Sunday, December 04, 2011

by Calculated Risk on 12/04/2011 10:45:00 AM

Yesterday someone sent me a column from John Mauldin that had the following graph. The graph shows the BLS Employment-population ratio, 16 years and older, and based on the trend, the author is expecting the employment-population ratio to rise to 65% in 2020.

I think this forecast is incorrect - and this gives me a chance to discuss the participation rate and employment-population ratio.

The employment-population ratio really increased in the '70s and '80s for two reasons: 1) favorable demographics as the baby boom generation moved into their prime working years, and 2) a rising participation rate for women. But that trend was about to change even if there hadn't been a severe recession.

Some definitions:
Participation Rate = Labor force / Civilian noninstitutional population

Employment-population ratio = Employed / Civilian noninstitutional population

Unemployment Rate = Unemployed / Labor Force

If we know the participation rate and the unemployment rate, we can calculate the employment-population ratio as follows:
Employment-population ratio = participation rate * (1 - unemployment rate).

This means that if the unemployment rate stayed steady, the employment-population ratio would follow the participation rate. This is important because the participation rate is impacted by changes in demographics - and we can forecast some of those changes.

The second graph shows the actual annual participation rate and two forecasts based on changes in demographics - both forecasts were made before the recent recession. Now that the baby boom generation is approaching retirement, the participation rate will decline.

Here are the two earlier papers with all of the author's assumptions:
• From BLS economist Mitra Toossi in November 2006: A new look at long-term labor force projections to 2050
• From Austin State University Professor Robert Szafran in September 2002: Age-adjusted labor force participation rates, 1960–2045

Those papers were written when the participation rate was in the mid-66% range. Based on demographics, Szafran had forecast the participation rate to fall to 64.6% in 2015, and Toosi had forecast the rate to fall to 64.5% in 2020. So some of the recent decline was expected - although it happened sooner and faster than either expected because of the severe recession.

And there might be reasons those forecasts were too high. First the participation rate of the 16 to 19 age group has fallen much faster than Toosi forecast (and might not bounce back much after the recession), and second, some people might have permanently given up.

I've made a similar calculation, and based on demographics, it is clear the participation rate would be falling even without a recession. Look back at the first graph - the projected increase in the employment-population ratio would require an increase in the participation rate (or a much lower unemployment rate than we've seen at full employment) - and that isn't going to happen.

The third graph shows the actual employment-population ratio. The two dashed lines are the calculated employment-population ratio using the actual participation rate - and two steady unemployment rates: 5% and 6%. You can see the impact of business cycles on the employment-population ratio (the participation rate is also impacted by business cycles, but much less than the employment-population ratio).

Using the participation projections from Professor Szafran and BLS economist Toossi, this would suggest an employment-population ratio in the 59% to 61% range in 2020 - not 65%.

Professor Krugman also used the employment-population ratio on Friday, but with one very important difference: he only used the 25 to 54 age group. From Krugman: Meh. And I Say That With Feeling

It could have been worse, but the basic story remains the same as it has been for 2 1/2 years: an economy that’s growing, but not enough to feel anything like a real recovery. The measured unemployment rate has trended down for a while, but it’s all basically reduced numbers of people actively searching. My favorite measure these days is the employment-population ratio for prime-age workers, which isn’t affected by changing demography. Here it is for the past decade; see the trend since the recession officially ended? Neither do I.
As Krugman suggests, the participation rate for the 25 to 54 age group has been fairly flat for the last 20 years so we can just look at the employment-population ratio.

This graph shows the participation rate for men, women and all in the 25 to 54 age group.

The participation rate for women had been increasing for decades, and really increased in the '70s and '80s. This is a key reason why the employment-population ratio was increasing (back to the first graph). But the participation rate for women has flattened out.

The participation rate for men has been slowly decreasing for some time. But the overall participation rate has been fairly flat - so for a quick look at the overall employment situation, the 25 to 54 employment population ratio is very useful.

However, when looking at the 16 and over population-employment ratio, we have to also analyze the trends for the participation rate. And the participation rate is expected to decline for the next couple of decades.