We are privileged to invite readers to a serious debate and discussion. Recently we published two pieces in which we tried to describe markets in terms of metaphors. We thank readers for their many thoughtful responses.
In the second “metaphor” piece we incorporated three charts prepared by Bob Brusca, an economist and longtime friend. Another longtime friend who is also an economist, Joel Naroff, replied and disagreed with Bob. The result is reproduced below, unedited by me. We thank both of our friends for outlining their views and for giving us permission to make them public.
Here is the link to the second metaphor piece, in which I used Bob’s work: http://www.cumber.com/markets-metaphors-labor-data/.
Below is Joel’s reply to that commentary, and following that is Bob’s response, in which he disagrees with Joel. We hope readers enjoy the articulation of these different views.
From Joel Naroff:
Thank you for writing this piece and thank Bob for the research. You have provided the platform for me to make my favorite arguments why the participation rate and U-6 are irrelevant.
Let’s start with the participation rate. If you believe that the decline in the rate explains the decline in the U-3 rate, then you have to be consistent when the participation rate rose.
Consider the period January 1977 to January 1981. The participation rate rose 2.3 points. Using the Brusca rule, there should have been roughly a 2.1 percentage point rise in the unemployment rate. But during that time, the unemployment rate fell as much as 2.3 percentage points and ended up, after four years, down 0.3 percentage points.
So, what are the data telling us about the economy between January 1977 and January 1981? First of all, after four years, you can reasonably say that the unemployment rate was down 2.4 percentage points (2.1 points from the rise in the participation rate and 0.3 from the actual fall in the U-3 rate). That is a really outstanding performance. Also, the unemployment rate averaged 6.6%, which is 1.3 percentage points below where it was in January 1977. Given the huge rise in the participation rate, that is an awesome performance.
In other words, using the participation rate, we now know the answer to Ronald Reagan’s famous question: “Are you better off today than four years ago?” The answer is unambiguously yes. Congratulations, you have made Jimmy Carter an economic guru!
And then there is the U-6, which some call the real unemployment rate and I call the really stupid unemployment rate. One of the major components of the U-6 is the number of people who want full-time jobs but cannot get them. To make comparisons over time, you have to assume there has been no structural shift in the usage of part-time vs. full-time workers. But we know that businesses have been moving toward a higher usage of part-timers as they can more effectively employ them at lower costs.
That more people cannot get full-time jobs is related to the production function, which is shifting. It is not a measure of unemployment. Since the shift has been to more part-timers, the U-6 is biased upwards, which explains the different movements in the two. Very simply, you cannot compare two periods where the production function has shifted, changing the relative usage of labor. Now you can complain that businesses are keeping the “real” or “really stupid” rate up by using more part-timers, but that is a wholly different argument that I doubt you, Bob or any of the others that actually believe the U-6 is meaningful want to make.
Which brings us to Fed policy. If the U-6 doesn’t mean anything, then you have to go back to the U-3 and that is signaling a tight labor market. The problem we face is that given the global labor force and technology changing the way labor is employed, we do not have a good handle on what is full employment. Thus, we don’t know how much slack there really is in the market.
I don’t know how many times I have written this diatribe, but hopefully you will reprint some or all of the argument. You have my permission.
From Bob Brusca:
I would like to thank you for featuring some of my work on unemployment. And I would like to thank Joel Naroff for his insightful comments. Now I would like to add a few words to this discussion.
First of all I would like to dispute Joel’s claim that the participation rate should have symmetric effects. I don’t think we should glibly assume the rate impact is symmetric without a closer look. The participation rate is the result of some pretty complicated behavioral relationships and the interplay of supply and demand in the labor market. While economists view the labor/leisure trade off as a fundamental choice of a person, in Adam Smith’s ‘first book’ of the Wealth of Nations (to use the short title) cited one of the main causes of the wealth of a nation to be the proportion of its population that is at work. The participation rate is central to the performance of an economy and when it shifts it can have pronounced and broad macroeconomic effects and so when it shifts we need to think carefully about WHY it shifted to understand what the shift means.
