U.S. Wages and Egyptian President Employment
First things first: you didn't miss yesterday's comment. I didn't write one. The ice/snow storm so impacted everything, from mass transit to market volumes and volatility to website hits, that I decided I didn't have a lot to say. Contrary to my usual practice, I therefore said nothing.
Today was more interesting, however. Bond markets and stock markets have more or less lost interest in the revolution in Tunisia, the amazingly persistent unrest in Egypt (today Mubarak told ABC that he is "fed up" with being President and wants to leave but is afraid it would lead to real chaos), the demonstrations in Yemen and developing unrest elsewhere in the general region. Yeah, I can't imagine what effect unrest in the Middle East could cause our economy.
That's sarcasm, of course.
Rising energy prices, if they rise for demand-related reasons, needn't be a major concern. Such a price rise acts as one of the "automatic stabilizers" and, while it pushes up consumer prices, it also acts to slow the economy. This helps reduce the need for the monetary authority to meddle (not that anything has stopped them any time recently). It doesn't need to respond to higher (demand-induced) energy prices, because those higher prices are serving the usual rationing function of higher prices vis a vis scarce resources.
But when energy prices (or, to a lesser extent, food prices) rise because of supply-side constraints - say, reduced traffic through the Suez Canal, or fewer oil workers manning the pumps in a major oil exporting region - then that's extremely difficult for the central bank to deal with. More-costly energy will slow the economy inordinately, and higher prices also translate into higher inflation readings so that if the central bank responds to the economic slowdown they risk adding to the inflationary pressures.
Equity investors, though, are just going with the momentum. Eat, drink and be merry! Since energy prices are only at $90/bbl, and haven't exactly reacted with nervousness to the Middle-Eastern unrest, stock market participants can be excused for refusing to try and see around corners. For now.
Bonds now have a double reason to decline. Recent signs of mildly percolating domestic growth on the one hand, but the potential of supply-side shackles on any Fed desire to tighten policy. 10-year yields today reached the highest level seen since last May (for real yields, the highest level since July at 1.22%), and I expect yields to move towards the post-crisis highs of 4% on the 10y note, Fed purchases or no.
Domestic growth is percolating, to be sure, but as yet only mildly. Initial Claims today showed at 415k, near expectations. I mentioned on January 13th that the underlying pace seemed to me to be "about" 420k, and repeated that last week. (http://mikeashton.wordpress.com/2011/01/27/what-to-make-of-this/) I point that out mainly in amazement, since I don't expect to have a lot of success with forecasting Claims and don't generally try to do it. But the week-to-week numbers aren't that important anyway; the important part is being able to identify when something big has changed. So far, it hasn't. The movement from about 460k to about 420k in Claims is an improvement, but hardly a "big" improvement.
One of the ways that we can restrain ourselves from getting too excited, too soon, about the upturn in employment is to reflect on the fact that surveys still indicate considerable uncertainty and pessimism among the people who are vying for those jobs (or clinging to the ones they have, hoping they don't have to compete for those scarce openings). This is illustrated by the apparent puzzle that Unit Labor Costs (reported yesterday) remain under serious pressure and Productivity continues to rise at the same time that profit margins are already extremely fat. Rising productivity is normal early in an expansion, but the bullish economists tell us that the expansion started a year and a half ago. We're about halfway through the duration of the average economic expansion (if you believe the bulls). And fat profit margins are not as normal early in an expansion.
Now, we don't measure Productivity and Unit Labor Costs very well at all. Former Fed Chairman Greenspan used to say that we need 5 years of data before we can spot a change in trend, and he may be low. But it seems plausible that there remains downward pressure on wages. Call it the "industrial reserve army of the unemployed" effect. While job prospects are improving, they are apparently not improving enough yet for employed people to start pressing their corporate overlords to spread more of the profits around to the proletariat.
Fear not, however, that this restrains inflation. The evidence that wage pressures lead to price pressures (and conversely, the absence of wage pressures suggest an absence of price pressures) is basically non-existent. Let me present two quick charts that make the point simply.
No surprise: tighter conditions in the labor market tend to be associate with wage inflation.
The chart above (Source for data: Bloomberg) shows the relationship between the Unemployment Rate and the (contemporaneous) year-on-year rise in Average Hourly Earnings. I have divided the chart into four phases: 1975-1982 (a period which runs from roughly the end of wage-and-price controls in mid-1974 until the abandoning of the monetarist experiment near the end of 1982), a "transition period" of 1983-1984, the period of 1985-2007 (the "modern pre-crisis experience"), and a rump period of the crisis until now. Several interesting results obtain.
