Thursday, August 17, 2017

Stop doing psychology and start doing econ

Brad DeLong:

[M]ore than a generation of inequitable and slower-than-expected economic growth in the global North has created a strong political and psychological need for scapegoats.

Well fuck you Brad DeLong. For inventing stories about my psychological need for scapegoats, fuck you.

DeLong continues, regardless:

People want a simple narrative to explain why they are missing out on the prosperity they were once promised, and why there is such a large and growing gap between an increasingly wealthy overclass and everyone else.

I can give you the simple narrative: Excessive private debt is the reason we are missing out on the prosperity we expect, and excessive private debt is the reason there is such a large and growing gap between an increasingly wealthy overclass and everyone else. End of narrative.

If Brad DeLong would stop doing psychology and start doing econ, he could see it for himself.

Wednesday, August 16, 2017

Too much fertilizer

The Symptoms of Over-Fertilizing:

Fertilizing plants encourages healthy growth and flowering, but too much leads to problems.

Some signs of over-fertilizing are easy to spot... And though fertilizer should encourage healthy growth, too much can stunt growth or stop it entirely.

Is it possible that the emphasis on free trade and globalization is just too much fertilizer?

Tuesday, August 15, 2017

Voluntary vs Involuntary Unemployment

Pettinger ("helping to simplify economics") offers this definition:

Voluntary unemployment is defined as a situation where the unemployed choose not to accept a job at the going wage rate.

That's it exactly. And by choosing to define voluntary rather than involuntary unemployment, Pettinger simplifies economics. Here is Keynes:

If, indeed, it were true that the existing real wage is a minimum below which more labour than is now employed will not be forthcoming in any circumstances, involuntary unemployment, apart from frictional unemployment, would be non-existent. But to suppose that this is invariably the case would be absurd.


Moreover, the contention that the unemployment which characterises a depression is due to a refusal by labour to accept a reduction of money-wages is not clearly supported by the facts.


For, admittedly, more labour would, as a rule, be forthcoming at the existing money-wage if it were demanded.

In other words, involuntary unemployment is a situation where the unemployed are willing to accept a job at the going wage, but do not find work. This meshes perfectly with Pettinger's definition.

And then there is Scott Sumner:

We all agree that there were lots of people without jobs. We all agree that lots of them wanted to be working. We all agree that lots of them were miserable. I call that “involuntary unemployment.”

Number one, Sumner gets the definition of "involuntary" unemployment wrong. It's not about being out of work. (That's just "unemployment".)

Number two: I omitted it, but Sumner prefaces these thoughts by saying, "But what is so obvious about involuntary unemployment, as defined by Keynes?" That's Sumner's emphasis on the words "as defined by Keynes", not mine.

The first three sentences after those four words all start with the phrase "We all agree". Sumner apparently agrees with Keynes three times. But he is putting words in Keynes' mouth, as we find out in the next sentence after, where Sumner admits he has been giving us his own definition of involuntary unemployment.

But Sumner's is not a definition of involuntary unemployment. It is only a definition of unemployment: without a job, wanting work, and miserable. And it's not even a technical definition of unemployment. It's just some social chatter.

"Involuntary unemployment" is a technical term, defined by Keynes. Pettinger simplifies the concept by choosing to define voluntary employment instead: a refusal to accept work at the going wage.

Funny thing: Sumner gets that wrong, too. He says:

I think they were unemployed because of sticky wages, and that if workers collectively accepted lower wages then we would have had full employment in 1936.

The Depression drove wages down, so that the "going" wage was lower during the Depression than before. Sumner is saying workers refused to accept the going wage (in 1936) because it was lower. This refusal, by definition, makes the unemployment in Sumner's story voluntary. Sumner calls it involuntary. Sumner is wrong.

Keynes establishes the definition:

The classical school [argue] that if labour as a whole would agree to a reduction of money-wages more employment would be forthcoming. If this is the case, such unemployment, though apparently involuntary, is not strictly so ...

