Is sunlight always the best disinfectant?

Ever since Brandeis turned the neat phrase “sunlight is the best disinfectant; electric light the best policeman” in praise of transparency, it has been virtually accepted wisdom in legal and policy making circles.  But the efficiency and efficacy of “sunlight” is not always so obvious.

Consider two examples raised by new SEC Chair Christopher Cox.  Cox said that two of the SEC’s priorities during his tenure will be requiring better disclosure of executive compensation and enforcing the new SEC registration rule for hedge funds, which is intended to help investors get more disclosure from funds.  Both of these are, at best, a misallocation of the SEC’s scarce resources, and, at worst, flatly wrongheaded.

On executive compensation, the SEC’s concern mimics recent media and academic hand wringing about levels and types of CEO pay.  The SEC apparently has bought the argument of academics Lucian Bebchuk (Harvard) and Jesse Fried (Boalt), that CEOs use their power over boards and shareholders to camouflage the true amounts of their compensation.  (My forthcoming paper on executive compensation of firms in financial distress shows that the managerial power theory of Bebchuk & Fried is wrong or incomplete in that employment contracts written in low agency cost environments (distressed firms with few owners) are the same in amount and form as those written in higher agency cost environments (healthy, public firms).  In other words, something more than managerial power must explain current compensation practices.)   To prevent CEOs from hiding the true nature of their pay, Cox wants “shareholders [to] have one number, [with] . . . the different kinds of executive compensation add[ing] up to a number that's comparable, executive to executive and company to company.”

While this seems unobjectionable at first blush, there is absolutely no evidence that it will be effective, while it will impose real costs on shareholders. Putting aside issues of whether shareholders actually want such disclosure  (if they do, why aren’t companies providing it already?) and whether the market is defective as Bebchuk & Fried suggest (and my paper refutes), the idea that disclosure is a panacea has already been discredited.   During the last epoch of executive compensation hysteria in the early 1990s (before the stock market boom made the debate seem quaint), President Clinton signed legislation amending the tax code to require, among other things, increased disclosure of compensation.  Proponents dragged out Brandeis’s old chestnut in defense of transparency, and the expected result was for shareholders to rise up and demand less compensation. 

What actually happened was exactly the opposite.  Shareholders remained passive investors, as the economic models predict, and managers’ compensation increased.  One explanation for the increase was that CEOs, by nature a competitive lot, had more data to compare themselves to their competitors.  If Susie running XYZ Corp. is making $5 million per year, then Harry running ABC Corp. figures he deserves at least that much.  Disclosure in this case led boards to set compensation levels using an “above industry-average” methodology, which had the effect of ratcheting up all compensation levels – after all, no one wants to be below average, and firms looking to attract talent certainly won’t pay below average.  Welcome to Lake Wobegon.  Without the facts about what other CEOs make, boards would be forced to use other, internal calculations about CEO worth when negotiating about the market-clearing price for CEO talent.  Instead of developing these techniques, executive compensation consultants and boards took the easy route provided by the disclosure.  In this state of the world, providing a single number, assuming that this can be done in a meaningful way at a reasonable cost (something I doubt), may only make this tendency worse.  The SEC, an “expert “ agency that is supposed to make empirical, fact-based decisions, should consider the potential unintended consequences of disclosure in a market like this before it acts.

The wisdom of more disclosure in the hedge fund industry is also questionable for several reasons.  First, investors apparently aren’t bothered by the amount of disclosure currently provided by firms.  Since 2000, investors have poured over $1 trillion into more than 10,000 new hedge funds, growing the industry to over $1.5 trillion.  If investors were spooked by the amount of disclosure these new funds or the existing funds are providing, we wouldn’t see the money flowing in.  (One potential counterargument here is that the SEC senses a bubble in the making, and is trying to burst it or prevent it from growing out of control.  Evidence of this might be the fact that in the past year assets under management grew almost 40 percent while the industry beat the S&P index by less than 2 percent.  If this is what the SEC is up to, and it isn’t at all clear, there are grounds for caution.  The SEC’s onerous disclosure requirements didn’t prevent the Internet bubble from forming or bursting, and thousands of investors lost billions of dollars notwithstanding the fact that Pets.com and its ilk complied with every SEC disclosure regulation.)    

