The changing relationship between unemployment and payrolls

It is somewhat worrying that the decline in the four-week average of jobless claims has leveled out in recent weeks and at a level that historically would not be consistent sustained job creation. However, if there are good reasons to suppose that the relationship between initial claims (which track layoffs) and net job creation (which measures the balance between gross job creation and job losses) may have shifted, the level of claims would be less of a concern. This brief note considers this issue.


Over the last three months, private payrolls have risen by an average of 139,000 per month. This is a stronger increase in employment than suggested by initial jobless claims (based on historical relationships), which have averaged 456,000 over the last three months. With household employment (population and payroll-concept adjusted) up an average of 419,000 per month over the last three months, it is difficult to dismiss the relative strength (compared to unemployment claims) of the private payroll data.

The apparent break in the historical relationship between initial claims and payrolls in the current cycle may be due to i) the sharp increase in the duration of unemployment, ii) the availability of extended claims programs, and iii) simplified filing procedures. In any given period, not all unemployed persons who are eligible file for jobless benefits. Indeed, according to the website, more than half of the unemployed do not claim benefits because they either think that filing is a complex process or they may not qualify. In recent years, however, the filing process has been simplified and the incentives to file for benefits have increased, which may have raised the level of initial claims relative to job losses.

First, New York State, for example, has provided an additional 73 weeks of unemployment benefits in addition to the 26 weeks of regular benefits—this availability of extended benefits may have prompted some people to file a claim who would not have otherwise.

Second, the significantly longer duration of unemployment in this recession may have prompted some people to file a claim who otherwise would not have done so because they expected to find new employment. Though speculation on our part (since we do not have data to support these theories), we think these are possible reasons for the recent environment of moderate private payroll growth against the backdrop of unemployment claims levels that are normally associated with continued payroll declines.


Between the week ending May 15th (the May employment survey week) and the week ending June 5th (the latest week for which we have data), the number of temporary Census workers has fallen by 168,000. In theory, therefore, laid off Census workers could be a factor boosting initial jobless claims (those claims filings may not be accepted as there are a number of criteria that must be met for a claim to be approved, but there is nothing that stops someone from filing a claim and being captured in the initial claims series).

However, a Labor Department official told us that i) there have been no comments from State unemployment insurance offices that would suggest laid off Census workers are boosting jobless claims and ii) the patterns in the raw claims data do not indicate that this special factor is affecting the claims data. Unfortunately, temporary Census workers would not be counted in the “former federal employees” series that is published with the claims data, so this cannot be used to judge the impact of the Census on claims. As an aside, weekly reports from the Census Bureau suggest that Census employment will provide a significant drag on nonfarm payrolls in June (we will get the Census worker data for the June employment survey period next week).

Source RDQ


Popular posts from this blog

October retail sales come in strong, especially auto sales

Tea Party Buffalo Pictures

How to spot a fake Tea Partier