Using Google to forecast a stock’s reaction to earnings reports – with Darren Roulstone

Announcer: You’re listening to Tradestreaming Radio with your host, Zack Miller. Expand your mind. Become a better investor with tools, tips and technology from the smartest investors on the planet.

Zack: Welcome to Tradestreaming Radio. I’m your host, Zack Miller. This is the place where we bubble up tools, tips and technologies to help investors make better, smarter investment decisions. Today’s guest definitely does that. It’s Darren Roulstone. He is the associate professor of accounting at The Ohio State University’s Fisher School of Business.

He’s also the author of a couple of influential papers. One of which, entitled “Investor Information Demand: Evidence from Google Searches Around Earnings Announcements,” which will serve as sort of focal point for our conversation today. He’s written a couple of other influential papers, one of which I quoted in my book, “Tradestream Your Way to Profits,” on insider trading that he wrote with a friend of his, Joseph Piotroski.

I wanted to thank Darren for coming on the program. It’s always great to talk to smart, creative minds looking to find new ways to use the tools at our disposal to make better investment decisions. That’s right up the Tradestreaming ethos and wanted to thank Professor Roulstone for joining us.
You can find the archives of this program at my website Tradestreaming.com. You can also find the archives of this program on iTunes. You can subscribe either way but definitely come to Tradestreaming.com to sign up for my free weekly email. You can stay abreast of all that’s going on in the Tradestreaming world that way, the easiest way to do that. Please, if you’re using us on iTunes, give us a rating or a ranking and let other people know if you’re finding value in this program. I’m trying hard to make this useful for you guys. Hope you’re enjoying it and we’ll catch you again soon.

Darren: Darren Roulstone, associate professor of accounting at The Ohio State University. My background a little, I grew up in Western Canada, actually. Eventually, I moved down to the States. I got an accounting degree at BYU. Went to Michigan, actually, for a Ph.D. People are a little shocked I ended up here at Ohio State after that.

Zack: Go Blue, right?

Darren: There you go. I get a little conflicted every now and then, especially in November. Spent some time in the middle between Michigan and Ohio State at the University of Chicago Booth School of Business. I was there for eight years before coming to here to Ohio State. It’s been a good experience being here. We’ve been in the Midwest for a long time now and really love it here. A lot of great opportunities here in terms of what we’re doing and the research, which we’re talking about here.

Zack: How did you first get into the field? What piqued your interest initially?

Darren: In Google or just in general?

Zack: In general, yeah.

Darren: It really goes back to my Dad, who taught at the junior college level. He just seemed to really enjoy what he was doing so I kind of naively got into it originally thinking, “Wow, it’d be fun to teach.” Then it was actually when I was at BYU, I was going into what they call the junior core, which is this the first year, full-time year of the accounting program.
I met a guy there named Doug Prawitt, who had just finished a Ph.D. from Arizona and had taken a job at BYU and was teaching auditing. I started talking with him. He was working on research and things. I told him what I thought about teaching sometime and he started kind of explaining the research world. I was just fascinated. I thought it was amazing. I had no idea. Most people say, “How do you do research in accounting? Trying to make better balance sheets or something?” It kind of opened up my eyes that, no, accounting information is a major piece of information and how it gets used by markets is an active area of research. That’s really what kind of did. I kind of went from there.
He kind of helped mentor me through. This was back in the old pre-Internet days, if anyone can remember that far back. He was really helpful at talking to schools, looking at schools. Fortunately, at one point, mentioned that Michigan was a good program and I ended up going there. Was there at a really great time. I’ve just been very fortunate. It was a great time. They have some great faculty, great students and really got a good basis in how to do this kind of stuff. I’m very fortunate.

Zack: Do you think your background in accounting has helped as you’ve sort of taken more of a finance route?