Pettinger simplifies it:

Voluntary unemployment is defined as a situation where the unemployed choose not to accept a job at the going wage rate.

It's not complicated. Sumner is wrong.

Monday, August 14, 2017

I hate politics

Replying to Mark Thoma, I wrote:

Half a dozen years back there was so much right-wing chatter about the threat of inflation (from quantitative easing) that it seemed the chatters wanted hyperinflation, just to show they were right about the economy.

... It's not quite so bad these days, not yet at least, and I'm not sure it's only a left-wing phenomenon, but these days the chatter seems to be about impending recession.

Why would "impending recession" be a left-wing phenomenon?

Why, indeed. On the day after the Presidential Inauguration last January, Adam Ozimek was at Forbes with The Best Case For The Trump Economy. Ozimek is not a fan of Trump, as his concluding remarks indicate:

So if everything goes smoothly and Trump doesn't do anything crazy, or anything really, he could potentially keep the current pace of job growth running for a few more years and maybe even for the whole term...

If I was Trump, I'd do the absolute minimum policy-wise and make job #1 avoiding a recession.... I'd also probably change the hairdo.

Not a fan, but not so full of hate for the man that he can't do econ. Many people are.

Anyway, in the same article Ozimek says

If you look over the long-term, the employment to population rate falls during recessions then slowly recovers. So if we can avoid a recession, it's not crazy to think we could get back to 2000 levels.

I quoted that yesterday, talking about Trump's 4% growth target. Ozimek's point is that for Trump to succeed in making America's economy great again, he's going to have to avoid recession.

And there, right there, is the reason the chatter about impending recession is a left-wing phenomenon.

Sunday, August 13, 2017

Trump's 4% growth target

Trump apparently wussed out, abandoning his 4% target for economic growth and embracing 3%. I read a couple things ridiculing him at 3%, as if even that is too high a target. Absurd.

Myself, I'd hold Trump to his original target: 4%. We can get there. Even if we don't make it, shooting for 4% is a much more honorable goal than shooting for 3%. And far more impressive, if we make it.

Back in January 2017, Menzie Chinn of EconBrowser looked at Trump's growth target. Quoting from the economy page of the Trump-Pence website, Chinn echoes

Boost growth to 3.5 percent per year on average, with the potential to reach a 4 percent growth rate.

The Trump-Pence link today reports a "page not found" error. Wuss.

Anyway, Menzie points out that reaching Trump's growth target will require a large increase in the size of the labor force or in labor productivity, or both. Unrealistically large, Chinn implies. He says

it seems unlikely to have acceleration of growth to the indicated rates, especially if policies are undertaken to deport some portion of the population


If we are deporting undocumented workers en masse, how is [the labor force] going to be expanded?

It's a fair question.

At Forbes (21 Jan 2017) Adam Ozimek said

If you look over the long-term, the employment to population rate falls during recessions then slowly recovers. So if we can avoid a recession, it's not crazy to think we could get back to 2000 levels.

Ozimek continues:

That means going from 78.2% today to 81.9%, an increase of 3.7 percentage points. If you apply this to the 125 million population, that's about 4.6 million jobs. If these jobs take the whole four years of the Trump administration to recover, that's 95,000 jobs a month above and beyond the natural growth in the labor force. When you add in that natural labor force growth, that should be enough to keep job growth in the range from the last three months, which has averaged 165,000.

That was January. Job growth in June 2017 surged "as employers surpassed the expectations of most economists by adding 222,000 jobs."

As that kind of job growth continues, Trump's 4% becomes more likely.

At FiveThirtyEight (4 August 2017) Ben Casselman writes:

Meanwhile, the U.S. doesn’t seem to be running out of available workers, at least not yet. Rather, the improving job market seems to be drawing people off the economy’s sidelines. The labor force grew by 349,000 people in July; the so-called participation rate — the share of adults who are either working or actively looking for work — has been essentially flat for the past year and a half. That’s an impressive trend given the ongoing retirement of the baby boom generation, which puts downward pressure on the participation rate.

