Trader compensation has been on the front pages this year. The issue was also part of last month’s G20 agenda. The question is: What kind of compensation doesn’t encourage excessive risk-taking? Reflecting the concern for avoiding another systemic financial crisis through the development of good practices and regulations, various studies have looked at gaps in compensation and the sharing of value added.
We will examine each of these issues and offer analyses from the perspective of the “OTC Conseil” interpretative framework – in other words, in terms of risk / return. We help companies in organize their business. An essential feature is decentralization mechanisms whereby executives’ targets are implemented by operational teams. Incentives and bonuses are tools for such mechanisms. In this month's newsletter we will take a closer look at how this works, starting with a necessary reconsideration of value-added sharing.

Several recent, quite convergent studies^₁ note a relative stability in the distribution of value added between compensation and profit over the last few years. Still, they emphasize growing gaps in compensation and an increasingly significant share of value added going to the best-paid employees.

The studies also show a growing portion of dividends in companies’ distributable income. Company recapitalization over the last few years has required such growth in order to maintain the same level of return. These tendencies are the by-products of the strong drop in nominal interest rates and the debt reduction that has followed.

In addition, the studies observe that the mechanisms for employee profit-sharing in company performance have taken a greater place in profit distribution, even though they remain relatively weak.

On growing wage inequality, the reports focus on the need to bring bonuses under control in the world of finance and on executive compensation. They emphasize in particular the need to base variable compensation on performance measured “on the long run.”

At the end of August, the French government and major French banking establishments proposed changing compensation by spreading out bonus payments over time, and by making future payments conditional on the confirmation of recorded earnings. However, these banks’ competitiveness would suffer should such compensation schemes be put in place if the same isn’t done worldwide, since they would inhibit the banks’ ability to hire and retain the most successful traders.

The G20 softened regulation temptations about bonuses but confirmed in September the “over-time” logic, and a need for mastering bonuses.

Indeed, “trader compensation” should not only be considered in light of the concern over equitable (and socially sustainable) distribution of revenue, but above all in view of the incentive it offers to excessive risk-taking, the latter often thought to be at the root of the 2008 financial crisis.

The studies thus tackle two rather different themes. From our point of view, both merit attention.

> The question of the distribution of value added (including growth in the proportion of high salaries) should include considerations bearing on the distribution of risk and compensation.

> As for the latter, executive and trader compensation turns on the question of incentive mechanisms. The old “principal-agent” theories of running a business don’t exactly fit, notably when it comes to the world of traders. Advances arising from behavioral finance have somewhat revived work on these issues dating from the 1970s.

As we shall see, however, at the heart of the incentives issue is the question of Value Added.

¹ AMF Report on compensation of the executive directors of listed companies and on the implementation of the French Association of Listed Companies (AFEP) / French Business Confederation (MEDEF) recommendations (July 2009); Information Report on compensation of executive directors and market traders – French National Assembly (July 2009); Distribution of value added and profits and the pay gap in France – Report to the President of the Republic (May 2009); Sharing the Fruits of Growth in France – Report of the French Council of Economic Analysis to the Prime Minister (May 2009).

Value added and Risk
__________________

The aggregates of national accounting are derived from a universe in which risk is not taken into account. There certainly exists the notion of return on (risk) capital relative to return on non-risk capital (debt interest). However, Value Added (GDP) does not take into account an economy’s improvement in terms of risk. This aggregate is regularly criticized because it also fails to measure gains in “quality” of life or environmental impact. We should emphasize here that, beyond passing over the question of risk, it further neglects other cumulative effects, including human capital.

Some will say that it is new wealth that “counts” and that value added is the measure of it. We believe that in this case the entire intertemporal dimension of the economy is neglected.

Value added of the financial sector

Promoting future growth today is indeed an immediate “enrichment.” Whether future growth is actually higher or lower – this will depend on other, unknown factors – does nothing to change the fact that action has been taken today and thus “added” to value. Diversifying risks, covering them, transferring them intertemporally are therefore actions that, in our opinion, are sources of value added. Making financing of the economy possible (starting with the transformation risk taken on by banks) is part of a set of analogous actions.

This epistemological question is not without consequence to the perspective from which one considers the banking and insurance industries. The standard calculation of the value added of their activity takes absolutely no account of their contribution to the economy. It is clear that implementing a complex derivative (because it transfers risk among different agents) will permit investments that otherwise would not be made; securitization (assuming that it is not implemented as thoughtlessly as it has been over the last few years) by way of risk dispersion allows for greater risk taking² ; even the trader, which theory takes as the foundation for market efficiency, must, in our opinion, be considered in terms of his or her value added.

Risk distribution and income distribution

More generally, analysis of the distribution of value added ought to be carried out specifically on the issue of risk. While this is not the place to enter a detailed discussion, taking into account human capital and employee risk in a world in which the job market is “flexible” makes less relevant those historical analyses that separate capital bearing risk and work secured by employees. Profit-sharing mechanisms, because they fail to diversify the risk borne by employees, must be prudently considered.

Finally, recapitalization of companies should have brought with it a decrease in return on financial capital through a decline in the risk premium. The growth in dividends doesn’t bear that out.

² That systemic risk increases correlatively does not preclude finding the right, “optimal,” measurement for the whole economy.

