Natural Unemployment

In response to Republican questioning on monetary policy and employment, Bernanke replied the Fed Already Follows Policy Rule.

The Fed already has a rule,” Mr. Bernanke said during a panel discussion at the Brookings Institution’s Hutchins Center on Fiscal and Monetary Policy. “It’s committed to hitting a 2% inflation target and aiming for the natural rate of unemployment. These are rules.

Of course, Bernanke failed to admit the rule makes no sense, and it doesn’t. Leaving aside the ridiculousness of a 2% inflation target while ignoring asset bubbles, it is impossible for the Fed to target two things at once.

For example: The Fed can set interest rate policy but then it has no control over money supply. The Fed can target money supply, but then it would lose control over interest rates.

And the Fed cannot target employment at all. Thus his comment on “rules” is complete silliness. Moreover, there is no such thing as a “natural rate of unemployment” given central bank and government interference.

If Bernanke wants to discuss rules, I propose

  1. Raise the minimum wage and employment goes down vs. what it would otherwise do.
  2. Hand out enough free benefits and people have no desire to work.
  3. The more disability fraud government allows, the less incentive people have to work.

Making Sure You Don’t Count

Logic would dictate the unemployment rate would go up as a result of the three points above. It doesn’t because the unemployment rate is carefully defined in such a manner that it doesn’t.

If you are in school and working you are in the labor force and employed. If you are in school and not working you are not in the labor force and therefore not unemployed. If you collect disability and are not working, you are not unemployed no matter how fraudulent the disability claim.

Not employed and looking for a job on Monster or in the want-ads? Sorry, that does not count. You are not in the pool of the unemployed. The list goes on and on, and it is hugely arbitrary.

The BLS makes every effort to ensure you are not counted as unemployed if you are not working. Don’t blame the BLS, they are just following the rules.

Right now, the BLS says the unemployment rate is 5.5%. Bernanke believes the “natural unemployment” rate may be lower than he thought previously.

Of course, if you hand out enough free benefits, the unemployment rate would drop to zero given the way the unemployment rate is calculated.

Employment vs. Non-Employment

One is either employed or not.

The “non-employment rate” calculation is easy enough: ((civilian population – employment) / civilian population) * 100.

Let’s define unemployment the same as non-employment. There are a few potential problems but numerous benefits with this method, so let’s discuss them.

Problems in Normalizing Unemployment Rates

  1. Changing definition of unemployment over time
  2. Changing demographics – Boomers and Boomer Retirement
  3. Women entering the labor force in increasing numbers over time 
  4. Increasing numbers of kids going to college and then on to higher education

Point number 1: The unemployment equals non-employment calculation is so simple that changes in the previous definitions of unemployment vanishes. Calculations of who wants a job, who is looking for a job, etc., vanishes, as does disability fraud. This is to the clear advantage of the non-employment = unemployment definition.

Point number 2: With the aging workforce, boomers are retiring en masse. People live longer than they ever have. Many keep working because they want to. Many others want to work past age 65 but are frustrated by the lack of opportunities and stopped looking. Still others have built up a sufficient nest egg and retired early. Untangling all that is very difficult under current rules as what constitutes unemployment. Thus, in the non-employment = unemployment model, demographic skew enters into play.

Point number 3: Starting in the 1960s, an increasing number of women entered the work force. Some women took jobs because they wanted to, others because they had to. Right now, it is safe to conclude we are not going back to an environment where following marriage, the male head of household works and the female stays home taking care of the kids.

Point number 4: More kids than ever before are going to college. Many go on to higher education. But what about people returning to college not because they want to but because they are out of a job and feel like they need to?  In the non-employment = unemployment model, education  skew enters into play. However, those who would rather be working than in school are properly accounted for.

Normalizing Unemployment

Is it possible to normalize the above? I believe it is. Point number 1 is self-explanatory. It automatically normalizes previous definition changes via simplicity. In fact, it corrects for previous errors automatically.

We address point number 3 with an assumption that we are not going back to an environment where following marriage the male head of household works and the female stays home taking care of the kids. A comparison of unemployment rates between 2010 and 1950 is invalid but comparisons from peak entry of women in the workforce going forward are valid.

