Is Small Business the Engine of Growth?

It is common for politicians and others to say that small businesses account for most job creation. This is somewhat misleading though. Many small businesses go out of business after a year or two, so that they are also responsible for a lost of job loss. A new working paper by Erik Hurst and Benjamin Wild Pugsley suggests that small businesses are small and do not account for a lot of new employment. They note that most small business owners are not the entrepreneurs that provide the "creative destruction" Schumpeter wrote about. Instead, most small business owners do not bring a new idea to the market and do not have a desire to grow large. Instead, many are starting small construction firms, a legal practice or small retail store or restaurant. The motive is often being one's own boss rather than starting the next Apple Corporation.

Most small businesses have no paid employees. For the state of Michigan, over half of all business establishments have fewer than five employees. The Small Business Administration provides regulations that determine the size limits for "small" businesses. A table of these limits for industries can be found here. The limits are given in revenues for some industries and in number of employees for some. For example, establishments in mining have employee limits of 500, while those in transportation manufacturing have limits of 1,000 employees.

Simple explanations of the economy or job growth generally miss the mark. But politicians often look for simple explanations and simple solutions. Promoting small business will not solve the employment problem. A better solution is to provide an environment that supports business activity regardless of the size of the firm.

More on "Jobs Saved"

Former colleague Victor Claar sent me a link to a news story about errors in the counting of saved jobs, including the treatment of raises as equivalent to saved jobs.

There is another dimension I had not considered as yet. There has been a disruptive construction project on an intersection in Holland through which I often drive. Signs indicated that funds came, at least in part, from the stimulus package. The firms that worked on the job then must estimate how many jobs the project saved or created. The project lasted three months. Will the companies report these jobs saved or created as if they were year-long jobs or will the length of the project be considered? In earlier blogs I wrote about how the methodology used by the government to estimate jobs created or saved focused on job-years. I believe the average citizen reading about jobs saved or created believes that the number of unemployed falls or at least doesn't increase by the number of jobs saved. But, that clearly is wrong, and there is really no way to tell by how much.

Jobs Again

There is an op-ed piece by Ed Lazear in today's Wall Street Journal. Lazear was chair of the Council of Economic Advisors under Bush. He writes that his final economic forecast while a part of the council predicted the economy would begin to recover in the third quarter, and this was without considering a stimulus plan. Of course, his forecast could have been overly optimistic, but it could also mean that the GDP growth in the third quarter was not due to the stimulus plan we have.

He also talks about the estimates of jobs created and the jobs retained as due to the stimulus. Recipients have to fill out a report providing the data. Lazear notes two problems with the reports. The first is reporting bias. A construction firm wanting additional funds may believe that it needs to err on the high side of estimates of jobs created or retained. Second, these programs are likely to count people who switch jobs as new hires.

I have written before on the methodology used by the government to estimate the jobs created and saved by the stimulus. It is a direct link to the spending, assuming so much money equates one job. It also measures job-years and not individuals. That is, a person who is hired as part of a two-year project would count as two jobs--one for each year. The Administration has been criticized for the way they are trying to estimate jobs created or saved, but they keep doing it. The process gives the illusion of precision that is only an illusion. The concept of a job saved is ok--grants sent to states that use the money as a stop-gap in cuts in schools will save some jobs of teachers. But to think we can measure these in a meaningful way is not legitimate.

Perhaps the members of the administration think that if you repeat something often enough, people will believe it must be true.

More On Jobs Saved and Created

In an earlier post, I outlined how the government is counting the number of jobs created or saved by the American Recovery and Reinvestment Act, or the stimulus package passed by Congress in February. Since the government used a formula that related so much spending to a job, it ignored the fact that some of the spending would be replacing other spending. For example, spending on green technologies means less spending on "dirty" techonologies. A job created in one area offsets the loss of a job in another rather than creates a new job.
The Wall Street Journal now has an article on estimates the states are reporting to the federal government about the number of jobs created or saved. The figures are much less than the estimates of the federal government.

How Are 3.5 Million Jobs Calculated?

The American Recovery and Reinvestment Act of 2009 is supposed to create or save roughly 3.5 million jobs. How was this estimate determined? Two documents provide the information. The first was written by Christina Romer and Jared Bernstein in January, before Barack Obama was inaugurated. It is entitled, "The Job Impact of the American Recovery and Reinvestment Plan." The second is a report prepared by the Council of Economic Advisors, of which Christian Romer is the Chair, entitled, "Estimates of Job Creation from the American Recovery and Reinvestment Act of 2009." I will rely primarily on the second report, but it relies a lot on the first report.

Probably the average citizen hears that ARRA is supposed to save or create 3.5 million jobs thinks that this means that 3.5 million people will have jobs that would not have had jobs in the absence of the program. This is not correct though. The jobs are actually job-years. Suppose a project lasts two years and results in an extra 100 people being employed. The Council of Economic Advisors is calling these 200 job-years—the 100 people are employed for two years so each person is counted twice. In my previous post, I mentioned a project in Washington that involves $1,635,000,000 to be spent over several years. The estimate for that project listed on the website is 21,255 job-years. This does not translate into 21,255 people with jobs who would not have had jobs otherwise.

There is still the question of how the 21,255 figure was determined. For this, the report provided by the Council of Economic Advisors states that they estimated that a $100 billion of government spending creates 1,085,355 job-years (p. 5). That works out to $92,136 of government spending needed to create one job-year. To use an example in the report, an $11 billion program, “…will create about 120,000 job-years during the President’s first term.” (p. 5). That is, $11,000,000,000/92000 = 11,956 or approximately 12,000 job-years. In the example above, the $1,635,000,000 project will create 17,772 job-years, which is less than the 21,255 job-years listed on the website. I do not know the reason for the difference.