First of all, there should be no surprise that that the supply curve for labor has the shape it does: when the pool of available labor is low, the price of that labor rises more rapidly; when the pool of available labor is high, the price of that labor rises more slowly. Labor is like any other good or service; it gets cheaper if there's more of it for sale! What is interesting as well is that abstracting from the "transition period," the slopes of these two regressions are very similar: in each case, a 1% decline in the Unemployment Rate increases wage gains by about ½% per annum. Including the rump period changes the slope of the relationship slightly, but not the sign. This may well be another "transition" period leading to a permanent shift in the tradeoff of Unemployment versus wage inflation.
But clearly, then, when Unemployment is high we can safely conclude that since there are no wage pressures there should be no price pressures, right?
The Phillips Blob
The second chart puts paid to that myth. It shows the same periods, but plots changes in core CPI, rather than Hourly Earnings, as a function of the Unemployment Rate. This is the famous "Phillips Curve" that postulates an inverse relationship between unemployment and inflation. The problem with this elegant and intuitive theory is that the facts, inconveniently, refuse to provide much support.
Why does it make sense that wages can be closely related to unemployment, but inflation is not? Well, labor is just one factor of production, and retail prices are not typically set on a labor-cost-plus basis but rather reflect (a) the cost of labor, (b) the cost of capital, (c) the proportion of labor to capital, and importantly (d) the rate of substitution between labor and capital. This last point is crucial, and it is important to realize that the rate of labor/capital substitution is not constant (nor even particularly stable). When capital behaves more like a substitute for labor, a plant owner can keep customer prices in check and sustain margins at the same time by deepening capital. This shows up as increased productivity, and causes the relationship between wages and end product prices to decouple. Indeed, in the second chart above the R2s for both periods is...zero!
This isn't some discovery that no one has stumbled upon before. In a wonderful paper published in 2000, Gregory Hess and Mark Schweitzer at the Cleveland Fed wrote that
It turns out that the vast majority of the published evidence suggests that there is little reason to believe that wage inflation causes price inflation. In fact, it is more often found that price inflation causes wage inflation. Our recent research, which updates and expands on the current literature, also provides little support for the view that wage gains cause inflation. Moreover, wage inflation does a very poor job of predicting price inflation throughout the 1990s, while money growth and productivity growth sometimes do a better job. The policy conclusion to be drawn is that wage inflation, whether measured using labor compensation, wages, or unit-labor-costs growth, is not a reliable predictor of inflationary pressures. Inflation can strike unexpectedly without any evidence from the labor market.
The real mystery is why million-dollar economists, who have access to the exact same data, continue to propagate the myth that wage-push inflation exists. If it does, there is no evidence of it.
So you can be assured that in tomorrow's Employment Report my eyes will not be glued to the Average Hourly Earnings number, except to record it as a datum that other investors might lean on too much. I will in fact try, probably too much, to downplay the importance of the number in my mind altogether for a number of reasons:
- It is a January figure, and difficult to seasonally-adjust.
- There was unusually bad weather; the economists I've surveyed see that as reducing Payrolls around 50-75k. The consensus estimate is 143k against 103k last month (which, recall, was a terrible disappointment), and since ADP was at 187k this is a reasonable guess. (ADP is not as affected by the weather since it counts people on the payroll while for Payrolls a person actually needs to be paid for work during the survey week). Presumably, the BLS will provide an estimate if an unusual number of people reported being "not able to work due to inclement weather," but then again...maybe they won't.
- The annual benchmark revisions are also due. In October the BLS pre-estimated that they would drop the level of Employment at March 2010 by -366,000, but there are nine more volatile months in there. A large negative number will blunt some of the positive news if there is an upside surprise.
- The Unemployment Rate should rise. It plunged from 9.8% to 9.4% due to a new 26-year-low in the labor force participation rate. If there are glimmerings of life in the job market, it should begin to show up in the form of increased participation, and that will have the perverse effect of keeping the Unemployment Rate up or increasing it (Consensus for tomorrow is a rise to 9.5%).
Finally, there is the fact that recently, economic data hasn't really mattered very much to market pricing. Global geopolitics and the evolution of the crisis in Europe (or its abrupt disappearance this week) have had bigger impacts on markets. And on that basis, the bigger employment news on Friday may be if the unemployment rate on Egyptian Presidents goes from 0% to 100%. I am still more comfortable being short fixed-income, though, since the market seems not to care about that any longer!