Darren: Yeah, that’s the thing. What I really do is a lot of stuff that would be seen as more on the finance side. I actually started doing a lot of insider trading stuff. My dissertation was insider trading and compensation contracts. I think the nice thing is that there’s a lot of stuff in accounting that can use tools from finance or use some of the settings of finance and bring them together and that’s kind of where I’ve been trying to do my work.
Some of the other research I’ve done was stock return synchronicity, for example. It’s a finance/economics concept that I and a coauthor kind of brought over into accounting. We put some of the accounting angles on it.

Zack: Before you get into your current paper, which is the focus of what I wanted this call to be about, can we talk about the insider trading research you had done? You did that with a friend, Joseph Piotroski, right?

Darren: Yeah. I had done a dissertation on it and then Joseph Piotroski, who was a student with me at Michigan and then we were colleagues at Chicago as well, we started working on some papers. We kind of wrote two main ones. One was the idea that, in the finance literature, it was kind of well-established that insiders are contrarians. They tend to sell when prices have gone up and when market-to-book values are high to glamour firms and they buy when their value firms, high book-to-market films, low market-to-book, and when returns have been low. They’ll come in and step in. We were kind of thinking, “On the accounting side that we think of information and we think about what do they know, are these things just masking each other?” We do know that value firms tend to kind of bounce back and they have kind of earning surprises met.
We kind of put the two together and wrote paper showing that controlling for all the contrarian stuff, future good earnings news was associated with insiders being more bullish, buying more than they’re selling. That was kind of one that was just kind of an information piece, accounting information.
The other was this idea of stock return synchronicity, the idea that firms returns, they’re always going to have some kind of an association with their industry and with the general market and then the rest is kind of firm specific. It’s how idiosyncratic they are. It’s taken of the measure of information. If there’s a lot of information about the firm itself, the firm should be more idiosyncratic in its stock returns. We just looked at it and said, “Well, within accounting, we’re always looking at the activities of analysts, the activities of institutional owners and the activities of insiders, what we were kind of looking at, and they should all get information in the price.” We just looked at if you have active insiders, a lot of institutional ownership, you have analysts following you, how does that affect the synchronicity of the firm? Is it more like industry market or is it more like an idiosyncratic firm?

I found with the insiders, when they’re active, the firm tends to move more idiosyncratically and less of the correlation with the industry and market. That’s consistent with the insiders. They know the firm when they’re trading, when they’re active. It’s firm news that they’re trading on, not something about the industry or the market.

Zack: I mentioned this to you in the pre-call. I actually quoted you and one of your papers in my book that I wrote in 2010. I wrote a whole chapter on insider trading and strategies that are replicable for institutions or individual investors, sort of piggy-backing on some of this activity. I’m always curious because it seems to me one of the no-brainer-type strategies if you’re an investor. The information is transparent so the insiders are filing this information with the exchanges, with their company. Why more people don’t use the insider activity as, it doesn’t have to be the sole input but an input into their investing decision. Do you have any color on that?

Darren: Yeah, I think a couple of things. One of them is so much of the activity itself fails, which there’s different compensation reasons. Many executives are competent in stock. They’re selling either routinely or not. You have to kind of burst through that cover first, separating out the ones that mean something given that the overwhelming majority is sales.
Another thing which is kind of a possibility, I’ll raise this, I’ve got a working paper, not a published paper, but one that I’ve actually had for quite a long time now where, among other things, we looked at the returns to insider trades. If you form portfolios every month on who’s buying and who’s selling and look at the returns, the returns are strongest in the really idiosyncratic firms. If you sort firms on their past idiosyncratic risk, how individually volatile their stock returns have been, there’s a very clear pattern where the really idiosyncratic ones. That’s where the insiders are making the most money.

If you’re following the insiders, it’s going to be pushing you. You’re going to make the money by investing in these very small, idiosyncratic, risky firms. Those are the firms where it can be tough to hedge your position. Maybe transaction costs are higher. There’s more information right where it’s hardest to exploit. That’s one of the typical finance/accounting economic studies of arbitrage are. It’s often where you can stand to gain the most is where it’s going to cost you the most and it could be a problem there. That’s one possibility is that people want to do this but it’s tough to trade on the stocks where you can actually get the highest return.

