Tuesday, September 19, 2017

You Can Start Disbelieving the Nonsense About the 1% Getting Richer

In the U.S., the 1% are getting richer, and no one else is. Right? That, at least, has been the narrative over the last fifteen years or so. The first pass research on this, by Piketty and Saez in 2003, helped make them famous.

The evidence for that has been based on individual tax returns. But tax rates change through the years. Looking at tax returns without considering the underlying rates (and how they might change behavior) is called the unadjusted approach. It shows the 1% getting much richer.

So what needs to be adjusted for?

  • The biggest tax reform of the last 50 years in 1986
  • Corporate profits, that are paid out as dividends at different rates depending on tax rates.
  • Value of employer provided health insurance.
  • Smaller household sizes, and lower marriage rates.
  • Government transfer payments

Now the tax experts have incorporated corrections for all of those. The big ones are that the treatment of corporate profits that were not paid out (but that rich people had access to, say, through corporate owned vacation property) was much larger in 1960, while it turns out that now the top 1% is actually supporting a bigger group of unreported dependents (rich families aren’t the ones getting smaller).

Piketty and Saez reported that the share of income taken in by the top 1% doubled since 1960. The new research by Auten and Splinter find that over 90% of Piketty and Saez’s increase is there because they didn’t adjust for that stuff. Basically, the 1% are a tiny bit richer. Not enough to worry about.

One thing that both studies find is that the share of the top 1% dropped significantly in the 1970’s. So their share is U-shaped. This is consistent with tons of evidence that business cycles hit the income of the rich the hardest, and there were 4 recessions in a 13 year period centered on the 70’s.

Over the last couple years, it’s also become common to remark that the share of income paid out to individuals has fallen, and that the share going to profits, interest, and rent has gone up. The adjusted “broad income” in this chart shows that’s not the case:

Income Shares Capture

What the new research corrects for is that about half of income is now coming from stuff other than wages and salaries:

Income Sources from Auten and Splinter Capture

Piketty and Saez were looking mostly as the center section.

This points to a good convention when thinking about news reports about inequality: if they focus on wages and salaries, they’re cherry-picking.

Wednesday, September 13, 2017

Second of Two Pieces on Income

Martin Feldstein is a macroeconomist from Harvard (probably their oldest). He’s been well-known since before I started college (he was one of Reagan’s economic advisors in the early 80’s).

In a September 8 op-ed piece in The Wall Street Journal entitled “We’re Richer Than We Realize” he argues that real GDP (and its growth rates) are understated.

The common assertion that middle-class households have seen no increase in real incomes for 30 years is simply not true. And contrary to a common fear, most members of the younger generation will have higher real incomes as adults than their parents had at the same age.

There are two reasons for this, and he argues that both of them are getting more severe.

First, government statisticians grossly understate the value of improvements in the quality of existing goods and services. More important, the government doesn’t even try to measure the full contribution of new goods and services.

On the first count, the government is basically “old school”. They view quality improvements as largely proportional to costs. This might make sense for home construction, but not so much for smartphones.

… In reality companies improve products in ways that don’t cost more to produce and may even cost less. That’s been true over the years for familiar products like television sets and audio speakers. The government therefore doesn’t really measure the value to consumers of the improved product, only the cost of the increased inputs. …

The official estimates of quality change are therefore mislabeled and misinterpreted. When it comes to quality change, what is called the growth of real output is really the growth of real inputs.

The second issue is about the introduction of new products. We’re OK at counting the value of the new products, but we don’t make much of an attempt to figure out how much richer we are from mitigating the problem the new product solved:

Think about statins, the remarkable class of drugs that lower cholesterol and reduce deaths from heart attacks. By 2003 statins were the best-selling pharmaceutical product in history. The total dollar amount of statin sales was counted in GDP, but the government’s measure of real income never included anything for improvements in health that resulted from statins—such as a one-third decrease in the death rate from heart disease among those over 65 between 2000 and 2007.

Think about that: one third of deaths from the biggest killer eliminated in 7 years. That’s a ridiculous improvement to leave unmeasured. But, of course, the techniques for counting GDP were developed before things like this happened regularly.

It is impossible to know how much the official statistics understate the true growth of real incomes. My own judgment is that the true annual growth rate could exceed the official figure by two percentage points or more, implying that the true annual rate of real per capita income growth during the past two decades has been much more than double the official 1.3%.

