Skewed Employment Impacts
Issue
The truth concerning some economic impact models is seldom revealed in the media, much less a publication such as The Washington Post. The Post recently printed an article (PDF) on organizations overstating their employment economic impacts. In the case cited, the numbers are made up to persuade gullible lawmakers to reduce regulation to create more jobs.
Employment Impacts
In all fairness to the economists who manufacture numbers, if lawmakers do not take the time to understand the data they use to make policy decisions, who is at fault? In the article some analysts published their model methods, and I applaud that. You may not agree with what they have done, but at least you can evaluate the model accuracy.
Employment impacts include:
- Direct – The actual number the company will supposedly hire.
- Indirect – The number of employees other businesses will hire as a result of direct hiring.
- Induced – The number of employees hired as a result of a rising economic tide in the study area.
Often the numbers are summed and presented as direct impact. In the article, one economist argues that induced jobs are accurate, and they are if you count food service and lawn care workers. But let’s not confuse those jobs with high-paying drilling jobs by providing a total without accompanying employment detail.
Assumption Pitfalls
It is the assumptions, then, that are at the heart of the debate. Two common mythological tweaks that significantly change employment impacts are geography selected (not discussed), and initial direct impact calculations. The article provides an example of a flawed direct assumption:
“The Wood Mackenzie study also makes assumptions about current policy. For example, it assumes that current regulations limit the number of Gulf of Mexico exploratory wells to 20 a year. But the number of exploration wells being drilled now is already well above that. As a result, gulf exploration would have little effect on job creation.”
This results in an estimate that is the equivalent to taking a 10 year old employment number, taking the difference from that time period to present and stating it as the new impact!
Conclusion
Currently firms that sell impact software have been unwilling to set guidelines for appropriate use of their software. A simple one is to require publishing model assumptions. It is true that some information is confidential to organizations. But on the other hand, if one is asking for public support to do something that benefits your organization, is it not reasonable to provided the data behind your model? Kudos to The Washington Post for publishing this piece. It is clear we need more transparency in the impact business.
Labor Force Forecast
Issue
Three data sets are needed to forecast the unemployment rate: labor force growth, employment and unemployment. One of these, labor force growth, is used to gauge the effects new employment has on unemployment. Forecasting labor force growth, therefore, is an important part of the process. It benchmarks the number of jobs the economy needs to create to maintain or reduce the current unemployment rate.
Process
Recently the Center for Economic Policy Research (CEPR) demonstrated this calculation. It seems straight forward enough. But these guys are good, so let me take the words out of the article and focus on data explained. The goal is to estimate the number of jobs the economy needs to add to keep pace with the labor force growth. The bogey is 90,000. What data sets do we need to arrive at that number?
Congressional Budget Office (CBO) Key Assumptions in CBO’s Projection of Potential Output
Table 2.2 Potential Labor Force Growth 2010-2014 = 0.7 percent/year
BLS Current Employment Statistics (CES) Payroll Employment January of 2008 total non-farm employment 137,996,000 (138)
142 million *0.7 percent= 994,000/12 equals approximately 83,000 jobs a month.
Conclusion
CEPR also uses data exclusively from the BLS – its growth estimate and employment to population ratio (EPOP) – to derive another estimate. One key calculation in the second estimation is the the number of self-employed workers. CEPR put this at 6 percent. The number is estimated from the Household Survey Table A8. Let’s look at September 2011: Self-employed 8.878 million divided by total employed, 128.565 million, equals 6.9 percent – close to the CEPR estimate. Both calculations suggest the labor force is growing something south of 90,000, but no higher. This is an important calculation because it gives a strong indication of how strong, or weak in this case, our economy employment growth is at the present time. Currently our economy is barely producing jobs at a rate of 0.7 percent (.007) a month, pretty sad performance.
