Is Inequality a Concommitant of Rapid Growth?

Greg Mankiw writes to the New York Times:

To the Editor:

Your chart about the percentage of income earned by the top 0.1 percent of taxpayers was fascinating, but "Richest Are Leaving Even the Rich Far Behind" failed to draw the obvious conclusions from it.

The data show that the rich take a rising share of income when the economy is booming, such as during the 1920's and 1990's. Their share declines when the economy hits hard times, such as during the Great Depression and the most recent recession.

The rich took their smallest slice of the economic pie during the 1970's - a period when productivity growth was low and unemployment and inflation were rising.

Here's the lesson: If policy makers' primary goal is to reduce income inequality, they should put the economy through the wringer. But if they want economic prosperity for all, they should avoid focusing on the politics of envy.

N. Gregory Mankiw

Cambridge, Mass., June 5, 2005

The writer, a professor of economics at Harvard University, was chairman of President Bush's Council of Economic Advisers, 2003-2005.

Well, let's see. Let's take the state-of-the-art data on income inequality from Emmanuel Saez and Thomas Piketty (2003), "Income Inequality in the United States, 1913-1998," Quarterly Journal of Economics 118:1 (February, pp. 1-39) (http://emlab.berkeley.edu/users/saez/pikettyqje.pdf), and plot it against the previous ten years' growth in GDP per capita from eh.net (http://eh.net/hmit/gdp/gdp_question.php):

The rich do take a rising share during the 1920s and 1990s, but growth income per capita was no faster in the 1990s as a whole than in the 1980s and 1970s--it was only the last half of the 1990s that saw rapid growth. And the fastest-growth decades of all are the 1960s--with a low share of income inequality--and the 1940s (driven by recovery from the Depression and the high-pressure economy of World War II).

The correlation between economic growth--defined as ten-year growth in GDP per capita--and income inequality that Greg Mankiw asserts exists? I certainly cannot see it in the data.

But maybe I'm wearing the wrong-colored glasses :-).

Comments

[anne]

Dear Brad,

What an important post :)

[howard]

important enough that i wonder, prof, if you might spend a few more minutes on the question and make sure you aren't overlooking anything important in giving us this first-order approximation.

in short, what might an honest conservative say about this? what might donald luskin? and what are the answers to both the honest conservative and the stupidest man alive?

[Buce]

I don't have data at my fingertips, but isn't my recollection correct that inequality actually declined in East Asia (esp. Korea, Taiwan) during the go-go 80s?

[Kuas]

You clearly haven't mastered faith-based economic data analysis. Perhaps Mankiw will be offering an online course at Harvard...

[bs23]

show me the money! (almost as good as: "survey says")

[Michael Carroll]

Is there any prospect of involving Dr. Mankiw in some sort of debate on questions like this? It seems like a lot of these guys (Andrew Sawicky included lately) have this pathetic habit of showing up out of the blue to parse people's grammar then keeping a low profile when the more important questions crop up.

What the hell does Mankiw have to say about the flat yield curve, the current account deficit, the RMB peg, housing price inflation, real income stagnation or half a dozen other tricky questions that defy conventional explaination. The silence is deafening.

Don't the investment excesses of the 1920's and 1990's sort of undercut the man's case anyway?

[Nate]

Without knowing how much due diligence Mankiw has performed or arguing one way or another, one possible source of discrepency would be that a 10-year backward looking GDP-per-capita measurement may not be the right metric to compare income inequality to. The lag time inherent in that statistic will be extremely large, and probably much larger than income distribution, sticky though wages are.

Unemployment, hours worked, and other factors will skew GDP per capita measurements versus productivity as well. Whether it's preferable to omit those from this comparison is beyond me, but they seem to introduce more noise.

[milo]

Is this something like the Laffer curve? It sounds good, so it must be true (all evidence to the contrary).

[dogfacegeorge]

I wandered over to Vox Baby because I knew that Samwick would want to highlight Mankiw's letter as an example of "shaping, slicing, and selectively citing" the numbers. Unfortunately, the post is not up yet. I'm sure that Samwick must be working on it hard - I can't wait to see how he skewers Mankiw!

[UCD Student]

Greg Mankiw = teh lol

[KenS]

Brad,

Try these colored glasses on for size:

"Partisan Politics and the U.S. Income Distribution," Larry M. Bartels
Department of Politics and Woodrow Wilson School of Public and International Affairs, Princeton University

Abstract
Census Bureau data reveal large, consistent differences in patterns of real pre-tax income growth
under Democratic and Republican presidents in the post-war U.S. Democratic presidents have
produced slightly more income growth for poor families than for rich families, resulting in a
modest decrease in overall inequality. Republican presidents have produced a great deal more
income growth for rich families than for poor families, resulting in a substantial increase in
inequality. On average, families at the 95th percentile of the income distribution have
experienced identical income growth under Democratic and Republican presidents, while those
at the 20th percentile have experienced more than four times as much income growth under
Democrats as they have under Republicans. These differences are attributable to partisan
differences in unemployment (which has been 30 percent lower under Democratic presidents, on
average) and GDP growth (which has been 30 percent higher under Democratic presidents, on
average); both unemployment and GDP growth have much stronger effects on income growth at
the bottom of the income distribution than at the top. Similar partisan differences appear in the
distribution of post-tax income growth of households since 1980, despite the fact that the
corresponding pre-tax income growth data for that period show little evidence of partisan
differences.

Paper available at:
http://www.princeton.edu/~bartels/income.pdf

[KenS]

Oops. Sorry about the formatting. Here's that abstract again...

Abstract
Census Bureau data reveal large, consistent differences in patterns of real pre-tax income growth under Democratic and Republican presidents in the post-war U.S. Democratic presidents have produced slightly more income growth for poor families than for rich families, resulting in a modest decrease in overall inequality. Republican presidents have produced a great deal more income growth for rich families than for poor families, resulting in a substantial increase in inequality. On average, families at the 95th percentile of the income distribution have experienced identical income growth under Democratic and Republican presidents, while those
at the 20th percentile have experienced more than four times as much income growth under Democrats as they have under Republicans. These differences are attributable to partisan differences in unemployment (which has been 30 percent lower under Democratic presidents, on average) and GDP growth (which has been 30 percent higher under Democratic presidents, on average); both unemployment and GDP growth have much stronger effects on income growth at the bottom of the income distribution than at the top. Similar partisan differences appear in the distribution of post-tax income growth of households since 1980, despite the fact that the corresponding pre-tax income growth data for that period show little evidence of partisan differences.

