Wednesday, July 17, 2013

Policy Impact and Red Herrings

Copyright, The New York Times Company

Red herrings are frequently inserted into policy discussions but can be readily identified as long as we remember a simple truth about public policy impact.

The impact of public policy on an economic outcome like employment is, by definition, the difference between employment with the policy in place and what employment would have been under an alternative “baseline” policy. Policy impact quantifies how things are different as a consequence of the policy.

Consider these statements:

“The Affordable Care Act will not reduce full-time employment because workers understand that full-time employment is the path to career advancement” (see my previous post).

“Unemployment insurance does not reduce employment because Americans fundamentally want to work and provide for themselves” (see, for example, this commentary on

These are all examples of red herrings, irrelevant statements that are attached to hypotheses.

Take the full-time employment example. It may be true that full-time employment is the path to career advancement, but that is hardly relevant to the Affordable Care Act as long as we assume that full-time employment would be that path regardless of whether we have that law.

That is, lots of people will choose full-time employment because of the career opportunities it provides, but they are counted as full-time employed under the policy and as full-time employed under the baseline policy (say, continuing as if the act had never become law). A policy impact estimate, by definition, counts only those for whom career advancement does not trump their decision to be in a full-time position.

As I explained in that an earlier post, the Affordable Care Act introduces funds and insurance opportunities for part-time employees that will be unavailable to most full-time employees. As long as there are more than zero people whose full-time vs. part-time work decision depends on funds or insurance, there is the potential for policy impact.

In my second example, it may be true that most people want to work and provide for themselves. But I assume motivation to work is the same regardless of whether unemployment benefits are paid for, say, 99 nine weeks or 26. What’s different between the 99-week policy and the 26-week baseline are the circumstances in which people find themselves.

As long as motivation is not the sole factor determining employment, there is the potential for unemployment insurance to have a policy impact, even in a country in which the people are fundamentally hard-working.

Nobody expects a government program to make everything different. So policy analysis is particularly useful in subtracting out the outcomes that would occur regardless of policy measures.

No comments: