New Revisions of Labor Share show economy at the effective demand limit
Labor share data going back many years was revised last week. According to these new revisions, the economy has been against the effective demand limit since the beginning of the year. With the stock markets retreating from their record highs, this should not be surprising.
Keep in mind that the updated graphs below are based on my own theory of effective demand…
The effective demand limit determines true slack in the economy. Productive capacity is not established by what we should be able to produce, but by the relative strength of demand to purchase production.
The effective demand limit says that the combined utilization rates of labor and capital will not surpass an effective labor share. Therefore, slack in the economy is determined by…
slack = effective labor share – (capacity utilization * (1 – unemployment rate)) > 0
The UT index
The UT index is a measure of the slack (Unused Total capacity of labor and capital). Here is a graph of the UT index since 1967.
The UT index wants to stay above 0%. Firms (in the aggregate) do not increase profits when slack goes negative, because firms will pay the increased cost of employing more labor and capital, instead of consumers paying that cost.
The UT index shows that slack has been at the zero lower bound since the end of 2013. It bounced on 0% in the 4thQ of 2013 and is now hovering above 0%. … So how has unemployment been able to decline in 2014 if slack is zero? Well, labor share has risen off of its low at the end of 2013. There is more relative consumer power from labor.
If labor share stays constant though, unemployment and capacity utilization will be capped. That is to say… the unemployment rate will only mildly decline, and capacity utilization will stay constant or decline. That has been a normal pattern in past business cycles.
I project that within the next 4 quarters, the UT index will start rising as it did in previous business cycles. You might wonder why slack would start increasing during a business cycle. The reason is that a recession is forming, profit rates are dropping and firms start employing less labor and capital.
Limit on Employment and Capacity Utilization
The down-sloping lines in this graph represent profit maximization limits of effective demand. (monthly data)
The scatter plot of employment with capacity utilization shows increasing profits moving Northeast. Then, when the plot hits the limit lines, profits will increase by moving Northwest along the limit lines. I project that if employment rises, we will eventually see a corresponding decline in capacity utilization, as long as effective labor share stays around 74.5%. If labor share rises, the limit lines will move to the right allowing greater employment of labor and capital. Firms have to sacrifice some of their profits though to raise labor share.
Productivity
The pattern in the past is that productivity stalls against the effective demand limit (where real GDP equals Effective demand.) It is happening again. The last 2 quarters of 2013 showed productivity started to rise. But the first two quarters of 2014 show productivity dropped back to its previous trend. Productivity is so far bound within the effective demand limit.
Path of Real GDP
The green line shows Real GDP rising since the crisis in terms of labor and capital utilization. Even though many think real GDP grew a lot in the 2ndQ of 2014, it is still close to the polynomial trend line. The red line of effective demand is bumping against real GDP. The pattern in the past says that real GDP will start rising up near the red vertical line as the utilization of labor and capital (TFUR) stops rising on the x-axis. The process warns that a recession is forming. I project that this vertical movement will be noticeable within 4 more quarters.
Conclusion
With the new revisions of labor share, the economy is reaching the effective demand limit. The projection going forward is that there is little true slack in terms of labor and capital. The business cycle is peaking. Profit rates are peaking. The stock market has little true upside left. I suspect the next year will be troubling to those who do not see the effective demand limit.
Edward,
How does globalization fit into this scenario as expansion happens outside the economic benefit to the US economy (and labor) ? I know that multi national companies success is not US based.
edward
Hi Dan,
GDP is domestic production, or better said, value-added domestic production that even comes from importing overseas production.
So success of multinationals outside the US still shows up in GDP numbers, which translate into domestic utilization of labor and capital.
Then labor share reflects the pay of domestic labor to move domestic production as well as foreign production which creates value-added domestically.
Consumption is then part of GDP (domestic value-added production).
By the way, it is also important to watch the balance of the trade deficit. This could have an effect on effective demand. So far the trade deficit is staying within a normal range. So the effective demand limit of the US will affect other countries like Japan, China and Europe.
