ECI Compensation Growth: Model Update and An Analysis

Keep in mind

Compensation growth is projected to be well above structural productivity and underlying inflation throughout the medium-term (2022-2025).  Nominal compensation growth does not appear to be consistent with the Fed 2 percent target and continues to be more consistent with price inflation averaging 3 percent going forward.

The incoming data

According to the BLS, over the three months ending in September, total compensation for private industry workers increased 4.3% QoQ (a.r.) and 5.2 percent YoY. Wages and salaries increased 5.2 percent for the 12-month period ending in September 2022, while the cost of benefits increased 5.0 percent.

Our model-based forecast

Based on the newly released Q3 figures, we have updated our model-based ECI forecast. (reminder: our ECI Phillips curve model mimics our “main” price Phillips curve model and uses the unemployment gap as a measure of slack but allows for structural productivity). The Q3 incoming data have been weaker than the model expected, although most of the downward surprise came from the volatile benefits component.

The inclusion of the third quarter in the sample resulted in a downward revision of the model forecast. The YoY change of ECI total compensation for private industry workers is now projected at 5.3 percent in 2022(Q4), 4.3 percent in 2023, 3.7 percent in 2024, and 3.3 percent in 2025 (see panel A of Figure 1).

Compared to the previous forecast, the current forecast is about 3-4 tenths lower throughout the forecast horizon. The downward revision is driven by a lower constant and a lower estimated persistency. The current forecast is nearly identical to the forecast at the end of Q1 as an upward surprise in the incoming data in Q2 was offset by a downward surprise in Q3.

Alternative measure of slack

Given the usual model uncertainty, we run an alterantive scenario using the vacancy per unemployed ratio instead of the unemployment gap as measure of “slack”. Results are shown in panel B of Figure 1. For this scenario we assume that V/U will drop to 1.0 by mid-2023 and remain at that level going forward.

According to the model, total compensation is expected at 5.4% in 2022, 4.6% in 2023, 4.1% in 2024, and 3.7% in 2025. In other words, the model using V/U as slack delivers a forecast about half of percentage point higher than using the unemployment gap. The reason is that while we assume that V/U converges towards unity by 2023, slack continues to put upward pressure on total compensation as the initial level in the forecast is well above the historical mean. The model can deliver a forecast that returns to pre Covid levels by 2025 but it requires V/U to fall well below unity (i.e. 0.5), a level historically associated to large recessions.

Figure 1. Phillips curve model-based ECI forecast

A. Unemployment gap as measure of slack

B. Vacancy per unemployed as measure of slack

Note: the figures shows the ECI Phillips curve model-based forecast. The charts refers to the year-over-year percent changes of compensation costs for private industry workers. The confidence intervals are based on the estimated parameters distributions. The left panel (“A”) shows the model forecast using the unemployment gap as a measure of slack. The right panel (“B”) shows the model forecast using the vacancy per unemployed as a measure of slack.

The Fed staff forecast

In our view and experience the Fed staff ECI forecast is informed by a model like the one presented above. However, at any point in time the Fed staff can deviate from the model because the Tealbook forecast is a judgmental projection. In the model, ECI growth is driven by three “fundamentals”: structural productivity, underlying inflation, and the contribution of slack defined as the unemployment gap. Below we show the ECI forecast that, in our view, the Fed staff has written in the November 2022 Tealbook and the decomposition (contributions) the staff has discussed with Powell and the Governors.

Figure 2. The ECI compensation Fed staff judgmental forecast and its contributions

Note: the figure shows our view of what the Fed staff has written in the November 2022 Tealbook for ECI (total compensation, private industry workers) growth. The black line shows the YoY percent change of ECI growth. “Other factors” refers to factors that cannot be otherwise attributed. “Slack” refers to the contribution of the unemployment gap. “Structural productivity” refers to the trend of labor productivity (output per hour, non-farm business sector) growth.

In our view, the Fed staff expects ECI growth to be about 5 percent in 2022 before stepping down to around 3 percent in 2023-2025, in line with the calibrated fundamentals. According to the Fed staff ECI decomposition, the contribution of underlying (business sector) inflation should remain above 2 percent in 2022-2025. Structural productivity is expected to add about 1 percent per year in the medium-term. Slack (as captured by the contribution of the unemployment gap) is expected to add about 25bps this year but gradually fade as the unemployment rate converges back to the natural rate in 2025. Crucially, the “other factors” (that is, the quasi-residual of the decomposition) are expected to remain positive and large in 2022 after two years (2020 and 2021) in which they were also large and positive).

The crucial role of “other factors” (and why the Fed staff might be in trouble)

The ECI decomposition shows that it is hard to explain nominal compensation growth in 2020-2022 using the calibrated fundamentals. Indeed, the model shows large and positive (quasi-)residuals in those years. In our view, this evidence has now become a big issue for the Fed staff because by being unable to explain nominal compensation growth, the staff is prone to large misses going forward.

The estimated forecast vs the calibrated decomposition

The estimated ECI Phillips curve model (Figure 1) suggests that nominal compensation growth might remain elevated going forward. On the other hand, in our view, the Fed staff is assuming ECI growth will return in line with the estimated fundamentals next year (Figure 2). The contrast between the estimated model and the calibrated decomposition is a crucial point. In our experience, the Fed staff assumes that the calibrated model (the decomposition in Figure 2) is correctly specified. As such, the forecast assumes that the contribution of “other factors” is zero for the years after the current. However, this is very risky in the current environment. Indeed, the Fed staff has already made this mistake back in 2020 and again in 2021 for both, price and wage inflation. In both years, the Fed staff has assumed that the yellow bars in the decomposition (the “other factors”) had no persistency. Ex-post, it was a big mistake. Unfortunately, it seems that the Fed staff is making the same mistake again in 2022, ignoring for the 3rd consecutive year the signals of the estimated models. For this reason, the Fed staff ECI (and price) inflation forecast is prone to be revised up again next year (and the following), as it did in 2021 and 2022.

How much upper pressure on core (PCE) prices should we expect from wage growth going forward?

The answer to this question can be found in Peneva and Rudd (2015). Figure 2 of the paper (see page 26 of the paper) shows the response of core PCE price inflation to an innovation in ECI compensation. According to the estimates of Peneva and Rudd, a 1 percent shock of ECI growth results in about 15-20bps on core PCE price inflation. Therefore, because ECI growth is currently about 2 percentage points higher than the estimated fundamentals, it is reasonable to expect about 30-40bps upper pressure on core inflation going forward from this margin.

Conclusion

Our analysis has shown that compensation growth continues to be well above structural productivity and underlying inflation throughout the entire medium-term (2022-2025). Assuming a higher pi* alleviates the tension in the forecast but results in a price (PCE) inflation forecast significantly above the Fed staff (and the FOMC) in the medium-term. As such, nominal compensation growth is projected to remain a significant and underestimated upside risk around the Fed inflation forecast.

Contact

Want something more tailored?

We provide tailored consulting on ad-hoc projects.