Quis custodiet ipsos custodies? Jean Tirole’s Proposal to Appoint Felons to Monitor CEOs
Tuesday, 28 March 2017 20:29
William K. Black February 18, 2017 Roma, Italia (5th in my series on Jean Tirole)
When in Rome, trot out a venerable Latin quotation from Juvenal: “Who will guard the guards?” I have “buried the lead” in this series of article about Jean Tirole by relegating my discussion of his proposal for fixing the problem of the criminal CEO – appoint a criminal “monitor” – to the fifth article in this series. His proposal is in his 2001 article titled “Corporate Governance.”
Tirole’s proposal for optimal monitoring of CEOs is supposed to prevent predation by CEOs against shareholders. Tirole begins his article with a catalog of some of the many ways that CEOs predate on shareholders. His list includes this passage.
They may collect private benefits by building empires, enjoying perks, or even stealing from the firm by raiding its pension fund, by paying inflated transfer prices to affiliated entities, or by engaging in insider trading [p. 1].
Tirole defines “insider trading” by the CEO as a form of “stealing from the firm.” Tirole’s proposed “natural” way to prevent the CEO from stealing from the firm through insider trading is to have the shareholders hire a “monitor” of the CEO who will be compensated through –insider trading. The insider trading by CEOs that Tirole deplores and the insider trading by the monitor that Tirole proposes both constitute felonies. Tirole refers to “insider trading rules,” so he knows that insider trading is prohibited [p. 13]. He does not, however, either note or discuss the fact that his proposal calls for shareholders to induce the monitor to commit a felony, making them co-conspirators.
Tirole’s optimal (and he thinks “natural”) method should have set off a series of warning bells in his own head. Tirole seeks to institutionalize the felony of insider trading. His euphemism for the monitor’s proceeds of his felony is a classic – “adequate incentives.”
A natural approach would be to hire a designated monitor and to provide this monitor with adequate incentives. Suppose for instance that the monitor is given at the initial stage s options at striking price equal to par…. That is, the potential monitor will be able to buy (before the final outcome is realized) s shares costing [par] each and paying dividend R each in case of success and 0 in case of failure [emphasis in original, p. 12].
Tirole then explains that the monitor and the CEO will simultaneously gain from insider trading if the CEO behaves properly.
Assume that the entrepreneur indeed behaves. The passive monitor's options are valueless if there is no monitoring. Their expected dividend is then … equal to the striking price. Suppose in contrast that the monitor incurs cost c, and thereby receives (privately) the signal. In case of a bad signal he knows that the shares are overvalued … and therefore does not exercise the options. A good signal implies an undervaluation and an expected profit … per option; so the monitor exercises the options, which reveals that he has received the good signal [p. 12]
Let me translate Tirole’s plan into English and explain its legal implications. Officers of a firm cannot “tip” other people through the release of “private” information about the firm and the person receiving private information from the officer in what they know to be a breach of the officer’s fiduciary duties cannot trade on that leaked information. Doing so constitutes insider trading, which is a felony. A conspiracy to commit a crime is itself a crime. The CEO under Tirole’s plan knows that the monitor intends to trade on private, insider information that the CEO is leaking to the monitor. Tirole’s plan is for the CEO and the monitor to conspire together to commit a crime.
The purpose of Tirole’s plan is supposed to be preventing the CEO from predating on the shareholders. Tirole’s answer to the problem of CEO predation is to add a predatory monitor to conspire with the predatory CEO.
What could possibly go wrong with Tirole’s plan?
It turns out that Tirole knows that his “optimal” plan is illegal and that it would fail and increase predation by bad CEOs against shareholders.
In practice, though, this natural way of creating passive monitoring is not frequently observed. This is perhaps due to the fact that the entrepreneur and the designated monitor have an incentive to collude. Suppose for example that the monitor commits, in exchange of a bribe, to always exercise the options. Incentives to monitor are then destroyed and so are the incentives for the entrepreneur to behave.17 (One possibility is that the bribe is paid from corporate resources (reducing the probability of success …but without any consequence for the entrepreneur, who receives compensation based on the exercise on the options) [pp. 12-13]).