Things are happening in the labor force.

// Related: You don't just show a graph and assume that the trend it shows will continue forever

Saturday, August 12, 2017

Defining Aggregate Demand Deficiency

Nick Rowe defines Aggregate Demand Deficiency: "It means that the stock of money is too small, or is circulating too slowly, to buy and sell all the things that people want to buy and sell."

It sounds like Nick is saying the velocity of circulation of money is not influenced by the decisions of spenders. I would say it is the decisions of spenders that determine the velocity of circulation.

Suppose Nick accepts the view that spenders determine the velocity of circulation. How does that fit into his definition?

"There are too few spenders, or they are spending too slowly, to buy and sell all the things that people want to buy and sell."

But "people" are the spenders.

"People are spending too slowly to buy all the things that they want to buy."

So it seems people are confused, wanting to buy things but not wanting to spend the money.

Most of us are familiar with that problem, I think. And if we want something but decide not to spend the money, demand is reduced because of our decision. Our wants and desires only count as demand if we actually spend the money.

Nick Rowe's definition again: The stock of money is too small, or is circulating too slowly, to buy all the things that we want to buy.

"Want to buy" is not quite right. Everybody wants a coach and six, Adam Smith said. Everybody wants a Corvette. But not everybody gets one. All of that "wanting" is not demand.

Okay, so maybe 1000 people were actually going to buy Corvettes but it turns out that only 750 of them can afford to buy one. There is a demand shortfall of 250 Corvettes, 1000 minus 750. The million people who "want" Corvettes do not come into the calculation at all.


That definition again: An "aggregate demand deficiency" happens when money is circulating too slowly to buy all the things that people want to buy.

But of course we can always just go out and borrow the money, and buy whatever we want. But maybe we don't want to borrow enough to get that new Corvette. Or maybe our bank is telling us we don't want to borrow that much.

An aggregate demand deficiency might happen because we are borrowing less. And that's not really the same as "the stock of money is too small, or is circulating too slowly". It is, and it isn't. Borrowing money is a way to make the quantity of money bigger (if you count bank money as money) or it is a way to make the velocity of circulation increase (if you don't). But borrowing money has other effects also: It makes our debt bigger. It makes our debt service bigger. It reduces our "after debt service" income. It reduces the money we'll have in the future for spending to buy all the things that we want to buy.

In the long run, borrowing money creates an aggregate demand deficiency. The U.S. economy slowed after 1973? Aggregate demand deficiency since 1966. But we ignored it and kept borrowing anyway, borrowing more and more, because policy paved the way for that. And then, after 2007, we didn't keep borrowing anymore. And suddenly the stock of money was too small and we had ourselves an aggregate demand deficiency we couldn't ignore. As a consequence of borrowing and of the effects of too much borrowing for too long a time.

We borrowed too much, early on (in the 1950s and '60s). It became a problem in the 1970s. They fixed the problem in the 1980s and '90s by making it possible for us to borrow even more. But they solved the wrong problem. The problem was not that we couldn't borrow all that we wanted. The problem was the growing cost of the growing debt. They tried to solve the problem by encouraging us to add even more to our debt.

Borrow more we did. Add to our debt, we did. Debt grew faster and faster. But the overall cost of that debt grew also. And paying the debt service took money away from "aggregate demand" and created an "aggregate demand deficiency". And the economy grew slowly because of it.

But all that ended after 2007, when people suddenly realized their debt was a problem. And people started borrowing less and paying back more. And the economy suddenly tanked.

So here is the big picture: In the early years we have too much borrowing (but not a lot of debt) so aggregate demand is high and economic growth is good. In the middle years we have too much borrowing (and too much debt) so we have an aggregate demand shortfall that slows the economy despite all the borrowing. And in the later years we have no borrowing (and still too much debt) and the economy tanks.

So there is a little more to the story than "the stock of money is too small, or is circulating too slowly". There is also the cost of finance.

