Note: This post first appeared on my Patreon page.
Last week Roger Lowenstein had a piece in the Post about GE’s hiring of a new CEO after the prior one served less than a year. According to Lowenstein, the new CEO’s contract will give him incentives worth $300 million over the next four years if he does well by the shareholders. He will walk away with $75 million if he does poorly. This follows the hiring of an inept CEO who was dumped in less than a year and long-term CEO Jeffrey Immelt, who pocketed hundreds of millions of dollars during his tenure while giving shareholders returns averaging 1.0 percent annually, according to Lowenstein.
This raises the obvious question: What is GE’s board is doing? I haven’t looked at their forms, but I am quite certain these people get paid well over $100k a year and quite possibly over $200k for a job that requires perhaps 200 to 300 hours a year of work. That comes to an hourly pay rate in the $300 to $1,000 range. The primary responsibility of directors is picking top management and making sure that they don’t rip off the shareholders.
How could you possibly fail worse in this job than GE’s board? Yet, my guess is that there has been very little turnover in the board.
As a practical matter, it is difficult for shareholders, even large shareholders, to organize to remove board members. More than 99.0 percent of the incumbents who are nominated by the board for re-election win.
This is the classic problem of collective action. It is almost always much easier to simply exit as a shareholder and sell your stock than to organize and try to change the way the company operates. For this reason, CEOs are able to make out like bandits, getting pay in the tens of millions of dollars, even when they do poorly by shareholders.
It is common for progressives to condemn the outrageous pay of CEOs. However, they rarely move beyond condemnation to point out that the CEOs are ripping off their companies. This means first and foremost the shareholders.
If a CEO gets paid $20 or $30 million, but does not produce returns for shareholders that are at least this large, or perhaps more importantly, doesn’t add, say $15 or $25 million, to what a CEO getting $5 million would produce for shareholders, then the company is being ripped off by the CEO. This means that the shareholders would benefit if they could organize to reduce CEO pay.
There is considerable research showing that CEOs are not worth their pay, much of which is cited in Lucian Bebchuk and Jesse Fried’s excellent book, Pay Without Performance. There are any number of studies showing, for example, that CEOs get richly rewarded for events they had nothing to do with, like higher profits at an oil company due to a jump in world oil prices.
It is possible to prevent such windfalls, for example by indexing compensation to relative performance. That would mean an oil company CEO gets paid based on how the company’s stock does relative to other oil companies, not just whether the company’s stock goes up. CEO pay is almost never structured this way.
It is also worth noting that incentives are always one way. CEOs get extra pay when the company does well, but they never have money deducted from their base pay when it does poorly. This can mean, for example, that a CEO may do very well in three years when the company’s stock price goes up, but they never have to give back their pay if the stock tanks in the next two years. Again, it is possible to write contracts that would require givebacks and penalties for poor performance, but it is almost never done.
Jessica Schieder and I did a short paper earlier this year that looked at whether the limit on the tax deductibility of CEO pay in the health insurance industry imposed by the Affordable Care Act (ACA), had any impact on their pay. The ACA prevented insurers from deducting more than $500,000 of a CEO’s pay from their profits for tax purposes. This meant that instead of CEO pay costing the firm 65 cents on the dollar (the tax rate was 35 percent at the time), it cost them 100 cents on the dollar, effectively raising the cost of a marginal dollar of CEO pay by more than 50 percent.
If health insurers are setting the pay equal to the value the CEO adds to the shareholders, the increased cost of pay to the company from the loss of tax deductibility should have unambiguously had the effect of lowering CEO pay, after controlling for other factors. We ran a large number of regressions, controlling for increases in revenues, profits, share prices, and other factors that could plausibly affect pay.
In none of them did we find any evidence that CEO pay in the health insurance industry had been lowered by this provision in the ACA. This would seem to support the view that CEO pay does not bear any relationship to the returns CEOs produce for shareholders.