[qwerty]

Well, duh. Top 1% derive more of their income from business profits than the population at large. So when the profits rise, so does their share. However it does not follow that rising the share of top 1% helps the profits grow.

[Winslow R.]

"Here's the lesson: If policy makers' primary goal is to reduce income inequality, THEY should put the economy through the wringer."

Greg would likely consider Chairman of the Council of Economic Advisers as a position of "policy maker". So HE is THEY.

Looking at the chart it is obvious Mankiw has figured out that he was working for a communist junta focused on reducing income for the top 1% after it gained control of the economy through a judicial overthrow in 2000.

Thank you Greg and NY Times for clarifying that.

[anne]

http://www.nytimes.com/2005/06/07/national/07pinehurst.html?pagewanted=all

In County Made Rich by Golf, Some Enclaves Are Left Behind
By SHAILA DEWAN

PINEHURST, N.C. - Golf has made Moore County rich. There are spas, country clubs and new $2 million homes. The United States Open, to be held later this month on the most famous of the county's 43 golf courses, is expected to bring $124 million to the state.

But as developers rush to provide 'resort quality' amenities in the newest subdivisions, some neighborhoods have been left behind - without sewers, police service, garbage pickup or even, in some cases, piped water.

These enclaves, Jackson Hamlet, Midway and Waynor Road, are virtually all black. They butt up against, or are even completely surrounded by, affluent towns that are mostly white: Pinehurst, Aberdeen and Southern Pines.

The 500 residents of these unincorporated enclaves are close enough to point out sewer lines that run past their properties en route to new developments, or to watch garbage trucks trundle past without stopping.

Though the towns have not annexed these hamlets about 60 miles southwest of Raleigh, and their residents cannot vote in municipal elections, they are subject to the towns' land use and zoning rules under what is called extraterritorial jurisdiction.

When asked about extending basic services, the towns' officials say they must take care of those within their existing boundaries before taking on new neighborhoods. The county, on the other hand, says that many of its rural constituents do not have the services the enclaves are requesting, and that the problems of these more densely populated areas can be better addressed by towns....

[anne]

http://www.nytimes.com/2005/06/07/national/07pinehurst.html?pagewanted=all

Excluding heavily minority areas from town boundaries is a common but little examined practice, particularly in small towns in the South, civil rights advocates and geographers say. With the U.S. Open beginning on June 16 on the Pinehurst No. 2 golf course, residents of the three black neighborhoods and their advocates are making a concerted effort for the first time to win more services, holding news conferences and giving tours.

Historically, they are the very people who provided much of the labor that built the hotels in the Sandhills, as the area is known, tended the greens at the golf courses or worked on the all-black crew of caddies, long since replaced by electric carts.

Ida Mae Murchison lives in Jackson Hamlet, a shady neighborhood of dirt roads hemmed in by Aberdeen and Pinehurst but claimed by neither. On one corner, a new apartment complex juts in; it was annexed by Pinehurst, which often expands to include areas where a developer has already paid for the infrastructure. At the rear, a place once called Buckety Ford, where Jackson Hamlet children fetched water, has been dammed to create a 200-acre lake surrounded by houses, also part of Pinehurst.

'I get the feeling that we're just forgotten, put on the shelf or the back burner or something,' Ms. Murchison said. 'But like I say, I don't want to offend anyone. I don't want to cause trouble.'

Ms. Murchison was the first black chambermaid at the Carolina Inn, the gracious centerpiece of the 110-year-old Pinehurst resort, where she worked for nearly 50 years....

[Cranky Observer]

I think at this point that there is no question that some system of differential rewards does much to stimulate human achievent. But at what point did the required differential go from 10:1 (or perhaps 100:1), something most human beings can live with, to 10,000:1? 100,000:1? If that isn't enough, what will be? One guy with 99.995% of the world's wealth and everyone else with breadcrumbs? Man, will HE be productive.

The absolute lack of reasonableness, and the transformation of our society from "reasonable compensation for effort + some differential rewards" to "a winner-take-all, monstrous-rewards-for-the-few" will I think be our destruction, and not very long from now. Say within 20 years.

Cranky

[Pancho Villa]

Forget Greg and praise his master Paul!

[Dave]

Brad-

I wonder here about your use of the 10 year GDP growth per capita as a measure of how well the economy is doing.

Your use of the data from Emmanuel Saez and Thomas Piketty seems to very much support Mankiw's opening remarks; In the 20's the wealthy's share of income was increasing as it was in the 90's. Further the 70's appears as the time period when this share reaches its minimum over this time period as Mankiw stated.

However, you quickly diverge and use this 10 year GDP growth as a proxy for how well the economy is doing in each of these time periods. You completely ignore Mankiw's comment on productivity, unemployment, and inflation. Is a 10 year growth in GDP per capita so telling of how well our economy is doing? Could you, or others, explain to me how these other factors such as productivity, unemployment, and inflation do not matter when we know the 10 year growth of GDP per capita. Be careful not to just brush these variables off. If these variables do not matter, then why does income inequality matter? Aren't all of these factors important in measuring our so called quality of life?

I look forward to a response.

[Nate]

Dave,

Delong doesn't mention this in his post, but I believe he's used real GDP per capita rather than nominal GDP per capita, so the effects of inflation are accounted for. However, I agree the choice of data series is a curious one.

[P O'Neill]

>>
But if they want economic prosperity for all, they should avoid focusing on the politics of envy.
>>

Is Mankiw presenting a new, or at least new to neoclassical economics, theory that stagflation was caused by government efforts at active redistributionary policies?

[Mike]

For some reason that combined graphical representation feels like a statistical bait-and-switch, but I'm not sure why. Just looking at the graph I'm left with the feeling that, for two allegedly correlated variables, the estimators chosen have remarkably different variances. I'm not saying that I agree with Mankiw -- indeed, I hadn't even thought about the question before reading Brad's post -- just that I'm lacking an intuitive understanding of the representation. Maybe someone could posit a bit more explanation?

[Greg Mankiw]

Brad,

I believe you can find the data the NY Times plotted here:

http://emlab.berkeley.edu/users/saez/TabFigOUPvolume2.xls

It seems that they used data in column 5 of table A3.

