Texas Employment Forecast

Incorporating January job growth of 2.6 percent and a rebound in the leading index, the Texas Employment Forecast suggests jobs will grow 1.5 percent this year (December/December), with an 80 percent confidence band of 0.1 to 2.9 percent. Based on the forecast, 191,000 jobs will be added in the state this year, and employment in December 2019 will be 12.8 million (Chart 1).

“In its annual benchmarked data released this morning, the Texas Workforce Commission revised its December/December 2018 job growth estimate from 3.2 percent to 2.3 percent,” said Keith R. Phillips, Dallas Fed assistant vice president and senior economist. “The Dallas Fed’s early benchmark data that we used for the forecast in January was very close to the annual benchmark, and hence our forecast for this year was little impacted.”

The improvement to the outlook came primarily from the rebound in the Texas Leading Index (TLI). The weakening of the TLI at the end of 2018 and then the rebound in January is consistent with business outlooks reported in the Dallas Fed’s Texas Business Outlook surveys. 

“It is clear that businesses are more optimistic now about 2019 growth than they were at the end of last year,” Phillips said

The Dallas Fed’s TLI decreased 1.5 percent over the three months ending in January (Chart 2). Growth was mixed across the components with four components making large negative contributions to the index and the other four having zero or mildly positive contributions. Two broad measures of labor market conditions—initial claims for unemployment insurance and help-wanted advertising—both suggested weakening in the months ahead. 

Energy indicators were somewhat mixed with oil prices down but permits to drill oil and gas wells increasing slightly. Manufacturing indicators were also somewhat mixed with the increase in the Texas value of the dollar putting downward pressure on demand, while a slight increase in hours worked suggested continued growth. The outlook for the U.S. economy as signaled by the U.S. leading index was neutral, while the sharp January rebound in the stock prices of Texas companies sent this indicator into slightly positive territory.

“Overall recent movements in the TLI are suggesting moderately weaker growth in Texas than existed last year but stronger than we expected in early January,” Phillips said. 

Next release: March 22, 2019

Methodology

The Dallas Fed Texas Employment Forecast projects job growth for the calendar year and is estimated as the 12-month change in payroll employment from December to December. The forecast reported above is a point estimate with 80 percent confidence bands; in other words, the true forecast lies within the bands on Chart 1 with 80 percent probability.

The Dallas Fed Texas Employment Forecasting Model is based on a transfer function that utilizes past changes in state employment along with past changes in the Dallas Fed Texas Leading Index (TLI). Changes in the TLI have an impact on employment with a lead time of three months, and the effect dies out slowly over time. The regression coefficients on lagged changes in employment and the TLI are highly statistically significant, and the model as a whole has been accurate relative to other forecasters over the past two decades.

The forecasting model has been in use at the Dallas Fed since the early 1990s, and the employment forecast has been published in the Western Blue Chip Economic Forecast (WBCF) since 1994. Phillips and Lopez (2009) show that the model has been the most accurate in forecasting Texas job growth relative to other forecasters in the WBCF. In particular, the model had the lowest root mean squared error and has been the closest to the actual the most times (nine of the last 17 years) out of five forecasters that have consistently participated in the survey.

For more details about the model and its performance, see “An Evaluation of Real-Time Forecasting Performance Across 10 Western U.S. States,” by Keith R. Phillips and Joaquin Lopez, Journal of Economic and Social Measurement, vol. 34, no. 2–3, December 2009.