Income Inequality

Greg Mankiw and Brad DeLong are debating the income inequality data from the Piketty and Saez study. Paul Krugman talked about rising income inequality in his last column. In response, Greg Mankiw says:

New Data on Income Inequality: In today's NY Times, Paul Krugman calls attention to the update of the Piketty-Saez data on income inequality, although Paul describes the data differently than I would.
Here is what I see: After rising substantially from 1986 to 2000, income inequality is essentially the same in 2004 (the most recent year of data) as it was in 2000.  Click on the data and see for yourself.

In response to Greg, Brad DeLong says:

New Data on Income Inequality: Greg Mankiw sees a little bit of good news in the latest income distribution data: after rising astonishingly rapidly from 1986 to 2000, income inequality in 2004 was no worse than in 2000...

I hope he's right, and that the trend of rising inequality has stopped--it is a very disturbing phenomenon, and further rises would be very worrisome indeed. But I can't be as optimistic as he is. He sees an essentially flat trend from 2000-2004. I see numbers for 1999 and 2000 that may have been transitorily boosted by high salaries paid during the dot-com bubble, and then a decreased in inequality from 2000-2002--a decrease that is then reversed in 2003-2004, which carries us up to bubble levels.

So my hope that we might not see 1999 and 2000 levels of income inequality again appears to have been vain.

We can debate what has happened the last few years and whether this year or that year had special circumstances such as the problems Brad notes with using 2000 as a base year. But in doing so, we shouldn't lose sight of the overall upward trend in inequality. Here's a few graphs from the study:

The Top Decile Income Share, 1917-2004

The Top 0.01% Income Share, 1913-2004

Average Real Income of bottom 99% and top 1% in the United States, 1917-2004

The little dip at the end is what the fuss is all about. But even if Greg's claim holds up to the types of qualifications Brad talks about, and there are good reasons to worry about using 2000 as a base year, the upward trend in income inequality since the 1970s is undeniable. And in any case, as the last graph shows, real income for the bottom 99% of the distribution has been flat since the early 1970s despite large gains in productivity.

Comments

[spencer]

the little dip in the end-- 2000 & 2001 -- is the impact of the bursting of the high tech bubble and
especially its impact on income from stock market gains.

I do not know if this data is adjusted for that but one thing that happened in 2000 & 2001 is that stock owners realized large capital loses that they could use to offset other income and reduce the income reported for tax purposes.

This does not negate the point that the stock market crash did reduce the wealth held by these people, but this report is about income, not wealth.

[Bruce Wilder]

The general challenge, here, is a narrative analysis of what has driven the trends, which Piketty and Saez have documented. I can imagine any number of competing narrative analyses, which might be derived from sound economic theory, and which might bring any number of factors into the discussion. Piketty and Saez cast their data as a general refutation of Kuznets and the general thesis of "technological determinism" for income distribution.

I trust that Piketty and Saez know that Kuznets and technological determinism is a bit of a strawman. The more obvious thesis, that politics have something to do with it, is treated obliquely.

Greg Mankiw, for ideological reasons, wants to shut down the search for a narrative analysis. And, Brad DeLong gives him a base hit, without so much as a throw to the first baseman. Why Brad wants so badly to engage the likes of Greg Mankiw in debate, I have no idea. There's no percentage in debating with someone, who is ideologically opposed to knowledge. And, this is a perfect demonstration.

A good debate between economists with differing points of view would start generating a list of theoretically sound relationships, a list of factors, or some dispute over how to weight factors, suggestions about how to trace changes in causal factors and to relate changes in "causes" to income distribution "outcomes".

Instead, we get, "the five-year trend looks flat to me." The appropriate response to that is NOT, "oh, I hope you are right." The appropriate response is, "f.u., too, you ignorant fascist lying bastard" or words to that general effect.

[Mark Thoma]

Bruce W.:

Good point. We should talk about underlying causes and the different theories for the rising inequality (e.g. Krugman's Graduates versus Oligarchs), not just time-series plots.

