In this update, we show that the published Unit Labor Cost (ULC) growth -possibly the best indicator to gauge pressures on firms’ prices- is probably mismeasured right now, and we provide a (less biased) counterfactual. At the end of the post, you can find a summary table of all measures of wages and total compensation which summarizes where the US economy is, and how it compares to the pre-Covid period.
(Note: as usual, reading this note on our website and not in this email ensures a better format. You can read it on our website by clicking on the title)
Keep in mind
Labor productivity has shown a puzzling behavior since the beginning of the pandemic. The level of labor productivity has probably been subject to a measurement issue, until recently. A simple counterfactual exercise reveals that published statistics of the unit labor cost are probably upwardly biased in recent quarters. Nevertheless, the big picture remains the same: it will take time to disinflate the US economy.
The ULC
The ULC describes the relationship between compensation per hour and labor productivity, or real output per hour, and can be used as an indicator of inflationary pressure on producers. Algebraically, ULC is defined as:
where “CPH” stands for “Compensation Per Hour”, and “LP” stands for “Labor Productivity”, that is real output per hour.
In theory, ULC is the single best indicator of inflationary pressure on producers. Indeed, the ULC numerator is a measure of total compensation (as opposed to wages only), and the denominator takes into account productivity gains. Unfortunately, the ULC does not enter in the public debate, but it is certainly worth understanding and monitoring its dynamics.
The measurement issue
The measurement issue we refer in this note comes from the ULC denominator: Labor Productivity (LP). Figure 1 shows the level of labor productivity (the black line), as published by the BLS. Figure 1 also shows a counterfactual exercise (the light blue dashed line) that we have constructed by extrapolating the pre-Covid trend (2016-2019 average growth rate).
In the three years pre-Covid, ULC grew at an average of 1.6% (a.r.), a bit stronger than the pre-2016 period. However, in 2020:Q2 (published) LP grew 17.9% (QoQ, a.r.), and in 2020:Q3 it grew 7.2% (a.r.). After this, LP moderated at an average of 0.8% (a.r.) between 2020:Q4 and 2021:Q4. Finally, since the beginning of this year, the level of LP has declined notably, although in the last published quarter (2022:Q3) it grew 0.8% (a.r.). In other words, as Figure 1 shows, LP originally deviated from its pre-Covid trend but it has recently traveled back towards it.
Figure 1. Published and counterfactual labor productivity
In this update, we do not investigate the source of the mismeasurement in published LP. We simply observe that it is hard to think that the US economy experienced such a large positive productivity shock in 2020, followed by a large negative shock two years later (for the record, according to published stats, 2020:Q2 is the second largest QoQ in history, very close to the highest ever: 1947:Q4). Rather, by looking at Figure 1, and given that the level of LP is now back to its pre-Covid trend, the natural question to investigate is the following: what would the ULC statistics show right now if measured productivity had grown linearly (that is, under the counterfactual of Figure 1)?
The counterfactual ULC
Using equation (1), we have constructed a counterfactual ULC under the assumption that labor productivity growth was linear in 2020-2022, as explained above. Our results are reported in Figure 2. Under the counterfactual scenario, the level of the ULC is higher than in published statistics because the level of LP is lower, although the gap has closed recently. The contour of ULC and its counterfactual are broadly similar in 2021 but the counterfactual has flattened out since the beginning of 2022. This has happened because while LP growth is constant under the counterfactual, published LP has contracted. In other words, the recent growth of ULC in published stats is mostly driven by the contraction of the denominator, and not by the numerator (indeed, as shown in our previous update, CPH has been running low in 2022, and it is already in line with its pre-Covid average).
Figure 2. Published and counterfactual unit labor cost (ULC)
Figure 3 shows the QoQ (a.r.) and the YoY of published ULC and counterfactual ULC. The notable difference is that in the last 3 quarters, the growth rate of the counterfactual ULC is lower than published ULC (left panel of Figure 3), as the growth of the numerator (CPH) has remained roughly constant but the growth of the published denominator (LP) has been negative, on average. It follows that the YoY of the published ULC (right panel of Figure 3) remains well above the YoY of counterfactual ULC but the baseline is that former should moderate going forward.
Figure 3. QoQ (ar) and YoY of published ULC and counterfactual ULC
Putting everything together: where is the US economy?
Because measures of wages and total compensation are sending conflicting signals (including published ULC vs counterfactual ULC), we have prepared a summary table. Table 1 shows the recent evolution of all measures of wages and total compensation available. Table 1 reports 8 measures: the Average Hourly Earnings (AHE), the AHE numerator (Average Weekly Earnings, or AWE), the Atlanta Fed Wage Tracker (AFWT), the Employment Cost Index (ECI), the Compensation Per Hour (CPH), the CPH numerator (aggregate compensation), the published ULC, and our counterfactual ULC.
All figures in Table 1 are percent changes. Column (a) shows the most recent published QoQ ar (or 3m/3m ar for monthly series), column (b) shows the most recent YoY, column (c) shows “deceleration” defined as the difference between column (a) and (b), and column (d) shows the 2016-2019 average growth. Finally, columns (e) and (f) show the deviation from the 2016-2019 average of the most recent QoQ ar (or 3m/3m ar) and YoY, respectively.
Table 1. Recent evolution of wages and total compensation growth.
There are several takeaways from Table 1:
- In the last year (months), wages and total compensation have grown at an average of 5% (4%). Therefore, wage and total compensation growth remains (well) above the level consistent with the Fed target.
- The “deceleration” column indicates that in the last few months, on average, the marginal variation has been below the average of the last year. Having said so, this number should be interpreted with caution given the volatility of some series and the fact that the “decelaration” in the ECI and AHE is small or null. Column (c) shows why we are skeptical there are convincing signs of deceleration and why we are even more skeptical there are signs of re-acceleration. For the record, the average growth across measures over the 3 to 6 months ago period (that is, one quarter before column (a)) is 4.3% as opposed to 4.0% in column (a), another confirmation that wages and total compensation are going sideways.
- The level of wage and total compensation growth consistent with the Fed target is around 3% (or a touch higher). Column (d) shows, in fact, that the 2016-2019 average across all measures is 2.9%. Over the same time period, core PCE price inflation averaged 1.8% and core CPI averaged 2.1%.
- Consequently, the average deviation of wage and total compensation growth from their pre-Covid average has been substantial in the last year (the color of the cells in columns (e) and (f) indicates the size of the deviation, with red indicating large deviations, and green small ones), although it might be -on average- smaller now (with the caveats mentioned above).
Conclusion
All told, it is generally hard to interpret incoming data in real time. It is even more challenging now because we are getting conflicting signals and because of measurement issues. On our end, we re-iterate that we need more evidence to make hard calls, one way or the other. At the moment, the safest thing to say is what we wrote in our last update: wage growth is persistent by nature and it is probably going sideways right now. It will take time to disinflate the US labor market.