Zack: Just curious, while you were at the University of Michigan, Professor Seyhun had written a whole book on insider trading strategies. Did you guys work together or collaborate on anything?

Darren: Yeah, he’s great. He’s really done so much. He got the field kind of going on the finance side and that. It’s kind of funny. When I was getting the idea for my dissertation, which was that idea about compensation contracts and how if you have insiders, they can trade and they can kind of make money that way, and how you would adjust compensation contracts for it. I went to talk to him about it and I asked him to be on my dissertation committee, one of the faculties advising me. He was working on a similar idea. He said, “Ethically, I can’t really be on your committee given that I’m working on something like this.” We talked and stuff and he was very helpful.
In fact, I’d written a paper earlier in the year. We had this requirement to write a paper in our second year. I’d written that I wanted to do it on insiders and he had given me a bunch of data for it. He was very supportive, as much as he could be, without being able to actually be on my committee because of some work he was doing at the same time. Yeah, he’s done some great stuff and has continued to do that. I saw him at a conference a few years ago. He had a great paper about when investment banks put people on boards of directors of other companies, they seem to get a lot of information from that. He’s continued to do a lot of work in that area. It’s been my regret that I was at Michigan and I didn’t get to work with him on things but he was very helpful.

Zack: That’s a great story. Just a question. I’m going to be blunt. Given what you’ve seen and the research you’ve seen on insider trading, and I don’t really know what you do in your individual investing practice, if you do at all, is it the type of information you would consider using? Is it that actionable?

Darren: I think there is actionable stuff there and, certainly, other academics have gone into it. [Carbetta 11:29] came out of Indiana. He’s had a foreign investment group that uses information from insiders and that. As far as what I do, I spent eight years at Chicago. I’m an index fund guy. It’s actually my Dad who is the one who does the investing stuff for fun. If I was going into something, if I thought, “Hey, I’m going to sit down and try to do something where I’m going to be looking at more individual stocks and that,” I’d definitely be looking at that kind of stuff. I think there’s strong stuff.
When I wrote the paper where we find that the idiosyncratic firms are the highest ones, overall, there are returns and there are things you can do. For example, if you look at firms where it’s, say, over the last six months only insider buys, no sales, or, alternatively, only sales, no buys, that pushes the returns up a little.

Zack: What they call “consensus”?

Darren: Consensus, exactly. It’s kind of a trivial thing. One thing that’s been kind of found is that, in general, sales aren’t followed by negative returns. The explanation is, “Well, there are so many liquidity sales. You’re just liquidating your stock portfolio.” What I found was that if you actually said, “Well, only sales and just sales, no purchases over a certain period of time, you actually get some significant negative returns.” Not super strong, but it gets stronger when you do that. There are some simple filters that you can put in. I think it would definitely be worth looking at. I have not, myself, tried to make money on that. Other people have. There are shops out there. You can Google search for it. Google search gets you everything. You do a Google search and there are people who are trying to exploit that in their trading strategies.

Zack: Cool. I guess that’s a great segue. You, yourself, said it. Google. Can we talk about the most recent paper you published?

Darren: Yeah. A while back, I was talking with a colleague of mine, Mike Drake, who came out of Texas A&M. We hired him at Ohio State. He also had a friend who was at the University of North Carolina and has just gone to the University of Washington, took a job as an accounting professor. We started talking about this paper in finance from Joey Engelberg, who you had the podcast with last year. He has this paper called, “In Search of Attention.” It’s this really great paper looking. . .

Zack: Good title too, right?

