The employment number last week caused quite the kerfuffle. According to the Bureau of Labor Statistics, the U.S. created 2.5 million jobs making it the best month on record. What made this so shocking was that economists had forecast a loss of 7.5 million jobs. 1 1 Close A loss of 7.5 million jobs would have been the worst month on record prior to this year. Forecasters were expecting the worst month and we got the best month. To put that in context, a forecaster who missed a payroll number by a mere 200 thousand last year would have a been a laughingstock. 2 2 Close I’m not sure what would have happened to a forecaster who missed a number by 10 million. They’d probably win an award. Since the minute the number was released, forecasters have scrambled to explain what happened. They point to technical issues in the classification of people who are temporarily out of work and to government programs such as the PPP which may have distorted the number. 3 3 Close “Economists Have Biggest Miss Ever in U.S. Jobs-Report Shocker.” Bloomberg, June 5, 2020. Some say it will reverse next month or, more regrettably, that the number may have been manipulated. 4 4 Close Cue voice “it wasn’t manipulated.”
None of the mitigating factors can fully account for the strength of the number. Both the household and the establishment surveys were strong, which makes technical factors an unlikely explanation because they usually affect one survey or the other. 5 5 Close For those of you who don’t follow the numbers closely, in the household survey the BLS surveys households, while in the establishment survey the BLS surveys establishments. But seriously, sometimes there are big discrepancies between the surveys because of self-employed people or because of timing of layoffs or other technical reasons. If they are consistent, then they are more reliable. Some of the biggest job gains this month came in the weakest sectors from the previous two months, exactly what you’d expect after such a sharp contraction. If that’s not enough, our friends to the North released an equally strong employment number on the same day, making it seem like a continental phenomenon. 6 6 Close Canada’s number looked remarkably similar to the U.S. number. It makes sense because the two economies are so tied to each other, eh? The real reason the miss was so big is that forecasting in this kind of environment is difficult. While we can’t discount this number, it raises some existential questions. How should investors think about data when it is so volatile and unpredictable? Can they use the same tools and approaches that they would in a more normal economic environment? And more specifically, what does one good month mean in the context of the largest job losses since the Great Depression?
Back in March, economic data was almost impossible to interpret. We didn’t know how much the lockdowns were affecting the numbers, and we didn’t know what the magnitude of the numbers meant. A percent or two difference in a GDP number wouldn’t change the medium-term outlook in any way. Investors already knew the economy was very, very bad and that was all they needed to know. As the economy recovers, the numbers are becoming relevant again. We want to know how quickly activity is returning and how close we can get to prior levels of growth. If the recovery does occur unevenly, as it likely will, we want to know which sectors are benefiting and which are not. We want to know if the recovery is stalling or continuing. Here are five pro tips (also known as clichés) to help you interpret the data.
Tip 1: Don’t be surprised by surprises.
One consequence of the volatility and uncertainty is that large surprises become more likely. But because forecasts are less reliable, the surprises have less meaning than they would otherwise. We can’t expect any kind of precision for individual numbers, so an upside surprise may reflect the challenge of coming up with an accurate forecast rather than underlying strength in the economy. As a result, the month on month change – simply whether an indicator is rising or falling – may tell us more about economic growth than how much it beat or missed forecasts.
Tip 2: Think longer term. Think 2021.
Often, we think that there is little visibility beyond this month or the next. As recent short-term data is more prone to distortions, investors should look past the next few months. Forecasts for next year may prove more accurate than those for the next week. In particular, the direction of revisions in next year’s forecasts may tell us how sentiment is changing and how optimistic economists are about the sustainability of the recovery.
Tip 3: Practice equality among numbers.
Investors tend to place more weight on certain numbers than others. This approach helps to manage the volume of information in an organized manner, but it is less effective in the current environment. So-called second tier data can confirm or contradict headline releases. The agreement or disagreement among the data can tell us more than the direction of any one number, no matter how closely followed it is.
Tip 4: Forget everything you know about leading and lagging indicators.
I know it’s dramatic, but I always like to sound like a movie trailer.
Investors like to categorize numbers as leading versus lagging, to try to get ahead of future trends. But as things have been unfolding so quickly, this distinction becomes less meaningful. Indicators which amplify small changes in trends may be too finely calibrated for such a blunt world. Investors should be aware that as the economy changes quickly, it can affect areas that were previously considered laggards before any of the leaders.
Tip 5: Use some common sense.
Many of the numbers you see won’t make sense. It’s inevitable in a crisis of this magnitude. Investors need to see what story the numbers are telling. Are its components consistent with each other? Does it stand up to scrutiny? Many times, it won’t. It doesn’t mean that the number is fabricated, but it does mean that you can get some weird data in unusual environments. While no numbers can be fully discounted, outliers are less likely to signal a change in trend than they would have in more normal times. 8 8 Close George Orwell might have said all numbers should be weighted equally. Just some numbers should be weighted more equally than others.
With these tips in mind, let’s return to the employment number. The magnitude of the surprise is not consistent with other employment data such as the ADP survey, jobless claims or the employment component of the ISM index. However, across these numbers and other related data, it appears that the labor market has likely bottomed. While the job losses have been enormous and the recovery may be slow, it appears that the worst is behind us. While there have been both positive and negative surprises, the general trend appears to be toward recovery. It isn’t clear yet how strong that trend is or if it will be enough to reach what we would consider normal levels. That should become apparent as the numbers continue to come in over the next few months. If someone asks you which number you think is most important, you can respond the same way you would if someone asks you for your favorite Michael Bolton song. "Mmm... I don't-- I don't know. I mean, I guess I sort of like them all."
What We Are Watching
China Industrial Production, Retail Sales, Fixed Asset Investment (Monday)
As the first country to implement, and later emerge from, a coronavirus containment lockdown, China’s experience may give an indication of what to expect as other countries seek to restart their economies. While April data showed continued YoY declines in Chinese retail sales and investment activity, industrial production exceeded economists’ forecasts, returning to positive growth much earlier than expected. In recent weeks, optimism that the global economy might experience a faster-than-expected recovery has contributed to gains in equity markets and many cyclically sensitive currencies and commodities. Should incoming Chinese data fall short of expectations, it could shake investor confidence in the near-term outlook for global recovery. This in turn could lead to reversals in some asset prices.
U.S. Retail Sales (Tuesday), U.S. Industrial Production (Tuesday), U.S. Housing Starts (Wednesday)
As noted above, the U.S. jobs report delivered a sizable upside surprise in May that points to a bottoming in the labor market. Moving on to other areas of the economy, next week we will get a new slate of data to inform us on the state of the housing, retail, and manufacturing sectors. With the U.S. economy in reopening mode, forecasters expect these sectors to post positive sequential growth, yet the magnitude is highly uncertain. Market reactions will probably be noticeable only in the face of substantial surprises in either direction. With that noted, market participants’ recent focus on second-wave risks ahead may somewhat dampen their sensitivity to these releases.
E.U. Council Meeting (Friday)
In most major economies, policymakers have responded to the coronavirus crisis with a mix of fiscal and monetary support. In the Eurozone, the fiscal half of the equation has been complicated by the disproportionate impact of the pandemic on Italy and Spain, two countries where concerns around fiscal sustainability have been prominent at points in recent years. In May, risk sentiment in the euro area received a boost when France and Germany signaled support for a generous E.U.-wide recovery program. The European Commission proceeded to deliver a detailed proposal for a €750 billion package including €500 billion in grants. This package must now receive unanimous support from all E.U. member states. The E.U. council’s (online) meeting this week will provide an important indication of how much resistance the current proposal will face. There has been particular focus on the so-called “frugal four” countries, who are expected to resist generous cross-border transfers. If it looks like these countries are open to constructive negotiation, it would be a positive sign that stimulus plans will not be derailed or delayed excessively.