Points number 2 and 4 can be addressed via elimination of retirement and education issues. We do this by careful selection of the “core age group” as follows.

New Definitions

  • Core Age Group: 25-54
  • Core Employment: Number of people employed in age group 25-54
  • Core Unemployment: Number of people in age group 25-54 who are not working.
  • Core Unemployment Rate: ((civilian population 25to54 – employment 25to54) / civilian population 25to54) * 100.

The definition of “core unemployment” allows for six years of education and retirement at age 55, both generous allowances.

Realistically, everyone else should be working except a relatively small (and constant) percentage of people who are genuinely disabled.

If the “core unemployment rate” is on the rise, it represents one of four possibilities:

  1. Disability fraud
  2. Overly generous benefits
  3. Structural issues such as lack of skills
  4. Cyclical issues

Six years into a recovery, I think we can safely remove point number four.

Core Unemployment Rate 1948 to Present

In the above chart you can see the clear effect of women entering the workforce. The absolute bottom in the core unemployment rate was in April of 2000 at 18.07%.

The second lowest rate was January of 1999 at 18.18%. The first dip below 19% was in February of 1998 at 18.97%.

Let’s zoom in on a time frame that catches peak entry of women in the work force.

Core Unemployment Rate 1985 to Present

Let’s now investigate some interesting developments since the year 2000. Specifically, let’s compare March of 2000 with March of 2015.

Core Employment March 2000

Core Employment March 2015

Core Employment March 2000 vs. March 2015

For age group 25-54, there were 1,791,000 more people employed in March of 2000 than in March of 2015 even though the population of that age group increased by 4,712,000!

This is remarkable given age that group 25-54 factors out retirement and allows for six years of education. On a population-adjusted basis, age group 25-54 has a staggering 6,503,000 employment shortage compared to the year 2000. This does not even factor in part time employment.

If we make the assumption that women in the workforce should stabilize, it follows that core employment should stabilize in the neighborhood of 80-82% and core unemployment should stabilize in the neighborhood of 18-20%.

Age Group 25-59 Analysis

In terms of normal retirement age, age group 25-59 might seem to be a better fit, but Fred does not have the Data.

The BLS does have the data, but we have to manually add age groups together. Let’s do a quick calculation of March 2000 compared to March 2015.

Employment Subtotals 25-59

25-54 Employment March 2000: 98,292,000
55-59 Employment March 2000: 09,020,000
25-54 Employment March 2015: 96,501,000
55-59 Employment March 2015: 14,653,000

Population Subtotals 25-59

25-54 Population March 2000: 120,237,000
55-59 Population March 2000: 13,324,000
25-54 Population March 2015: 124,949,000
55-59 Population March 2015: 21,389,000

Employment and Population Totals 25-59

25-59 Employment March 2000: 107,312,000
25-59 Population March 2000: 133,561,000
25-59 Employment March 2015: 111,154,000
25-59 Population March 2015: 146,338,000

Unemployment Rate Age Group 25-59

  • Age Group 25-59 Unemployment March 2000: 19.65%
  • Age Group 25-59 Unemployment March 2015:  24.04%

Structural Unemployment

Before the housing bubble burst, cyclical unemployment with each recession eventually corrected itself.

We now have huge, unemployment (non-employment if you insist) that is not accounted for. To date, we could only see this in the declining labor force and participation rate.

Yet, the participation rate itself is a poor measure. It is declining in part because of aging boomer demographics and in part because of the excess number of people who dropped out of the labor force.

Disability fraud also entered the picture. Starting with a simpler definition of “unemployment” smooths all of those factors out.

What we are left with is an economy that is roughly 6,503,000 jobs short of where it should be following six years of recovery (in age group 24-54 alone) .

That’s how bad things are, at a minimum, and for the core group alone. And that’s precisely what is wrong with existing measurements and economic analysis touting the “recovery in jobs” and “new highs in employment”.

Many people tell me this does not “feel” like a recovery.

For 6.5 million people in “core age group” 25-54 there has been no recovery. And that number does not account for other age groups, for underemployment (e.g. engineers working in restaurants), for part-time employment, or for declining real wages for all but the top 10-20% or so.

Mike “Mish” Shedlock