[spencer]

Economic theory says rising income inequality should lead to greater savings and investment that generates a bigger pie that makes everyone better off.

But over the last quarter century the rise in income inequality has been accompanied by falling savings rate
and weaker investment by individuals, partnerships, etc., as compared to corporations and nonprofits.

The results has been just the opposite of what theory says should happen. this is the message that we should
focus on, and keep asking the defenders of these results to explain why it isn't working.

[anne]

The terrific Piketty-Saez data are in no way assuring, rather they are not even a surprising demonstration of just how much income and wealth inequality is gaining and how how much more to expect from fiscal policy that has been increasingly designed to the benefit of the wealthiest. Wealth generates income, and tax law is insuring that increasing amounts of income will be tax favored. We are sadly in the beginning years of this newly revived move to inequality.

[anne]

Wealth is a prime source of income and becoming increasingly more of an income source and tax policy has been increasingly designed to preserve income and wealth for the wealthiest. So, we should not be the least surprised that income inequality is increasing strikingly after a cyclical decline during the bear market of 2000 to 2002.

[save_the_rustbelt]

Mankiw is looking for a nugget of gold in a pile of manure.

[anne]

The question of why personal saving is not increasing even though income inequality is growing is interesting with no answer obvious to me. What am I missing in thinking about this question?

[anne]

Suppose there are 2 of us and we have equal incomes and wealth and start by saving the same. My income declines and yours increases. I save less and you save more. But will my lower saving be compensated by your increased saving? Possibly not. You will add to your consumption and save more but not simply save as much as I give up in saving as my income is lowered. Hmmm. Suppose then income inequality lowers saving. How does this sound?

[anne]

The wealthy save more, but not as much as the less wealthy cut saving as income inequality increases. I wonder....

[Bruce Wilder]

"Wealth is a prime source of income"

Of course it is, and the Reagan deficits and Volker disinflation combined to increase financial asset values and totals, just as the Bush I/Clinton disinflation/deficit reduction continued that trend. Keep in mind that the P/S data don't count income from unrealized capital gains. Bush II has significantly pushed the tax favoritism shown to investment income, restoring the cuts introduced by Reagan/Bush I, but partly reversed by Clinton, and going way beyond them.

I am concerned, of course, that Capital is not balanced, politically, by organized Labor, but I have nothing against the legitimate claims of Wealth, per se. I accept that inflation in the 1960's and 1970's had depleted the stock of financial assets and combining deficit spending with Volker's disinflation probably made some rough sense, and necessarily involved some shift of income toward Wealth.

I would suggest that higher anticipated costs of energy after 1973 probably contributed to a dampening of individual marginal labor productivity growth, which contributed to a leveling off of baseline, individual labor income. Feminism also played a part, as well.

What worries me most is the increasing share of income going to top corporate executives. P/S document that "high wages" account for a very significant proportion of the increased diversion of income to the top 1%.
This is neither the product of honest labor, nor an income paid to Capital. The Corporate Executive Class is, in effect, stealing massively, from Capital AND Labor alike. The present Administration and Congress have become the corrupt creatures of this New Class.

This phenomenon is new, unprecedented, involves the use of pure power to make a massive shift in income distribution, and, as far as I can see, has absolutely no justification or redeeming economic value. Its only economic motive is unrestrained greed, and its only political goal is authoritarianism.

The wealthy have some skin in Society's game -- their wealth -- and so, can be expected to be concerned about the Society's long-term health, even if their view of the Public Interest may be skewed by self-interest. The Corporate Executive Class are exploiting a power vacuum, created by a policy decision to both restrict the control Capital can exercise over large, public corporations, and to relax the regulatory regime imposed by Public regulatory agencies. The fraudulent stock options and bonuses and lavish pensions impose no down-side risk on the top Executive. He is being given carte blanche to impose risks on the corporation's business, to steal from the stockholders and the unorganized employees, as opportunities arise. And, since the claim of this New Class on such a sizeable fraction of the nation's income depends delicately on the power vacuum created by the rules limiting stockholder power and the vigor of the regulatory agencies, they have every incentive to impose authoritarian government, backed by a corporate, right-wing propaganda machine.