Edward,
I shall not go on further about your essentially theological assertion that the natural rate equals the NAIRU, but I will note that your measure of slack continues to use the unemployment rate rather than the employment rate. I would think that at least some of the 3% lower labor force participation rate should be counted as part of slack. I do not think we are quite as up against the limit as you are claiming.
I had a moment of clarity as I looked at a couple of your charts in combination.
You are creating an indicator which will tell us whether the growth of the economy is demand constrained. It 2007-2008 consumers had increased aggregate demand to record levels as they spent part of the bubble equity in their homes. When the bubble popped, there was a sudden decrease in demand. After a period of adjustment, demand returned in such a way that businesses could grow the economy by adding workers even though these returning workers were not producing an incremental increase in demand per work hour.
When GDP equals effective demand you have to move into a new growth regime. If, in the process of adding the worker, demand does not increase enough to pay the worker, the worker will not be added.
We have had different growth regimes in the post-war economy. The Baby Boom grew the economy even without productivity increases. Changing from net exports to net imports also changed growth factors. It seems to me that you should expect the constant in your equation to be different if the factors of growth are different. But they change slowly and you can only really identify effective demand if you actually know that there is a demand constraint.
Hi Arne,
There have been different growth regimes as you say… but the constant for effective labor share has held up in all of them.
Yet, you make an interesting comment… “When GDP equals effective demand you have to move into a new growth regime.”
Your statement is profound and i think true, from what happened in the late 90’s after real GDP hit the effective demand limit. The economy continued to grow upon an increasing labor share. Eventually that new growth spurt decayed and a recession formed. But it was definitely interesting how the economy resurrected itself from the effective demand limit with that growth spurt.
Hi Barkley,
The employment rate is simply… (1 – unemployment rate)
So they are the same really. It is depends on if you look at the part of the glass with water, or the empty part.
You measure a 3% drop in labor force participation as slack. Yet, if those people come back into the labor force, unemployment rate would go up. Then if you employed those people, the unemployment would go down. There is a balance between the inflow and outflow from the labor force that would tend to keep the unemployment rate moving slowly downward.
I see the unemployment rate reaching its natural rate, so I project that as hiring continues, unemployment rate would tend to go down, but more people will come back into the labor force. The net effect will be a very slowly moving unemployment rate around its natural rate. From my look at past data, the unemployment rate could come down more, but it will be a slow process.
Have we seen the bottom of the drop in labor force participation? I think yes. Will it try to rise some going forward? I think yes too.
But the 6.0% to 6.2% natural rate of unemployment that I see would be supported by people coming back into the labor force as hiring increases.
It is normal for labor force participation to rise in a business cycle even as the unemployment rate is coming down.
http://research.stlouisfed.org/fred2/graph/?g=Hvi
So people are hired faster than they come into the labor force. That is a normal process. But the effective demand limit still held when this happened in the past.
In this graph, you will see the unemployment rate bottom out around where the UT index of slack goes to zero. The late 90’s was a special case of increasing labor share, decreasing capacity utilization and increasing productivity.
http://research.stlouisfed.org/fred2/graph/?g=Hvj
Barkley,
One more comment… When you look at this graph…
http://research.stlouisfed.org/fred2/graph/?g=Hvi
You will see labor force participation rising from 1980 to 2000. But you will also see average unemployment trending lower. That is somewhat of a contradiction. If a larger % of the population is included in the labor force, one might think that the unemployment rate would rise. One might suppose that there would be more unemployed on average. However, jobs were created faster than people came into the labor force in general over that period of time. The main idea is that the unemployment rate has its own dynamic of equilibrium. Therefore the unemployment rate that we see, 6.2%, is the true rate to represent the dynamics in the labor market.
This is really the first time that labor force participation has fallen alongside unemployment. Before, as unemployment fell, labor force participation rose. We are seeing a strange collapse of the labor market with truly deep dynamics.
Edward,
I agree that there are some peculiar dynamics going on in the labor market. I note as a simple technical matter that the employment rate is not (1-UR). It is the percentage of the working age population that is employed. That does not equal what you say it does. The UR is measured based on who is in the labor force, which is my point, which in fact I think you understand.
