A belief that market makers at institutional brokerages are “naïve” providers of liquidity — uninformed players operating from behind a firm Chinese Wall — may itself be an uninformed presumption.

That’s the conclusion reached by two Wharton finance researchers in a recent paper exposing the “leakage” of insider trading information within brokerages by market makers, and suggesting that the practice may need additional regulation. The paper, titled “Who Are the Beneficiaries When Insiders Trade? An Examination of Piggybacking in the Brokerage Industry,” is co-authored by Christopher C. Géczy, director of the Wharton Wealth Management Initiative and an adjunct finance professor, and Jinghua Yan, a research analyst at Tykhe Capital. Both are Fellows at the Wharton Financial Institutions Center.

Market makers are specialized intermediaries, registered with the New York Stock Exchange and the NASDAQ, who provide liquidity through the buying and selling of large volumes of securities. The Wharton researchers, in a detailed parsing of four years of insider trading at 15 of Wall Street’s largest brokerages, find that market makers executing insider trades at these firms appear to act on information gleaned from those trades.

The evidence can be seen in the more aggressive prices they set for the company’s stock following an insider trade. Put another way, compared to their peers, market makers affiliated with the brokers used by insiders post more aggressive ask quotes during periods when insiders trade. The study was undertaken using information from trades made between March 1999 and November 2003.

“Academics and, to some degree, those who trade in the market, might assume that market makers are there simply to take the other sides of trades and provide liquidity, whereas it looks as though they may have, and may act on, information,” says Géczy. “What we found is that there is a leakage somewhere along the lines in the information transmission channel between the investor — in this case, company insiders — and ultimate trades, and the way information is transmitted into the market in the form of buy or sell orders.”

The term for this behavior is “piggybacking.” Géczy and Yan found that the less information or certainty around a given company, the greater the impact of insider transactions on market makers’ behavior. The ensuing piggybacking is affected by the number of analysts covering the company (or the amount of public information available); the spread of analysts’ company forecasts; the bid-ask spread of the stock, and post-event stock returns.

The forecast spread reflects analysts’ degree of certainty about a company’s future performance. The wider the bid-ask spread, the greater the uncertainty, and therefore the liquidity, of that stock. Piggybacking diminishes, the authors found, when the firm of the broker-dealer making markets has had a prior investment banking business relationship with the company doing the insider trading.

“In a classical theory framework, market makers who provide liquidity to a market with informed and uninformed investors do not possess superior information. However, the model implications are not entirely consistent with some recent empirical evidence,” states Yan. “One important implication of our results is to call for a more careful study of the market structure. What if the market maker is informed? And how should we understand the market structure differently? I think, ultimately, the fairness and transparency of the market are crucial to making a competitive capital market.”

Yan’s remarks point to the significance of the timing of the research. In 2000, well into the study’s four-year period, the Securities and Exchange Commission adopted Regulation FD (short for “full disclosure”), which banned the selective disclosure of information by corporations to large shareholders and securities analysts. Géczy and Yan report that their work provides evidence that Regulation FD is working, at least with respect to the information they examined, and that information leakage slowed after the SEC’s action.

More Policing Needed?

However, their results also suggest that there may be more avenues for conflicts of interest with regard to insider trading than previously thought. Regulators may need to go further than the traditional policing of the Chinese Wall between research analysts and bankers who work for the same institution.

This could be a game-changing result of the findings, especially at a time of intense regulatory scrutiny in the wake of the failures of all five U.S. investment banks in 2008 and the financial meltdown of equities markets around the world. Asked what the most direct result of the researchers’ work might be, Yan says: “Disclosure.” After Regulation FD, “the window for piggybacking [got] shorter.”

Before and after Regulation FD, Géczy adds, “you see a change, and that may be generally correlated with the associated greater disclosure statement. Potential advantages that arise from anomalous behavior are obviated by disclosure.”

The authors present two hypotheses. Their “Information Hypothesis” is that insider-affiliated brokers’ clients or traders, possessing privileged knowledge that the stock price may fall after insider sales, could place limit sell orders with lower limit prices. The “Market Stabilization Hypothesis” refers to instances when insiders trade in large quantities, and affiliated market makers then greatly influence the stability of the market — by the more frequent and more aggressive nature of their quotes.

The authors acknowledge a few significant limitations to the sample selection of more than a half-million trades between March 1999 and November 2003, but they could be offset by conservative assumptions. For example, the exact time of day when insider trades took place is not known (many regulatory filings aggregate several trades from the same day, or even over a period of several days, usually when volume is large).