Incentivizing risk-taking
___________________

In the 1970s, the “principal-agent” or “center-periphery” theory was aimed at modeling information – or criteria – asymmetries between shareholders and company executives. It also applied to decentralization mechanisms within companies, when employee performance is not directly observable. The models describe the form of incentive mechanisms to implement, leading the company to a second-level optimum³.

The models take into account the existence of uncertainties but are not fully adapted to decentralization and management in a risky environment. Work on the decentralization of RAROC criteria or management in risk-environments has begun to be published, without providing a simple reference model⁴.

Still, the essential issue in business is defining the “risk budget” specific to each operational unit: then local actors perform to their best, while leaving to the group level the task of measuring and controlling risks. At the end of the day, the operation of a company always hinges on mechanisms incentivizing risk-taking. Even if, of course, overall control of risks and management of these mechanisms through specifically adapted policies are necessary.

Banks and Basel II don’t seem to have found the perfect policy – which is the least that can be said – but it would be difficult to deny that the question is at the very center of the Basel Committee’s concerns. What’s the story?

³ A second-level optimum because the informational constraint (agents do not have all the information) engenders, despite the incentive mechanisms, a loss of efficiency.

⁴ "Economic capital allocation under liquidity constraints" - Fernando Mierzejewski, 2006

The abnormality of risk profiles

It is not necessarily “irrational” to play the lottery, even if one’s hope of winning is slim. The structure of winnings and probabilities is such that it is possible to make out with a considerable jackpot. As the jackpot is likely to change the winner’s life radically, it is rational for him or her to pay for the possibility (the same reasoning is behind an insurance premium and the possibility of a catastrophic event). Such considerations form the basis of experimental economics or behavioral economics and earned the psychologist Danny Kahneman a Nobel Prize in economics in 2002 and have made him a popular invitee to the White House⁵. It is therefore understandable that the pursuit of profit by traders entails risk-taking that VaR fails to measure.

That is not to say that traders “play the lottery,” but that one should accept the idea that the structure of probability and the risk profile of traders themselves are poorly suited to a theoretical representation designed according to a normal law, even if the latter is made more sophisticated. The diversification of products and traders may offer a bank a more standard risk profile, but it is better to think of “incentives” as risk limits consistent with the reality. Above all, it is advisable to reexamine the hypothesis according to which it is the abnormalities in risk profile that truly enable the creation of value within risk-transfer mechanisms (product design, sales and hedging). Kahneman’s work on prospect theory may allow us to rethink the principal-agent theory within a framework relevant to the trading floor.

⁵ The New York Times columnist David Brooks has also made his work a popular subject.

The wrong way to think about liabilities

Beyond this immediate risk profile, one should consider the specificity of the banking industry. Its production cycle is reversed. The “product” (financing as a structured product or intermediation in certain cases) is sold prior to actually being “produced.” We have examined this point elsewhere: production itself is in fact the chain of imperfect hedges which creates the possibility of a margin⁶. That being the case, the amount of the margins realized – or not – cannot as yet be known at the time bonuses are calculated. Payment of bonuses spread out over several years, with the possibility of adjusting payments of subsequent years in case of losses, is quite consistent with this reality⁷. But it is still an oversimplification, for the risk generated throughout the chain of hedges is not really measured. The issue is complex and only genuine conceptual advances – a product-centered approach within corporate and investments banks – will allow us to better define the basis for profits and each actor’s contribution to value creation.
The issues of risk and time horizon are at the heart of the need to reformulate the “principal-agent” model with the necessary convergence of the horizons of the Principal (General Management) and Agents (Front Office).

⁶ Cf. "Reinventing the economic model of Corporate and Investment Banks" study

⁷ This is what French establishments have done.


Social policy and company policy

The question of a convergence of interests also arises between front office agents and support teams. We have proposed elsewhere⁶ organizational strategies that rebalance motivation and responsibilities within corporate and investment banks. Imbalances between the front office and support teams are indeed the source of major operational risks. More generally, here we strike upon a fundamental issue for business management, connecting the accumulation of human capital and the proper functioning of internal processes.

Of course, an internal policy cannot be developed without a compensation policy, and the latter, if it is to be fair (compensation based on something like marginal productivity⁸), must also be in line with “market salaries.” The risk of overbidding between businesses always exists. One often cites the agreement signed by US professional basketball teams not to lay off their players in order to avoid wage inflation. But such practices are quite far from free-marketers’ references to the “market” practices of banks.

It is worthy to note that the distribution of bonuses over several years brings with it a certain loyalty on behalf of traders, and might in this way offer a double response to the question. However, it seems to us that the issue fundamentally remains one of determining and recognizing the right measurement of the “value added” of front officers.

If business management must reexamine “fair compensation” of employees, the same should be done of the “management” policy. From intervening in the modes of bonus payment to direct taxation, everything is currently being studied in order to move the “question of compensation” forward. To denigrate traders and their bonuses neither explains nor provides a solution to the crisis. But the role society has played – through public indebtedness – in the rescue of the financial system makes a profound examination of the financial system necessary. That the latter stands accused of having taken the real economy “hostage” should force us to direct our attention to its role, its contribution, and thus its value added within our economies •

⁸ With a larger perspective that takes into account generated risks