Second, as recent survey data from Hedge Fund Manager magazine shows, about 80 percent of hedge funds provide investors with data about performance and other non-proprietary characteristics of the fund.  But shouldn’t we be concerned about the 20 percent of firms that aren’t providing such data?  Not really.  We would be concerned if firms that are likely to commit fraud or meltdown are those not providing disclosure, but the survey evidence shows that new, small firms--the ones most likely to be frauds--are forced to provide disclosure to attract business, while the (marginally less risky) large, well-known firms with strong performance rely more on word-of-mouth reputation.  This is a market working as it should, with firms basing disclosure decisions based on cost-benefit analysis, and with reputation playing a strong role.   

Third, current regulations keep the pool of investors limited to sophisticated investors who, under the SEC's own precedents and governing philosophy, don’t need the protection of the securities laws.  At current, only qualified investors with greater than $5 million in net worth (about 200,000 lucky Americans) and institutional investors (mutual funds and pension funds that the rest of us hold) can invest in hedge funds.  It is difficult to argue that these investors need the help of the SEC to get more information about hedge funds, especially when the SEC admits that it doesn’t fully understand the industry, and is planning to use the new regulations to get up to speed with the investment community.

Finally, the recent enforcement actions against fraudulent funds like the Bayou Group, Rocker Partners, and Pontus show that the government is able to punish and deter wrongdoing through good old-fashioned fraud investigations based on tips from insiders, disgruntled investors, and enterprising analysts and journalists.  According to industry experts, these successes have already increased demands within the industry for more disclosure and self-policing.  While we should expect an increased awareness of the risk of fraud to lead to more regulation, these successes show that more enforcement of existing laws may be a more efficient result.

So while Brandeis was definitely on to something, it isn’t clear that the benefits of disclosure always outweigh the costs.  As Richard Epstein says about publicly financed sports stadiums, the costs are real and the benefits often illusory.  So too with disclosure.

Comments

[Paul Gowder]

As to executive compensation, what are these costs which you declare to be "real?" You only cite one cost to increased disclosure, and that's increased competition between corporations over CEO salaries. Other than that, the only problems you cite with increased disclosure requirements are questions about whether those requirements would be effective. That's as may be, but they land on the benefit side of the equation rather than the cost side of the equation.

As for the one cost you do cite, it seems both unrealistic and not necessarily problematic. Realism: is it really believable that CEOs and corporate boards of major corporations will not, in the absence of mandatory disclosure, know approximately what other corporations are paying their CEOs? The CEOs, in particular, have an incentive to obtain that information if it can be used to browbeat their boards into raising their salaries. Problem: isn't it good to permit compensation decisions to be made in a free market with minimal information costs? I'm hardly a capitalist, but it seems strange for someone defending the market against government intervention to insist that it's a bad thing when corporations are forced to compete with each other in an open market for talent.

I don't see how the hedge fund argument you make works at all. Each of your four points is a snapshot of various data suggesting that the current system is working well. Be that as it may, that doesn't raise any questions about either the costs or the benefits of the proposed changes. If I eat a chocolate bar, I am happy. But maybe if I eat two chocolate bars, I'll be more happy. Similarly, investors may be happy with current levels of disclosure. But they may be more happy with increased disclosure, and nothing in your post suggests a way we could measure that happiness and compare it against whatever costs there might maybe be.

[toddhenderson]

Responding to Paul Gowder. You are right, I don't focus on costs, as they appear to be obvious. Take the hedge fund example. The SEC estimated that funds would have to expend over 300 hours per year on compliance with the regulations, and reasonable estimates of costs exceed $1 million per year. In fact, funds are required to hire Chief Compliance Officers, a position that funds expect to pay over $1 million in salary alone. Even assuming relatively modest costs of compliance, funds may be forced to raise management fees by 20-30%, which will deter entry by new funds on the margin. This doesn't even include the opportunity costs of compliance (e.g., management attention), and the increased risk of litigation that comes with regulatory oversight and disclosure. For evidence of the costs of disclosure, ask anyone who works at a public company or any lawyer or investment banker about the costs of compliance. It is so large and so obvious that I didn't spend time harping on it.

One the executive compensation side, it is true that one would expect the market to work better with more information. But what if the information is not entirely accurate or is misleading or is incomplete, as the single number Cox wants will be? Don't get me wrong, I'm a believer in a functional market for CEO talent. My point is that disclosure won't result in a decrease in compensation levels. It will leave them unchanged, assuming a well-functioning market, and may put pressure on them to rise on the margin, assuming that there may be some perverse impacts from make compensation levels more salient and encouraging even more comparison-based assessments instead of internal, shareholder value-focused assessments.