If you look at these data, you will find four episodes when the share going to the top 0.1 percent changed substantially over a five year period (where I somewhat arbitrarily define “substantial” as more than 4 percentage points): the periods ending in 1921, 1928-29, 1933-34, and 2000. The share going to the rich decreased in the first and third episodes, and increased in the second and fourth.

As an economic historian, you will surely recognize these dates. The first and third are associated with economic turmoil and financial market panic, while the second and fourth are associated with a booming economy and stock market.

It is an interesting question what the right time-series model is to explain the dynamics of inequality and the macroeconomy. But the simple scatterplots you posted on your blog don’t seem to capture what is clearly in the data.

Greg

[Jay C]

Jeez, Brad, why even bother with trying to parse Mankiw's numbers? The entire thrust of his letter can be summed up in the last three words: "politics of envy. This, along with the equally tripish punishing the successful is just the standard Republican mantra duly trotted out to try to deflect any and all criticism of their fundamental economic principle: that the rich deserve more (if for no other reason than just because they're rich), and everyone else can go ----- whatever.

[Stephen Kriz]

The devastation of the American middle class and five years of stagnation in the stock market should convince people what I have known for years - Republicans are simply bad for the economy!!!!!!!!!

[P O'Neill]

>>
The first [1921] and third [1933-34] are associated with economic turmoil and financial market panic, while the second [1928-29] and fourth [2000] are associated with a booming economy and stock market.
>>

The 2nd and 4th also see stock market crashes and recessions right afterwards. Marx pops up in the strangest places!

[bakho]

Greg writes: "Here's the lesson: If policy makers' primary goal is to reduce income inequality, they should put the economy through the wringer. But if they want economic prosperity for all, they should avoid focusing on the politics of envy."

Just WTF is that supposed to mean? If one looks at the Clinton era, then it is clear that the wealthy did better than the poorest Americans, but the poor made very significant gains under Clinton. The poor made these gains because Clinton had a policy of "making work pay". Clinton increased the EITC. The last minimum wage increase was signed into law by Bill Clinton. Clinton signed the family and medical leave act, provided support to working mothers and health care to children of uninsured workers. Under Clinton the unemployment rate for African Americans dropped below 10% for the first time in decades. A single mother moving from welfare to work saw available resources increase over 5 fold under Clinton. The poor greatly increased their benefits and the wealthy did not do too shabby. Despite paying a higher tax rate, the wealthy increased their wealth faster than the poor under Clinton.

Under fiscal policies that increase wealth for the poorest Americans, the rich can still get richer. Let's rephrase that Mankiw statement:

If Americans want economic prosperity for all, the politicians and the rich should avoid focusing on the politics of GREED.

[Unstable Isotope]

Funniest.DeLong post.ever

I'm glad you took on Mankiw. I thought my head exploded when I read that letter this morning. We have to remember that Mankiw was the Bush official who said that outsourcing is good for the economy.

[Nate]

Greg,

Thanks for coming here to elaborate on your points. While I don't like the data Delong chose to use, I don't necessarily agree with yours either. Those are obviously times of great turmoil, and other than 1921, each associated with the formation or aftermath of a bubble. I suspect it says more about the nature of income distribution during and following speculative mania than natural patterns of accumulation.

Can you draw similar conclusions from more general patterns in less exceptional times by lowering your threshold?

Nate.

[Mike]

Professor Mankiw,

First, I just wanted to thank you for answering Brad's criticism of your article. A little back-and-forth tends to generate better information than isolated soliloquy.

I wondered, however, whether there's a difference between a generalized statement about the correlation of change in fractile income and change in GDP. I went ahead and took averages of change over five years for the two series (from the sheet to which you linked) and looked at the numbers. After staring at it for a while I finally figured out that my earlier objection to the significant variance difference issue in Brad's scatterplot was the problem. So I used the 99.9 fractile percentage to get the total inflation-adjusted dollars dedicated to the fractile. Doing a five-year running average for those numbers yielded much more satisfying results visually (and a 50+% R^2).

I can't post an image or hyperlink in Brad's comments section, but you can view it here:
http://img.photobucket.com/albums/v159/michaepf/Graph.bmp

I decided I was too lazy to format and upload a scatterplot of the 1-year changes in 99.9 fractile income against 5-year changes in GDP. I do have a final to study for, after all....

[Mike]

Of course, in retrospect it would've made more sense to average the five-year AVERAGES than simply to sum them. (Obviously GDP didn't increase by 110% in one year during the mid-1980s or fall by more than 100%, er, ever. Whoops. Darn Excel for not reading my mind!)

[howard]

ya know, my original thought when i saw the mankiew letter was that this is wall street journal editorial quality, the arbitrarily lumping together of two things to tell some kind of story.

Then i looked at the prof's first-order study and i was confirmed: no one knows exactly what - if any - relationship there is.

And then i read what i will take (until someone tells me otherwise) to be the real Greg Mankiew gracing our little conversation, and he says, and I quote: "It is an interesting question what the right time-series model is to explain the dynamics of inequality and the macroeconomy."

While perhaps i am being dense about this, i can't see how this translates to anything other than (to quote myself): "no one knows exactly what - if any - relationship there is."

Which makes me wonder what the point of the letter was.

[howard]

er, sorry, "mankiw," not "mankiew."

[ogmb]

This is the key Mankiw lie (or rather, Bushism, as it is not technically incorrect but designed to mislead):

"The rich took their smallest slice of the economic pie during the 1970's - a period when productivity growth was low and unemployment and inflation were rising."

But the data (Table A3, column 5) shows that the share of the top earners was virtually unchanged (between 9 and 11%) from 1952 to 1980, a time frame that saw some of the strongest economic growth and prosperity in U.S. history. So his conclusion of "If policy makers' primary goal is to reduce income inequality, they should put the economy through the wringer" is supported by partisan cherry-picking of data rather than anything resembling analysis.

