Employment: tortoise slow, tortoise steady?

Since Summer 2011 job growth has generally outpaced population growth, adjusting for the retirement of the baby boomers. However, it's a small mountain that we need to climb, given the severity of the Great Recession. As a result, the economy remains several years away from normal levels – an optimistic projection shows we might be back to normal as early as summer 2017. More realistically, we're looking at late 2018 or early 2019, given headwinds to the economy. These include slowing global growth and a strong dollar, and the end of the oil boom, which is hurting investment faster than lower gasoline prices are adding to consumption. In any case, the economy remains 6 million jobs shy of where we need to be. That's reflected in many things, large and small. To give one example, I sit on the board of the local United Way of Rockbridge. We hear that local non-profits that attempt to meet emergency needs for utilities, food and rent see more rather than less need, with more working poor showing up than two years ago: jobs are failing to provide income sufficient to keep up with long-run needs.

Let me reiterate that the economy continues to improve, bit by bit. One indicator that I follow (which underlies the "normal job" level calculation) is the employment to population ratio. This avoids the challenge of counting discouraged workers, which over the past 8 years has swayed the unemployment rate in ways that make it a weak indicator of the job market. Now this participation rate likewise is imperfect as a measure, as it doesn't allow a distinction between part-time and full-time work, and one feature of the Great Recession was a large increase in those put on short hours. The BLS began collecting that data in 1994, and while it peaked at 6% during the depths of the recession, the current 4% rate remains higher than at any point during 1994-late 2008. The third graph provides that data for younger workers. Employment fell by 6% for prime-age workers during the Great Recession. Given noise in the data, it was unclear in spring 2013 that that rate had improved.

During the past two years, however, the share working has clearly been recovering, though it is still 3% below normal levels. The exception is among the young. Some 8% of those age 20-24 have yet to start their work-lives, relative to the stable rate prior to the Great Recession. While according to annual data from the BLS the majority of the difference appears to be accounted for by an increase of those in school, from 8-9% prior to the Great Recession to 13% in 2014. It's unclear to me as an economist what change in the economy would suddenly lead to education becoming more valuable. Instead, it's the opportunity cost that's changed: if (good) jobs aren't available, and in particular if career-oriented entry-level jobs aren't available, there's much less downside to remaining in school. (For those age 16-20 the shift is more dramatic: labor force participation among this group of young Americans fell from roughly 50% to 40%, which is a drop of 10 percentage points (or a 20% decline). Almost all of this appears offset by an increase in schooling. (See the table at the bottom.)

One component of slow growth is the lack of recovery in the housing sector. Now the rate of new housing starts shows a long-run decline, reflecting a decline in the birthrate, the aging of the population and a consequent decrease in the rate of new household formation. I've not tried to model that, and a quick search did not find any papers doing quite what I wanted. It is clear that household formation falls during recessions. For example, FT Alphaville notes the rise in children living with parents; there's no reason to think this is other than a response to the recession. (Kwan Ok Lee & Gary Painter (2013) "What happens to household formation in a recession? Journal of Urban Economics 76:1, 93-109 model this statistically.) What is clear is that housing starts remain very low and for the last year have shown no tendency to rise. So not only construction jobs but housing-related consumer durables suffer.

Will this component of our economy rebound, and offset the headwinds? Theory is unclear, as there are many margins of adjustment, tied to incomes of the young but also the age composition of the population, shifts in the nature of rental housing, and the price of owner-occupied housing. The empirical record is one of volatility. So there's no grounds that I can see for projecting change.

Finally, all this ties into interest rate policy. There's a 6 month lag between a change in interest rates and the start of the impact of higher rates on the economy, and the full impact is not felt for 12-18 months. If the economy will normalize in 24 months, then the Fed needs to start gradually raising interest rates later this year. Given the tendency of the Fed to move in increments of 25 basis points and FOMC meetings generating roughly 8 decision points per year, a 2016 start could lead to short-term rates of 2.5% by sometime in 2017. But if my analysis is accurate, we are still 3-4 years out, and there's no rush. Since it's easier to use monetary policy quell inflation than to spur growth, that reinforces the argument for delay: if you have to make a mistake, it's better to be too late than too early.

	Age 16-20			Age 21-25
Year	Labor Force Participation Rate	In School or Training	Either Work or School	Labor Force Participation Rate	In School or Training	Either Work or School
1998	56.2	31.0	87.3	80.3	7.6	87.7
1999	55.6	31.8	87.4	79.8	7.8	87.6
2000	56.0	32.1	88.1	80.0	7.5	87.5
2001	54.0	34.0	88.0	79.4	7.9	87.3
2002	51.8	35.9	87.7	79.1	8.3	87.4
2003	49.4	38.7	88.1	77.8	9.0	86.9
2004	48.7	39.3	88.0	77.4	9.4	86.8
2005	48.6	39.2	87.8	77.1	9.1	86.3
2006	48.5	40.1	88.5	77.4	9.5	87.0
2007	46.3	42.1	88.4	77.5	9.6	87.1
2008	45.5	43.0	88.5	77.2	10.0	87.2
2009	42.8	44.9	87.7	76.3	10.4	86.6
2010	40.7	46.9	87.6	75.1	11.4	86.5
2011	40.1	47.8	87.9	74.7	11.9	86.6
2012	39.9	47.8	87.7	74.3	12.2	86.5
2013	39.8	48.1	87.9	74.0	12.1	86.2
2014	36.0	53.2	89.2	73.2	13.3	86.5