Daphne Teh – INSEAD
Question 1. Your paper makes rich contributions to the literature on social movements and non-market strategy, theoretically as well as empirically. It’s evident that this paper draws upon both of your expertise, and I’m just wondering what inspired you to start doing joint work?
TW: Even before the two of us coming together to work on this, this paper originated, I guess, in the summer of 2012. I was reading an article on the website Politico that was entitled “How Corporate PACs Deal with Bad PR.” I was reading it, and it seemed full of testable propositions. As somebody who studies corporate political activity I found it really fascinating, but then just put it on the back burner as a research idea, until Mae came to give a talk at McCombs in the spring of 2013. I said “Hey, you study boycotts, here’s this article, and I have all this data on what firms are up to politically. Why don’t we merge our data sets and see what happens?” Lo and behold, a lot of these relationships that politicians were speculating about, we found empirical evidence for.
MHM: Now that I’m on the editorial board at ASQ, it’s interesting to me because we really value dissertation quality work, and we prioritize publishing dissertation work. Although this paper wasn’t Tim’s or my dissertation, it’s a fun project because it actually combines the data that I put together for my dissertation with the data that Tim had put together for his book project, so it’s kind of like putting two dissertations into one for a paper.
I think it’s the kind of project that would be impossible without us both having put together those data sets already, because Tim wasn’t your data two years in the making? And so was mine.
TW: Yes, at least two years.
Question 2. How do you think associative risk affects political stakeholders differently, for different relations with target firms, and why is that?
MHM: I think the different types of political strategies that firms use to try to reach politicians are either more or less direct in signaling exactly what politicians they’re interacting with. So if you’re looking at something like the government awarding a firm a procurement contract, it does create an observable link between government officials and firms, but it’s not as direct a link as, say, accepting campaign contributions, where it’s clear exactly why a politician is in agreement with interacting with the given firm. So I think that those tactics that not only create an observable link between politicians and firms, but also make it clear exactly what politicians are willing to interact with firms, are probably the ones that are most conducive to the associative risk.
TW: Yeah I would agree with that, in the sense that one of the differences between the campaign contributions, the rejection of them that we analyze, and say the awarding of appearances or contracts is that the first category—the rejection—it’s a much more active thing the politician has to do to convince themselves than say not award a contract or not invite a firm to testify.
Question 3. I found the rigor of your paper impressive, with the extensive data collection, difference-in-difference methodology, and selection model. What was your biggest insight on developing a strong identification strategy?
TW: I think that the nice thing about the way in which we identify the effects of social movement challenges is that we have this set of boycotts that Mae had collected, and if you look at the substance of what the boycotts are about, there are all sorts of contentious practices that firms engage in that aren’t related to politics, so it gives us an opportunity to say, “Hey, what is the effect of this shock to the firm’s reputation on its ability to access political stakeholders?” without the disruption being due their prior political activity. It allows us to avoid indoctrinating.
At the same time, it allows our mechanism to work because if the firm is controversial, the politician will still want to distance themselves from it, because either the public will be paying attention or the politician’s challenger in a future election will be. So there is a reason to believe that the boycotts won’t matter as substantively for politics, yet still it will allow us to meet the assumptions that we need to for identification purposes.
MHM: Since our paper came out, I have been a reviewer on a number of papers that are trying to use a similar strategy, and it is crucial to this empirical strategy that the treatment you’re using is exogenous, arguably, to your dependent variables to the greatest extent you can ensure.
I think a mistake that it seems like a lot of researchers inspired by this and other work, wanting to engage in some kind of a quasi-experimental method, is that the treatment that they are using is not exogenous to the context that they’re studying. It would be similar to us, for example, using a sample of boycotts that were about a firm’s political behavior. Luckily, that was something that, it’s happening more recently, but it wasn’t happening in the time period that we’re studying in our paper. But you can see how that would be a problem, to use that as a treatment, because it’s not exogenous to the context per say.
Question 4. What did you learn from bringing this paper through the R&R process? What was the biggest challenge?
