A hard look at soft dollars.

Rather than dissipating investor's wealth, soft dollars subsidize portfolio research by brokers, guarantee execution quality, and promote specialization.

Momentum is slowly building among U.S. and British regulators to restrict the freedom of money managers to accept a novel kind of brokerage rebate known as "soft dollars." The proposed restrictions are based on the mistaken belief that those who invest in mutual fund, pension funds, and other institutional portfolios need even more regulatory protection than they already have.

But further regulation will only make matters worse, most profoundly by stifling organizational innovation. Indeed, what investors may really need is protection from regulators.

SOFT DOLLAR BROKERAGE

Those who invest in institutional portfolios hire money managers to research and identify profitable portfolio trades. Once having identified a likely trade, the manager must hire an independent broker to perform the trade, since by law neither the manager nor an affiliate can do so.

The broker's charge is to work the tr; by searching for better prices and tl exucuting it at the best available pric making sure to minimize the price imp of his activity. In a typical soft dollar d a broker prepays researh expenses behalf of a money manager in propor to the future trades and trading commissions the manager promises the broker. The manager receives today research products from specialty "research originators" and - if all goes as planned - pays for it later by sending the "soft dollar broker" the promised trades.

Soft dollar use has grown considerably during the past decade, amounting by some estimates to more than $1 billion annually in the U.S. alone. This growth has been met with considerable objection from regulators, the financial press, and many of the established Wall Street brokerage houses.

NO FREE LUNCHES

What most of these observers find objectionable about soft dollars is that they "bundle" the costs of portfolio research into cash trading commissions and thereby compromise a manager's judgement. This view was recently echoed by no less a journalistic heavyweight than The Economist (Jan. 11).

According to the magazine, bundling muffles price signals and obscures the management costs of institutional investments apparently dissipating investor wealth in the process. The argument goes something like this:

Since there are no free lunches, the manager must "pay up" for trades to compensate for the research he receives from the soft dollar broker. But since trading commissions go into the basis of the stock, they are paid implicitly by the managers clients. "Being only human," the magazine notes, the manager "spends other people's money less carefully than he does his own." Soft dollars tempt the manager to churn his portfolio to generate the commissions necessary to pay the research bill he should properly pay out of his own pocket.

What's more, having received research in advance, the manager may feel obligated to direct the promised commissions to the soft dollar broker when he otherwise would not. By "bundling" research together with the execution of portfolio trades, soft dollars diminish the manager's willingness to switch brokers in the event he discovers the broker has failed to execute trades at the best possible price. This, in turn, dilutes the broker's incentive to provide quality execution. Thus, soft dollars should be "banned," The Economist concludes, in "America as well as Britain."

AGENCY COSTS 

This argument relies on what can be characterized as a standard problem of "agency costs." The problem arises because an agent, such as a money manager, has only a partial stake in the profitability of his principal's enterprise. As a result, the agent's interests are less than perfectly aligned with those of his principal. There is a little doubt that agency costs limit the advantages of professional portfolio management, or that in their absence investors would be somewhat wealthier than they are today. But there is also little doubt that every agency cost problem presents the parties with the opportunity to increase their joint wealth if they can organize their enterprise in a way that efficiently reduces the problem.

Indeed, a case can be made that soft dollars constitute such a solution. To see this, it is first important to separate two issues that invariably get confused. First, why are research and execution bundled together at all? Second, why use such a novel arrangement as soft dollars to account for bundled research?

Perhaps with the exception of a few discounters, virtually all brokers bundle research and execution. This is true of soft dollar brokers, who act as conduits in providing the manager with research produced by others. But it is also true of traditional full-service brokers, who produce the research "in house" and make it available to managers on an informal basis. In all fairness, any restriction on bundling, per se, should apply uniformly to all brokers.

But forcing brokers to unbundle research and execution would be a mistake. If managers were required to pay the entire research bill out of their own pocket, they would have too little incentive to make well-researched trades. This is because most managers receive just a small fraction of any increase in wealth they generate on their clients' behalf. "Being only human," they tend to manage other people's money less carefully than they do their own.

Here we see a second source of agency costs. One way to reduce the attendant losses is for investors to subsidize the manager's use of research. Bundling provides the ideal subsidy. It quite correctly increases the manager's incentive to trade, but it also provides him with the tools to identify profitable trades.

