How Big Data Connects Bankers, Politics And Insider Trading
Wharton Accounting professor Daniel Taylor on politically connected bankers and insider trading
In the tense days before the big bank bailouts of October 2008 under the Troubled Asset Relief Program (TARP), highly unconventional deals worth hundreds of billions of dollars were being cut between then-Treasury Secretary Hank Paulson and the nation’s top bankers on a single sheet of paper. The U.S. economy seemed to hang in the balance. The Dow was in the midst of plunging some 50% from year-earlier highs and financial shocks eventually cost some 9 million jobs.
The ExodusPoint Partners International Fund returned 0.36% for May, bringing its year-to-date return to 3.31% in a year that's been particularly challenging for most hedge funds, pushing many into the red. Macroeconomic factors continued to weigh on the market, resulting in significant intra-month volatility for May, although risk assets generally ended the month flat. Macro Read More
Yet, leading up to these events, it turns out, there was a large spike in stock trading activity by the bank insiders who were the most politically connected to politicians and regulators, according to new research by Wharton accounting professor Daniel J. Taylor, who used Big Data and algorithms to reach his conclusions in the paper he co-authored, “Political Connections and the Informativeness of Insider Trades.” In this Knowledge@Wharton interview, Taylor describes how the methodologies outlinedin the paper could shine a “bright light” into areas where investigators could uncover potential wrongdoing. His co-authors are: Alan D. Jagolinzer, University of Colorado; David F. Larker, Stanford University; and Gaizka Ormazabal, University of Navarra. An excerpt from the abstract of the paper appears below, followed by an edited version of the conversation with Taylor.
“Examining insider trades around the announcements of TARP infusions, we find evidence of significant trading thirty days in advance of the announcement, and that these trades predict the market reaction to the announcement. Notably, we find these relations are present only for the trades of politically connected insiders. Overall, our results suggest that politically connected insiders had an information advantage during the Crisis and traded to exploit this advantage.”
Knowledge@Wharton: I’d like to welcome Daniel Taylor to Knowledge@Wharton — he’s a professor of accounting here at Wharton. And he’s written a very interesting paper on the links between political connections and insider trading during, or just after the financial crisis, and during the period where TARP was under consideration. TARP is the Troubled Asset Relief Program.
Please us a short summary first, and then we’ll get into more specific questions about what your paper found.
Daniel Taylor: What we did in the paper is, we looked at the trading of corporate insiders. So by corporate insiders we mean officers and directors of publicly traded corporations. All of these individuals have to file what’s known as Form 4 with the SEC and the public record that discloses their trades in their firm’s shares or their firm’s stock. So they’re allowed to trade in their firm’s shares and they file that information with the SEC.
We gathered all of the information that’s out there on these Form 4s — on their trading. And we looked at the trading before the financial crisis, during the financial crisis and after the financial crisis. So this is the trading of executives and directors in their own firm. We looked specifically at financial institutions. One of the big questions that comes out of the financial crisis that’s interesting on Wall Street and on Main Street is: Did anyone know — did anyone see the crisis coming?
“One of the big questions that comes out of the financial crisis that’s interesting on Wall Street and on Main Street is: Did anyone know — did anyone see the crisis coming?”
We initially started the project by looking at: Okay, can we see whether insiders, corporate insiders, officers and directors, traded in advance of the crisis? So at banks that did poorly during the crisis, did their executives sell shares before the crisis hit? And then we looked at the trading during the crisis — so around the bank bailouts, the TARP monies that you alluded to. And then after the crisis, did trading stabilize?
We found that there didn’t seem to really be any evidence of trading before the crisis. There was no evidence that insiders traded in anticipation of the crisis. But we did seem to find some evidence that insiders traded during the crisis in the period in which the TARP funds were dispersed. And then we investigate that a little bit more and what we find is, interestingly enough, that it’s only the insiders that had political connections, and by political connections I mean connected to a bank regulatory agency or the current House or Senate at the time.
So we looked at bank boards who had a director or officer who had work experience, current or past, at a bank regulatory agency, the Senate or the House, and we found that the boards of those banks that had those political connections traded more heavily during the financial crisis. Their trades had higher predictive ability of outcomes during the financial crisis. So they predict, for example, the market reaction to the amount of the bailout that the bank would receive. So we basically find that during the crisis there is some trading in anticipation of bank bailouts.
