Offline I have been peppered by questions about the “uselessness”, “fudge factor”, and “down right dishonesty” of the Administration’s goal to create or save 4 million jobs.

Friends and colleagues quote Mankiw, Krauthammer and Hennessey who all say that there is no way to measure a job saved. We measure jobs created yes, but jobs saved is incalculable.

This view, I argue, is wrongheaded. Either both terms are immeasurable or neither is.

The nerdier among you might satisfied by my simply noting that “zero is a number.” For the rest here is a longer rundown.

When we say a job was “saved” by the stimulus package we mean the following are all true:

  1. The job existed before the stimulus package was passed
  2. The job exists now
  3. The job would not exist now if the stimulus package had not passed

Now, what do we mean when we say a job was “created” by the stimulus package? We mean the following are all true:

  1. The job did not exist before the stimulus package was passed
  2. The job exists now
  3. The job would not exist now if the stimulus package had not passed

 

The only difference between these two sets of conditions is the first line. However, the first line is no less measurable under the conditions for “saved” as it is for the condition of “created”

In both cases the hard to measure condition is the third, whether or not the job would exist now if the stimulus package had not passed. However, this is just as hard to measure for jobs created as it is for jobs saved. Either both terms are meaningful or neither is.

I think that the confusion comes because the term “saved” inherently carries the notion of “by the actions of the government.” In this case the stimulus package. While one could think of jobs “created” generally, without the concept of cause.

However, as soon as we have asserted the condition: “The stimulus package will” we are in the world of causation and in that world condition (3) exists for both created and saved.

That is to say, you can either measure the impact of the stimulus package or you cannot. Whether the baseline is a positive number (above zero) or a negative number (below zero) is neither here nor there.

 

Note: Those who find Hennessey’s analysis compelling should note that he is exactly correct right up until the end where he says

Had he said that his Administration would create 4 million jobs, then we would have a simple metric:  134.3 million + 4 million = 138.3 million.  Each month, we could compare the nonfarm employment level to 138.3 million to see if it was higher or lower than that goal.

Hennessey is implicitly assuming that zero jobs created is a meaningful baseline in the “created” case. However, zero ain’t nothin’ but a number. That is, there is no reason why we should assume that in the absence of administration policy exactly zero jobs would be created.

Note on Note: Yes, the additive identity is often a special number, just not in this case.

About these ads