[Lee A. Arnold]

The question is, what will be done with the money. One has only to talk to some really rich people off-the-record to learn how preposterous the notion is that a few bucks either way matters to them. There are only so many hours in the day for them to be productive, and there are lots of other people who can be productive too, with a decent education, a clean environment, and medical coverage. We can only hope the Dems play their cards right, so that it is the Republicans who are driven to raise taxes on the rich--and then we can vote the bums out anyway in the next election, just for wasting our precious time by threatening Social Security. I hope that someday soon someone with the scholarly chops writes the psycho-social history of the entire "rational self-interest/incentives" canard, from its early adumbration in Adam Smith through its 19th-century sophistical mechanics to its recent ascendancy as the one-size-fits-all explanation of human behavior. It really is one of the main diseases in the world, yet the etiology and epidemiology are unlimned. It appears to be rather directly related to two main intellectual failures of the modern era: (1) in religion, the casting aside of the mystical tradition and the real basis of altruism (very far from envy); and (2) in science, the inability to predict complex systems and the substitution of single-variable explanations to make it look like you know what you're talking about. Of course, most people still think it's necessary to believe in one camp or the other--and even after that, they blow it!

[Carol]

Mankiw's comment about choosing the right time series reminds me of the famous joke in scientific circles: a mathematician posits that all odd numbers are prime. He tests his theory by going through all the odd numbers: "1, prime, 3, prime, 5, prime, 7, prime, 9, bad data, 11, prime, 13, prime, 15, bad data...."

[bakho]

The real SMEAR in Mankiw letter is the implication that the goal of Democrat's fiscal policy is to reduce "income inequality". Reducing income inequality has NEVER been a FISCAL POLICY goal of ANY DEMOCRATIC ADMINISTRATION. That is the BIG GOP LIE that Mankiw snarkily inserts into his letter. Expanding the economy and expanding the income of the poorest Americans is the goal of Democrats and has always been the goal of Democrats. Corrupt GOP politicians like Bush only care about getting re-elected and handing out corporate welfare and our tax dollars to wealthy supporters. The issue is not Income Inequality. The issue is ECONOMIC JUSTICE, making sure that those who work hard and play by the rules are justly rewarded and compensated. Of course, to Mankiw and Bush Economic Justice is "Class Warfare". Bush policy is to screw the poor so he can reward his rich buddies.

The GOP tries to SMEAR BILL CLINTON by saying that President Clinton did not do enough to address income inequality and that President Clinton did not deliver to the Democratic base economically. Bullshit. The best welfare program is a job. Clinton delivered jobs to communities that need them. The GOP ignores the problem and delivers more unemployment. Blacks overwelmingly support Clinton and the Democrats because under Clinton fiscal policy unemployment dropped below 10%. Now the Republican Bush is in office with his fiscal policy of looting the treasury to pay off his wealthy campaign supporters and giving the finger to American workers. Unemployment in the black community is back over 10% again. No wonder blacks support Democrats like Bill Clinton and won't vote for corrupt GOP politicians like Bush that screw the black community economically.

The rich can take care of themselves. That is what they do best. The rest of us who work hard and play by the rules expect Economic Justice. A higher tax rate is the least the wealthy could pay for the priviledge of living in the country that made them wealthy.

[Nate]

I'd ask everyone to take a deep breath and go back and look at Mike's excellent post, and particularly the graph contained therein. It indicates that, with some exceptions such as the inflationary 70's, distribution of income to the top .1% tracks pretty much as a high-beta version of GDP growth. For all the implications about bankrupting society to bankrupt the rich, or the response from Brad that the pattern doesn't even exist, this isn't that surprising.

This implies both Mankiw and Delong are somewhat partisan. I'm shocked.

[Michael Carroll]

Mr. Mankiw,

The use of the periods 1928-29 and 2000 as the end points of periods of economic boom scares me. It should scare anyone who undestands what happen to markets in 1929 and 2000. If there is ANY correlation between the "substantial" increase in the share of income goint to the top 0.1 percent and the severity of the correction that ended that those periods we should ALL be scared.

If you want people to agree with you, stop scaring them with self parodying references that undercut your basic theme.

Did it occur to you that the two periods of economic boom that you reference are both associated with investment bubbles?

Did it occur to you that this current period of income disparity is also associated with a widely perceived investment bubble?

Should we expect the story to end any differently this time?

Here's one not-too-reassuring difference: This time (i.e. 2000-2005) the REAL income growth of the lower 80 percent has been "substantially" flat in contrast to 1921-1929 or 1987-1999.

[ogmb]

Nate: "It indicates that, with some exceptions such as the inflationary 70's, distribution of income to the top .1% tracks pretty much as a high-beta version of GDP growth."

1. Mike tracks inequality *CHANGE* vs. GDP change. This is not the same as your claim.

2. If you look closer (e.g. by graphing one series over the other rather than both over a timeline or by tracking Z-scores), this apparent correlation evaporates quickly.

3. If you lag the GDP change against inequality change timelines by (arbitrary) 5 years, the correlation becomes negative (translate: "reduction in income inequality precedes GDP growth").

[Ottnott]

As Michael Carroll (above) and P O'Neill ("The 2nd and 4th also see stock market crashes and recessions right afterwards. Marx pops up in the strangest places!") both point out...

...an alternative to Mankiw's interpretation of the data is that, when the share of income going to the top 0.1% has risen "substantially", a crash results.

Punditry is so easy and fun when you don't take the time to find out if the data really shows what you say it shows.

[Mike]

Nate - Yay, someone read it.

OGMB - I'm glad someone else took the time to fiddle with the numbers. :) My guess is that the data for both sets is trended -- and not linearly, from looking at the data. Furthermore I'd wager that both variables are autocorrelated as well, although I'm not going to download the dataset again to check.

Working with macroeconomic variables effectively is just extremely difficult. It's much too easy, as this comment thread has made abundantly clear, to interpret the data as one pleases. Without longitudinal data it'd be pretty much impossible to pull out any of those biases from the comparison. It's fun to play statistics, though :)

[Nate]

Some politely frustrated new blood my 501(c)3's recently hired suggests that this is all mush, and further, we've ignored data points we don't like.

I'll turn to those with much more knowledge than I have, the IMF:

http://www.eldis.org/static/DOC4760.htm

who find that the following variables strongly improve income distribution, holding demographics and population static:

1) Rapid GDP Growth
2) High income levels
3) High investment rates (e.g. low consumption)
4) Real depreciation
5) Improved trade

Interestingly, they found inflation did not really have an effect. The first point immediately demonstrates that there's others who have found Mike's correlation.