And then after the crisis … things go back to normal and we don’t really find that those trades or those individuals’ trades have any more information than other ones.
Knowledge@Wharton: So this would seem to be the definition of insider trading on the face of it. I want to point out that this study had a very large sample. It looked at 7,300 corporate officers … across 497 publicly traded, TARP-eligible institutions.
Taylor: That’s correct.
Knowledge@Wharton: This was a very wide net that you cast.
Taylor: One way to think about this paper is to think about it as using Big Data and computer algorithms to sift through public information on the trades of officers and directors, and to see which trades of the officers and directors are correlated with future outcomes. We would say if a trade is correlated with a future outcome, if it has a high predictive ability, it’s more likely that that trade might have been based on [private] information about that future outcome.
But I do want to clarify that … in the popular press and legal scholars, they use the term insider trading, they’re thinking of illegal insider trading. When we use the term insider trading, we’re thinking of the trading by officers and directors. So when we say insider trading, we mean trading by corporate insiders. And there are two types: The first type is the legal trading — the board or the CEO can trade in his shares, just like anyone else can.
“What we’re doing is we’re reporting correlations on … a large sample and casting sort of a suspicious eye in the direction of insiders that have political connections.”
Knowledge@Wharton: With proper disclosure.
Taylor: With proper disclosure … as long as they do not have any private information. The illegal kind is when they have private information that they have not disclosed to shareholders. This relates back to a set of rules called “disclose or abstain.” If you’re a CEO or a manager, and you have private information, it’s your duty to either disclose it to shareholders or abstain from trading. And so in this setting it’s a little bit more murky because when we’re looking at political connections and connections to bank regulators, it’s not really clear how the manager would disclose any information that they would have gotten from their connections.
Knowledge@Wharton: Because it’s not definite information?
Knowledge@Wharton: [An insider is telling a banker:] “It looks likely that this is going to happen.” So legally he doesn’t really have an obligation to disclose something that’s not really certain, is that right?
Taylor: That’s correct. I think there’s been a spate of recent insider trading cases … [which] the government has prosecuted [and lost], and out of those insider-trading cases, the definition of illegal insider trading has become more murky.
Taylor: Yes, more slippery. And so it would be wrong to suggest that this paper has sort of a smoking gun. What we’re doing is we’re reporting correlations on, as you mentioned, a large sample and casting sort of a suspicious eye in the direction of insiders that have political connections. If you go back to all of the books and the popular press during the period, the bailouts were effectively decided in back room meetings with sort of insiders. There’s a great book, you know, Too Big To Fail by [Andrew Ross] Sorkin that details all of these interviews with people.
And so a natural question is, “Well, if the bailout is being decided in private and in consultation with bankers or treasury officials or government officials, how does that information leak out?”
Knowledge@Wharton: It’s also interesting to note that this was not the result you were expecting when you started out on this path. Could you talk about that briefly, about how this result surprised you?
Taylor: Yes, absolutely. And frankly we’re still kind of stunned at the results. And so when we started the project several years ago, you know, as I mentioned it was sort of like a Big Data study. We sat down, the co-authors and I, and talked about, okay, this would be interesting. Does it make sense? Because there was all of this concern about “did they anticipate the crisis?” So we began the paper thinking, okay, we would shed some evidence on whether they anticipated the crisis or not.
And so five years ago that would have been a very timely finding: They did anticipate the crisis or they didn’t anticipate the crisis. But it took us a little longer than we expected to gather the data and to do the analysis…. When we started doing the analysis and we realized, okay … let’s add some more data, let’s add the post crisis, the bailout period data, let’s add the data from 2010-2011 into it and run the analysis. We found somewhat surprisingly that there was this blip, for lack of a better term, in the correlation between insider trades and future performance during what we call the bailout period. And that’s the period from when TARP began being dispersed in October  through June .