Friday, August 11, 2017

A weapon against the nation-state

Economic Theory: Limitations and Biases at Conversable Economist, 6 July 2017:

Arnold Kling tackles the hardy perennial topic "How Effective is Economic Theory?" in the Summer 2017 issue of National Affairs. His overall approach is to focus on "five interlocking subjects in particular: mathematical modeling, homo economicus, objectivity, testing procedures, and the particular status of the sub-discipline of macroeconomics." He then compares and contrasts what economists were saying about those subjects in 1966 and 1980, compared with his views on current patterns. For details, read the essay! But here are few excerpts that caught my eye and may give some flavor of his discussion:

"Economists are not without knowledge. We know that restrictions on trade tend to help narrow interests at the expense of broader prosperity. We know that market prices are important ...

There is a lot more, excerpts from Kling and thoughts on the excerpts. I want to cut it off right at the start and deal with the first piece of knowledge presented: Restrictions on trade tend to help narrow interests at the expense of broader prosperity.

It is a general statement, offered as something more than a rule of thumb. A lot more than a rule of thumb: Restrictions on trade are costly. Period, end of story.

It is presented as a given. Something economists know, and know for sure.

I have trouble with the period, the given, the certainty. But most of all, I have trouble with the incompleteness of that particular piece of knowledge. When the "narrow" interests are those of nations and the "broader" prosperity is that of global corporations, that harmless little piece of knowledge becomes a weapon against the nation-state.

How do the world's biggest companies compare to the biggest economies?

How will they compare when government is small enough you can "drown it in a bathtub"?

Thursday, August 10, 2017

Gotta love that focus on interest rates

From Bridging the Gap: Forecasting Interest Rates with Macro Trends, an FRBSF Economic Letter by Michael D. Bauer:

... interest rates affect household finances, business funding costs, government debt, and ultimately the health of the overall economy.

Yup. And everybody focuses on interest rates. But nobody worries about the size of accumulated private debt, which is the money on which interest has to be paid. Public debt, too; but if I didn't specifically say "private" you'd think I meant only the public debt. (Public debt is the least of our worries, or anyway the lesser.)

Wednesday, August 9, 2017

"the largest effect of the mortgage deduction is on household financial decisions, inducing them to increase indebtedness"

From the NBER access page for Do People Respond to the Mortage Interest Deduction? Quasi-Experimental Evidence from Denmark by Jonathan Gruber, Amalie Jensen, and Henrik Kleven:

Using linked housing and tax records from Denmark combined with a major reform of the mortgage interest deduction in the late 1980s, we carry out the first comprehensive long-term study of how tax subsidies affect housing decisions. The reform introduced a large and sharp reduction in the mortgage deduction for top-rate taxpayers, while reducing it much less or not at all for lower-rate taxpayers. We present three main findings. First, the mortgage deduction has a precisely estimated zero effect on homeownership. This holds even in the very long run. Second, the mortgage deduction has a sizeable impact on housing demand at the intensive margin, inducing homeowners to buy larger and more expensive houses. Third, the largest effect of the mortgage deduction is on household financial decisions, inducing them to increase indebtedness. These findings suggest that the mortgage interest deduction distorts the behavior of homeowners at the intensive margin, but is ineffective at promoting homeownership at the extensive margin and any externalities that may be associated with it.

(emphasis added)

The mortgage interest deduction encourages the growth of debt. We could have instead a policy that offers a tax benefit for making an extra payment or two during the tax year. Such a policy would discourage the growth of debt. And it could be designed to offer approximately the same tax benefit as the existing policy.

Tuesday, August 8, 2017

Parallel lines of thought

Douglas Campbell shows a graph I've not seen before -- total nonfarm employment "now something like 23% below the long-run trend" -- and says

Of course, there are caveats here. Population growth did naturally slow a bit, and the absorption of women into the labor force was a one-time event that was mostly played out by the 2000s; 9/11 exogenously reduced immigration, and thus job growth, and the Boomers have been retiring, etc... [But] why should exogenous negative shocks to labor supply cause wage growth to slow?

