HBR Blog Network

Who Should Actually Have Say on Pay?

by Justin Fox  |   8:00 AM May 30, 2013

It's say-on-pay season at American corporations. What shareholders have been saying, in overwhelming numbers, is yes! At 74% of the 1,471 companies that have voted so far in 2013, according to Equilar's say-on-pay tracker, the "yes" percentage exceeded 90%. That's up from 69% in 2012 and 2011. Only 31 companies (2%) have gotten sub-50% no-confidence votes in 2013.

One key reason for shareholders' positive tone is that the stock market has been doing well. Since say-on-pay hit the U.S. in 2011 (it was part of the Dodd-Frank Act), academic researchers have found that the chief determinants of how shareholders vote appear to be (a) stock performance, and (b) the voting recommendations of proxy-voting advisors ISS and Glass-Lewis, which are based in part on returns to shareholders over the previous three years. To a large extent, say-on-pay — which was introduced in the UK in 2002 and has spread to several other countries, most recently Switzerland — is a simple exercise in bandwagon-following.

That's not all it is, though. The size, growth, and design of paychecks do play into both the voting recommendations from ISS and Glass Lewis and the votes of shareholders. There is evidence that say-on-pay votes have led British companies to make executive paychecks more sensitive to poor performance. Say-on-pay votes do have an impact. The question is, what kind of impact?

Say-on-pay is part of a big shift in recent years toward giving professional money managers more tools to affect the governance of (and in some cases discipline the managers of) corporations. Most of these theoretically increase the power of individual investors, too, but for the most part individuals aren't a factor in corporate elections. Professionals appear to control somewhere around 60% of the shares of American corporations, and have an even higher percentage of the vote in corporate elections. (Individual investors tend not to vote, and while brokerage firms used to vote the shares of customers who didn't get around to voting themselves — almost invariably siding with management — the SEC stopped allowing that practice three years ago.)

Driving these changes is a widespread belief that more needs to be done to hold CEOs and boards accountable. That's understandable. But it's far from clear that professional money managers have what it takes to play the role of effective watchdog. When it comes to executive pay in particular, these people are a deeply compromised bunch.

In the latest issue of the Journal of Economic Perspectives, economist Burton J. Malkiel argues that most of the gigantic growth in asset-management-industry profits since 1980 "is likely to represent a deadweight loss for investors." His reasoning, as I discussed in an earlier post, is that active money managers as a group underperform the market indices and that, while active management does play a key role in setting stock market prices, there's no evidence that today's gigantic active management industry is doing that job any better than its much smaller precursor of three decades ago. American corporations outside the financial sector may have many flaws, but I'm pretty sure their increase in profits over the past few decades hasn't been a "deadweight loss" to the economy.

What's more, the asset management industry — in particular the alternative-asset subset of hedge funds and private equity — has exported many of its pay practices into the corporate sector. The idea was to get away from paying CEOs "like bureaucrats," as Michael C. Jensen and Kevin J. Murphy urged in a famous 1990 HBR article. It was a successful campaign: CEO paychecks came to consist mostly of stock options.

This shift to financial-markets-based compensation had some of the promised impact — CEOs did become less risk-averse (bureaucrat-like) in their decision-making. But it also inflated what Mihir Desai has dubbed a "giant financial incentive bubble". In Desai's telling:

Financial markets cannot be relied upon in simple ways to evaluate and compensate individuals because they can't easily disentangle skill from luck. Widespread outsourcing of those functions to markets has skewed incentives and provided huge windfalls for individuals who now consider themselves entitled to such rewards. Until the financial-incentive bubble is popped, we can expect misallocations of financial, real, and human capital to continue.

Say-on-pay has done nothing to deflate this bubble; executive pay has kept going up in the U.S. and UK. Which makes sense — most asset managers have a shared interest with CEOs in keeping top-of-the-scale paychecks high. If we wanted to have a real impact on executive pay levels, we should probably have employees vote.

While highly paid hedge fund and mutual fund managers set the tone for the CEO-pay discussion, though, they do not as a rule get involved in the details of pay packages and say-on-pay votes. Instead, they mostly outsource the decision-making to Glass Lewis and ISS. The people who set the compensation policy guidelines at these two firms are not paid like CEOs or hedge fund managers, and lots of thought and empirical research go into their recommendations.

They have, however, bought into the argument that the main metric of executive performance should be shareholder returns and that most executive pay should be in the form of stock. They're supposed to represent shareholder interests, so this seems logical. But beyond the compensation bubble that stock-based pay has helped create, its incentive effects are also potentially perverse. As Roger Martin argued in his book Fixing the Game, stock prices are all about (often incorrect) expectations of future earnings. Linking top executives' pay to stock prices thus rewards them more for creating high expectations than for running their company well. With banks there's an even bigger problem: shareholders provide only a tiny percentage of their funding, and are thus motivated to encourage risk-taking that endangers depositors and taxpayers. So paying bank CEOs mostly in stock is a recipe for a financial crisis.

The proxy advisers do attempt to counteract these forces somewhat, by frowning upon stock and option grants that aren't linked to other performance metrics. But it's not clear that their approach yields better results. One recent study by David F. Larcker, Allan L. McCall, and Gaizka Ormazabal found that the stock market reacts negatively when companies adjust their compensation policies to adhere to the proxy advisory firms' recommendations. I'm not convinced that really proves anything one way or the other, but I do think the current state of knowledge about the impact of executive pay on corporate performance is muddled enough that standardizing pay practices to conform with what ISS and Glass Lewis think is best is probably a bad idea. Sometimes a board of directors will have a better sense than the stock market or a proxy advisory firm of how well a CEO is performing. Do we really want to make it impossible for boards to exercise discretion?

It's not that say-on-pay is necessarily a disaster. Unlike some other corporate-governance reforms, it hasn't imposed major regulatory burdens on anybody (public corporations were already holding annual shareholder votes), and for the vast majority of companies it has been a nonissue. The votes are non-binding, and there's at least a chance that they're changing pay practices for the better.

But it's worth remembering that the explosion in American executive pay over the past three decades coincided with and was in part driven by an increase in shareholder clout. It may be that shareholders just had the wrong tools in the past, and say-on-pay will allow for a more surgical approach to governing CEO compensation. It's also at least possible, though, that the shareholders have been the problem all along.