Darren: Yeah, it really is. He’s got great titles. Among the other things that he does very well, which is pretty much everything, he has great titles for his papers. He had this paper about weekly Google search and how it’s associated with things that we think of for attention-grabbing stocks. Stocks that have strong returns, a lot of volume, there’s a lot of Google search and you find good correlations there. They do a lot of things to kind of show that this seems to be that there’s attention paid to these stocks. Google search is one measure of that attention.
We were thinking about this and we’re thinking, “Okay. What can we do with this? What’s an accounting angle on it?” One of the things we were thinking about was that we know when people are following a stock and there’s going to be information coming out about the stock. The biggest piece of information firms usually have, on a regular basis, is their earnings announcement. What they should do is they should try to find out everything they can in advance to be ready for that news. Some surprises are going to come out. The more I know about it, the more I can predict that surprise, the better I’m going to do.
We were looking at the idea of would people be searching more for information about a stock that until the earnings announcement comes? We can kind of test that idea that they will get ready for the announcement. We wanted to look at if Google search kind of go up around these information events, especially earnings announcements. Then, if it does, do you actually get better prepared for the announcement?

The way we wanted to look at that was if there’s some news coming out, say there’s some big positive surprise coming, if that kind of leaks out in advance, then returns should be positive before the news comes out. If you’ve got a big bad news surprise coming, returns should be negative. The price should be dropping ahead of the news because people kind of figure it out or get early information or make good predictions about it. We thought, “Hey, we can use Google search to do this. It can be a proxy for people searching for information about the stock.” We really liked that idea.

We talked about arbitrage. We talked about this process of people figuring it out if a stock’s over- or undervalued and then trading and making money on it and, in the process, pushing the price to where it should be. Whenever we see arbitrage working well, we think there’s some problem. Post-earnings announcement drift where there seems to be an under-reaction to earnings surprises or the accruals anomaly where firms with high accruals, low cash flow components of earnings tend to underperform. Those kinds of things seem to violate market efficiency. We always say, “What’s wrong with arbitrage? Why isn’t it fixing it?”

One of the ideas is, well, there’s trading costs. Again, maybe the firms where you’d actually see the biggest violations of efficiency, they’re hard to arbitrage. They’re costly. Their cost is short, something like that. We’re thinking, well, another thing is people have to actually go get information. We usually assume information, especially what we call public information like SEC filings. It’s public. Everyone just knows it. In reality, people don’t know it until they actually look it up and read it.

Zack: You still have to find the information, yeah.

Darren: You’ve got to find the information, yeah, which in some cases is pretty easy but in some cases you might have to hunt for something. This has gotten a lot easier in recent times. The SEC with EDGAR on the web, you can just go and look at filings and that. Until someone goes and looks up a filing for the first time and reads it and does something about it, that information really can’t do anything.
We saw Google search as a way to kind of proxy for this process of arbitrage, the information part of arbitrage. People go out, they get information and they act on it. That’s how arbitrage works. That’s what we did.

We looked in the paper and we’ve got two halves. The first half is just kind of, “When do we see Google search ramp up?” We looked at things like earnings announcements, announcements of acquisitions, dividend change announcements, forecast revisions by analysts, management forecasts of earnings and found that all of these are associated with spikes in Google search. The biggest ones were earnings announcements and acquisitions.
Then, in terms of focusing on the earnings announcements, within a couple weeks to a week before, it starts going up and then spikes at the announcement. There’s kind of this process of Google search getting more intense as we know there’s an announcement coming. Many announcements are pretty much scheduled. You know when to expect them. This is something where people could know, “Hey, IBM’s about to release earnings. What do I know about IBM? What happened last quarter? What happened last year? I want to get ready for this announcement.”
Then the second half, which is kind of the more important part, is what we find if you have a firm that’s a week away from announcing a big, positive earnings surprise. What we see is when there’s a lot of Google searches, the returns that week are positive. When it’s a big, negative surprise and there’s a lot of Google search, returns are negative. The price starts moving in the right direction ahead of time when people are searching on Google. Google search is kind of proxying for people going out and getting information and getting a better expectation of what’s about to happen.
That’s kind of a nice thing where it really does seem to enhance market efficiency, this ability to go and get information. It’s not necessarily that Google search is doing it but there’s just information that the Google search kind of tells us that people are using that information to know what’s going on.