They shouldn't. As if in answer to Campbell's question, J.W. Mason writes

If employment is falling due to demographics, that should be associated with rising productivity and wages, as firms compete for scarce labor.

Monday, August 7, 2017

Explanation versus evidence

The reddit title caught my eye: Intangible capital making output gap & Phillips curve irrelevant? There is a good opening:

Here’s my thesis upfront: The growing dematerialization of capital accumulation has material consequences for financial markets and for the conduct of monetary policy: It helps to explain why there is a corporate savings glut that contributes to ultra-low interest rates, and why time-honored concepts used by generations of economists such as the output gap and the Phillips curve are becoming increasingly, well, immaterial.

Next, they explain what they mean by "immaterial":

Our economy is becoming ever more intangible. A rising share of our consumption is services rather than tangible goods. And to produce these services and goods, firms nowadays typically invest more in intangible capital than in physical capital. New ideas generated by smart people, patents, copyrights, brand, software and cloud space matter more than bricks and mortar, machines and inventories. In short, production processes are dematerializing – more sales are generated with less physical capital.

This caught my interest. Not because it's right, particularly -- it's way too soon to make a judgement call like that -- but because I never heard the story before. New ideas are rare and precious, even when they might be wrong.

It caught my interest. So I kept reading:

A growing body of research suggests that the progress of intangible capital has been the most important fundamental driver of the secular uptrend in corporate cash holdings over the past few decades...

Sentences like that unfailingly fool me into thinking I'm seeing evidence. However, such sentences are only stories about evidence; the conclusions they reinforce remain unsupported.

The paragraph continues:

The reason is that, unlike physical capital, intangible capital cannot easily be pledged as collateral for loans.

This sentence does not present any of the evidence we've just been told about. Rather, it plunges deeper into the story, giving us a "reason" for "the progress of intangible capital". Please note: I'm not saying it's all bullshit. I like the story, despite myself. It's just that, on second read, I'm seeing a story supported by explanation rather than evidence. So far, it remains only a story.

The rest of the paragraph builds on the unsubstantiated analysis:

Therefore, to maintain the financial flexibility to invest in intangible capital – spending more on R&D, buying start-up companies with smart people and products, acquiring or developing brands and patents, investing in artificial intelligence etc.—firms optimally hold larger cash balances. Of course, other factors, such as a U.S. tax system that only taxes corporate profits generated abroad once they are repatriated, also contributes to swelling (foreign) cash holdings. However, the dematerialization of capital investment seems to be the single biggest factors explaining the corporate savings glut.

As I say, it's a good story. And here's how the next paragraph starts:

This corporate savings glut, along with a higher desire by private households to save more in the face of rising longevity and with strong demand for safe assets from emerging economies, is the main reason why the natural or equilibrium rate of interest is very low.

Wow, see? It even explains why the natural rate of interest is so low. (Told ya it was a good story!)

But, well, the corporate savings glut, I've heard of that. And the strong demand for safe assets. And of course the low natural rate of interest. I am not yet sure how these factors fit my own economic theory, but at least I've heard of them.

But I never heard of the "higher desire by private households to save more in the face of rising longevity". Never saw a story about a connection between longevity and savings. It seems to make sense, yes, but it's just a story. One small piece of a larger story. But this storybook connection is something I can actually look at: Is it true that increasing longevity has brought a rise in the personal saving rate?

FRED says No:

Graph #1: Life Expectancy and Personal Saving
Life Expectancy and Personal Saving run in opposite directions. When life expectancy was low, the saving rate was high and rising. As life expectancy went up, the saving rate turned and went down. And now life expectancy is high and the saving rate is low, try as it might to increase.

So I don't see "a higher desire by private households to save more in the face of rising longevity". I can easily believe there is a higher desire to save, a desire that is never actualized. But if it doesn't show up in the numbers, it's only a story.

By putting longevity and saving together in a single thought, they gave me a relation to look at. So I looked. And, hey, I looked at only the one graph. But it's not there. The relation is not there.