I don’t know if I’d go that high, but I wouldn’t even debate a 1% mismeasurement.

First of Two Pieces on Income

The Census Bureau reports this week that real median household income hit a record high.

There’s been quite a bit in the news over the last decade or so about stagnating incomes. This is the piece of evidence that is primarily used to justify that position. In this case, real median income for households just passed the previous peak from 1999. The implication is that incomes fell and recovered in that intervening period:

The discontinuity towards the right notes a change in how the data was measured.

Data like this is used to justify the position that the economy is weak.

That may be, but it’s also important to consider what the data is missing. In this case, it’s the definition of a household. We’re in the midst of a multi-decade phase of people living in smaller households: parents have fewer kids, couples get divorced more, and it’s quite a bit less likely for working adults who are not related (as in TV shows like Two Broke Girls) to live together than it used to be.

If households are getting smaller with the passage of time, this means that the data in the graph above is understating growth in the data, and that understatement gets more severe as you go to the right.

Officially, the Census Bureau does not make that adjustment. But it’s fairly easy to find on the internet (here’s an article from Forbes from a few years back). It combines it with a second adjustment for the price index used to deflate nominal incomes (they advocate using the PCE).

While the Census Bureau estimates suggest that median household income rose by just 10 percent from 1969 to 2013, when the PCE is used for inflation adjustment and incomes are adjusted for the number of adults in a household, the increase was 30 percent. The 10 percent rise the Census Bureau estimates translates into a $4,800 increase. A 30 percent rise in unadjusted terms would amount to $14,400—quite a difference.

These issues are not hidden. But they are on the difficult side.

An appropriate interpretation is that the Census Bureau is providing a first pass estimate that indicates there may be a problem. Dig a little deeper, and the second pass says we should worry less. That part doesn’t get covered in the legacy media enough.

Friday, September 8, 2017

Optional: Traditional (Illiquid) Asset Tokenization

Have you heard of Bitcoin? If not, you should have.

Bitcoin is the best known application of a far more important technical advance called  blockchain.

A blockchain is a secure way to store and update a database.That database might hold the information on something of value. That value can be broken up into little pieces called tokens. Bitcoin is a currency that is not backed by government fiat, and whose tokens (called bitcoins) have value.

The more secure the blockchain, the easier it is to trade the tokens because their value is clear.

So … what if the something of value was an expensive asset that isn’t liquid … and you put its ownership in a blockchain … and then sold the tokens? For example, you might have just created fractional ownership of something like an artwork, or a building, or a ….

This is where the world is heading. It’s not clear how this will impact macroeconomics.

One More Explanation for Increasing Wage Inequality

Wage inequality is increasing.

In America, contemporary discussion often presents this is in political and social terms: market-oriented policies pushed by Republicans (and some Democrats) over the last 35 years have increased inequality.

This argument has a serious weakness, that is well-known outside the realm of politicians and pundits. This is that wage inequality is getting worse in many countries around the globe, particularly in the developed ones. It’s difficult to see how market-oriented policies within the U.S. could have produced that result outside the country.

There’s new evidence from Sweden that it’s about higher pay for non-cognitive skills.

Cognitive tasks are technical and/or quantitative. Think math. Non-cognitive skills are sometimes called soft skills or people skills.

The theory is that we are offloading cognitive tasks onto computers, making those skills less valuable. So demand for cognitive skills is shifting left, freeing up funds for a shift in the demand for non-cognitive skills to the right, increasing their price.

This increase occurred primarily in the private sector, among white-collar workers, and at the upper-end of the wage distribution.

… Workers with an abundance of non-cognitive skill were increasingly sorted into occupations that were intensive in: cognitive skill; as well as abstract, nonroutine, social, non-automatable and offshorable tasks. Such occupations were also the types of occupations which saw greater increases in the relative return to non-cognitive skill. This suggests sorting on comparative advantage …

Via Marginal Revolution.

Friday, August 25, 2017

Evidence Piling Up that Might Explain Sustained Low Growth In the U.S.

Let me state out front that I do not have a predisposition to believe this set of arguments. But I’m watching the evidence pile up higher.

First, growth rates in the U.S. have not been as good as we might like for about 40 years.

Second, there are theories in which growth can be stifled by monopolies accumulating profits that they use to stifle more efficient entrants.