Zack: Is the ramp in search volume contemporaneous with the stock price movement? Is there a strategy there?

Darren: Yeah, that’s a really good question. That’s, again, something we haven’t looked at yet. We kind of document that people seem to use the search to know things, but we haven’t looked in at can then an investor know what’s going on? Right now, everything is kind of, “Hey, the market acts in a sufficient way. Use the search to get information.” We haven’t taken it to the step of, “Now, if I was looking at the market and observing Google search for firms, could I actually do something with that?” It’s possible that maybe you could see Google search ramping up and figure out that, “Price is going up. We’ve had some positive returns. The market seems to know what’s going on.”
A big part of it, of course, is timing. When do you get the Google search data? They make the weekly stuff easily available. The daily is often a little harder to look for. We had to really do some programming stuff to get the daily stuff. The weekly is much more easily available. As a result, you might not be able to kind of get this in time. It might not be totally implementable. We haven’t looked at it. Our first focus was just kind of, “Hey, does this show up as making prices more efficient?” It’s because people out there put the time, energy, and investment in and they’re going to get the benefit from it. Just looking at Google search isn’t necessarily going to tell you anything on your own.

Zack: How are they using Google? They’re typing in the stock ticker?

Darren: Yeah. The way we looked at it was stock tickers. I think the Engelberg paper also with stock tickers. Some people have done it where they look at names of companies but we put the stock ticker and. . .

Zack: Then you know they’re pretty much interested in the stock itself and not necessarily in the company?

Darren: Yeah, that’s the idea. The general idea is that someone might put in “Apple” and they’re looking for iPods or the Apple store locations. If we put in the ticker, maybe it’s a little bit more of finance or stock market focus. The other thing is we don’t know what they’re actually clicking on. Once they do that search, lots of things pop up. Their stock price will pop up. There might be a link to different websites that analyze it. The firm’s website can come up, the investor relations section. We don’t know that. In a way, it’s a black box. That’s one of the big criticisms that we’ve gotten as we’ve talked about this paper is, “Well that’s great that people are doing something. They’re typing a ticker in but what are they finding?”
We talked to someone who knew someone who knew someone at Google and thought they might be able to get data on what’s getting clicked on. Be surprised the people at Google are keeping track of these things and trying to figure out what they can pull from this. We don’t have that ability at this point. All we know is what the Google trends releases which is, “This is roughly how much people are searching for this term,” in this case, a ticker. One of the big pushbacks on our paper was, “Interesting, but we want to know what they’re actually looking for, what they’re searching for.” That’s what actually led us to some new research we’re doing now.

Zack: I was just going to ask about that. Are there other databases or other data sets that you can use to actually figure out what that second click is?

Darren: Yeah. What we did was we were actually presenting this at Ohio State and Anne Beatty, who’s one of our big capital markets people there, kind of just said, “What you should do is go to the SEC and see if you could figure out how people use the EDGAR database, the EDGAR website, Electronic Data Gathering, Analysis and Retrieval.”

Zack: Haven’t they moved away from EDGAR? Wasn’t there a big launch or supposed to be a couple years ago to a whole new system? Did that ever happen?

Darren: Well, what we’ve got with the EDGAR system… We have this idea of, in the past, if you wanted to get SEC filings, you had to go to the SEC reading room or have them mailed, microfiche, stuff like that. Things were revolutionized in the mid-90s when they went on the Internet and had the EDGAR website. What Anne Beatty was telling us when she was saying, “Hey, if you could go and see people pulling from that system, then you’d know exactly what they’re looking for because people click on specific things. You could get the server logs, essentially.” That’s what we went out and did.
What we found is that there are essentially two ways to get it from the SEC. You can go to the website and pull stuff. You could also, if you want, and this is what bigger institutions will do, is subscribe to a feed, an STP feed, called the EDGAR Public Dissemination System and it’s from Keane Federal Services, a company that kind of handles that process. What you can do is you can just sign up and they will just send these things out to you. As filings come in, they will just send a stream of filings to you. You have to have a computer somewhere that can just accept these filings.