The poorest fit of the data, as I noted earlier, is the 1965-1980 period, which saw a significant increase in equality followed by a rebound in the ROI of income distribution to the wealthy around 1975 despite persistent poor GDP growth. I don't know if anyone else can fit something to this better than the inflation data. Anyone want to try?

ogmb, pardon my imprecise language; given that the other series is clearly a delta, I assumed others would take distribution of income to be a delta too. Clearly that was incorrect, sorry.

[Sam Spade]

Egad sir. There you go again, laying waste with your gigantic economics wang.

[Julian Elson]

First, I've also looked at data in international comparisons. No where have I seen persuasive evidence that income inequality increases growth.

That said, I think there is an idea to Mankiw's letter that MAY be valid, but which he didn't elaborate on. It isn't the idea that income inequality per se is good for the economy, but rather that specific, state policies which can reduce inequality, such as progressive taxation and transfer payments, are bad for the economy. If technology, culture, preferences, etc, all shake out so that the economy has very little inequality, great, but if the state imposes a lot of taxes and creates a lot of transfer payments, or raises the minimum wage, it creates a lot of deadweight losses.

It's very easy to see this when you're drawing supply and demand curves and looking at deadweight loss triangles from taxes. It's less easy to support when looking at empirical, real-world data to see whether growth is associated with the kind of low-tax, low-transfer, low-minimum wage state which libertarians and conservatives would like to see.

[Oskar Shapley]

OMG, it turns out Clinton was a Republican all along! ;)

[ogmb]

From Nate's link: "The macroeconomic variables that are found to have a significant negative effect on changes in income inequality (that is, are associated with an improvement in income distribution) are higher growth rate (...)"

Nate, it seems like the IMF guys found the opposite of Mike: *negative* correlation between inequality growth and GDP growth. Not to mention the issue of causation, which seems to be at the center of Mankiw's claims.

[Nate]

Ugh. I meant to point out that the fit is good, but it's the inverse correlation as compared to Mike's conclusion. This just bolsters that new guy and ogmb's claim that either the US is just different, or this is, indeed, all mush, no matter what variables you choose.

Think twice before clicking Post this late at night, or posting...

[Mike]

Well, Nate, since anything seems to be better than studying for a Finance final, I've taken a look at the paper to which you linked.

Given the amount of work that obviously went into blending the data sets for this paper and the estimators generated, I'm reluctant to tear it down too badly. I know that I certainly could have done no better. At the same time, I think the problem is very clear: insufficient, heterogeneous data.

Immediate problems: the author's data comes from repeated cross-sections. Now, he does attempt to correct for this by averaging his data into a single cross-sectional set, but this is still unconvincing in groups where N=3, and it reduces his degrees of freedom by a spectacular amount. He never gives us the final degrees of freedom estimator, and I don't remember exactly how it's affected by averaging repeated cross-sectional data, but it makes his claims to significance difficult to judge. Is it so difficult to include p-values with your regression output when using SAS or STATA?

Furthermore, different sets of cross-sections are taken from different time periods. He notes at the end that, at least for the gini coefficient estimator, he included a time dummy to check for significant time trends, but never did so on the adjusted gini coefficients by themselves. Again I'm pretty sure that that adjustment by itself further removes degrees of freedom from the calculation.

Furthermore, the estimator itself assumes that change in the Gini coefficient (which itself is an imperfect estimator of income inequality...) is linear while -- as the appendix describes -- it may very well be nonlinear.

Finally, I don't know if it's just because I'm tired, but some of the independent variable estimators seem very obtuse. The terms of trade effect is measured using changes in growth versus changes in trade-adjusted growth; growth is significant but less so economically than the terms of trade effect, which to me seems strange. The investment coefficient seems economically insignificant as well (10% increase=gini increase of 1). The income variable itself is measured in levels of deviations from the log of the mean of the time series in 1985 dollars, so a difference of three standard deviations is correlated with an 1-unit lower Gini. Most striking for me came the information that the adjusted ginis had been affected most strongly by the differences in sampling technique.

Frankly, if I had turned in this topic for my Applied Data Analysis class, my professor would've returned it to me with a note saying, "You can make no meaningful claims about this data."

[Mike]

On the other hand, the author is very forthright about admitting the need to drop lots of observations and countries, and he gives clear reasons for doing so. I am no econometrician, I haven't finished graduate school, and I have no strong background in studies of inequality. So please take my comments with a grain of salt unless someone more knowledgeable agrees with me!

[kharris]

Even if Mankiw had been right about correlation between income equality and slow growth (why is it that Mankiw seems to keep being wrong about economics?), his statement –

"If policy makers' primary goal is to reduce income inequality, they should put the economy through the wringer. But if they want economic prosperity for all, they should avoid focusing on the politics of envy."

-- presents a false dicotomy. Policy makers usually have multiple goals. If greater income equality is one of several goals, then putting the economy through the wringer would not be the first choice of intermediate targets. We could couple the goal of greater income equality with the goal of (environmentally sound) growth, for instance. It is hard to find evidence at anything like our current marginal income tax rate that growth would be undermined by higher tax rates.

Then, sorry folks, but "the politics of envy" and "class warfare" are just veiled name-calling. Mankiw and his masters don't mind doing something about income distribution, as long as the direction if right, but they don't want to talk about it. If the other sides tries to talk about it, they slap ugly labels on the discussion. This one time "great economist" is now working overtime to stifle discussion. He does it when asked (planted question, perhaps?) about Krugman and he does it when he insists that discussions of income distribution are the politics of envy. Pure scum.

[anne]

KHarris

Lovely response. I will add when bird free :)

[weco]

First of all, first rule of statistical inference: correlation does not equal to causation. So, even if there's evidence of correlation between income inequality and growth, Mankiw has no justification to say:

"Here's the lesson: If policy makers' primary goal is to reduce income inequality, they should put the economy through the wringer. But if they want economic prosperity for all, they should avoid focusing on the politics of envy."

One can equally well argue, due to the strucutre of the US economy, the top 1% benefit disproportionately from economic growth (and also suffer greater percentage drop in income in bad times, lest I be accused of class envy).

That does not mean we can't raise the marginal tax rate or tax dividend to balance the budget or even :gasp: just put some money back to the so-called SS 'black hole'.

