Why Trump Doesn't Trust Jobs Revisions, While Economists Do | WSJ

Key Concepts:

• Jobs Report (Employment Situation): Monthly report by the Bureau of Labor Statistics (BLS) detailing changes in the US economy.
• Establishment Survey: Survey sent to 650,000 workplaces to collect data on employment, payroll, and hours.
• Household Survey: Survey of 60,000 people to calculate the unemployment rate.
• Revisions: Adjustments made to the initial jobs report data based on late responses and more complete information.
• Nonfarm Payroll Employment: The number of jobs, excluding farm jobs.
• Completion Rate: The percentage of establishments that have completed and submitted the survey.

Data Collection and Revision Process

The BLS releases the Employment Situation report monthly, which includes data from two surveys: the household survey (for unemployment rate) and the establishment survey (for nonfarm payroll employment). The establishment survey is sent to 650,000 workplaces, requesting data on employee count, payroll, and hours. The initial report is released roughly two weeks after the survey period, with only about 60% of establishments having responded. The BLS revises the data three months later when over 90% of establishments have responded. Further revisions occur twice a year using unemployment insurance data, and even 10 years later with census data.

Reasons for Revisions

Revisions occur because not all establishments respond in time for the initial report. Common reasons for late responses include business closures, accountants being on vacation, or simply being too busy. Larger-than-normal revisions often occur during economic transitions (coming out of or moving into a slow time). For example, late responses may skew more negative than expected if the economy is slowing down. Another reason for large revisions is a difference between the way the world works this year and the way the world worked last year.

Example: August Revisions

The August revisions were attributed to late responses skewing more negative than expected. Employment in schools and the federal government was particularly low compared to the previous year, potentially due to schools running out of COVID-era funding.

Political Influence and Data Integrity

Experts inside and outside the BLS believe that political influence is not a factor in the revisions. One former BLS employee stated, "I could not change the numbers. I had no access to the input data, to the programs, to the preliminary results, to any of that, none of that." Another stated, "The upper bound of my influence was zero. I didn't see the data until Wednesday morning. By that time, there was no way I could have influenced that data."

Improving Data Accuracy

The best way to reduce the size of revisions is to increase the completion rate for the initial report. While responding to the survey is not mandatory under federal law (but it is in some states), incentives and improved delivery systems could help. However, there is concern that response rates could decline if businesses believe their data might be used for political purposes.

Importance of BLS Data

Eroding trust in the statistical system is detrimental to the country because this data is infrastructure that supports many of the most important decisions that companies make.

Notable Quotes:

• "There is information in the preliminary release and you don't wanna sit on it. People wanna know right now. The Federal Reserve wants to know right now. But two months later if the data's much better, you don't wanna sit on that either. You wanna tell people a much better guess."
• "It's normal for BLS not to explain those differences, because then they're doing a job outside of the job they're supposed to do, which is to take the data and statistically make modifications to the estimates based on the data."
• "Anything that erodes trust is really bad for the statistical system. But the most important thing is it's bad for the country, because this is infrastructure that supports many of the most important decisions that companies make. So this is a big deal."

Synthesis/Conclusion:

The BLS jobs report is a crucial economic indicator subject to revisions due to incomplete initial data. While large revisions can raise concerns, they are often attributed to economic transitions and data collection challenges rather than political interference. Maintaining trust in the BLS and improving response rates are essential for ensuring the accuracy and reliability of this vital data, which informs critical decisions for businesses and policymakers.