And third, we’re starting to get some evidence pointing to this being the case. Noahopinion summarizes:

... I see the case of the Market Power Story - or any big economic story like this - as detective work. We're collecting circumstantial evidence, and while no piece of evidence is a smoking gun, each adds to the overall picture. IF the economy were being throttled by increased market power, we'd expect to see:

1. Increased market concentration (Check! See Autor et al.)

2. Increased markups (Check! See De Loecker and Eeckhout)

3. Increased profits (Check! See Barkai)

4. Decreased investment (Check! See Gutierrez and Philippon)

5. Increased prices following mergers (Probably check! See Blonigen and Pierce)

6. Weakened antitrust enforcement (Check! See Kwoka)

7. Decreased output (Not sure yet)

So, as I see it, the evidence is piling up from a number of sides here. Economists need to investigate the question of whether output has been restricted.

Now, do note that there are many other papers that could be cited under each number, and many criticisms of each one. But there’s a story building up here that’s getting hard to ignore. More importantly, the criticisms are piecemeal — there doesn’t seem to be a a small set of arguments that take down that whole list.

How does he suggest we test # 7:

I think the answer is that it's very hard to know a counterfactual. How many more airline tickets would people be buying if the industry had more competition? How much more broadband would we consume? How many more bottles of shampoo would we buy? How many more miles would we drive? It's hard to know these things.

Still, I think this question could and should be addressed with some event studies. Did big mega-mergers change output trends in their industries? That's a research project waiting to be done.

If you’re thinking about graduate school in economics, this is going to be a hot topic over the next few years. As I write this, the cite in # 2 is just out, so there’s a lot of talk amongst economists that’s reminiscent of paleontological searches for missing links.

Fakers? Not All of Them

It’s a common thing to view welfare recipients as unmotivated. (Personally, I don’t hold that view, but it’s common enough for me to put it in the title of this post).

That has some overlap with lazy, but it isn’t quite the same thing: you could be one without the other, or both.

However, unmotivated could also be a function of background or environment. So it’s possible that if a welfare recipient is unmotivated to get a job, it’s because they were previously unmotivated to get the skills to be gainfully employed.

Whatever. Can we test for this with data from the real world? Manal Dashpande, a recent Ph.D. student from MIT, now on the faculty at the University of Chicago, has been researching this (here is an older draft of her paper, and here is a summary more accessible to undergraduates).

To do work like this, you look for a “natural experiment”. In short, this means that somehow the circumstances of the people you’re interested were changed, and they had to respond to this change, but they weren’t capable of influencing how or when their circumstances were changed.

Deshpande looked at the change in welfare in 1996. There was a “bright line” cutoff that made the natural experiment possible: children were treated differently based on whether or not they turned 18 before or after August 22, 1996. For a particular form of welfare (SSI), prior to that cutoff, if you got it before the cutoff you automatically got it afterwards. After that cutoff, if you got that form of welfare as a child, you had to be reevaluated when you turned 18. About 40% of childhood recipients were denied as adults. The motivation for that was that some conditions, say mild mental retardation, may incur extra expenses in childhood, but don’t affect a person’s ability to get and hold a job.

One comparison Deshpande made was the income performance of people whose benefits were removed at 18 versus three other groups. One was those who stayed on SSI after age 18 (they were not denied benefits at 18). A second were those who applied for the first time as adults but who were denied (you might call those “fakers”). The third were children whose families had received AFDC (that program was replaced by TANF, which we still have), a program targeted at low income families rather than disabled children (you might call these “poor yet able”). Here’s a chart of their income performance:


The “poor yet able” are the red points: their income rose through time. The “fakers” are the purple points: their incomes also rose. In both cases the rise was up to an average of about $15,000 per year in income. That’s not a lot, but since it’s an average, it indicates that some portion of them were able to do better than minimum wage employment.

The people who were retained on SSI are the dark blue dots at the bottom. Their incomes outside of welfare stayed close to zero, but were supplemented by those continued welfare payments.

The light blue dots indicate that welfare reform may have been to aggressive. Children who were denied benefits when they turned 18, on average, had income that never approached those of more advantaged youth. Deshpande estimated the present value of their income loss at $76,000 over the ages of 18 to 34.

Basically, welfare reform was a big tax on some low income people.

Via Marginal Revolution.