But not only that, from what I understand from my Macro courses, the relationship between economic growth and income inequality is a very active area of research, which means the relationship is best described as uncertain.

Also while one can legitimately argue that in a developing country like China, people might prefer higher GDP growth over eqality, there's no reason why people in the developed world should have to make such a choice. The correlation, as the above data has shown, is filmsy and can go one way or the other.

I'm sure Prof Mankiw knows all these. So why did he write to NYT as though the issues of income inequality and growth is settled?

[Carlos]

The interesting thing is that Mankiw sees four periods when inequality increased and identifies them as high growth periods but does not try the reverse; locate ALL high growth periods and see if inequality increased. That would show that during the highest growth period of all inequality diminished.

[Dave]

http://www.truthandpolitics.org/top-rates.php

A nice little timeline of the top marginal tax rates.

It seems to correlate very well with rising income inequality. When hyper-progressive taxation was reduced in the 1980's, income inequality as shown by the top percentiles, correspondingly increased. If anybody denys this connection, let me know.

How income inequality changes an economy isn't a concern for me. Looking at the time of the lowest income inequality: the 1970's. We can all say that the 1970's was a booming time thanks to the high taxes and price controls. We can also say the 1970's was a horrible time for our economy. For the last 80 years that are looked at in any data, there are so many things going on in the US and World economy. Income inequality alone will never fully explain how well the economy is doing. Neither extreme seems good for an economy. If everybody has equal income, there is no incentive to do anything. If only one person or a few people have all the income, people aren't able to get ahead. We need to keep an eye on income inequality but not get so worked up about it. We should concentrate on social programs that make it easier for workers to move between jobs. Mobile pension funds, social security reform, and a more transparent health care system such as Medical Savings Accounts or Unniversal Health Care.

This question remains: Should a tax system be used for income redistribution? Perhaps this is just a matter of opinion. Taking earned money from the rich with the specific intent of giving it to the poor seems wrong. I believe a tax system should be used for raising government revenues. If a government wants to implement social programs and help labor markets move more efficiently then they should raise income through taxes to introduce a program to help people find jobs. I support a somewhat progressive tax system, perhaps a little less progressive than our current system though. Through the end of this week you can still comment on possible tax reforms at:

http://www.taxreformpanel.gov

We can subsribe to Howard Dean's belief shown in a recent quote that "[republicans] have never made an honest living in their lives." If we believe this then we need to make sure that people are honest and not defrauding or stealing from their workers and customers. What we don't need to do is steal the money back through an overly progressive tax system.

[theCoach]

KHarris,
Just to add a little to your post, one of the goals might be to ensure that affordable housing across the entire income spectrum is available in all areas, I think a laudable goal.
Some of the fallout from that goal definitely involves income distribution, and maintaining that condition is probably a factor in maintaining long term growth.

[C Roast]

Geez…all of this econometrics is hurting my head. I think I liked Political Economy before the rise of that buggah’ Marx required the separation of church and state.

Let’s keep the Big Picture in view here. We live in a free market country where everything has its price, and both the wealthy and the panhandler are allowed to buy the laws, amendments, statutes and regulations of their choice that will facilitate transfer payments and tax exemptions allowing them to enhance the accounts of themselves and alternately their golfer friends or wino buds’. If I’m not mistaken, intellectual justification has its price as well.

Typically, the Dems price is quite high and the marginal return to the wealthy is less than they would wish. So, the Dems (though infinitely malleable) present the golfers with a relatively inelastic demand curve. On the other hand, the GOP (being sublimely corrupt) are relatively elastic in their approach to capital accumulation leading to the above discussion.

Supported by a “winner take all” voting system, if the winos have enough packet change, they too can buy their way onto this free-riding juggernaut for a joy-ride to the bank. And if they don’t…well, in my old neighborhood, the golf cart replaced the caddy long ago.

[bakho]

Mankiw is writing in response to a NYT article that takes the curtain off the rapidly exapanding income gap between wealthy and wealthiest. Mankiw creates a straw-man argument about "economic inequality" and snarkily suggests that economic inequality could be brought eliminated by class warfare that laid the economy low.

No one wants to achieve income equality by bringing down the US economy. No one. This is a stupid straw-man argument. (Would you rather have corrupt, irresponsible Bush fiscal policy or another Great Depression? That question is stacked. it is not relevant.) The relevant question (the one Mankiw replaces with his straw-man argument) is "Should we return to fiscal sanity by having the wealthy pay a higher tax rate (their fair share of taxes)? The data clearly show that the wealthy can pay higher tax rates without tanking the economy. This is the issue that Mankiw is trying to avoid. Lowering taxes on the wealthiest was reckless and irresponsible fiscal policy by the corrupt Bush administration that Mankiw advised. The NYT piece on income inequality adds to the evidence that Bush fiscal policy is reckless and irresponsible. Mankiw is in a no-win position of either admitting that the corrupt Bush administration did not follow his advice or signing on to a corrupt, irresponsible fiscal policy that creates huge deficits.

Mankiw tries to change the subject with a stupid straw man argument about blowing up the economy to achieve income equality. We are not amused. Mankiw can feel good that he did not follow a program of deliberately tanking the economy to achieve income equality? I guess it beats having to address unjustified support for a corrupt and irresponsible fiscal policy.

[Jonathan Goldberg]

Dave:

Re:
"I support a somewhat progressive tax system, perhaps a little less progressive than our current system though. "

I hope not, since our current tax system is not progressive. Income taxes are, a bit, and that's all Republicans ever talk about. However, payroll, sales, and real estate taxes are all regressive. The last time I saw numbers this all added up to about zero. Since the Bush tax shifts I wouldn't be surprise if the total tax system is regressive, but I don't know and am too lazy to find out. Perhaps someone else can tell us.

[Chuck Roast]

Jonathan and Dave:
Check out this Citizens for Tax Justice page for a real eye-opener on our current "flat-tax" system.
http://www.ctj.org/html/whopay.htm

[Dave]

Jonathan-

You bring up a good point. You say that our overall tax system is not progressive, but then later say that you can't really be sure that it isn't regressive. While your partisan blinders are pretty obvious, you do make a very important point; our system of taxation is not very transparent. Pushing for more transparency in our tax system is important. Tax simplification is a major goal of the Bush administration right now and I want to encourage you and everybody else to participate in that process. If liberals don't get involved the republicans may simply choose to further obfuscate our taxation to their advantage.