[Commenterlein]

The argument that observed CEO compensation contracts and levels are not a form of rent extraction because you observe similar contracts and levels in situations in which CEOs have little or no power is flawed. Firms compete in a common (or at most mildly segmented) market for CEO talent, and hence the only feasibly equilibrium outcome is that all contracts pay more or less the same and provide more or less the same level of benefits and protections (adjusted for firm characteristics like size). This common level of pay could still be way in excess of what would obtain in a world without CEO power.

To be clear: Bebchuck and Fried have not proven that CEO power is a major driver behind CEO compensatin levels, but neither have you proven their claim wrong.

[Paul Gowder]

In terms of the hedge funds, I think I must be missing something. Maybe I don't know enough about how hedge funds operate. 300 hours a year is a significant cost? How many people do those hedge funds usually employ? In most businesses, 0.15 FTE, which is what 300 hours a year comes down to, isn't the kind of cost that makes one bleed from one's hours.

As for the hiring of a chief compliance officer, perhaps that is a flaw in the SEC regulations, but it doesn't necessarily mean that they should be thrown out altogether. It seems a little bizarre to require the hiring of an officer for a function that requires only 300 hours a year. (Frankly, I'd sell my soul for a million dollar a year job that required me to work 6 hours a week. Can lawyers be the Chief Compliance Officer? Hell, if you know any hedge fund managers, you tell them I'll do it for only $500,000/year! HALF market rate! And I'll pay for my own medical insurance out of that!)

So perhaps the Chief Compliance Officer position should be scrapped, or companies should be allowed to combine it with another job that already exists in house (general counsel, CFO, etc.). Or perhaps they should be allowed to contract it out to a firm of lawyers and accountants at a mid-upmarket $400/hour (for lawyers, that's probably way over for accountants) for $120,000/year -- a far cry from a million bucks a year.

That 300 hour figure can't be right, or if it is, the regulation can easily be fixed by wiping out the CCO position.

[toddhenderson]

These are great comments. Keep it up!

In response to Commenterlein: Why don't you wait to read the entire paper before you discount it? As to your specific objection, there are a couple of avenues I explore in the paper along these lines. To give just one example, let's assume that the market for CEO talent is fairly homogeneous, that the market for CEO talent is the same for healthy and distressed firms (something explored in the paper), and the market-clearing wage for a CEO is $5 million per year in total compensation. Firms can choose how to deliver the utility of $5 million per year -- Firm A can give all cash equivalent to $5 million; Firm B can give cash plus non-indexed stock options, which when discounted for risk and portfolio effects equals $5 million in cash (N.B., it is likely to appear to be more than $5 million); Firm C could give cash plus restricted stock; Firm D could give cash plus indexed stock options, etc. Firms will use a mix of cash, options, loans (until SOX banned them), perks, stock grants, retirement benefits and so on. And we observe in the market firms of all shapes and sizes innovating about how to deliver the market wage. B&F argue that many of these mechanisms are used only because managers have power and use them to camouflage their "real" compensation. But managerial power is greatly reduced when diffuse shareholders are replaced with sophisticated vulture investors and banks holding huge stakes in the firm. These new owners don't suffer from the agency problems B&F identify as the source of CEO power, and as a matter of course rewrite CEO contracts, with judicial and creditors' committee oversight. So you may be correct that the firm still must deliver $5 million, but these owners can choose to alter the way it is delivered. For example, B&F say that the use of non-indexed stock options is explained by CEO power. If so, then why do employment contracts still use non-indexed stock options when CEOs are weaker (i.e., they don't dominate the board or control the compensation consultant)? Firms in these circumstances could say to the CEO, we will pay you the going rate, but we are going to use indexed stock options, which is will more closely link your pay with shareholder performance. They don't, and that is telling.

The paper also presents various arguments and data about whether the market for CEO talent is in fact the same for healthy and distressed firms, as well as several other arguments along these lines. But this is a great comment, and one that I get frequently when I present the paper. It surely is the key question. Thanks.

[toddhenderson]

Responding to Paul Gowder. Yes, this is a bit confusing. The 300 hours figure is an SEC estimate, and is way too low. The figures about CCO wages are actual numbers taken from the behavior of funds in the market. So the SEC way underestimates compliance costs. When the risk for funds is going out of business or going to jail, they will over-spend on compliance.

As for the 300 hours figure, funds are leanly staffed. A few guys (and gals) can operate a multi-billion dollar fund out of a single office or house. There is almost no overhead, and the investment managers usually wear many hats. So this is a non-trivial time expenditure (even if correct), and will likely eat up investment manager time.

[Commenterlein]

Todd,
Fair points, thank you for your detailed response. Before replying further I will indeed read the paper.