Lost in the euphoria over second quarter GDP revisions is the ongoing corporate struggle in terms of profits and how that suggests a more than reasonable proportion of GDP’s ability to measure the economy is overstated. To this point, I had focused more so on the productivity problems as they related to a potential over-optimism of the labor market but there is a good deal of inequality within “output” measures that suggest the same condition (which would make the labor statistics even more exaggerated).
This is not conjecture on my part, at least in the case of the recent past. With GDP benchmark revisions last month, the 2012 slowdown that was only believed a minor and transitory denouement was transformed into a near-recession in its own right. Those revisions, with the “advance” estimate for Q2 2015 GDP, now include corporate profits.
Before getting to that, some background is necessary. GDP is the expenditure side of the economy, at least as formulated by orthodox thinking about how to measure such things. As orthodoxy takes it, for every dollar spent there is a dollar “earned” so GDP is matched by GDI, spending to income; or should be. While there is fair amount of theoretical weakness in that framing already, the focus here is further inward upon how even on these terms there is far less validity than otherwise it may seem. That starts, as always, with trend-cycle subjectivity.
There is typically some appreciation of GDP as to what it attempts to accomplish, but almost none for how it goes about it. Even the literature that the BEA publishes in an attempt toward transparency, and it is mostly laudable, good work to that end, is probably impenetrable to the layperson. That density is often matched by sheer volume, which is understandable since the task here is to attempt single measurement of the entire US economy.
Given those, however, there are windows of easier opportunity for understanding the construction of the statistics especially from a more general point of view. Trend-cycle, for all the mathematical babble, is really quite simple – economists estimate how much of variation is due to expected contribution (loosely stated) of the state of the economic cycle. This is not something actually measured but rather a statistical attempt by pure, gross bias so that GDP matches what economists (mostly) perceive about the business cycle that there is a “smoothed” formulation for public view.
The reason for this is that statistical thinking about variation has derived economic trend as the greatest source of it; which does make some sense, given that moving from “boom” to recession (or back again) would produce the most variation in economic accounts. It does not immediately follow, however, as to why the BEA should interject itself into that point regularly, nor is it clear at all how they actually do so.
We do have some quantitative elements to this partiality within GDP, as the BEA itself notes that at the third estimate for each quarterly update 12% of GDP is based on “trend” estimates alone and more than one-fifth additionally upon “indirect” sources.
Third estimates of both GDP and GDI are published for each quarter. At the time of the third estimates, about 67 percent of the source data for GDP components are based on comprehensive data or direct indicators, 21 percent on indirect indicators, and 12 percent on trend-based estimates. In comparison, about 14 percent of the source data for the third estimates of GDI components is from direct sources, 56 percent from indirect sources, and 30 percent from trend-based estimates. As a result, considerably more judgment goes into the construction of the third estimates of GDI than into the third estimates of GDP.
In a nominal economy of $17.9 trillion, having $2.15 trillion come out as nothing more than subjective “trend” analysis should engender quite a little bit of skepticism especially in the case of this current “cycle.” Notice, also, that the “trend” component of GDI, the income side, is far greater at nearly one-third.
This is why the last benchmark revision was so sharp and downwardly severe surrounding 2012/early 2013. It was clear in more real-time against other economic estimates (retail sales, industrial production, global trade, etc.,) that there was a slowdown in the economy that had not abated beyond the single quarter assigned by the BEA’s trend bias. The GDI side, however, has by the annual revisions far less “trend” components surviving (compared to the quarterly updates) due to the annual IRS’s Statistics of Income database. The BEA notes that by the second annual revision, there is almost no “trend” calculation left (or even “indirect” sources).
With that in mind, it wasn’t surprising to see overall GDP revised into the 2012 slowdown as the GDI side particularly in terms of untouched (by further adjustments) corporate profits and cash flow had stopped growing toward the end of 2011.
Not only had the economy, from this perspective largely free of trend-cycle, stopped growing by 2012 it has been remarkably unstable ever since. The GDP side has certainly picked that part up despite all attempts to reduce it to “residual seasonality”, but still preserving, as you would expect of orthodox tendencies, the over-optimistic bias. We can see that quite clearly within the “suite” of profit estimates that are “massaged” to varying degrees by GDP crossover imputations (accounts like IVA and CCadj which attempt to put GDP components together with GDI statistics).
Where corporate cash flow and profits from current production received very little adjustment in this last major benchmark revision, other profit measures were severely downgraded with GDP.
These profit accounts take on the now-familiar 2012 flatline where growth and recovery was once counted as assured. That, of course, immediately raises the questions about 2014 and especially 2015; if 2013’s rebound and growth was nothing but bias erased with a hindsight revision against prior “trend” estimates, why wouldn’t the same apply to 2014 and 2015? This is particularly relevant for this year given that so many economic accounts away from GDP and its “trend” bias are even worse than in 2012; to the point of some serious contractions that just don’t seem to fit the GDP illusion. It was, after all, the vision of the economy from those other measures, the “bad” view, that GDP has now essentially confirmed and adopted retroactively only upon being stripped by benchmark revisions (of more “direct” data; thus less “trend” and “indirect”) of its prior subjectivity.
To add more doubt about GDP so far this year, corporate cash flow and profits from current production continue to sputter and decline; year-over-year both are down about half of a percent in Q2 after managing small gains even in the seasonality of Q1. And that small contraction comes after a revision in cash flow, in particular, from the middle of last year:
While the rest of the world cheers the GDP revisions, the GDI side will get far less attention especially since it was just +0.6% in Q2 compared to nearly 4% for GDP! That means for the first half of 2015, real GDI shows only +0.5% and that includes a 30% “trend” over-optimism component. We know that by past history once that bias is removed into the annual revisions, as the BEA itself shows a pretty stark upward bias in GDI corporate profits just prior to the last two recessions.
This is all not to say that the second quarter in the US economy wasn’t better (which isn’t much of standard) than the first, but given a comprehensive review of the actual situation and the manner in which it is largely and mathematically described, it is once again (as 2012-13) very likely being highly overstated which is why the media and economists set themselves into repeated confusion about what comes next. The most that can be said of all this is that GDP has signaled the very clear unstable nature of the economy particularly after 2012. The rest is just subjective faith and inventory.