[anne]

Interesting supposition to think about, but we need to ask in detail why the convergence in income from the New Deal to the 1970s and divergence from then. Interesting comments.

[Bruce Wilder]

"The question of why personal saving is not increasing even though income inequality is growing is interesting"

Income distribution is intimately tied to the distribution of risk. Saving is a response to risk. That's one of the things that makes any scheme for social insurance potentially problematic; to the extent that you provide "free" insurance, that's, in effect, giving people an incentive to reduce saving.

Wealth is a form of, and foundation for, insurance. A wealthy person is likely to be better insured against risk, by a broader portfolio, as well as a greater margin between desired consumption and expected income. A wealthy person can take on risky investments, without taking on risk, because the wealthy person can afford to diversify the risk away, so the wealthy investor will realize larger returns on investment.

A person earning a high income may feel they face a high risk in relation to the continuation of that income, and be inclined to save. (It is easier, as well, to "save" in the sense of "not consume" when income is high relative to an acceptable rate of consumption.) But, wealth, per se, is actually a disincentive to saving from current income, since the risk of not having sufficient income in the future is small. The extremely wealthy often find satisfaction in massive conspicuous consumption and philantropy -- saving is not high on their list of things to do today.

The passive accumulation of wealth, which many experienced in the stock market bubble and the housing bubble, have predictably depressed personal saving.

For a poor person, "personal saving" and "buying insurance" are the same thing. If the poor person is deprived of opportunities to buy efficient insurance, her decision space can move in one of two directions. She can hunker down, become more conservative in her investment decisions, reducing her expected future income by more conservative investment decisions (e.g. better not borrow to finish that M.A.), and also reduce her current savings to a proportion more appropriate to her expectations regarding her permanent income as well as her diminished opportunities to earn an income from investments.

Or, she can accept that she has no opportunities to efficiently buy insurance (i.e. save and invest with a reasonable expectation of return), and, instead, devote a smaller percentage of income to extremely high risk "investments" (e.g. buy lottery tickets, or take up a high-risk hobby, like selling illegal drugs, robbing convenience stores and the like).

Either way, increasing the risk experienced at the bottom of the income pyramid, by reducing the opportunities to obtain efficient insurance, will also, predictably, reduce personal savings.

At the risk of alienating everyone with my hobby horses, I will point out that P/S are attributing a large portion of the shift in income distribution to (what I would call) the predation of the corporate executive class. The corporate executive class has gotten that money by shifting risk onto lowly employees, eliminating pensions, traditions of life-time employment and the like.

In a very real sense, the corporate executive class has stolen the savings of millions at the bottom; it is not surprising that whatever increase in nominal savings make be taking place at the top, the overall net savings rate is substantially depressed, by the transfer.

[Bruce Wilder]

spencer: "Economic theory says rising income inequality should lead to greater savings and investment that generates a bigger pie that makes everyone better off."

I could certainly use economic theory to tell such a story, but it is only one of many stories I could tell, using economic theory to make relationships in the story plausible and coherent. Rising income inequality could very well depress investment, for a wide variety of reasons. It is not an accident that there's a peak in the P/S data in 1929!

[Stormy]

Well put, Bruce, well put.

[Tymbrimi]

Besides envy, which one should get counseling for, who cares about income equality? And why, besides envy? :)

[Tymbrimi]

Since the growth in productivity is at decades high levels, obviously the Bush tax cuts are performig as expected, and those measurements that say otherwise are flawed. And, in fact, it is known that the investment statistics are flawed in today's knowledge-based economy. :)

[Dave Schuler]

Mark, I believe that income inequality is increasing and my experience tells me income inequality is increasing. However, something I've wondered for a long time is how much of what's observed since the early 70's an artifact?