“Is not frequently observed?” The true statement would be “is never observed.” Even Tirole knows that Tirole’s plan would be disastrous – and he understands only a tiny portion of the reasons it would increase CEO predation and harm the economy. Tirole’s admission that his plan would fail and make CEO predation worse does not discourage him. He lives so deep down the rabbit hole of orthodox economics that his every urge is to keep digging seeking the elusive pixie that will make laissez faire work.
A market has more integrity. Any participant in a stock market for example de facto has call (as well as put) options on the shares of the firm, in very much the same way our designated monitor had call options. But with a market (cum insider trading rules) it becomes much harder for the entrepreneur to capture the passive monitoring process. This may explain why in practice managerial compensation is based on the value of the firm's stock and thus on "anonymous passive monitoring" rather than on the exercise of options by a designated monitor. More work on the relationship between market monitoring and the "optimal collusion-proof passive monitoring scheme" is warranted, though.
Why does a “market” have more “integrity” than a monitor? Tirole’s answer is that the transaction costs of the CEO bribing many shareholders are far higher than bribing a single monitor. Tirole even drops a footnote to explain that CEOs could probably bribe the monitor very cheaply, which does not say much for the monitor. Tirole does not stop to analyze the implications for other supposed corporate monitors such as outside auditors, appraisers, and credit rating agencies. Shareholders while harder to bribe, should be vastly easier to deceive than an expert monitor with access to insider information.
Tirole, however, assumes that shareholders can “monitor” CEO performance. Worse, he assumes that the stock price is the accurate product of that monitoring and that it is therefore appropriate to base CEO compensation on the current (short-term) stock price. Remember, he is writing these things shortly after the collapse of the largest bubble in history – the stock market bubble in dot-com shares and after a series of crises driven by CEOs inflating the stock price to enrich themselves. Naïve is too weak a word to use to describe Tirole.
Tirole is also self-contradictory, for he premises his criminal insider-trading plan on the explicit assertion that shares are systematically mispriced in the markets due to the shareholders’ inability to know whether the CEO was “behaving” or “misbehaving.” He says that the reason that the monitor will profit is that the markets “undervalue” the firm’s shares if the CEO is “behaving” and “overvalue” its shares if the CEO is misbehaving. Those premises require that the stock markets are not able to discern whether the CEO is a fraud or “good” person. Two paragraphs later, he inexplicably (and implicitly) makes the opposite assumption.
It is the last sentence of the quotation above, however that is most telling about Tirole’s crippling dogmas. He asserts, without any attempt at logic or evidence that what economists need to be doing is “more work” on his “optimal collusion-proof passive monitoring scheme.” No, economists need to stop this search for the holy grail of laissez faire. There is no such thing as a “collusion-proof” scheme for CEOs. Because a “scheme” for CEOs cannot be “collusion-proof,” orthodox economists also need to stop talking about “optimal” CEO “schemes.” Precisely because there is no reliable way to prevent CEOs from colluding, there cannot be an optimal CEO “scheme.”
Assumption is the Mother of Error
Tirole’s proposal to encourage the CEO and the monitor conspire to engage in illegal insider trading implicitly assumes out of existence one of the most important drivers of CEO fraud – the magnitude of the reported profits. He relies on the explicit, but false, assumption that shareholders can “verif[y]” the firm’s true profitability. This allows Tirole to make the implicit, but false, assumption that the CEO’s reports of the firm’s profits are irrelevant because the the CEO’s reports cannot deceive shareholders. Tirole then makes the explicit, but false, assumption that if the CEO “behaves” the firm’s chance of being (truly) profitable increases. Tirole implicitly, but falsely, assumes that he need not determine whether if the CEO behaves the firm will falsely report profitability because Tirole has explicitly assumed that the CEO’s inflation of reported firm profits cannot influence shareholders. Tirole then implicitly, but falsely, assumes that the magnitude of true profits and the magnitude of the CEO’s inflation of reported profits have no effect on the monitor, the CEO, or the shareholders’ actions. The magnitude of the gain from fraud, however, is critically important to CEO and monitors under Tirole’s own model (if one inserts reality into his assumptions). Each of Tirole’s assumptions is contrary to reality and multiple fields’ literature findings.