Barkley,
Now I see. Yes, technically you are right.
Employment-to-population ratio is also the employment rate. I defined it as the flip-side of the unemployment rate, (1 – UR)…
Edward:
I believe Barkley is correct. What was your reasoning for: “defining it as the flip-side of the unemployment rate, (1 – UR)”
I like your analysis, but I’m wondering where the 0.762 in your FRED chart equation comes from. Is this a structural thing? Did you determine this number empirically?
Run,
In the equation for slack, you see (1 – UR). In the graph under the section “Limit on Employment and Capacity Utilization” I put employment on the y-axis because that shows increasing utilization of labor and capital going in the same direction.
One thing to realize… I never used the term “employment rate” in the post, which is the term Barkley is talking about. I only used the word “employment” by itself. I think his critical eye tried to find an error in the terminology, but maybe through selective perception saw an error that really wasn’t there.
Kaleberg,
There is a known relationship between capacity utilization and labor share. They move in a circular way. As one goes up, the other will go down. Then the other will go down. Then the other will go up. Then the other will go up. And the cycle starts over.
This circular movement has a center. That center moves on a trend line with a 0 y-intercept. So the trend line has a slope. So you get an equation like this for the central tendency trend line…
capacity util = slope * labor share index
The slope for the US data turns out to be between 0.76 and 0.77. I use 0.762. You could use 0.765 for example to increase effective demand if you choose, but 0.762 seems like the knife’s edge for me. So I go with it.
You will notice that 0.762 * labor share index gives you a value of capacity utilization. That value is effective labor share. That value also represents the center of the business cycle.
That center of the business cycle is also the proper definition of potential GDP. So comparing capacity utilization to its center tendency in terms of labor share, I can determine the output gap. That is why my model coincides with the output gap in the past.
Here is a graph…
http://effectivedemand.typepad.com/.a/6a017d42232dda970c01a3fd446082970b-pi
Thanks for the explanation of the 0.762. It’s derived from the data. I was wondering if it was a structural thing. For example, every unit of consumer spending requires 1/3 a unit of capital spending.
Also, I noticed the negative values in the chart in the 1960s. Could that have to do with productivity improvements being passed through to labor back then? (i.e. Labor share was more linked to the employment percentage. Our current employment “boom” is based on lower take home pay.)
HI Kaleberg,
Back in the 60’s there was a mindset to push unemployment down low. They inadvertently pushed beyond the natural level. This eventually contributed to problems of inflation that Volcker set out to control.
Now that we have inflation well under control, the UT index has not gone negative like that since.
The UT index would go negative depending on how willing policy is to push the economy past its natural level and risk inflation. The current debate is how inflationary the economy will become. Like today, with job openings hitting a high, some feel that inflation is on the verge of rising fast. We seem close to the natural level. However, wage growth is still slow. So the debate centers around how much slack is there really in the economy before we go beyond the natural level.
Lambert
I swapped out your Capacity Utilization for one that goes back to 1948. It’s CU for Manufacturing, not Total Industry, so it’s a little different. But it extends your first graph back enough to show the 1950s and 1960s.
http://newarthurianeconomics.blogspot.com/2014/08/farther-back-in-time.html
Arthurian,
That is interesting to see.
I posted a comment on your post.
Just to notice:
The UT Index moves down very quickly at the ends of recessions. Its behavior prior to recessions is much less synchronized.
So, from an investing perspective, a good temporal indicator of when to buy, but a much trickier indicator of when to sell…
Steve,
Good observation…
Also my memory of older RGDP/TFUR graphs is that they can jink and jank around the ED limit for a very long time — even years. This may be a result of widespread credit availability replacing labor share for a while.
Household credit is showing some slight rise right now, but it’s hard to imagine it can keep a finger in the dike for long. Lenders are perfectly aware that labor share sucks and shows little sign of mending, so they’re not throwing loans around freely… Mortgage rates are low but standards and reqd downpayments are high compared to 10%-down days…