MHM: Originally, our paper was only looking at campaign contributions. Campaign contributions are limited under regulations, you can only give a certain amount of money in each election cycle. That also limited the effect sizes that we could observe, with that as the dependent variable, and so the primary difficulty that we had with the reviewers is in convincing them that these arguably small effect sizes we were seeing were meaningful. So the reviewers pushed us to look for other indications that this was meaningful, and that inspired us to go out and find the data on congressional appearances and on government contracts. And I’m so glad that we did, and that the reviewers pressed us in that way, because it does do a much better job of demonstrating how generalizable this effect really is across the full set of variables that signal a firm’s involvement in politics.
TW: One of the things you learn in going through the R&R process is when you’re working in an area, say, where a journal hasn’t published a lot in recently—so in political activity, there hasn’t been a ton in ASQ in recent years—you have to do a bit of educating of the reviewers, right? So like Mae was saying, the effects that we were getting didn’t seem economically significant just for the campaign contributions, because the dollar amounts were small, but they were large in the context of that space.
That’s something that I learned in this process was that, when you have a set of reviewers who know about the topic generally but not, say, the nitty gritty of the law in an area and how that might be constraining the effects you might get, it’s your job as an author to provide that context during the review process.
MHM: One of the wonderful things about ASQ is you can be fairly certain going into a review process that the reviewers are going to help you make a much better and more compelling paper.
Question 5. As you discussed in your paper, targeted firms lose political influence due to associative risk. In what way does this relate to or is distinct from stigma by association?
MHM: I think that’s a great question. The distinction that I make in my mind is that associative risk is deduced from, so it’s not separate from stigma by association, but it works in a little bit different way. Stigma by association is retrospective; so a scandal occurs, and immediately people look at the retrospective ties that individuals or organizations had with each other, and then they use those ties as information to decide who to distance themselves from. If you look at, say, Elizabeth Pontikes et al.’s paper on mere association effect, they’re finding that the Hollywood red scare led to all actors that had worked with actors that were blacklisted, having negative effects in their likelihood of being put on movies on an ongoing basis. So it’s a retrospective mechanism that drives the mere association effect in that way.
What we’re arguing is that associative risk is a forward-looking mechanism. So basically, a scandal happens, and actors that have not yet had a tie with an organization then are less inclined to make ties with them going forward, because they’re afraid of those spillovers if another scandal or bad event occurs in the future.
TW: I get the sense from the work on the Hollywood blacklist that stigma by association is essentially almost a permanent condition once it’s applied, whereas a lot of our effects hold for, say, a couple of quarters or a year or so, but the risk isn’t a constant once it’s applied. The risk is variable going forward in a way that the stigma is not.
MHM: I don’t actually think that’s been tested directly in the mere association literature, so I’m not sure if that would be true of the retrospective stigma either. If it’s true that once you are tainted, you’re tainted forever, or if the longer period of time you’ve had since the stigmatizing event would actually mitigate those effects.
Question 6. Do you have any tips or advice for young scholars thinking of doing interdisciplinary work?
MHM: My advice would be to find an amazing collaborator. [laughs] I really lucked out in finding a Tim Werner and lucking into this project with him. We’re going to work on more things going forward, because we’ve just had such a great and fruitful collaborative relationship.
I think a lot of folks feel pressured when they’re doing interdisciplinary work to do it all alone, but the nice thing about finding a collaborator who comes from a different discipline is that you don’t have to be a master of more than one trade. You can still just come from your discipline and bring the best tools that your discipline prepares you to use. But then you learn how to combine those tools with another person who has mastered them in a different discipline.
TW: Kind of the key thing is that complementary of skills, of data, of knowledge of the literature in the different fields that you’ll be addressing. We had two different sets of skills that worked out perfectly here in matching and working on this project.
Bigger picture, I think one of the key things for interdisciplinary scholars is to keep in mind where you’re likely to send them [projects], the reception they’ll receive at various journals, and tailoring your work to some degree to the journal and its audience, but also thinking about the broader audience you want to reach—if you’re a junior scholar, the senior scholars working in your area who might be reviewing your work for tenure, etc. I think interdisciplinary work is really appreciated but sometimes can be more difficult for folks to evaluate as a whole.
The last piece of advice I would give is to think about each project you work in, whether it’s disciplinary or interdisciplinary, within the context of your broader research agenda and make sure that it fits.
Pontikes, E., Negro, G., & Rao, H. 2010. Stained Red: A Study of Stigma by Association to Blacklisted Artists during the “Red Scare” in Hollywood, 1945 to 1960. American Sociological Review, 75(3): 456-478.