published by the Bureau of Labor Statistics (BLS). For most of 2012, this index has alternated back and forth with little net increase to show for the year. In February, the TSI rose 0.5%, but it fell back 0.8% in March. April brought a modest rebound of 0.2% but in May the Index was unchanged. June and July saw offsetting 0.1% changes.

These indices display the bedrock virtues noted above: stability and congruence with general economic activity and with each other. One of the rising economic forecasting stars of recent years was the Ceridian Pulse of Commerce Index (PCI), compiled by respected econometrician Edward Leamer of UCLA. After recording some lackluster increases earlier this year, the Index ceased publishing somewhat mysteriously following its May release. Ceridian has responded to queries by saying that although the Index does not now publish its results, it is contemplating a return on a subscription basis and is soliciting indications of interest among potential customers.

Is the DRI a Leading, Lagging or Coincident Indicator?

Undergraduate students of economics are taught that economic indicators come in three flavors – leading, coincident and lagging. These indicate whether changes in the indicator lead, accompany or lag changes in the general level of economic activity. Since the unknown future preoccupies our attention, leading indicators are especially studied and prized.

There has long been a casual presumption that trucking indices are, or at least should be, leading indicators. Disruptive phenomena like layoffs and unemployment are presumably necessitated by the accumulation of unsold goods. Since trucks carry goods and the inputs necessary to produce them, freight shipments should register the incidence of cutbacks in production and materials. This is the sort of logic that supports the categorization of trucking as a leading indicator.

Early in its life, though, the DRI was observed to be a lagging indicator. We rationalized this as the caution of recruiters, who – unsure and suspicious of the depth and duration of any increases in freight supply after the Great Recession – waited to verify the persistence of demand before incurring the fixed costs of hiring. In fact, similar behavior had been recorded in connection with the Index of Classified Advertising, an economic index that bears a strong family resemblance to the DRI.

Since leading and lagging indicators are at opposite poles, this plants a seed of skepticism about the traditional taxonomy of economic indicators. Further consideration should nourish that thought.

The inherent logic of leading economic indicators says that they can be used to predict the economic future – or at least the turning points of business cycles. And this is simply impossible.

Anybody who can accurately predict the future onset of a recession – or, for that matter, the end date of one – can earn a fortune by so doing. Anybody who can repeat the performance reliably can become fabulously rich. Business forecasters are not fabulously rich. Ergo, they cannot predict the beginning or ends of recessions. Not reliably, anyway.

This must mean that leading economic indicators do not really lead. Maybe they don’t even indicate. In any case, something isn’t quite kosher. The question is: what?

The Theory of the Business Cycle – Such As It Is

Something else that all undergraduate students of economic statistics and econometrics learn is that economic theory and logic form the basis for all empirical work. Without theory, we cannot know what data to collect or what relationships to posit between the variables that make up the data. So an economic model is borrowed or created to embody the relationships upon which data is collected. Only then can the data gathering and testing begin.

But in business forecasting that runs us smack up against a formidable obstacle. There is no generally accepted business-cycle theory. The categories used by the National Bureau of Economic Research to codify episodes of the business cycle – expansion, peak, contraction and trough – were developed by institutional economist Wesley Mitchell in the early 1900s, based on years of observation and study of past recessions. Unfortunately, observation is not theory.

The national income and product accounts used to compile U.S. economic data were developed later, based largely on the economic categories developed by English economist John Maynard Keynes in his influential work The General Theory of Employment Interest and Money. But Keynes explicitly denied that the General Theory contained any theory of the business cycle. He simply declared that capitalist economies suffered from a chronic shortage of aggregate demand – e.g., private spending by households and producers. He never said why the shortage existed. Government should make up for this shortfall, Keynes maintained, by running budget deficits that increased the aggregate total of aggregate demand or spending. Essentially, government should commit to purchasing whatever volume of output is left unbought by households and producers, so as to insure full employment. Keynes didn’t supply a theory to account for the business cycle, only a purported remedy to cure the symptoms.

Without knowing where the shortfall in spending lies, we cannot predict where to look for leading indicators at any stage of the business cycle. There is one school of thought whose business-cycle theory offers general advice on this point. That is, it doesn’t offer a list of specific industries, sectors or indicators as such, but instead provides advice about where to look for them in general cases.

The Austrian theory of the business cycle pinpoints monetary expansion by government – probably supervised by a central bank – as the proximate culprit behind recessions. By driving interest rate below the “natural rate of interest,” the rate that would equate the saving households and producers want to do, the money creation will make interest rates artificially low. This makes long-lived, capital-intensive production processes artificially attractive to producers. In turn, this creates a “bubble,” or artificial excess prosperity, in the sector thus favored. While it lasts, this bubble can seem deliriously prosperous, almost too good to be true. That is because it is too good to be true, as the drawn-out period of housing prosperity in the U.S. proved. But when the bubble bursts – owing either to a rise in interest rates or a rising burden of debt – the artificially-prosperous sectors are the first to crash. They are the leading indicators of the coming recession.

In general, then, leading indicators are those pertaining to the sectors receiving the artificial encouragement. In the U.S. during the 1990s and early 2000s, this would have been the housing sector. But this does not mean, as many have implied or outright insisted, that the housing sector should be the first to recover its balance. And it does not mean that nobody can recover until the housing sector does. Indeed, the reverse is more nearly true – housing should be the last sector to recover fully. It also means that the stubborn attempts to stage-manage recovery in housing by holding interest rates low and artificially raise housing prices through government purchases of mortgages and securities are counterproductive. After all, this is exactly the process that caused the problem in the first place – repeating it merely enlarges the backlog of adjustments that must occur before recovery can take place.

An Austrian Look at Economic Indicators

The effect of this theory on forecasting practice is striking. Housing becomes a leading indicator for the downturn phase because it is a long-lived production process, extremely sensitive to interest rates. But it is a lagging indicator for the upturn; it cannot recover until all of the bad investments made during the bubble phase are liquidated and their resources reallocated.

What about trucking? Well, trucking feeds the housing industry its materials and some of its manpower, so trucking shares this dual forecasting status. Sometimes it will be a leading indicator, sometimes a lagging one. Moreover, trucking feeds other industries as well, so it simply cannot be pigeonholed by a simplistic taxonomy. A sophisticated approach to business cycles requires us to abandon our primitive definitions of economic indicators. First, we must classify industries and sectors according to their status in the production and consumption hierarchy. Second, we must recognize the difference between recession and recovery.

We have still not exhausted the reserves of analysis, since there is still the possibility of inherent cyclical movements in economic activity that are not driven by monetary mistakes made by the authorities. The fact that money substitutes for barter in allowing human beings to trade the product of their labors solves huge problems, but it also creates smaller ones. These subtle problems may well mean that we have to live with an unavoidable element of cyclical instability in our economic life. And this may demand still more adjustments in the terminology of forecasting.

Progress Report on the DRI

The DRI is completing its fourth year of operation. It has passed the standard tests of stability, reliability and usefulness that apply to economic indices. Of course, it has not unlocked the door to wealth and fame available to any economic index that could actually forecast the future – but, as we have seen, that is a chimera. Explaining the past and recognizing the present is tough enough and a worthy goal for any economic index. Many a worthwhile aspirant has fallen short of even this limited objective; Ceridian’s PCI may be the latest addition to this list.

Nearly four years in, the DRI is still trucking.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s