And so when we saw the blip, the next question is, “Okay, where is the blip coming from? Why is the blip there?” Then we found that the blip only occurred in banks that actually received TARP money, okay? So think about an investigative journalist. Dig deeper. Okay, so now the blip is only in those banks that received the TARP money. Okay, so then dig deeper. Look at the connections of the individuals or of the bank. Okay, well now the blip is not only in banks that received TARP money, it is also only, within that set, [those that] received TARP money and had ties to former or current bank regulatory agencies like the Fed, the Treasury, the Office of the Comptroller of the Currency, the FDIC. And so that’s when we said, “Wait a second, is this actually where the blip is?”
“A natural question is, ‘Well, if the bailout is being decided in private and in consultation with bankers or treasury officials or government officials, how does that information leak out?’”
My co-authors and I are actually pretty suspicious and skeptical people by nature. In terms of data analysis, we threw everything we could at it to make it go away. And so I wouldn’t be here if it went away. This isn’t a case where we’re trying to cast some suspicion on something that may or may not go away. This is a very robust result.
I also want to mention that the data on political connections — it was not like we collected this data or that we generated this data. Everything in the paper is generated from publicly available information. So it’s perfectly conceivable that somebody could go out there and investigate it themselves, [and collect] the data on political connections. Where do you get data on political connections? Well, in the bios of directors and officers in the annual reports, they list what their current and former work experience was. So everything that’s in the paper is out there in the public domain.
Knowledge@Wharton: So you weren’t using any insider information?
Taylor: No. We were not using insider information.
Knowledge@Wharton: What are some of the practical implications of these findings? It’s really interesting, as you pointed out, that any alleged criminality of all of this is questionable because the information they had wasn’t necessarily actionable in the way that they had to disclose it. But if someone was whispering in their ear, and it was good information, that seems a little bit sleazy, right, to put it mildly.
Knowledge@Wharton: So what practical implications could come from this? For example, this speaks to the damage that can be done with the revolving door of government officials having come from industry and then going right back into industry without a proper amount of gap in time. But I’ll leave it to you, what are some of the practical implications here?
Taylor: I think there are probably … two main practical implications. The first is we need to give some more thought to what exactly constitutes illegal insider trading. I think right now, in light of recent court cases, there’s a big gray area out there. And I think a lot of people would be surprised at how gray the area actually is. You need a smoking gun to be able to prove an insider trading case. And it’s very hard to get that smoking gun. You have to get the tipster. The tipster has to confess. There has to be a series of dots that you can collect.
Knowledge@Wharton: A tipster who may have benefited in some way from this to begin with.
Taylor: Correct. It’s difficult to do that using a large sample with correlations. But this suggests areas where people may want to shine a light. So think about it as insider trading is a dark room, and we’re shining a light. It’s not entirely illuminated, but we’ve got a good area where people should look.
The second one is, like you mentioned, the revolving door. There is a series of studies in the academic literature on political connections, and specifically political connections of financial institutions. So for example, there’s a recent study in the Journal of Financial Economics [“The value of connections in turbulent times: Evidence from the United States, by Acemoglu et al.], and they found that — this was before Tim Geithner was made Treasury Secretary — that when he was announced to be the Treasury Secretary, the banks that had connections to him, either board seats or what not, experienced massive increases in shareholder value [12% over the subsequent 10 days]. So what this is saying is that the market is aware that there is this connection between bank regulators and the banks which they are regulating.
For example, many people are surprised when I say go to the Federal Reserve’s website, and you [will see that] during the crisis, Lehman Brothers CEO [Richard Fuld, Jr.] was one of the board of directors of the [New York] Federal Reserve. Jamie Dimon (president, CEO and chairman of JPMorgan Chase) was on that board as well.
Knowledge@Wharton: And of course Paulson, himself [Henry Paulson, former Treasury Secretary and CEO of Goldman Sachs], was heading that backroom deal you alluded to earlier.
Taylor: Correct. So you’ve got this, “marriage” is too strong of a word, but you have these very strong connections between the regulators and the people that they’re regulating. Now that’s good and bad. On the one hand you want somebody knowledgeable about the industry and how banks work to regulate banks. But where it becomes questionable is, are there conflicts of interest? And our paper suggests, yes, there are conflicts of interest and these conflicts of interest are potentially pervasive. And I don’t think that’s really been given enough consideration out there … there’s a good side to having a banker regulate the banks, and there’s also a bad side.
Knowledge@Wharton: Did you attempt to quantify any of this, like how much some folks may have benefited?