I tried to participate. I went looking for evidence. I tried to confirm one little thing they told me. The FRED graph contradicts the story. So I lost interest in the story they tell, because it's only a story.

Sunday, August 6, 2017


Reuters: "U.S. employers hired more workers than expected in July and raised their wages, signs of labor market tightness that likely clears the way for the Federal Reserve to announce a plan to start shrinking its massive bond portfolio... Average hourly earnings increased nine cents..."

Sure. But what about "wages have been stagnant since the '70s"... Isn't there some catching-up to be done? Does the Fed need to kill this thing as soon as it starts?

Just a reminder. In April of 2016 I wrote:

I predict a boom of "golden age" vigor, beginning in 2016 and lasting eight to ten years. It has already begun. In two years everyone will be predicting it.

The debt service ratio has been a little slow to pick up, I'll give you that. But we are seeing more and more news like the Reuters story above. These are indications.

Saturday, August 5, 2017

Federal Employees as a Percent of All Employees

I marked up a graph from yesterday's post:

Graph #1: Federal Government Employment as a Percent of All Employment
Or look at it this way. Take that blue line, showing Federal employment as a percent of total employment, and put it on a graph along with the growth rate of Real GDP. The employment line makes a remarkably good trend line for RGDP growth:

Graph #2: Real GDP Growth (red) and Federal Employment as a Percent of All Employment (blue)
Weird, huh?

Friday, August 4, 2017

Another look at Federal employment

Back in September of 2011 at Business Insider, Doug Short looked at The Incredible Shift From Manufacturing To Services In America:
I spent some time this morning studying a topic I've occasionally mentioned: The shift in the United States from a manufacturing to a services economy.

 The Department of Labor's Bureau of Labor Statistics has monthly data on employment by industry categories reaching back to 1939. The first chart below is an overlay of the compete series of employment numbers for the two major categories, manufacturing and services.

When I say major, I'm referring to the complete domination of the labor market by these two industries. To illustrate this fact, I've also included the total of the two categories and a dotted line showing total nonfarm employment.

An impressive chart. Manufacturing is the red line down at the bottom. Services is the blue line up near the top. The dotted gray line is manufacturing and services together. And the green is all employment. Scarcely any room between green and gray. Most jobs are either in manufacturing or in services.

Most of them in services, according to the blue line. Not much going on with manufacturing, dragging along the bottom there, red with embarrassment.

I got wondering how the Federal employment numbers compare to manufacturing employment. Both are shown relative to the size of the economy, same as I showed yesterday. Have a look:

Graph #2: Manufacturing (red) and Federal (black) Employment-to-RGDP Ratios
Manufacturing (red) appears to have fallen much more rapidly than Federal government employment. But Federal employment (black) is a much smaller number, so appearances may be deceiving. To get a better idea of the relative decline, we can look at manufacturing employment relative to Federal employment:

Graph #3: Manufacturing Employment relative to Federal Government Employment
Now there isn't such a steep decline. The ratio runs almost flat, between 6 and 7, from 1950 to 1980. It falls below the 6 level in the mid-1980s. But manufacturing recovered in the 1990s, and the ratio climbed up above 6 again. Stayed there until 2001.

Manufacturing employment has fallen significantly faster than Federal employment since 2001, the graph shows. But for the whole second half of the 20th century, it was almost neck-and-neck. Manufacturing employment and Federal employment fell at almost the same rate. So where Doug Short speaks of an "incredible shift" away from manufacturing, we could as easily speak of an incredible shift away from Federal employment. But no one speaks of such things.

Remember, too, that Graph #3 shows the ratio running largely between 6 and 7. That means manufacturing employment was 6 to 7 times the size of Federal government employment, consistently, until the 21st century.

It was pretty easy to duplicate Doug Short's graph at FRED. I even tried to make the colors match. Then I added a dashed black line to show Federal employment. It runs way down at the bottom, a mere fraction of manufacturing (red) employment. If I didn't point it out, you might not even notice it:

Graph #4: Federal Employment (dashed black) Barely Above Zero
Since the end of the Korean war, Federal employment doesn't even wiggle.