Zack: It’s just a reverse chronological feed?

Darren: I think the idea is that, as they come in, they’re kind of processed and sent out. That’s my understanding. My understanding is there’s like 2,000 to 4,000 filings a day that get sent out on that thing. That’s how many filings there are. It’s just overwhelming, the information flow that goes through there.
What we did is we went to the SEC and we started calling people. Fortunately, we have Ph.D. student at Ohio State who’d worked at the SEC and he knew some names to call. We were very fortunate to talk to people there that had actually looked at something like this.

They had had a talk. I believe it was Christopher Cox, the former commissioner and chairman of the SEC, who had kind of asked about how these things were being used. How’s EDGAR being used? How much is it used? They started looking at it. They actually pulled their server logs for about six months and had done some analysis on them. It was kind of a long process. Obviously, this comes from public funds. It’s government information and you have to be very careful about disseminating this kind of thing.

After a long process going through the Freedom of Information Act and talking to lawyers and everything, we were able to get those server logs for the six months. We basically had, for six months, everyone that goes to the website to go click and pull names down. We have those. We have all the clicks and all the requests. This is for any SEC filing that any firm has sitting on there during those six months.

Zack: I’m sorry to interrupt, but do you have the ability to identify who’s actually pulling the information off?

Darren: No. What they did was, they obviously had the IP address and they blacked out the last octet of the IP address. One thing we haven’t really done much at this point, but you can run these through geo-IP databases and get at least a city location. We haven’t actually analyzed that much yet but there’s a lot of interesting things you can do with that.

Zack: Right. Meaning if there’s a lot of activity in Stanford, Connecticut, that means something different than in South Beach, Florida.

Darren: We figured Connecticut, New York, San Francisco, London or a lot of these kind of things. That’s one thing we’d like to do in the future. We can’t really get that. Obviously, there are privacy issues with that. That’s why they take off the last octet to kind of keep it a little broad in terms of where they’re coming from. Other than that, we know exactly the time and what they’re pulling. We know the firm, the actual filing. If they pull a 2008 10-K for Apple, we know that’s the exact document they wanted. It’s not that they requested something about Apple. It’s, “Here’s the exact document.” It’s an amazing database and we were very fortunate. The SEC has just been wonderful in this in helping with this.

We’ve actually talked with them. One of my coworkers presented the Google paper at the SEC, got ideas from them, talked with them while we were doing this process of getting this other data. We’ve been in talks with them to kind of go back and, hopefully, present the new research there.

It’s a really fascinating database in terms of, we look at it and we have, I think it is, several hundred million requests over this time period. What we’ve kind of found, in general, is that there’s a lot of demand for 10-Ks. That’s kind of the big one. No surprise, a lot of information in there. 10-Qs, 8-Ks and then, coming back to the insider stuff, the insider filings. That’s the next one. That’s the next biggest one. People like those insider filings and they like them quickly. They follow that in real time.

What we were interested in is how old is the information when they pull it? Do people pull 10-Ks immediately and then they never look at them again because they already have them and they don’t go back to old stuff? It turns out that, for 10-Ks, it’s often many months before they get pulled. Form 4s get pulled very quickly. The median time is like a day for pulling up a Form 4. People are out there looking at what the insiders are doing very quickly. They want to get that information right away. That makes sense. People have tried to see if you can profit on insider information. The idea is you need to know this right away. You need to know this very quickly. The information will get stale and useless pretty quickly.

Where it seems like 10-Ks, which have a lot of information for a lot of reasons, might keep their value for a long time because you could use it. For example, if next year’s 10-K is coming up, we can look at last year’s to get a base of what to expect.