Taylor: The benefits are tricky because what we’re looking at is what we call opportunity profits. So when an individual trades, if their share price goes up, let’s say 6% over the next month, then that means that they’ve earned a 6% return on their money if they sell. But we can’t match buys and sells. So we can’t actually compute a measure of realized profit. All we can say is, “Well, you bought this month. Stock prices went up 6%. That’s 6% more than you would have made if you had waited a month.
In this case we look at the trades in advance of the TARP infusions. We know the date of the TARP infusions, we know the date of the trade. And we look at the trades, and surprisingly we find a significant amount of trades, both in volume and in dollars, clustered 30 days before the bank actually receives a TARP infusion.
Now I have done some work on insider trading policies in my prior research. And what was surprising to us is that most companies have something known as trading blackout windows. And a trading blackout window is imposed by the company in the code of ethics on the officers and directors and it says, “You shall not trade N days,” typically 20, 30, 40 days, “in advance of a material event.” So for example, for an earnings announcement that the firms make on a quarterly basis, insiders are prevented from trading several days in advance of that earnings announcement.
In this case if there was a blackout window [that] applied to the TARP infusion, those trades [30 days before the infusion] should not have existed. So what this suggests is that either the event, the TARP infusion, was unanticipated, so a blackout window could not apply; or that they sort of skirted their firm’s blackout windows.
Knowledge@Wharton: I want to focus on the “no smoking gun comment” that you made at the outset. It seems [there is] pretty clear evidence that there was insider trading … as I read it, very strong evidence. And yet this isn’t pointing to any individuals, necessarily. You were sort of doing this globally. And the information that you have, while interesting, isn’t the kind of information that someone takes to court. It’s the kind of information that an investigator would use to say, “Here are the most interesting places to look if I were going to try to find a wrongdoer.”
Taylor: Right. I would think about this as sort of an academic paper on the cost and benefits to political connections. Most of the academic literature has focused on the benefits that the shareholders in the firm receive from having [an insider with] political connections on the board. So it may very well be that you as a shareholder in the firm, you want your bank to be politically connected because you’re more likely to get bailed out, you’re more likely to get special favors. And so most of the academic literature suggests this is a good thing for shareholders, not necessarily for the government, for the public, but for a narrow set of shareholders.
What we’re coming along and saying — this is sort of the academic debate — there is also a cost. What is the cost? Well, the cost is that if you have these political connections during the crisis, [politically connected insiders] may engage in some, for lack of a better term, shady trading. That shady trading comes at a cost to shareholders. So who loses out from the insider trading? If the executives traded with private information, it’s other shareholders who potentially are on the other side of that trade, and they lose out.
So this is really about, if you think about academics, political connections have a benefit to the firm, and we’re documenting one cost, which is sort of extracting rents — what we say is rent extraction from the shareholders.
“So who loses out from the insider trading? If the executives traded with private information, it’s other shareholders who potentially are on the other side of that trade, and they lose out.”
Knowledge@Wharton: And also the potential cost or the risk of being exposed and what that could do to the banks’ stock.
Taylor: Right. And so one shouldn’t ascribe any sort of policy agenda or anything like that to the paper or to the findings. We’re just basically saying: There’s this cost of political connections. And by the way, this is not a trivial cost. This seems to be a big deal, especially in light of all of the debate about TARP, how the bailout monies were doled out, the decisions about how those things were given out. And so hopefully something like this doesn’t happen again, but it does suggest that these conflicts of interests do have real costs for the bank shareholders, even if the bank gets bailed out, the managers or the officers and directors may have traded on private information in advance of that actual bailout.
Knowledge@Wharton: Any idea of how many people were involved in those potential violations?
Taylor: It’s hard to say. Since everything is based on correlations, I can’t go around and say, “You, you, you, you, you. This firm, that firm, this firm.” All I can say is [this is] the set of firms and the set of individuals where the correlation is particularly high.
Now those trades may have been conducted entirely legally, you know, within sort of the purviews of the existing insider trading rules, and by chance there was a very high correlation. But the fact that we find this correlation so pervasive as you mentioned, and specifically focused within those that are politically connected suggests that it’s not just one or two or three individuals — they would not be able to drive a result on 7,300 insiders. So it’s sufficiently pervasive that it’s able to generate a large sample result.