May jobs report not as strong as expected, but lower than average temperatures are partly to blame

A worker stands outside the construction site.

The Bureau of Labor Statistics (BLS) employment report released today shows that 138,000 new jobs were added in May. In this blog post, I will put forward three alternative projections for job growth in May 2017, each of which was calculated using methodology outlined in my past research. Applying a more stable seasonal adjustment to the raw data, and accounting for the effects of weather on employment yields 139,000 jobs, almost identical to the BLS official number.

Calculating the Alternative Seasonal Adjustment

Monthly job gains and losses can indicate how the economy is doing once they are corrected to account for the pattern the BLS already expects in a process called seasonal adjustment. The approach for this seasonal adjustment that is presently used by the BLS puts very heavy weight on the current and last two years of data in assessing what are the typical patterns for each month.

In my 2013 paper “Unseasonal Seasonals?” I argue that a longer window should be used to estimate seasonal effects. I find that using a different seasonal filter, known as the 3×9 filter, produces better results and more accurate forecasts by emphasizing more years of data. The 3×9 filter spreads weight over the most recent six years in estimating seasonal patterns, which makes them more stable over time than the current BLS seasonal adjustment method.

To produce the Alternative Seasonal Adjustment, I calculate the month-over-month change in total nonfarm payrolls, seasonally adjusted by the 3×9 filter, for the most recent month, which you can see in table below. The corresponding data as published by the BLS are shown for comparison purposes. According to the Alternative Seasonal Adjustment, the economy added 116,000 jobs in May, 22,000 less than the official BLS total of 138,000.

Calculating the Seasonal and Weather Adjustment

In addition to seasonal effects, abnormal weather can also affect month-to-month fluctuations in job growth. In my 2015 paper “Weather-Adjusting Economic Data,” Michael Boldin and I implement a statistical methodology for adjusting employment data for the effects of deviations in weather from seasonal norms. This is distinct from seasonal adjustment, which only controls for the normal variation in weather across the year. We use several indicators of weather, including temperature and snowfall.

Temperatures in May were lower than average and depressed job gains by about 26,000. Controlling for weather yields a higher Seasonal and Weather Adjustment estimate of 164,000 new jobs.

Combining the Alternative Seasonal and Weather Adjustments

My Alternative Seasonal Adjustment shows lower job gains than reported by the BLS, but the Seasonal and Weather Adjustment shows higher job gains. One can indeed do the two adjustments jointly—both adjusting for weather effects and using a longer window. This is shown in the far-right column of the table labeled Combined Alternative Seasonal and Weather Adjustment.

Making both adjustments, the employment change in May was an increase of 139,000 jobs, almost identical to the BLS Official number of 138,000. Notably, however, the Combined Alternative Seasonal and Weather Adjustment shows no effective difference in job growth for April and May (140,000 vs. 139,000), in contrast to the BLS Official numbers that suggest slower growth in May than in April (174,000 vs. 138,000). Under the Combined Alternative Seasonal and Weather Adjustment, job growth over the first five months of 2017 averaged 154,000, about in line with the prior year’s 163,000 average.

Thousands of jobs added BLS Official Alternative Seasonal Adjustment[1] Seasonal and Weather Adjustment[2] Weather Effect[3] Combined Alternative Seasonal and Weather Adjustment[4]
2017-May 138 116 164 -26 139
2017-April 174 159 142 +32 140
2017-March 50 24 111 -61 85
2017-February 232 253 205 +27 226
2017-January 216 205 194 +22 181
2016-December 155 169 160 -5 163
2016-November 164 175 170 -6 181
2016-October 124 129 125 -1 136
2016-September 249 235 264 -15 243
2016-August 176 173 164 +12 171
2016-July 291 307 290 +1 321
2016-June 297 301 278 +19 269
2016-May 43 39 30 +13 14

Note: Changes in previous months’ numbers reflect revisions to the underlying data.

[1] Applies a longer window estimate of seasonal effects (see Wright 2013).

[2] Includes seasonal and weather adjustments, where seasonal adjustments are estimated using the BLS window specifications (see Boldin & Wright 2015).

[3] BLS Official number less the Seasonal and Weather Adjustment number.

[4] Includes seasonal and weather adjustments, where seasonal adjustments are estimated using a longer window estimate.

The author did not receive financial support from any firm or person for this article or from any firm or person with a financial or political interest in this article. He is currently not an officer, director, or board member of any organization with an interest in this article.