Zack: Right. I guess there’s like a lot of learning information in 10- K, learning about the business, the dynamics of the business, competitors. Do you this is just sort of like a piece of research, a form of research we’re just beginning to exploit, sort of like the click stream on the Internet? These are inherently non-financial, I guess, processes. You’re not looking at balance statements. You’re not necessarily comparing P/E ratios. I’m thinking of people going to Amazon and researching products there in the capacity of using that as part of their research. As that stuff becomes more transparent or you have other ways to access it, do you think there are going to be other stories there to tell also? I don’t know if that made any sense.

Darren: No, that’s how we do it. That’s what I’m thinking about is, actually, another one of Joey Engelberg’s papers, one called, “In Search of Earnings Predictability.” I just love the paper because what they did in there was they looked for searches, not on the firm, but on the products of the firm. They’ll look in, for example, “iPod” instead of “Apple.” What they find is that when they see a ramp up in information for the products, it predicts, for example, revenue surprises to the firm. The firm’s going to have good sales because people were interested in their product.
It’s like you say, there’s kind of like these measures where it’s not that you’re looking at financial information. You’re not trying to figure out some fundamental analysis to the firm. You’re trying to figure out just, “Are people paying attention to the firm? Is there something in here that’s going to correlate with demands of the firm’s products? Is there something in here that’s going to be correlated with activity in the market for the firm’s stock?” For example, in our Google paper, that people were figuring out information so that they can price the stock at the earnings announcement. I think there are a lot of unknowns there, but this information is becoming more common, more available. We’re just getting to being able to figure out what the market’s thinking.
There was another paper that I was looking at the other day, which looked at Twitter and how the mood on Twitter can predict the stock market.

Zack: Is that Johan Bollen’s paper?

Darren: Exactly.There are some colleagues of mine at Michigan, Hal White, Greg Miller, and Beth Blankespoor, who have a paper on Twitter where they show the firms can use Twitter to disseminate information and that affects the market. Firms are more liquid when they’re putting information out through Twitter. There are just a lot of ways now. Technology’s given us ways now to figure out how people are getting information. It might not be the information itself. It’s just we know they’re getting the information or we know information’s being disseminated by the firm. Even without knowing exactly what the information is, it’s just a good thing to have information flowing. I’m sorry. Go ahead.

Zack: No, I didn’t mean to cut you off.

Darren: I was just going to say real quick that I think back to when conference calls started becoming big. Conference calls were another example where people would look. There’s this great paper from late 90s, around 2000, Franklin Johnson Skinner. We looked at conference calls and showed the market reacts to conference calls. People seemed to be trading on the information they’re getting at the conference calls. Again, it’s not so much that we have the information but we know the market’s getting a source of information and can do something at that point in time.
It’s pretty exciting how technology is giving us different ways to figure out what the market’s doing and understand better how this… Again, taking it back to the arbitrage process, people get information, they decide what to do with it, and then they act on it. We’re getting a lot more on the information part.

Zack: Apropos to that, when I did step in and cut you off accidentally, that whole conference call thing is really interesting. One of the things that I wrote about my book was I used to run business development at Seeking Alpha. We launched this service to provide preconference call transcripts. Previously, you had to go and pay for a feed somewhere to get those, pay a division of Thompson-Reuters or whatever for that feed. Seeking Alpha was disseminating them for free. There’s a search function there where you can actually search for terms across transcripts. If it’s the new Airbus or the new Boeing Dreamliner or whatever, you can do a search for “Dreamliner” and see how many companies in the ecosystem are talking about it and what they’re saying about it.
As an investor, that’s really cool. You can sort of get a view for how a company’s competitors view its products or its business or whatever. I’d love to see some information on stuff like that. Again, it’s this ability to follow the click stream cutting across all different types of media and whatever. I think we’re just beginning to scratch the surface there.