My former argument still stands though. Overly progressive income taxation is very transparent and often encourages tax avoidance and evasion. Until we have a more transparent system overall, we can't justify taking excessively from one income class just because we have other regressive taxes on the books. Don't just cover up problem taxes with a catch all progressive tax, remove the problem taxes. If there is a regressive tax, get rid of it or alter that tax so it is no longer regressive. More trite, "Two wrongs don't make a right."

[Dave]

Chuck-

Thanks for the link. Very good overall information.

I'll wait 'til Brad's post on consumption tax regresivity before I comment on the topic though. For now I'll just say that it is good to look at the overall system of taxation. But, to reiterate my earlier points we need to fix the regressive taxes not make the progressive taxes more progressive to make up for it. This just further seperates the American taxpayer from understanding their true cost of government, allowing upward spiralling spending by either party. Don't allow career politicians in either party to obscure the price we pay for government services.

[piotr]

Modest conjecture: rich people are different than you and me.

One difference is that most of your income and mine income come from wages, and little from capital gains. For the rich, the opposite is true.

This difference has a curious implication: you and me make money when we have too, and the rich when they choose too. If you keep investments for a number of years, the years of high income may be exactly the bad years for your estate. More precisely, when the investments are doing nicely, you may choose to keep them, no capital gains, but when they start to wobble and either go down or they seem to be increasingly risky, you sell, and you register gains.

This could explain why spikes in the income of the rich were leading indicators of economic downturns -- smart money were leaving the market.

Alternatively, the rich are in the position to make a killing during speculative bubbles, e.g. they may be represented disproportionally among the sellers in California real estate market. Bubbles are a lagging indicators of growing economy and a leading indicator for downturns.

About those rich who actually work for the living: is a CEO who makes 10M/year ten times more productive than a CEO that makes only 1M/year? I guess that one can define the productivity that way.

[DoBeDo]

It's interesting to look at this question in the economies of other countries. The fastest growing economies in the last fifty years have been those of the Pacific Rim countries. One of the characteristics of these countries is that they have relatively low levels of income inequality. Some of the slowest growing economies in the last 50 years have been in Latin America. These countries have extremely high levels of income inequality. This suggests that the idea that high income inequality leads to high growth may need some tweaking.

[a different chris]

This rang true to me:

If you lag the GDP change against inequality change timelines by (arbitrary) 5 years, the correlation becomes negative (translate: "reduction in income inequality precedes GDP growth").

I'll kick in my just-so story:

I think what happens is simple - an important technology shift drives a growth period. The innovators (Bill Gates) get rich. And richer. And richer. Unfortunately, they stop being innovators and start becoming rentiers (Bill Gates). So all the money pools and freezes at high altitudes.

Well, the next wave of innovators need that money. Eventually government steps in and unfreezes the money and lets it flow down to the masses, where the innovators can get it. And growth starts again. WWII was a really good, if unfortunate example of this.

Consider:

1- Are personal computers (A) a cutting edge tool, or are they now (B) a commodity?
2- Are Bill Gates children likely to (A) equal him in the contributions his PC revolution made to the economy, or (B) pretty much be nobody special
3- Is Bill Gates/Microsoft going to (A) invest and do ground breaking work in the next technological wave (biotech?) or (B) continue to make computer crap?

If you answered, like I would, (B) to all three then you see why government has to wade in *occasionally* and tax the hell out of the rich. Greg Mankiw is pretending that he doesn't know the old economics joke "Every year the questions are the same, only the answers change." Steep taxation on the rich is, IMNSHO, necessary at this point, but that doesn't mean permanent. But leadership means you gotta see that innovator->rentier transformation and act accordingly, before Mars steps in and does it for you.

Mankiw, Greenspan, even Cheney know this somewhere deep down in their bones, but they won't face it. So we get huge deficits and low prime rates in hopes it will provide the nourishment our next generation of innovators will need that Bill Gates simply isn't going to give them. But that's like feeding a tiger twinkies, it's just wrong wrong wrong.

Ah well, the American Century sure was short, no?

[bakho]

dave misses the point. Rich people get rich because the laws favor them getting rich. The rich are receiving the most benefit from laws as they are currently written. The rich should not expect the rest of us to subsidize them with corporate welfare and infrastructure that overwelmingly benefits the rich. The rich SHOULD pay a lot more to support the system that makes them wealthy. There are laws that set the minimum wage at little more than $5 per h. The rich take advantage of that to become more wealthy. The rich OWE minimum wage workers (who work as hard as anyone else) access to health care and decent retirement because those workers were not paid enough wage to save for retirement. The rich are welcome to their excesses. However, we should not overendow the rich when basic needs are unmet for large portions of the population. That would be an unjust system that leads to social problems.

Mankiw tries to relabel the demand for economic justice as "envy the rich". That is grossly misleading. Average Americans don't give a FF how much money rich people have as long as the needs of average Americans are being met. Mankiw is pandering to the rich and trying to protect a corrupt Bush administration by ignoring the economic deterioration of the poorest Americans as a direct result of corrupt and irresponsible Bush fiscal policy. Americans don't want to take from the rich out of "envy". Americans want people who work hard and play by the rules to have their basic needs met. Americans want a just distribution of the economic pie, not a winners and losers game of Monopoly. The rich in America OWE the rest of us that much. The rich are welcome to keep the rest.

[Tim Worstall]

Hhhmm. Maybe I’m stretchinghere, but didn’t Mankiw look at the figures the NYT had used? And Brad use some different figures? What happens to Brad’s plots if one uses the NYT figures?

[The NYTimes uses the Saez-Piketty top 0.1% series. I have the Saez-Piketty top 1.0% series up on my computer. That's the difference. The correlation between the two is high. (In fact, IIRC, the 0.1% series is extrapolated assuming that everything follows a Pareto distribution.)]

[Auros]

I've tried to post a polite reply to the one conservative blog that had the balls to TrackBack their critique of this post. If anyone with better credentials wants to add on to that, I encourage it.

Commenting on the plot provided by Mike: I take it your plot is not adjusted for population growth. This suggests that if the correlation exists at all, it may be that the share belonging to the people at the top stays steady, but they don't breed as fast and they don't let new people into the "club" during recessions, while as a whole the population continues to grow.

