Thursday’s release of ADP’s estimate of US private payrolls for November looks set to deliver another sign that US economic activity has peaked. The projected year-over-over increase (due on Dec. 6) still aligns with a healthy pace of jobs creation, but the mild deceleration trend appears to be on track to continue in the months ahead. A downturn in payrolls growth will align with numbers from other corners of the economy that point to a softer macro trend for the foreseeable future relative to the strong gains posted earlier this year.
The Capital Spectator’s point forecast for the one-year change in ADP’s employment report for November is 1.9%, which is down from October’s 2.0% rise – the highest annual gain in 2-1/2 years.
The forward estimates for the one-year changes in ADP’s estimate of payrolls suggest that a slow-but-steady deceleration in growth will roll on for the near term.
The projections are based on averaging forecasts from eight models, each using a different methodology and offer a different set of pros and cons. Numerous studies over the years advise that combining forecasts from different models generally improves the accuracy of the estimates relative to the projections from any one model. (For an example of the literature on the topic, see this summary.) Here’s a summary of the eight models used to generate the forecasts above. Seven of the models use univariate methods – analyzing the indicator under scrutiny in isolation – while the eighth model draws on multiple indicators. All the analytics are generated in R.
Exponential smoothing state space model: the average forecast is used from 100 simulations based on bootstrap aggregating via the forecasting package. The data set is the historical record for the target indicator.
Autoregressive integrated moving average model: the average forecast is used from 100 simulations based on bootstrap aggregating via the forecasting package. The data set is the historical record for the target indicator.
Neural network model: the average forecast is used from 100 simulations via the forecasting package. The data set is the historical record for the target indicator.
Naïve model: this forecast simply extracts the last data point and assumes that it will prevail for the next 12 months.
Cubic Spline model: a local linear forecasts using cubic smoothing splines via the forecasting package. The data set is the historical record for the target indicator.
Facebook’s Prophet forecasting tool. The data set is the historical record for the target indicator.
Theta method forecast model: the methodology is a simple exponential smoothing with drift via the forecasting package.
Vector autoregression model: this multivariate methodology (via the vars package) uses the following datasets:
ADP estimate of private payrolls
US Labor Dept estimate of private payrolls
Personal consumption expenditures
10-year Treasury yield
Effective Federal funds rate
Retail sales (headline)
Consumer Price Index (headline)
University of Michigan Consumer Sentiment Index
Disposable personal income
University of Michigan consumer inflation expectations