Darren: Yeah, I think it is. One of the other ways that this is kind of happening is that people are taking a lot more attention to qualitative data. The idea of being able to search for words and terms and using those instead of just numbers, we’re seeing that all over.
I’ve actually started a paper with a couple colleagues at Ohio State, Zahn Bozanic and Andy Van Buskirk. We’re looking at firms can often put out forecasts of earnings, but they can also have forward-looking statements. They can talk about the future in general qualitative terms without actually giving the number. There might be times that they want to do that. If things are really uncertain, they might be a little weary about giving you a hard and fast number that they might be in trouble for if they don’t make it. When things are uncertain, people want information. They might feel better about, “Hey, let me talk to you about the future. I’m not going to give you a forecast, but I can tell you what I’m thinking, what I’m feeling.” We actually find that when there are proxies for uncertainty, they’re much more likely to want to talk about the future than actually give a quantitative forecast.
Things like that are just fascinating now, where we’ve kind of gotten measures of this ability to kind of try and figure out what the qualitative, what the words are saying, in addition to what the numbers are saying.

Feng Li, who’s a Chicago grad who’s at Michigan now, has been kind of one of the pioneering guys on that. He’s written several papers looking at the idea of, “Can you analyze a 10-K and look at how readable it is, how complicated the language is, how much it talks about the future and what does that say about the future?” We’ve got a lot of wonderful tools now that are kind of coming online, which, no doubt, are going to be exploited by people trying to make money off of them and also an awful lot of researchers just trying to figure out how the market’s processing these things and what information there is in these disclosures beyond just the numbers.

Zack: This has been an incredibly diverse conversation. One thing I’d like to ask all the participants on the podcast is resources that you use maybe in coming up with ideas for research, tools that you use that you find yourself coming back to, keep bumping into. These aren’t necessarily recommendations, but are there some things you use in your own sort of research process that you find valuable that you could talk about?

Darren: Yeah, a big part of it is just kind of talking to other people and just kind of getting their ideas for things. The lifeblood of academic research is you go to other schools and you give talks. They come to your school and give a talk. Then we’re always reading each other’s research there. There’s the SSRN.com system, which I remember starting. That’s how old I am. I remember when that started back in the mid-90s. People can put all their papers up. Every day, you get a stream of papers and you kind of look through them and you see ideas and get ideas for, “We can do this.”

Zack: I guess somebody should do research on SSRN, right? See what people are doing with it.

Darren: Yeah, the pulse of academic research. Just watch the stream of papers come through. Another thing is just paying attention to the developments in the marketplace and in the world. Reading the Wall Street Journal, you can kind of see all of these things that are going on. In particular, you just get ideas for how things are changing in the structure of markets and that.
We just had someone, Xue Wong from Emory was just here a few weeks ago presenting a paper about the Euronext mergers with the exchanges and how that changed. Suddenly transaction costs, information costs across national trading platforms changed. They look at how that affected home bias in investment. A lot of times, what we’re doing is looking for things that changed in the marketplace and we can use that change to tell us about how markets work. Just paying attention to what’s going on.
Obviously for my stuff, Google comes along. Many years ago, EDGAR came along and we got some data on that. You just see these new developments and they just give us an opportunity to go out and look at this.

I think there’s also a bit of it where we try to draw from psychology. There are all sorts of research about behavioral biases and making mistakes in that. Maybe we can learn something from there about what we should be looking for when people go into the stock market.
I think the big thing is you just try to be engaged with as many people as you can talk to. You try to watch as many people and look at as many people’s research. You see what’s happening in the market. What you probably want to bring it back to the big questions we have. There are probably not a lot of new questions, but just old, important questions we haven’t quite answered yet. You look at all these things out developing in the world and you try to say, “Can I use those to answer some of the big questions?” The big one for me is always just, “How does information get in the stock price? How does the market become efficient?” Google and EDGAR are two recent ways I’ve been looking at that.

Zack: Cool. Darren Roustone, thanks for joining us on the show.

Darren: Oh, thanks. It was great talking with you.