Tymbrimi, I don't care all that much about income inequality unless it's due to rent-seeking and I think there's pretty good evidence that's the case. Consider this, for example.

[Tymbrimi]

I did. Don't care about stock options. If there is wrong-doing here, someone should report it. :)

[Tymbrimi]

The savings rate is a not a problem. We are awash in liquidity. A higher savings rate would merely be a recipe for stagnant growth, as it is in many developed nations. Our rate curve is fairly flat.

[Jonathan Lundell]

MT: "And in any case, as the last graph shows, real income for the bottom 99% of the distribution has been flat since the early 1970s despite large gains in productivity."

So, what changed in the early 70s? Or was 1933-1973 an aberration, and if so, how do we do it again?

[Bruce Wilder]

"Besides envy, which one should get counseling for, who cares about income equality? And why, besides envy?"

I don't think anyone should. I think the writers of economics textbooks, in the tradition of Paul Samuelson, did us a great disservice, by framing arguments about egalitarianism in terms of "tax-and-transfer" schemes, aimed at greater equality of income.

I think the institutional structure of the economy necessarily does and should provide incentives to work and invest rationally and productively in a system of distributed decision-making. In that context, I think it is important that people have adequate insurance against economic risks, so that their economic choices are not afflicted by excessive risk-aversion. I do not want people holding back from rationally making the extra-market investment decisions to start a business or extend education because of risk they cannot insure against, nor do I want the relatively poor to be driven to such extremities that they pursue high-risk activities, which are destructive to themselves and others (e.g. crime).

So I support all kinds of schemes -- Federal Reserve management of the business cycle, Social Security, unemployment insurance, public education, confiscatory inheritance taxes, progressive personal income taxes, corporate income taxes -- which have the effect of insuring against risk, because I think it improves both the economy and society to do so.

Too great an inequality of income distribution seems an indicator to me of too great a concentration of power, a concentration of power, which is fundamentally incompatible with the continuation of a free society (a society of highly decentralized decision-making). Such a concentration of income and power in the hands of very few can only be achieved by pushing costs and risk onto to the many, oppressing the many to benefit the few.

[anon/portly]

So from 1967

(99% * 35,000) + (1% * 300,000) = 37,650

to 1997

(99% * 36,000) + (1% * 620,000) = 41,840

overall income growth was 11%? Over 30 years that seems kinda low. Per capita real GDP growth can't have been anywhere near that low, can it? (I refuse to look it up, just in case it was). So that suggests the difference going to the government, but I thought the size of the government had been flattish also....

The Mankiw - Delong discussion seems a touch pointless, since even if the trend is flattening out (it has to sometime, you'd think), that doesn't make the current level of inequality any less of a problem. I guess worrying about the trend takes your mind off other things.

As good of an economist as Mankiw no doubt is, will we ever see him come up with something half as inspired as that last Ben Stein piece? His response to the Krugman column may or not be insightful, but it isn't exactly drunk with passion.

[Dirk van Dijk]

Money is like water, income is like a river and wealth is like a lake. One of the key features of the Bush administration's tax policy is to insure that wealth, or income derived from wealth (i.e. cap gains and dividends) are not taxed at all or are taxed at the lowest poossible rate. (Recall that origionally they wanted to make the cap gains and dividend rates zero. The inheratance tax is a pure form of wealth taxation, and they consider its eliminataion to be the top economic priority.) Thus, essentially we have a situation where Chicago is dying of thirst, but it is not allowed to take any water from lake Michigan and only a little bit of water from the Chicago river which flows out of the lake.

[James Kroeger]

...over the last quarter century the rise in income inequality has been accompanied by falling savings rate...

