Regressive Progressivism

Just in case any of you are stuck under a rock, there has been a lot of discussion about inequality of wealth and income recently. This discussion of course includes the usual hand wringing about hollowing out the middle class and needing more "progressive" taxes. In fact Brad DeLong thinks at least part of the reason is "Further increases in inequality as the tax and transfer system becomes less progressive." In a blog-off between Russel Roberts and Heather Boushey published at, Heather asks "why should we not tax them to create greater income equality?"

Brad, I think the problem is too much "progressivism" in U.S. policy. Heather I do not believe your means, taxes, leads to your desired ends, income equality.

I am going to use CEO's of major multi-national companies (just "CEO" hereafter) as a proxy for the highest income earners. The following analysis applies similarly to others such as sports and entertainment stars.

Let us start off by determining the likely demand curve for CEOs in the market. We know that by law (both legislated and by practical requirement) each company must have a CEO or equivalent. We also know that the cost of a CEO to a multi-national company is very tiny compared to revenues, operating budgets, and market capitalization. What these two things mean is that the demand curve for CEOs is nearly vertical and inelastic. The actual supply curve does not matter for this analysis.

Progressive policies typically increase costs to CEOs, whether it is a straightforward increase in income tax (as Heather suggests) or increased liability in being a CEO (such as Sarbanes-Oxley). At the margin cost equals price, so with a new or additional progressive policy, the supply curve shifts upwards. Since the demand curve is nearly vertical, most of the additional cost will be born by the company or more accurately its shareholders, and yet demand will not decrease by any real amount. Finally then we see that CEO income increases significantly.

More significantly according to IRS data, the Adjusted Gross Income (AGI) share for the top 1% appears to have a positive correlation to the progressiveness of policies. The AGI share for the top 1% has decreased with the Bush "regressive" tax policy, from a high of 20.81% in 2000 to 16.77% in 2003 (the last year I found data for). More research is needed here, but I think I can be confident that the data will bear out the above theoretical analysis.

If everone concerned themselves with their own absolute wealth rather than things that just don't matter, I wouldn't have to think about these things.

Share this

Brandon and Eric, I know at

Brandon and Eric, I know at least one CEO making less than minimum wage for the last two years, that is why I stated CEO of "major multi-national companies". These guys make millions, even though their salary can be very small (stock and options being the prefererred primary compensation these days).

Also, isn’t leverage a perfectly valid explanation for high CEO salaries?

Up to a point. As I understand the data and analysis, CEO compensation has been rising faster than revenues, budgets, and other corporate financial metrics.

Using CEOs as a proxy for

Using CEOs as a proxy for the "utlra-high" income distribution percentile is not entirely wise for a simple reason: CEOs tend to persist over time, while the total populations of the "ultra" percentiles do not. The fact that CEOs are the exception rather than the rule is precisely the point we want to make.

The whole reason why income distributions are not a useful statistic is because the percentiles are not static; we do not live in a caste society where the rich are always rich and the poor always poor. This is especially true at the "ultra" ends of the spectrum.

The better way to debunk these numbers is prercisely to remind people like DeLong that there is tremendous turnover in their composition: today's ultra-high-income are very likely not to be tomorrow's ultra-high-income, since most are in the "ultra" category due to windfalls.

High tax rates at the top

High tax rates at the top are merely going to increase the number of people under the cap. We saw it with CEO salary "reform" under clinton. They capped salary so CEOs started getting a bulk of their compensation through stocks and options. The result? Enron, MCI, Tyco, etc.

If you tax income over $500,000 then there will be very few people that make over $500K, they will get compensation like housing, autos, and other off-book transactions that the IRS can't touch.

While many will point out

While many will point out that the overall tax burden as a percentage of income is greater among the rich than among the poor, something I tend to doubt, at least for the very rich, because of the greater availability of tax shelters to the rich, overall marginal tax rates are not actually higher as income increases. The peak marginal tax rate is around 110-120k. And let's not forget AMT which tends to strike middle class people trying to become upper class people, resulting in their being stuck in the middle class.

Most taxes also tend to distribute wealth from the poor to the less poor even if they are progressive. For example, wealthier people tend to live longer, so they can collect more social security. David Friedman talks about this in Machinery of Freedom, and I've heard it elsewhere. Taxes for higher education are the same; lower class kids usually don't go to university, public or otherwise, and even private universities get a lot of tax money. Even in the case of public primary schools, richer kids end up in better public schools that spend more per child, and the effect is almost certainly greater than just property tax would account for.

And certainly lobbyists, politicians, and defense contractor executives don't tend to be the beneficiaries people have in mind when proposing new redistribution schemes, but they sure get a huge amount of the benefit.

No, more taxes or more progressive taxes aren't the answer. A simpler, easier to understand tax code, sound money, and minimal government spending would go a long way toward rectifying the situation, but it won't happen until the government is bankrupt, which is why we should all vote Republican... they are obviously trying to bankrupt the government :)

Catallarchists continue to

Catallarchists continue to read my mind. I was just having a conversation about CEO pay with a friend yesterday, and we were wondering why their salaries were so persistently high. I hadn't put two and two together, but now it seems obvious. Thanks, David.