Compare the lowest line to the highest line: Compare Federal employment (dashed black) to All employment ("Total Nonfarm Payrolls", green):

Graph #5: Federal Employees as a Percent of All Employees
Set World War Two aside, and the Korean war. The ratio runs reliably between 4% and 5% from the late 1940s to 1970. But it is gently trending down, and reaches the 4% level around 1970. The downtrend continues; after the very early 1970s the ratio is never again close to the 4% level. Since 2000 it has been running at the 2% level -- half what it was in 1970. And remember, Federal employment has been falling damn near as fast as manufacturing employment: "incredibly" fast.

Two percent of the workforce is employed by the Federal government today. Four and a half percent of it, give or take, was employed by the Federal government in the 1950s and '60s, back when our economy was good. If there is a problem with our economy, that problem cannot be the excessive size of the Federal government. The problem lies elsewhere.

Thursday, August 3, 2017

Less than 200 Federal employees per billion dollars of Real GDP

"I'm pretty sure that by 'government' they mean all levels of government, not just Federal." That's what I said yesterday, showing this graph of the decline in the size of government relative to the size of the economy.

Graph #1
But I was only "pretty sure" they meant all levels of government. Because I hadn't seen it yet. So I went looking. And I found a series named All Employees: Government: Federal which tells me that the series named All Employees: Government counts employees of state and local governments in addition to Federal.

To actually see it, I looked at "Government: Federal" as a percent of "Government":

Graph #2: Federal as a Share of Government Employees
Half of all government employees during the Second World War were Federal employees. 50 percent (and that's not counting military personnel). Something less than 40% were Federal employees at the time of the Korean war. From there, the number continued to drop: 35% around 1954, 30% around 1957, then lower yet; and since the latter 1990s, less than 15 percent.

The first graph shows that the number of all government employees fell, relative to the size of our economy. The second graph shows that the number of Federal government employees fell as a share of all government employees. So I'm thinking that the number of Federal government employees must have fallen a lot more, relative to the size of our economy, more than the fall we saw on the first graph.

The number of Federal employees relative to the size of our economy:

Graph #3
The graph peaks near the 1.5 level in the early 1940s. That's about half the level shown on Graph #1 around that date. 50 percent: That agrees with what we saw on Graph #2.

It peaks again, just above the 1.0 level in 1952, during the Korean war. From there, it falls below the 1.0 level and continues to drop.

On Graph #1, "below the 1.0 level" puts us off the chart: off the bottom of the chart. For 1953 and after, everything visible on Graph #1 is state and local. Federal employment, between 1.0 and zero, is lost somewhere in the blue border that runs across the bottom of that graph.

On Graph #3, the most recent level shows that the ratio is less than 0.2. That's 0.2 times a thousand people, or something less than 200 Federal employees per billion dollars of Real GDP. That's what the graph shows. (Read the Y-axis label if you don't believe me.)

To be sure, that doesn't include the U.S. Military and the CIA and the NSA and all. But it is less than 200 Federal employees per billion dollars of output, compared to 400 in Reagan's time, 600 in Nixon's time, and 800 under President Eisenhower.

Wednesday, August 2, 2017

"Government work"

The size of government in long-term decline, relative to the size of the economy (where the size of government is measured by the number of government employees):

Graph #1
A slight drift downward from the latter 1940s to 1975, and a definite decline since 1975.

But there are a couple issues with this graph:

1. I'm pretty sure that by "government" they mean all levels of government, not just Federal. But only pretty sure. And

2. According to the notes below the graph at FRED: "Government employment covers only civilian employees; military personnel are excluded. Employees of the Central Intelligence Agency, the National Security Agency, the National Imagery and Mapping Agency, and the Defense Intelligence Agency also are excluded."

