In the years since the financial crisis of 2008, there have been periods of outrage every time there is news of a banker being paid a large bonus. How could those who seemingly caused the crisis continue to be paid exorbitant sums, especially when their actions have led to such huge losses for others?
Bonuses are built into their contracts. That is, bankers are simply being paid what has been agreed beforehand. So the real question is, how could bankers be given employment contracts that involve large bonuses despite adverse performance? Is this wrong? If so, how do we prevent this from happening again? A study of the work of this year’s Nobel laureates in economics might prove instructive. The > 2016 Prize has been jointly awarded to Prof. Oliver Hart of Harvard University and Prof. Bengt Holmstrom of the Massachusetts Institute of Technology. They have been felicitated for their contributions to contract theory.
The optimal contract Contracts are legally binding agreements governing who will do something in exchange for something under some circumstances. Employment contracts are best seen in the context of principal-agent problems.
In this case, shareholders of a bank (the owners, or principal) employ bankers (agents) to run the bank. Shareholders are interested in returns. However, it is the banker who is responsible for the day-to-day decisions that generate those returns. In turn, the banker is to be compensated for his or her services.
Typically, the principal cannot observe the agent’s effort. If agents are offered a fixed salary regardless of effort, there is not much incentive for hard work. We see this in government jobs. Thus, in many industries, including banking, we instead see performance-based pay, where better outcomes for the shareholder (greater returns) are to be rewarded with better compensation for the banker.
There are other tricky aspects too. The owners might be keen on long-term performance, but if bankers’ pay is tied into the short-term share price, this might generate a conflict of interest. Bankers might work to boost the immediate share price at the cost of long-term investment.
In other contexts, the outcome of the efforts of the agent might not be measurable. Teachers, for example, are instrumental in the overall development of students, which is hard to measure. However, scores on tests are easy to measure. A contract that links teachers’ pay to only measurable outcomes might lead to teachers focusing on tests alone, ignoring more general learning.
What is an optimal contract under such circumstances? First, the bad news: Professors Hart and Holmstrom established that there can be no perfect contract. However, their large body of work showed ways to improve outcomes.
Lessons from Laureates In a seminal co-authored work in 1979, Prof. Holmstrom argued that if agents are paid based on the share price of the company alone, rises and falls in their pay might actually be determined by dumb luck. Instead, their pay should also be linked to external outcomes (to the relative share price against other similar companies, for example). This leads to better measures of relative performance, and therefore effort.
Of course, this is when effort is measurable, and can be linked to performance — both dubious propositions in the case of banking. In many other industries, though, such as manufacturing, it works.
In any event, the implication of this result is immediate: the harder it is to observe the manager’s effort, the less the manager’s pay should be based on performance. In particular, in high-risk jobs, payment should skew in favour of a fixed salary, while stable jobs should attract performance-based pay. Wall Street’s signature achievement was in persuading regulators that its bankers were in a low-risk environment, with contracts designed accordingly, when in truth they were gambling with vast sums of borrowed money. No wonder the incentives were so spectacularly misaligned.
Often, however, parties cannot specify detailed contract terms in advance. Nobody knows what will happen, and so outcomes cannot be specified. This is the problem of incomplete contracts, and Prof. Hart’s biggest contributions lie in this area, also with relevance to financial markets.
In several co-authored papers, Prof. Hart argued that a contract that cannot explicitly specify what the parties should do in future eventualities must instead specify who has the right to decide what to do when the parties cannot agree. Depending on the eventuality that arises, the party with the corresponding decision rights has more bargaining power, and can secure a better outcome for itself. Decision rights thus substitute for performance-based pay.
The upshot of Prof. Hart’s work, particularly as it relates to bankers, is that bankers should hold decision rights when performance is good, but the rights should favour shareholders when performance deteriorates. Crucially, a Hart contract should see banker pay slashed when outcomes are adverse. With great upside should come at least some downside.
But while shareholders in American banks have belatedly woken up, and banker bonuses have occasionally been reduced, punishment is still rare. Compensation almost never decreases. And in any case, it is too late to modify existing contracts.
In other work, Prof. Holmstrom studied the trade-off between complete insurance contracts and the resulting moral hazard. A car owner who knows that she will receive a full payout in case of damage might not be as careful as she would be if she were liable. Socially optimal contracts ensure that the car owners too have skin in the game, as it were, by being made to pay a part of the cost.
Another application of Prof. Hart’s theory of incomplete contracts is the question of whether providers of public services, such as schools, hospitals, and prisons, should be privately owned or not. Such situations lead to a trade-off between investments in quality and cost reduction.
Prof. Hart’s work suggested that the private sector might focus too much on cost reduction, but proper contracts can provide incentives for quality as well. The public sector, on the other hand, has little incentive to do either.
Several of Prof. Hart and Prof. Holmstrom’s insights have indeed found their way into modern contracts. But the example of Wall Street suggests that regulators must pay more attention to their findings, particularly in high-risk industries.
But it also begs the question: what is an optimal contract for a regulator? If we are lucky, we won’t have to wait until the next crisis to find out.
Madhav Raghavan is a PhD in quantitative economics from the Indian Statistical Institute, New Delhi, and specialises in markets for education and employment.