The question of why personal saving is not increasing even though income inequality is growing is interesting

Whenever economist-types begin to discuss the decline in the Personal Savings Rate, I feel I need to point out the following:

In calculating the Personal Savings Rate, BEA economists assume that money saved is income that is not spent:
SAVINGS = INCOME – CONSUMPTION

The Personal Savings Rate is simply the percentage of total income that is saved:
PSR = SAVINGS / INCOME

What the BEA actually presents to us is a calculation of the percentage of total income that is not spent:
PSR = (INCOME – CONSUMPTION) / INCOME

Notice that if expenditures are held constant, a higher INCOME number will result in a higher calculated Personal Savings Rate.

The reason why the nation’s calculated Personal Savings Rate does not provide us with useful information on national savings is because BEA economists do not include in their calculation of total personal INCOME the income that households earn from capital gains. They have their reasons for doing this but they are not ultimately good reasons, for capital gains income can be either saved or spent on consumption, just like any other kind of income. If they were to include capital gains income as part of total personal income, the result would be a significantly higher calculated PSR.

From this, we can see why the calculated PSR has declined while the Republicans have been running the Federal government. As they have cut the income tax rates of the wealthy, a greater share of the nation’s total income is comprised of capital gains income. This is because the very wealthy are most likely to use the huge gifts of disposable income that the Republicans have given them to buy assets, like stocks and real estate.

The PSR is a calculated [per household] 'average' that tells us nothing useful if our desire is to ascertain whether or not Total Savings levels are high enough. What statistical measurement could we depend on that would tell us---if only it were accurate---when we are not saving enough as a nation? The Unemployment Rate.

The only time an economy is in need of more total savings (less consumption) is when there is zero unemployment and hyperinflation is threatening. Any increase in total savings when there is any level of unemployment is simply going to eliminate jobs (because decreased aggregate spending is going to be necessary in order to increase aggregate savings.)

It may be true that the Personal Savings Rate has been declining over the past quarter century, but we really need to stop making the mistake of construing from this that the economy is not producing a sufficient amount of savings.

[Blissex]

«BEA economists do not include in their calculation of total personal INCOME the income that households earn from capital gains. They have their reasons for doing this but they are not ultimately good reasons,»

Now, not to pick on your words, but this is the sort of utterly meaningless comment that is based on a complete misunderstanding («total personal INCOME») of the national accounts.

The national accounting scheme calls certain quantities with names like "savings" or "consumption" or "income" instead of "green" "red" or "yellow" almost by chance.

One should not assume any given or fixed relationship between those accounting categories and what is commonly understood by those words outside national accounts, they are just terms of art.

It is a bit like "depreciation" in company accounts, where its meaning vaguely resembles but it wholly distinct from its intuitive meaning.

The reason why capital gains are excluded is because they do not contribute to *product* in the (term of art) sense it is used in the national accounts.

BTW in the BEA link you provide reasons #1 and #2 are totally bogus, because excluding income on the basis of illiquidity and volatility just does not make sense.

The real reason is hinted in their #3 explanation: that the "product" in GDP and GNP is meant to be something like ''value added'', and capital gains as such do not add value *in the aggregate*.

Sure they do contribute to some people's income and expenditure, but as a rule changes in the price of capital have only a redistributive effect, between owners and non-owners of capital.

That redistribution may have an effect on "investment" and "consumption", but the effect is captured independently anyhow.

Why does the BEA FAQ give bogus reasons for the non inclusion of capital gains? Well, it could be that they try to talk down to the little people, or perhaps more likely that the rationale for why the national accounts are the way they are is little known even among practising economists.

National accounts are the way they are because many decades ago some guys decided they should be like that mostly out of (Keynesian) intuition, and what was easy to measure at the time, and now it is too late: changing the categories reported would break many important series and would require some big big changes in the statistical apparatus of government.

But it is important to remember that those labels are purely terms of art, that national accounts measure aggregate quantities that have been defined somewhat arbitrarily and the not just the definition but the behaviour of aggregates can be very different from what is at the micro level.