Part of the problem is that Tirole fails to understand that what it means under orthodox economists’ twin dystopian assumptions of self-interest and rationality for a CEO to “behave” is that he will lie, cheat, steal, maim, and murder. This contradicts the definition of “behave” that humans use and it has powerful and perverse consequences for Tirole’s plan to encourage felonies in order to prevent CEO abuses that he ignores.
Three Real World Examples of the Frauds that Tirole Ignores
First, Tirole assumes that the monitor will receive $0 if the CEO sends the “bad” signal – if the monitor sends the accurate (bad) signal to the shareholders by refusing to exercise his stock options and buy the firm’s overvalued shares. Tirole’s plan, however, is premised on the monitor’s incentive to become rich by engaging in insider trading when the private information he receives is the good signal that the CEO is “behaving.” Why would the monitor not profit by insider trading when he received the private information that the CEO was “misbehaving?” The monitor simply has to short the shares – secretly so that the shareholders do not learn what he is doing.
Second, Tirole ignores what would happen if the monitor learned that the CEO was bribing EPA inspectors to ensure that the firm could get away with illegally disposing of toxic waste in a manner that would harm public health. In Tirole’s terms, the toxic CEO is “behaving” rather than “misbehaving” because it is far cheaper for the firm to pay the small bribes in order to gain a huge cost advantage over honest competitors that safely dispose of their toxic waste. The monitor would therefore send the “good” signal to shareholders by exercising his stock options. The public would realize that the stock was undervalued and would react by biding up the stock price. The CEO’s stock would surge and he would be wealthy. Tirole says that he designed his “corporate governance” proposals to help all stakeholders, but his actual proposals would encourage the CEO and the monitor to collude to predate on the public and honest competitors.
Third, what if the monitor found that the stock price had surged due to the CEO running a different kind of fraud in which he predated primarily on the creditors and shareholder by using accounting fraud to inflate the firm’s reported profits and hide its real losses and insolvency. The advantages of this form of accounting control fraud include the fact that it is a “sure thing” and it leads the firm to report record profits that will cause a sharp increase in the firm’s stock price. Under Tirole’s twin dystopian assumptions, the monitor would exercise his stock options when he discovered the CEO’s fraud. The monitor would publicly praise the firm and its CEO. The monitor would later, gradually and secretly, sell his shares in the company before it collapsed. The CEO does not need to bribe the monitor to send the “good” signal that the CEO is “behaving.” The monitor will eagerly send the false signal because exercising his stock options and covering up the CEO’s accounting control fraud are the two things he needs to do to become wealthy from sharing in the CEO’s fraudulent “sure thing.” If the monitor were to take a bribe from the CEO it would materially increase the risk that both could be prosecuted. Tacit collusion would be the optimal strategy whenever the CEO and the monitor were corruptible.
These three examples of how Tirole’s proposal to encourage the CEO and the monitor to conspire to commit the felony of insider trading illustrates would produce an epidemic of elite fraud illustrate the saying about what a “tangled web” we weave when first we practice to deceive. The Swedish Central Bank, of course, did not find it remotely disqualifying to award their “Nobel” prize to Tirole. The recipients of the world’s top prize in economics can get the economics hideously wrong and propose to make bad ethics and criminality ubiquitous in the C-suite without any negative repercussions. The thing that could be disqualifying would be for to propose effective regulation of CEO compensation to reduce the perverse incentives that are the norm