Kip, you raise two points,

Kip, you raise two points, first are CEO's good proxies for the ultra rich. CEOs and many of the high income people are in inelastic demand markets with very limited supply. This is why they are high income. I don't see any problem using CEOs here. Also income amongst CEOs tends to be highly variable depending on when they exercise options or sell stocks.

The second point is one of which argument to raise. Yes, you are right, turnover amongst the ultra-rich is quite high. That is a different argument. The argument I present is that "progressive" policies are actually regressive. It is just another argument against Brad and Heather's stated preferences.

Catallarchists continue to

Catallarchists continue to read my mind.

We've been having trouble coming up with our own ideas, but that chip we implanted in your brain is really starting to pay off.

Also, isn't leverage a perfectly valid explanation for high CEO salaries? As a non-manager, I control only what I do, and there's only so much I can do by myself. My salary is limited by that fact. But our CEO directs 50,000+ people. If he were replaced with someone even slightly less effective, profits could fall by millions of dollars.

By the way, I've heard that the median CEO salary is only something like $200,000. It's only the CEOs of very large corporations that make millions of dollars per year.

VentureVoice had a podcast

VentureVoice had a podcast with John Bogle of Vanguard. Around the 40 minute mark John spoke about the changes in stock ownership and how corporations are now manager driven instead of share holder driven. He said that in the middle of the 20th century most stocks were owned by people and 9% were owned by financial institutions. Now that is completely reversed with most stocks owned by finacial institutions.

Since most corporations are owned by financial institutions, corporations no longer have demanding share holders who want big dividends and care about the company. This has resulted in huge increases in the salaries of the managers, like CEOs.

I think the point John was trying to make is that people need to start buying shares and acting like owners. This would cut the salaries of CEOs and other managers as share holders start demanding that expenses get cut so they can be paid big dividends.

Brandon, I don't know the

Brandon, I don't know the exact numbers, but most CEO's are not paid millions. Typical corporations have annual revenues of $100 million, or less. IIRC, the median, excluding small businesses, is around $20 million. Companies of that size cannot, and will not, pay their CEO's 5% of their annual revenues. What they will do, however, is give their CEO's stock options + a reasonable salary (~ $200K). The CEO is motivated to improve company performance by the stock options. I know several CEO's of mid-sized companies. All of them are earning in the 200 to 300 K range, not the million plus range. On the other hand, their car is usually provided by the company, they travel on company expense, and their Amex is covered by the company. Their true compensation is probably 100K higher. If the company performs well, their stock options are usually worth millions.

"He said that in the middle

"He said that in the middle of the 20th century most stocks were owned by people and 9% were owned by financial institutions. Now that is completely reversed with most stocks owned by finacial institutions."

I wonder, might that be because companies are more stable nowadays? If they are less likely to fold then money managers are more likely to utilize them.

Another possibility, of course, might be the elimination of deflation from the economy (the dollar doubled in buying power from 1850 to 1910) which drives down the value of owning something, such as a company, thus discouraging money managers.

"We’ve been having trouble

"We’ve been having trouble coming up with our own ideas, but that chip we implanted in your brain is really starting to pay off."

[Having completely missed his cue, Matt belatedly steps in with a joke.]

Okay, just as long as you don't start making posts about my crush on the cute bespectacled brunette I sometimes work with. Because that would be embarrassing, and I'd have to do a Pi job on my head with the powerdrill.

Using CEOs as a proxy for

Using CEOs as a proxy for everyone with a high income is problematic for a couple of interesting reasons that may also be reflected in your data for the years 2000 and 2003. Yes, CEOs often get very large salaries. However, for the most part they derive a large portion of their income from bonuses, often in the form of stock options. Assuming the conditions for awarding the options in the first place and the pricing of those options is appropriate that should be good for both the CEO and the company. The CEO gets to take the income as capital gains rather than wages, lowering the tax burden. He also makes more by increasing the share price, which if it is more than a temporary increase and is done honestly benefits the shareholders, managers and employees of the company.

Other categories of high earners do not necessarily follow that pattern. Investors derive their income from the performance of their investments. Those may be stocks, but could also be commercial real estate or something else. They will generally receive very little in wage income.

Entertainers are going to get wages. Those will be tied to some estimate of the audience they will be able to draw and the revenue that will help to generate.

My point in elaborating this is that different high earners have significantly different amounts of exposure to the performance of the stock market. Furthermore, CEOs will be tied heavily to the performance of one, or a small number of companies.

Name withheld - My analysis

Name withheld -

My analysis does not depend on the type of compensation the high earners receive. It only depends on the high earner participating in a labor market that is inelastic.

As I stated, the AGI data would need further research. Specifically, proper controls for different forms of income, controls for other societal and economic trends, additional income data over longer time periods, etc. Not an impossible job for someone with the tools and time available, perhaps as a doctoral thesis? But not something I'll be getting to anytime soon.