[Lord]

Other than cyclical effects, it really doesn't seem real incomes have gone anywhere since 2000. Even the top 1% may be running out of steam now.

[Movie Guy]

Lord,

What data are you citing?

[Paul Botts]

I'm an economic progressive who's never voted for a Republican for president and doubts he ever will, but my highest political value is honesty and rigor in examination of the issues. In that spirit it's worth pointing out that Piketty and Saez's conclusions are nonsense unsupported by complete data.

In the first place, Piketty and Saez are counting not income, but tax returns. (They may be "sensitive" to this distinction but choose to ignore it in practice.)

Furthermore they don't count transfer payments (Social Security checks and the like) as income. Since such transfer payments have risen by 50% as a share of personal income in the U.S. since 1980, and since they are a proportionately much larger share of poor and middle class incomes than of large incomes, excluding naturally forces the top-1%-share graph to move upward.

In the third place, in the late 1980s and 1990s a series of changes in U.S. tax law moved lots of high-end incomes out of business tax returns and into individual tax returns. Since Piketty and Saez used data only from individual tax returns, naturally the large incomes they added up rose due to those changed tax rules -- not because any rich incomes actually grew but simply because they shifted columns and started being counted by Piketty and Saez.

In the fourth place, Piketty and Saez insist on measuring income per tax unit rather than per family or household or individual. They apply a blanket assumption that income per tax unit is on average 28% smaller than income per household. But since there are many more two-earner couples sharing a joint tax return among high-income families than among lower-income ones, this choice once again serves to exaggerate inequality per family or per worker.

The whole thing is quite silly, and adjustment for these and other errors reveals that in fact the share of personal income received by the top 1% has virtually unchanged from what it was in 1980. Which, among other things, answers the conundrum regarding the savings rate: there hasn't been any particular rise in inequality and hence unsurprisingly no rise in the savings rate.

[Mark Thoma]

That's a pretty good summary of Reynold's recent editorial in the WSJ. You ought to cite him - and read more widely. He has a history of erroneous debunking. Are you sure he was looking at the right tables this time?

Take a look at the variety of studies out there, there's plenty of evidence to support widening inequality.

[evagrius]

It's amazing the lengths people will go to to deny what's in front of them.

[Paul Botts]

Actually I looked up the tables myself, and read several of the oft-cited studies claiming a rise in inequality; turns out that Reynolds while obviously an ideologue has the virtue of having his facts largely straight on this subject. (Not in every detail which is why I didn't mention all of his points, above.)

Whether or not he has been wrong about other subjects (wouldn't surprise me at all) doesn't change the actual data on this one. Indeed that the response to an explanation of fallacious reasoning is simply to try to discredit the source of the explanation, is rather revealing.

[Mark Thoma]

Here's a place for you to start.

On the discrediting, when sources have a history of being wrong and are making claims that are contrary to most other work on an issue, their previous record is worth pointing out. So, I agree it "is rather revealing" that this particular source has been wrong before.

But, more seriously, if you can counter Piketty and Saez, that's publishable in a good academic journal. Let me know when your paper comes out and I'll post the abstract.

[evagrius]

Thank you Mr. Thoma. A nice rejoinder. Is the debate now closed?

[Paul Botts]

Sounds like the minds and hence the debate here are closed, which is a shame. Thanks for the pointer to Brad DeLong's blog, I've posted a couple of specific questions there in hopes of learning more about this subject.

As an aside, I'm not familiar with Reynolds' previous work or reputation and even if I was that wouldn't inform a judgement about claims regarding a factual matter. He may be the biggest right-wing prick on the planet for all I know but that's not really the question. My reaction is simply due to having read Piketty-Saez and looked up the relevant data myself and finding, thus far, that Reynolds' versions of the facts check out. If his logic from those facts is flawed then I want to know that, which is why I've posted a couple specific questions at DeLong's journal.