If CEOs are ripping off shareholders, then we should look to shareholders as allies in reducing CEO pay rather than as actively colluding in exorbitant pay. This isn’t a question of doing justice by shareholders. I am well aware of the enormous skewing of stock ownership, although there are many more non-rich people who would benefit from lower CEO pay than there are non-rich CEOs who benefit from higher pay. Most middle-income people do own some stock in their retirement plans. And, we still have tens of millions of people enrolled in traditional defined-benefit pension plans that hold stock.
More importantly, CEO pay helps to set pay structures throughout the economy. In a context where CEOs of large companies routinely earn $20 or $30 million a year, the heads of large non-profits, such as foundations, charities, and universities can typically command compensation of more than a $1 million a year. They can truthfully say that they would be earning far more if they were running a private company of the same size.
And, excessive CEO pay affects their immediate subordinates. If the CEO is getting paid $20 million, then the chief financial officer and other top executives might be getting in the neighborhood of $10 million. The third level of executives could still be crossing the $1 million threshold. As fans of arithmetic everywhere know, the more money that goes to those at the top, the less is available for everyone else.
Imagine that run of the mill CEOs got $2–$3 million and the really outstanding ones (in producing returns for shareholders) got $4–$5 million. In this scenario, the chief financial officer probably gets a bit more than $1 million, and the third tier of executives are looking at pay in the high six figures. This leaves a lot more money for everyone else.
In my better world, we would limit the top pay for anyone at non-profits to the president’s salary $400,000 a year. (They lose their non-profit status if they pay more.) This means that provosts, vice-presidents, deans and other administrators would get less. This would leave much more money for faculty and staff.
The key step in my view is to find ways to make it easier for shareholders to rein in CEO pay. There was a very small step in the Dodd–Frank financial reform bill. It required a “say on pay” vote by shareholders every three years in which they vote up or down on CEO pay. The vote is non-binding. As a result, there is generally little organizing around it and more than 97 percent of pay packages are approved.
I have argued for putting some teeth in the vote. Suppose directors sacrificed their own pay if a CEO pay package was voted down. This would give them some real incentive to think carefully about whether they could get away with paying their CEO less money or whether they could get another CEO who was as good for half the price.
The neat thing about this sort of measure is that it could be adopted at the state level, where states imposed it on companies incorporated within the state. Companies could, of course, change their state of incorporation, but this matters little to the state. They get little revenue as a result of companies being incorporated in the state. Also, it might be an embarrassment to the company to say that they have to change their state of incorporation because the directors are worried about losing a say on pay vote.
In fact, shareholders could push companies to adopt this rule voluntarily. After all, how many directors want to argue that they are worried about being in the bottom 3.0 percent of corporate boards?
I realize the power of inertia, so nothing in politics is ever as easy as it should be. But we do have good cause to worry about CEO pay, which is grossly out of line both with its past levels and with respect to pay in other countries. And rich shareholders are an ally in this story.
There also is another aspect to the story that has largely escaped notice. If we go back forty years, most of the shareholders’ return came from dividends. These were typically in the range of 3–4 percent annually. In the last four decades, there has been a near doubling of the price-to-earnings ratio. This run-up has allowed stocks to offer acceptable returns (although less than in the 1947 to 1973 Golden Age), even as the dividend yield has fallen sharply.
The problem in this story is that the run-up is unlikely to continue. We are not likely to see price-to-earnings ratios of 40 or 50 to one. If stock prices just rise in step with profit growth, which in turn is likely to be roughly the rate of economic growth, then stock returns will be much lower going forward than in the past.
This also means that CEO pay that depends on rising stock prices will be much lower. If CEOs are still going to pocket $20 to $30 million a year, then companies will increasingly be forced to pay them directly rather than through grants of stock options. This will make the cost to shareholders more apparent. That may still not be enough to reverse the massive run-up in CEO pay over the last four decades, but there is always hope.