In part, profitable trading requires a manager to ensure that the brokers who trade on their clients' behalf receive best execution. But quality of execution is nearly impossible to discern in the short run. An inept, indolent, or opportunistic broker could cheat the manager (and his clients) by doing a shoddy job of execution - in the process saving the added costs of properly working the trade.

Here we see yet a third source agency costs. How can the parties reduce the attendant losses? Once again, bundling provides the ideal solution. Having already received research at the broker's expense, the balance of trading commissions a manager owes the broker bonds the broker's performance. Since the manager's promise to use the broker's services is legally unenforceable, the broker risks being terminated with his account balance unpaid if he cheats the manager. Rather than diluting execution quality, bundling actually guarantees it.

INNOVATION AND SPECIALIZATION

Why use soft dollars to explicitly account for bundled research? Under the traditional system, the broker's research obligations are only loosely tied to the trading commissions he expects from the manager. This is perfectly reasonable, since over time the parties can be presumed to know whether their expectations are being met.

But one consequence of such loose bundling is that the broker's research obligations are too indefinite to be delegated to other firms; the research must be produced by the broker "in house." Unlike the traditional broker, the soft dollar broker's research obligations are explicitly tied to the cash value of the trading commissions he expects from the manager. By precisely metering these obligations, soft dollars assure that they are definite enough to be delegated to specialty research originators, who can perform them at much lower cost. With research being produced by specialists, the soft dollar broker is free to concentrate on what he does best.

Thus, soft dollars appear to be an innovative form of organization that achieves the benefits of bundling wile allowing reseach and execution to be efficiently produced by specialized firms. Since they reduce costs and increase competition, soft dollars cannot possibly make investors worse off than under the traditional system.

Meanwhile, there is one additional reason to doubt whether soft dollars dissipate investor wealth. Any number of large, aggressive funds successfully appeal to the market by advertising their disdain for high management expenses. Why not advertise a similar disdain for soft dollars? If soft dollars really dissipate investor wealth, surely at least a handful of the managers who used them over the past few years would have lost their jobs for poor performance.

Wouldn't we expect these managers to have tried the "market solution" first? Moreover, why do investors stay with managed funds when they are free to seek refuge in virtually transparent and unmanaged indexed funds? Even more obvious, why don't investors simply manage their own accounts?

THE COMMISSION ON COMMISSIONS

Apparently lacking confidence in the market's ability to solve tha agency cost problem. The Economist proposes that soft dollars be entirely banned. In fact, the process has been slowly underway for well over a decade in the U.S.

In 1975, Congress provided money managers with a statutory "safe harbor" from suits by their clients for "paying up." But with few exceptions, the U.S. Securities & Exchange Commission has narrowed the protection afforded soft dollars while leaving traditional brokerage unrestricted.

For example, in 1986 the SEC found that pension plan sponsors will receive no protection from the safe harbor if they, as opposed to the managers they hire, attempt to capture the research benefits of soft dollar brokerage. Far from reducing the agency cost problem, this ruling appears to make it worse.

Moreover, in July 1990, at the request of the Department of Labor, the SEC ruled that the protection afforded soft dollars by the safe harbor applies exclusively to "broker" trades (those performed on an agency basis), and not to "dealer" trades (those performed on a principal basis), despite its language specifically referring to "brokers or dealers."

With this single feat of interpretational legerdemain, the SEC caused the growing soft dollar trade in fixed income and OTC securities, which are uniformly traded on a principal basis, to virtually disappear. Now, the SEC is entertaining suggestions that soft dollars - but not tradional brokerage - be subject to burdensome disclosure requirements that could easily bury the soft dollar industry under a mountain of paperwork.

A COMMON COMPLAINT

Rather than more regulatory protection, what investors really need is protection from regulators. A good place to start is with the restriction that prohibits money managers from having their broker affiliates execute portfolio trades on the exchange floor.

Under current regulations, a broker affiliate is free to introduce the trade to the floor, but the execution must be performed by an independent broker. One common complaint about the existing law is that it allows other members of the exchange a free ride on the research the money manager did to identify the trade.

The effect of such free riding is to increase the price impact of the manager's trades and therefore to reduce the return from portfolio research. Eliminating this regulation would no doubt go a long way toward increasing investor wealth.

It would be comforting to live in a world where agency costs were completely absent - where money managers' interests were perfectly aligned with those of their clients, where the quality of brokers' execution could be instantly discerned, and where innovating novel business arrangements to avoid these "problems" was unnecessary. But that is not the world in which we live.

A hard look at soft dollars reveals that they actually benefit investors. Therefore, the call for further regulation should give us pause.