[ogmb]

"Hhhmm. Maybe I’m stretchinghere, but didn’t Mankiw look at the figures the NYT had used? And Brad use some different figures? What happens to Brad’s plots if one uses the NYT figures?"

They're here (link from NGM's post). You can use any of the columns, they're essentially collinear.

http://emlab.berkeley.edu/users/saez/TabFigOUPvolume2.xls --> Table A3

[ogmb]

"Commenting on the plot provided by Mike: I take it your plot is not adjusted for population growth."

Mike uses per capita numbers, and he tracks inequality CHANGE. And of course the five year span he looks at doesn't involve "breeding". The point of Mike's graph is simply, as he points out later, that you can always find something in the data if you restrict yourself to looking at them from only one angle.

[howard]

Dave, i don't honestly know what you are trying to get at.

I favor a highly simple, progressive tax system. That state taxes are highly regressive, and contribute to a situation where the overall tax burden in america is quite flat, does not seem to me in the slightest to be an argument against simplifying and increasing the progressivity of the federal income tax.

Nor do i understand how doing such a thing would "obscure" the costs of government, unless you subscribe to the WSJ "lucky duckies" school of thought that we all benefit equally from the goods and services government produces, so we should all pay equally; under this view, i suppose, a minimum wage worker with no benefits whose income taxes wouldn't rise might support programs that he or she wouldn't if he or she were better off and felt an increased tax bite.

i am rather underwhelmed, so perhaps you have a better argument than the one i just ginned up?

[Mike]

OGMB -- Thanks for sticking up for me, haha. I'll have to make graphs more often, this is fun. ;)

[Dave]

bakho- "The rich OWE minimum wage workers (who work as hard as anyone else) access to health care and decent retirement because those workers were not paid enough wage to save for retirement."

Where in the United States Constitution does it say that the rich owe anything to anybody?
Workers enter a willing contract with their employer for whatever payment they agree to. If employers are colluding against workers then they should be prosecuted as such.

[Dave]

piotr- "This could explain why spikes in the income of the rich were leading indicators of economic downturns -- smart money were leaving the market."

I couldn't agree more. Smart people make money. Suckers lose money. Solution: Teach the suckers not to play games with their money. Stock Market has become white collar gambling. People that can't afford risk of investment shouldn't invest.

[Dave]

Howard- To clarify, I was not making "an argument against simplifying and increasing the progressivity of the federal income tax."

I very much support a simpler federal tax, be it on consumption, wealth, or income. I include consumption and wealth here under the belief that a properly administered tax can be made progressive and remain simple in theory, regardless of who and what the legal burden of the tax is on. The cluttered system of deductions for this thing or that in our current income tax is problematic. How flat or progressive a tax should be is still open to debate. My argument is simply against using the regresivity of state taxes as justification for excessive progresivity in federal taxes. Make a federal tax as progressive as it should be for the federal budget and fix state budgets as a seperate matter.

[weco]

bakho: "We are not amused. Mankiw can feel good that he did not follow a program of deliberately tanking the economy to achieve income equality?"

No, he followed a program of deliberately bankrupting the federal government to achieve income inequality ...

[bakho]

Dave, a guarantee of pension and health care coverage in old age is not in the Constitution. Our Constitution was written during a different set of social circumstances.

However, current wage and benefit contracts are negotiated under a Social Contract that guarantees SS retirement income and Medicare benefits. Workers are willing to accept lower wages based on good faith that SS retirement and Medicare will be there for them. If the rich have no intention of fulfilling the social contract, they are negotiating in bad faith. In which case the minimum wage needs to be closer to $10 / h and all workers need to negotiate much higher pay rates.

However, since current wages are negotiated based on the addition of SS and Medicare benefits, the rich OWE this to the workers. Anything short of fulfilling the promises would be bad faith negotiation and cheating the workers.

Mankiw doesn't even address the issues of social contract, guaranteed future benefits or the fact that Bush has been corruptly using SS taxes, intended to build a Trust Fund to pay for future benefits as his own personal slush fund to grant tax cuts and corporate welfare to his buddies. The income gains of the rich must be eyed as a source of paying for future benefits because a good chunk of change from our SSTF has been misused by corrupt Bush fiscal policy and given to the rich. This does not amount to "Envy the Rich" as Mankiw maintains. A better analogy is victims claiming what is owed after suffering embezzlement. Mankiw should be ashamed for doing nothing to stop it from happening.

[Dave]

Bakho- Your points are well taken.

Failing pensions and faith in social programs contributes to a situation where the rich can benefit more than the poor. If one side is consistently negotiating in bad faith, the other side should stop trusting them. Pensions are just another risky investment that the middle and lower classes should start avoiding. The way our government handles punishment of failure is unacceptable. Organize your coworkers and demand your pay in the year that you work not when you are 65. It has become foolish to trust a company to pay you many years later. People should learn to keep their money safe, close, and liquid.

If we are unwilling to punish the rich that fail our trust, then we need to stop trusting them.

[Bakho]

Great new Krugman column. PK "ratfucks" the GOP with their own words. "politics of envy" "slicing and selectively presenting data in an attempt to mislead". What else could PK do?

http://www.nytimes.com/2005/06/10/opinion/10krugman.html?n=Top%2fOpinion%2fEditorials%20and%20Op%2dEd%2fOp%2dEd%2fColumnists

[Mr. Econotarian]

May I say that there is a big difference between income equality achieved through market processes, and income equality achieved through taxes, and that it is difficult to compare the two. One involves a large deadweight loss to the economy, and the other does not.

There also is a big difference if the income inequality is being powered by immigration. Non-native born Americans are making up an increasing proportion of people in poverty, which is not surprising. They are generally better off "poor" in the US than their existance in their native countries.

[Ari]

Econo

Where is the analysis or data?

[liberal]

Mr. Econotarian wrote, "May I say that there is a big difference between income equality achieved through market processes, and income equality achieved through taxes, and that it is difficult to compare the two. One involves a large deadweight loss to the economy, and the other does not."

Not necessarily.

For example, there is no deadweight loss when land is taxed. And taxing land heavily (based on its rental value) would tend to narrow income inequality.

It's also difficult to see that there'd be much, if any, deadweight loss to a flat tax on wealth, which would narrow inequality in wealth (and wealth is a far better metric for measuring "inequality" than income).