Wikipedia shows what looks like a decline in the number of U.S. military personnel for the years 1950-2003. There are two peaks early (for the Korean War and the Vietnam War). Setting those peaks aside, the decline is smaller. Much of it occurs in the Clinton years. And what happened after 2003 is unclear.

The CIA doesn't have to disclose the number of personnel it employs.

The number of NSA employees is "somewhere between 37,000 and one billion".

According to Wikipedia, the National Imagery and Mapping Agency is now known by a different name. The word "personnel" is not found in the article.

Again Wikipedia: The Defense Intelligence Agency's "personnel numbers are classified."

So I'm not gonna find more than that on government employment numbers. But no matter how much U.S. security has grown, I cannot imagine that it would be enough to make the government-employment-to-RGDP ratio run up hill. That "one billion" number, that was a joke. C'mon guys, that was a joke.

Tuesday, August 1, 2017


In comments at Economists View, cm said

I noticed that the old definition of the working age population, 16+, has been discontinued and replaced with the new definition, 15-64.

Well that's a pretty big deal. I half-noticed a change of definition myself recently, but I thought it was my own confusion. Is cm right about a definition change?

In Technical Notes (a PDF from the BLS), subtitled "International Comparisons of Annual Labor Force Statistics, 1970-2012", we read:

The lower age limit of the working-age population according to U.S. concepts is 16 while most foreign countries collect data on the working-age population ages 15 and older. In addition, some countries may have an upper age limit.

Yeah: "some countries may have an upper age limit." Like everybody but the U.S.

That excerpt seems to support cm's "old definition". However, the PDF is undated; I can't tell how old it is. And it might be pretty old, as it doesn't work well with CTRL-C and CTRL-V. So I'm not getting a good answer to my question.

I see FRED has a series titled Working Age Population: Aged 15-64: All Persons for the United States© (LFWA64TTUSM647S). But that series is from the OECD and is compatible with the "foreign countries" approach: the count begins at age 15, not 16; and there is an upper age limit.

Ah-ha! FRED also has the BLS data. But it's discontinued: It ends in 2012. Actually, that's what cm said: discontinued.

Here are the two series together, for comparison:

Graph #1: Two Measures of the U.S. Working-Age Population
Start with the BLS data, the blue line. Add to it people who are 15 years old; this makes the blue line higher. Then subtract everyone who is 65 years old or older. That brings the blue line down to where the red line is. Theoretically.

In the notes below the graph at FRED, something interesting about the BLS data:

FRED has a monthly series that’s a continuation to this discontinued data series.
The series can be found at

Add that series to the graph and it's a good match to the discontinued series. So what do we learn? That the U.S. working-age population is essentially (i.e. by definition) the same as "Civilian Noninstitutional Population". That definition, from the "CNP16OV" series notes:

Civilian noninstitutional population is defined as persons 16 years of age and older residing in the 50 states and the District of Columbia, who are not inmates of institutions (e.g., penal and mental facilities, homes for the aged), and who are not on active duty in the Armed Forces.

So the working-age population and the civilian noninstitutional population are the same (for the U.S.). Okay. Age 16 and older, still today. So maybe, what with the discontinuation of the U.S. "working-age" series and the availability of the OECD version of it, maybe that's what cm was talking about when he said the old definition was replaced with a new definition. Yeah. Exactly.

At first I thought cm meant the definition (in words) was changed. That doesn't seem to be the case.

I wonder if U.S. data is available for how many people are 15 years old, or for how many are 65 or over.

Actually, I can see the U.S. excluding 15-year-olds from the "working-age" count. Maybe child labor laws are tougher in the U.S. than in OECD countries. If they're less civilized than us, so be it! But I can't see counting every old duffer that hasn't died yet as still of working age.

Bur maybe that's part of my pre-2008 mentality. In that comment at Economists View, cm did point out another change we need to account for:

... we don't want to ruin the employment rate with all these people who are supposedly retiring in droves, never mind that many of them keep working (or seeking to work) past 65, if they can.

I dunno how to account for that.