Also, information on inventory positions for each broker, or the daily volume of any given stock traded by a particular broker on any given day, is not known. However, as footnoted in the report, insider trades may be completed by more than one broker at the same company on the same day, but the authors record them as “independent events” — a conservative assumption which they assert could reduce their test power but, at the same time, increase the significance of their results.

The 8.5 Minute Difference

On insider trading days, company-affiliated market makers “on average quote at the inside [asking price] more frequently than their peers by 2% of the time (or 8.5 minutes). There is no significant abnormal quoting behavior in bid prices,” the researchers write. Suggestive as these results are, a caveat is that “they do not necessarily serve as direct evidence” since “under pressure to fill sell orders for their clients, affiliated market makers may have to ask more aggressively in order to complete their trade orders.”

By using strict controls, though, the authors’ research accounts for various ratios of insider trading volume, sorting them according to monthly trading volume by insider-affiliated market maker. As it happens, “large-size trades do indeed result in more aggressive asking from affiliated market makers. Both the economic and statistical significance increases as relative size increases,” according to the paper. “In order to execute these large-sized orders from insiders, insider-affiliated market makers ask more aggressively than normal over a longer period of time. For example, for the trades in the largest [sampling], aggressive asking by affiliated market makers persists for seven days.”

However, the researchers point out that trade size does not explain all abnormal quoting behavior by insider-affiliated market makers compared to their peers. The content of the information may play just as big a role. To observe this, Géczy and Yan used other controls for trades by “informed” and “uninformed” insiders (“uninformed” insomuch as they are insiders who must report their trades to the SEC, but who more than likely do not have “material inside information”). 

With the exception of the largest sampling, statistics show that “trades from informed insiders coincide with greater magnitudes of aggressiveness in asking [quotes] from insider-affiliated market makers. Most importantly, in the smallest relative size [sampling], we find that when insiders are uninformed, the difference between insider-affiliated market makers and their peers becomes small and insignificant,” the paper states.

The amount of “information asymmetry” is another factor in the correlation between informed trading and broker behavior, the study shows: “Insider-affiliated market makers ask more aggressively when the number of analysts following the company is small or when the degree of information asymmetry is high…. When the degree of information asymmetry is low and the trade size is small, insider-affiliated market makers do not post significantly more aggressive quotes vis-à-vis their peers.”

This leads to another unique aspect of the study — whether the information content in insider trading is new to insider-affiliated market makers. “In other words,” the authors ask, “can there be other channels through which the insider-affiliated market makers acquire the information?”

Again, Géczy and Yan set specific controls to answer the question, dividing their samples into two groups based on whether or not the insider-affiliated market makers’ investment banks had had prior business relationships with the firms of the insiders who were trading.

“When the size of the insider trade is small and when the bank has had prior investment banking business with the firm, the aggressive asking behavior completely diminishes,” the report states. “This, too, is consistent with the Information Hypothesis in that when there are other channels available for an investment bank to acquire information from the company, the insider brokerage affiliation becomes less important. We view this as indirect evidence that information may leak from the investment banking (or corporate finance) division to traders within the same brokerage house. This raises new concerns about the strength of the Chinese Wall in brokerage houses.”

Regulating ‘Information Leakage’

The authors say the unofficial name for this project has been their “Martha Stewart paper,” and they refer to the famous home design and media mogul in their introduction and conclusion. Stewart was convicted in 2004 of making false statements to federal investigators, among other charges.

“The type of information leakage (or piggybacking) present in the Martha Stewart case may actually be more generally present in the brokerage business than that single case may imply,” the paper states. “In fact, our results suggest that current regulations, or at least their enforcement, may not fully contemplate this sort of information transfer. In the unsettled debate about whether and how insider trading should be tightly regulated, those who endorse a laissez-faire approach may believe that insider trading improves market efficiency.

“However, the insider information leakage documented in this paper leads to almost no improvement in market efficiency through insider-affiliated market makers, because the public discovers the information about insider trading within a few days.”

In reflecting on the other scandals that rocked the markets since the beginning of the period that they studied so closely, from Enron to Bernie Madoff’s alleged pyramid scheme, Géczy says his and Yan’s findings mean that regulators “need to continue to take a close look at the structures of markets. The conflicts of interest are really very serious and still exist all over the place.”

All these things are “part of a cloth that we have woven for ourselves in which regulatory authorities have a very difficult job with heavy resource demands, but one that we rely upon getting done well, efficiently and effectively,” Géczy says. The end result here, he adds, “may be that we find ourselves paying more for the regulatory function. The SEC may ultimately claim that it didn’t have enough resources to have caught Madoff, that it would have cost too much to guarantee that outcome. Well, as a society, we have to weigh the costs against the benefits of such oversight.”