I like David Frum. I agree with him 90-plus percent of the time. But I really disagree with the implications of his recent post on the financial crisis:
By and large, the former smartest guys have not paid much of a price for their ignorance. None of those who contributed to the financial crisis has been prosecuted nobody has even been much disgraced.
The complaint that nobody went to jail has been a trope of the fever swamps of the left netroots. It's also part and parcel with the broader problem of criminalizing agency costs.
I wrote about this problem some years ago in a column for what was then TCS Daily, explaining that:
"Criminalizing corporate agency costs? That's just a fancy way of saying that the government has misused criminal laws in this case to punish questionable business transactions and bad business decisions." ...
"Agency costs" are a central concept of modern corporate governance and finance. These costs arise because public corporations are characterized by a separation of ownership and control: The firm's nominal owners, the shareholders, exercise virtually no control over either day to day operations or long-term policy. Instead, control is vested in the hands of professional managers, who typically own only a small portion of the firm's shares.
As Professors Berle and Means observed in their landmark book The Modern Corporation and Private Property: "The separation of ownership from control produces a condition where the interests of owner and of ultimate manager may, and often do, diverge . . . ."
Economists Michael Jensen and William Meckling coined the term "agency costs" to refer to the consequences of Berle and Means' divergences. Jensen and Meckling defined these costs as the sum of the monitoring and bonding costs, plus any residual loss, incurred to prevent shirking by agents. In turn, shirking is conventionally defined to include as any action by a member of a production team that diverges from the interests of the team as a whole. As such, shirking includes not only culpable cheating, but also negligence, oversight, incapacity, and even honest mistakes.
A sole proprietorship with no agents will internalize all costs of shirking, because the proprietor's optimal trade-off between labor and leisure is, by definition, the same as the firm's optimal trade-off. Agents of a firm, however, will not internalize all of the costs of shirking: the principal reaps part of the value of hard work by the agent, but the agent receives all of the value of shirking.
Economists Armen Alchian and Harold Demsetz offered the useful example of two workers who jointly lift heavy boxes into a truck. The marginal productivity of each worker is difficult to measure and their joint output cannot be separated easily into individual components. In such situations, obtaining information about a team member's productivity and appropriately rewarding each team member are very difficult and costly. In the absence of such information, however, the disutility of labor gives each team member an incentive to shirk because the individual's reward is unlikely to be closely related to conscientiousness.
Accordingly, an essential economic function of management is monitoring the various inputs into the team effort: management meters the marginal productivity of each team member and then takes steps to reduce shirking.
The process just described, of course, raises a new question: who will monitor the monitors? In any organization, one must have some ultimate monitor who has sufficient incentives to ensure firm productivity without himself having to be monitored. Otherwise, one ends up with a never ending series of monitors monitoring lower level monitors. Alchian and Demsetz solved this dilemma by consolidating the roles of ultimate monitor and residual claimant. According to Alchian and Demsetz, if the constituent entitled to the firm's residual income is given final monitoring authority, he is encouraged to detect and punish shirking by the firm's other inputs because his reward will vary exactly with his success as a monitor.
Unfortunately, this elegant theory breaks down precisely where it would be most useful. Because of the separation of ownership and control, it simply does not describe the modern publicly held corporation. As the corporation's residual claimants, the shareholders should act as the firm's ultimate monitors. But while the law provides shareholders with some enforcement and electoral rights, these are reserved for fairly extraordinary situations. In general, shareholders of public corporation have neither the legal right, the practical ability, nor the desire to exercise the kind of control necessary for meaningful monitoring of the corporation's agents.
How then do we deter agents of a public corporation from shirking? Economic theory tells us that actors are deterred when the expected punishment exceeds the expected benefits of committing the crime. In turn, the expected punishment is determined by multiplying the nominal sanction (the penalty imposed if convicted) times the probability of being caught.
An example might be helpful. I know from painful personal experience that the nominal penalty for speeding on Sunset Boulevard in West Hollywood is a $271 fine, plus the cost and opportunity costs associated with going to traffic school to keep the citation off my record. Because the odds of getting caught are very low, however, the expected sanction probably is just a few pennies. Being a rational actor, I therefore speed.
As I explained in an earlier TCS column, "Crime? And Punishment?", however, the criminalization of bad corporate governance has increased the expected sanctions faced by corporate executives. The probability of conviction has gone up because ambitious prosecutors can bring to bear the full panoply of coercive measures available only in criminal litigation, while the nominal sanction has gone up dramatically. All of which raises the question of whether we have cranked up the expected sanction to the point of over-deterrence.
The problem is that business decisions rarely involve black-and-white issues; instead, they typically involve prudential judgments among a number of plausible alternatives. Given the vagaries of business, moreover, even carefully made choices among such alternatives may turn out badly.
If prosecutors, judges, and juries are unable to distinguish between competent risk taking and criminal mismanagement, however, the threat of criminal sanctions may discourage managers from taking risks.
Yet, risk] taking is precisely what shareholders want managers to do. As the corporation's residual claimants, shareholders don't get paid until all other claims on the corporation are satisfied. Because risk and return are positively correlated, the high returns on corporate investment necessary to leave something for the shareholders require high risks.
The problem of criminalizing agency costs should now be apparent. Obviously, shareholders want managers who steal from the company or defraud investors to be punished. Yet, they also want managers to take risks. If prosecutors and juries can't tell the difference between the two, however, criminal sanctions may leave shareholders worse off by making managers more risk averse.
Mistakes were made. Massive mistakes so large that to call them a mistake seems wholly inadequate. But hard cases cannot be allowed to make bad law. Using the financial crisis to justify the further criminalization of agency costs is just plain bad policy.
