Teddy DeWitt (UMass) interviewing Nathan Wilmers (MIT) and Maxim Massenkoff (UC Berkeley), Winners of the OMT Best Paper Award, for “Wage Stagnation and the Rise of Merit Pay, 1974-1991”
[Note: this interview is an edited transcript of an email exchange with the two authors. Both authors are therefore listed together in all the responses.]
Teddy DeWitt (TD): Thank you so much for agreeing to speak with us about this great paper. To kind of set the stage, how did the two of you come together to collaborate?
Nathan Wilmers & Maxim Massenkoff (NW & MM): Thanks for reaching out to interview us about it, Teddy. We actually met in Washington DC, before we subsequently went on to graduate school. We’ve both had a long interest in labor market and wage issues.
TD: If one hears the phrase “merit pay,” one might think there is a likelihood of fair compensation for effort in your career. But your paper finds surprisingly that the introduction of merit pay has actually contributed to a stagnation of worker pay. Can you say a little bit more about this surprising finding?
NW & MM: Yes, that’s right. In our data, workplaces that pay their workers with a merit system actually pay less than comparable jobs that have pay set according to seniority or pay that is invariant within job titles. We also find that when workplaces adopt merit pay, real wages decline on average. There are a few things going on here. But we think an underappreciated aspect is that merit pay systems can make employers less likely to pay so-called efficiency wages. Merit pay allows them to provide performance incentives to workers, rather than paying all their workers—high and low performers—a little extra for retention purposes.
TD: Do you think that this is an unintended consequences of merit pay systems, or is it that as these systems were implemented companies understood the potential for these systems to more easily shift compensation dollars away from lower-level workers to higher-level workers?
NW & MM: Great question and tough to say with our quantitative data. We read a lot of practitioner publications from the 1970s to the 1980s (the period we’re studying). Certainly managers were trying to reduce compensation among blue-collar workers during this time. And they saw merit pay systems as a way to increase incentives among employees. We saw less evidence that merit pay was implemented with this explicit justification, though. That would definitely not be wise implementation messaging!
TD: It seems like that, in order to see the stated effects of merit pay in a given occupation, both within and between establishments, you would need access to a pretty unique dataset. What is this data set and how did you negotiate access?
NW & MM: It is a survey of private-sector employers of blue-collar workers, conducted by the Department of Defense. We are really excited about the data because it is rare to have occupation-by-establishment data in the US from this critical early period of wage stagnation. So, these data were gathered for an idiosyncratic purpose (setting pay for Department of Defense employees). But we use it to shed light on this big, important question about the dynamics of wage stagnation during the period.
TD: You mentioned in your work that researchers have been generally unable to use macroeconomic trends to explain wage stagnation. Was that lack of a macro-explanation a motivating factor for the this study or was there something else driving this exploration?
NW & MM: Yes, this is a huge question. It’s at the root of a lot of concern about rising inequality: the problem isn’t that the top 1% have done amazingly well over the last 40 years. The problem is that the bottom 50% have had very slow pay growth. We know about some important macro changes during the 1970s/80s period – stagnant minimum wages, business cycles, declining unions, economic globalization – but it’s a lot harder to study organization-side changes (due to data availability). We thought this is a great opportunity to do that.
TD: How would you turn these findings into potential practice recommendations for both unions and worker associations who might want to reform pay systems, but also managers who see that “merit pay” systems may not be in the best interests of employees?
NW & MM: We imagine that there’s a lot of heterogeneity in how performance-based pay systems are rolled out and implemented across companies. It’s hard to get a sense of those details in our data and we’re hoping that future research will do more on this. But, our guess is that the details of implementation matter a lot for figuring out how to use incentive systems without undermining worker bargaining power.
TD: What did you enjoy most about this research project?
NW & MM: Dusting off these old, exciting data. Discovering that the data had identified firm names, that we had repeated panel observations of firms and jobs … all of this stuff was critical to the success of our project. And we had no idea, when we originally received the data, what it would include.
TD: Would you like to share any challenges you faced during the research and writing process? If so, how did you overcome them?
NW & MM: This project lies at the intersection of labor economics and organizational theory. Nate is a sociologist and Maxim is a labor economist. So, a big challenge has been figuring out how to find the audience for this work and how to present the findings in a way that speaks across academic disciplines. We’re still working on this, so no great solution we can share. But, a first step, as with any audience issue, is just getting feedback from a diverse array of readers.
TD: And last but not least, what are the next directions that you intend to follow after this paper? Did it prompt more questions that you wish to pursue? Do you think there are other questions that you can potentially address with this dataset?
NW & MM: Definitely. We’ve been thinking about a couple extension projects with these data. In one, we’re working with matched EEO-1 data [information required to be submitted to the Equal Employment Opportunity Commission], so we can see something about how demographic changes and race/gender integration in these industries also affected wage dynamics.
TD: Thank you so much for your time! It was really informative to hear about your process, the dataset, the motivation for the work, and the unintuitive finding. And I know that the right audience is waiting for this. We look forward to seeing it in print!