In the US at least three important themes are impacting the participation rate. One is the use of opiates that has been documented to have reduced labor force participation (or a redlining of users by potential employers), the second is aging which is on auto pilot except over very long periods of time, and the third is technological/Global displacement and what it does to the displaced. Since we can’t impose the ‘Benjamin Button’ effect there is no chance that participation rates in the US are going to rise because old people are ‘youthing’ and participating more vigorously in the labor market as they ‘youth’- that won’t happen. Old people are participating in the labor market with more ‘vigor’ that their predecessors in the old-age cohort but that is mostly because this is the first generation to retire on 401Ks rather than on defined-pension benefits and few have saved enough. Even so, the participation rate falls so fast as we age there is virtually no chance that it will rise enough to make it level with the rate of the previous age cohort. So overall participation drops. And early retirement is becoming a feature so the labor force is losing people who disproportionately used to be employed. Now there are more people in this group that retires (and more retire early) and that will bias up the unemployment rate. Still, that does not mean that the lower rate is not ‘real.’
To understand the meaning and impact of the participation rate drop the question we need to ask is this “ is there some force acting on the unemployed today making it more likely that they will leave the labor force rather than go back to work with the same propensities as in the past?” If that is true that the unemployment rate will be lower than it would have been in the past at the same employment level.
The pace of innovation is brutal and as Joel remarks, the shift in the production function is to absorb more part time and less full time labor. These are events that have put more labor into unemployment or under-employment than in the past. It is robots and software too and all sorts of things that have cut out middle class jobs and other jobs, not just reduced hours.
At the same time the administration of disability has become more lax. There are many more people on disability than in the past – that affects the effective labor-leisure tradeoff.
Being in mid-career and out of work is stressful. If you do not go back to work and if you stop looking for work you are no longer unemployed, your work skills atrophy. The longer you are out the less marketable you are at anything like your former wage. I think there is good reason to see dislocation leading to higher unemployment and eventually to labor force reduction and eventually lower unemployment rates, as workers drop out and stop searching for a job. Once marginalized like this you do not get symmetry and roll the clock back; few have an epiphany go back to school for retraining get a job and fire their old boss who once fired them.
I grew up in Detroit. I worked on the assembly lines there to make money for college. When those jobs disappeared huge portions of Michigan (and other industrial areas suffered this fate) generated vast pools of unemployment this chain reaction was put in gear. Michigan has an ‘export economy.’ It made cars and trucks shipped them out of state and the money was shipped in to pay workers. Once that stopped there was no injection of money for that local community and no way for it pull itself up by its own bootstraps. So the cycle of labor market unemployment, discouragement, and drop-out kicked in and played out. These forces do not impact the employed and unemployed evenly and so there is a biased impact on the rate of unemployment. Joel is right. The production function has shifted. It has moved overseas in some cases!
This also gets to why I am vocal about FREE TRADE. I am in favor of it, of course. But I am opposed to the current world trading systems since China (and others) are like the roach motel: jobs go in and none come out. Our industrially or technologically generated unemployed do not get a crack at jobs ( apparently we have no comparative advantage which is quite a concept if you try to think about – Danger DO NOT ATTEMPT TO DO THIS WHILE DRINKING!). This is another reason why the unemployed pool in the US labor market is in heretofore unseen numbers and why a so many people simply ‘exit’ the labor market effectively impacting the participation rate instead of the unemployment rate.
All this happens in the face of the JOLTS survey with job openings data which are spectacularly hilarious. Someone should see if you could get the JOLTS report to host Saturday Night Live, I’m sure it would be a hoot. According to JOLTS there is a RECORD number of job openings. Apparently they are all for jobs which Americans lack the skills to fill. More likely it is that the openings are for jobs that will only be filled by H1B visa holders willing to work for even less and live six-to-an-apartment.
This comment gets to the point that you can call a ‘spade a spade’ but if it’s really ‘a club’ it will not fill out your flush in a poker game. The problem with JOLTS “job openings’ and ‘unemployment’ is that they are not static things. They may not even be the same things over time. Just because you call something by the same name does not mean it has not changed (if you called Hyde Jekyll would he change back?). It’s one of the reasons I will not buy pants or shoes on the internet. Everybody has a different idea of what ‘size is’. Women’s clothing is a hilarious example where you see women shopping for size ‘zero’ or ‘double zero’ or ‘one.’ It’s as though if they print a small number on garment you won’t be as big.
Well what I have discovered is if you also run regressions of average hourly earnings (wages) on the quit rate you find that more quits do raise AHE but not as much as they used to. So there some sense in which a quit is not a quit in the same way a unit of unemployment is no longer a unit of unemployment and a size zero is not the same thing either.