Jobless Claims: A Vital Number For Macro & Markets Analysis

Tomorrow’s weekly update on initial jobless claims will draw the usual crowd in search of clues on the outlook for the business cycle, but there’s another reason that new filings for unemployment benefits deserve careful scrutiny. Claims data provide another robust perspective on projecting the near-term outlook for the stock market.

There’s no shortage of methodologies for developing assumptions about expected return for equities, but one relatively easy system that’s worthy of regular reviews is comparing the inverted one-year percentage change for jobless claims as a proxy for the one-year change in the stock market (S&P 500). As the chart below shows, these two metrics tend to track each other fairly closely. The connection isn’t perfect, particularly over a few weeks. But nothing else is perfect either, which is why it’s essential to model ex ante returns from multiple angles for deeper perspective than any one forecast or time series can provide. By that standard, keeping an eye on claims and stock prices deserves a spot on your short list.

For example, I regularly look at the claims data on a rolling one-year basis for insight on whether a sliding stock market may be more than short-term turbulence. Both data sets bounce around a lot across a few weeks and so it’s always wise to remain skeptical at the first sign of what appears to be trouble. But if both claims and equities are in sync for any length of time, and for bearish reasons, that grabs my attention. By contrast, a sharp downturn in stocks without a corroborating move in claims suggests that Mr. Market’s simply having another one of his hissy fits and so it may not mean much for the economic outlook.
Equities generally are considered a barometer of expectations of economic conditions. But it helps to have context. Stocks, after all, sometimes anticipate recessions that never materialize. But to the extent that layoffs are trending higher, that’s a dark sign for the business cycle… if it persists for any length of time and it coincides with declining stock prices. Why? Because twin warnings on this front are a clue for thinking that the primary driver of equity bull markets—higher earnings—is vulnerable to the cycle’s dark side. Markets and macro are linked, but not every data set offers the same degree of timely warnings.
You can, of course, find a similar relationship between stocks and other economic data series, such as industrial production and payrolls. But jobless claims are highly prized because the numbers are published relatively frequently and with minimal time lags. Tomorrow’s update, as usual, will reflect the state of the previous week’s layoffs. By contrast, last month’s industrial production report is still a mystery and will remain so until next Tuesday, when the June numbers are published.
True, we already have June’s payrolls report, but the July update is nearly a month away. By August 2, when the government releases the next jobs report, most of July’s claims data will already be revealed, and so the implications for stocks will probably already be priced into the market.
In other words, you can’t afford to ignore weekly claims data—for several reasons. It’s a valuable data set for thinking about the outlook for stocks. Claims are also a useful series for analyzing the business cycle, which is why I include these numbers for calculating the Economic Trend & Momentum indices, a set of benchmarks for assessing business cycle risk.
It’s hardly a shock to find that claims data tends to bottom out at business cycle peaks. On the flip side, claims usually top out around during the darkest days of a recession’s bite.
All the standard caveats apply, of course, starting with the obvious one: looking at claims data, or any data set, in isolation is asking for trouble. Although this series generally offers constructive insight for macro and market analysis, it’s subject to short-term noise, just like everything else. But in a world of poor choices for divining the future, jobless claims are surely a no-brainer in terms of looking for the lesser of econometric evils.
Okay, so what are jobless claims telling us these days? More of the same, i.e., modest growth for the economy and a respectable return for stocks. Running a simple regression analysis on the two data sets implies that equities will earn around 12% over the next year.
We should take that forecast with a grain of salt, since statistical noise infects any one forecast at a given point in time. A better way to use claims numbers to model equity returns is by running regression-based forecasts on a weekly basis and watching how the predictions evolve through time. Are the forecasts rising, falling, or more or less remaining stable? That information is quite a bit more valuable than any lone point forecast.