Do companies pay their CEOs more than they pay in taxes?
You might well have seen, this morning, the news that 25 of the 100 highest paid US CEOs earned more last year than their companies paid in federal income tax. The Reuters version of the story was linked to by the WSJ and retweeted by David Leonhardt; the NYT version already has 120 comments. Both versions, it seems, were based on embargoed copies of this report from the Institute for Policy Studies; because the reporters were given a copy of the report before it went up online, they were unable to link to it from their stories.
But if you do manage to find the IPS website and follow the links to download the full 46-page report, you’ll see that there’s less to it than meets the eye. Certainly it doesn’t come close to demonstrating that its title — “The Massive CEO Rewards for Tax Dodging” is justified. Yes, CEOs get paid vast sums of money. And yes, a lot of corporations pay very little in taxes. But what the report doesn’t do is demonstrate that CEOs who reduce their corporate tax rates get paid more. This kind of thing, from the NYT story, notwithstanding:
The authors of the study, which examined the regulatory filings of the 100 companies with the best-paid chief executives, said that their findings suggested that current United States policy was rewarding tax avoidance rather than innovation.
There are lots of ways that the authors of the study could have tried to back up that assertion. For instance, they could have taken a set of CEOs and split them into two groups: those who are paid more than their companies pay in taxes in Group A, and those who are paid less than their companies pay in taxes in Group B. Then they could have compared whether CEO salaries in Group A were higher than CEO salaries in Group B.
But they didn’t do that.
Instead, they did this:
Of last year’s 100 highest-paid corporate chief executives in the United States, 25 took home more in CEO pay than their company paid in 2010 federal income taxes.
These 25 CEOs averaged $16.7 million, well above last year’s $10.8 million average for S&P 500 CEOs.
Do you see what they did there? The initial set of CEO was the 100 highest-paid CEOs in the country. They then took 25 of those CEOs, and instead of comparing their pay to the pay of the other 75 CEOs in the group, they compared their pay to the average pay for a CEO in the S&P 500. This proves nothing: any subset of the 100 highest-paid CEOs in the country is going to have higher average pay than S&P 500 CEOs in general.
As for the central conceit of the paper — the one which made the Reuters and NYT headlines — that’s pretty silly too. 25 CEOs make more than their companies pay in taxes? Wow! Except, it turns out that only five of those 25 companies are paying any taxes at all, by IPS methodology. The lowest-paid janitor, at those 25 companies, makes more than the company pays in taxes. The driving force behind the IPS result is entirely a function of how IPS calculates the corporate effective tax rate, and the ease with which that can go negative. It has nothing at all to do with CEO pay. (The IPS ignores deferred taxes, which is justifiable; it ignores taxes paid to foreign governments, which is less so, in an era of global corporations operating in dozens or even hundreds of tax jurisdictions.)
This is one good reason, then, for every news organization to link to reports they’re writing about — doing so gives their readers the opportunity to see for themselves whether the report stands up to scrutiny. After all, the world of embargoed reports is clever that way. If you’re a think tank, you send them out to lots of journalists. Some will look at them and see little news there; they will ignore the report. Others will buy it, and write the report up. So the only stories you see about the report are from journalists who buy into its thesis. That’s a bias right there. And always linking to the report is one good way of helping readers and news organizations overcome that bias.