Madeline King – New York University, Stern School of Business
Derek Harmon – University of Southern California, Marshall School of Business
Question 1. Integration is a longstanding topic with decades of research already in place. What inspired and/or motivated you to reengage these questions from a relational or network perspective?
To be honest, the original idea we had for this project was a bit different. We wanted to use the merger event as an exogenous shock or natural experiment, to see how networks change when people are exposed to new potential ties and a new social environment. That idea is still present in the paper, as described in the Network Change section of the discussion – but it’s not the main framing of the paper.
Instead, as we got into the project, our focus shifted to thinking about how we could shed light on the human side of the post-acquisition integration process. This is critical for merging law firms (as our interviews revealed) – as well as most other types of merger & acquisition (M&A) events.
For this purpose, law firms make a great strategic research setting, because we can quantitatively capture the human integration process using data on lawyers’ client sharing networks. When we drew pre- and post-acquisition network diagrams for the law firms in our study, they revealed just how much organizational integration is indeed a relational phenomenon. If an organization’s network is fragmented with unconnected components, then the organization is not integrated.
Plus, focusing on integration allowed us to start our paper with a reference to Chester Barnard.
Question 2. You reference “quantitative ethnography” as your methodological approach. Could you speak to the relationship between your quantitative and interview/ethnographic data? Did these interviews help in the development of your hypotheses? Did they supply supporting evidence that enabled you to tell a richer story?
The short answer is, yes, field interviews were central to our research process, playing a role at multiple points in the research process. Of course, interviews gave us a deeper understanding of the study context, and we iterated between interview data and our quantitative data as we sought to clarify and interpret our findings. But interviews also helped in three other critical ways that are perhaps less obvious:
First, the original spark for even thinking of an acquisition integration study actually came from interviews. Early interviews with leaders in our case study law firm revealed that, despite being part of an unprecedented merger wave in the industry, these extremely smart and experienced folks were puzzling about how to get newly merged partners to share clients and collaborate with one another. Corporate law firms were essentially caught in a cycle of feeling the need to grow through acquisition – but not having a clear view of how to realize benefits from integration. So here was an interesting tension in the real world, deeply connected with organizational theory, and in a setting where we could easily measure the integration/collaboration process.
Secondly, the very process of conducting interviews, and discussing the industry’s challenges with members of the case study firm, served as an occasion for building trust and understanding. This turned out to be critical in order for the firm to later agree to share their quantitative data for research purposes. Without investing in those many hours of interviews and meetings, we doubt the study would ever have been possible.
Thirdly, our later follow-up interviews turned out to be useful for interpreting findings that we were not expecting. Specifically, interviews we conducted with current and former law firm associates helped us realize some factors that linked inter-unit client sharing with negative human capital development outcomes. In essence, work on these inter-unit client sharing projects tended to be very stressful, and partners and associates engaged in them often scrambled to keep up, potentially offsetting the potential benefits of new learning and growth opportunities.
We would advocate for more quantitative researchers to get themselves into the field. For us, talking with people in real organizations helps inspire, validate, and sometimes even redirect our work. It also makes the work feel more relevant. And, it’s just plain fun. (By the way, this kind of field research is the core tenant of an expanding professional group called the Industry Studies Association, which evolved out of projects at MIT during the 1980s exemplifying high-engagement field research.)
Question 3. Your paper teased apart the distinction between interunit and intraunit effects. How helpful (or necessary) was this distinction to provide you with more tractability when uncovering the microprocesses of integration?
The distinction between interunit and intraunits effects allows us to see where integration takes place in the organization. Intuitively, we probably tend to focus on interunit connections in the merger-acquisition setting, since we clearly hope to see people in the acquiring unit working with people in the acquired unit. But, if interunit connections are formed at the expense of intraunit connections, then the organization may still not be well integrated. We found that people spanning structural holes in the past are likely to form interunit connections but cut intraunit connections. The finding provides food for thought about the micro-processes of integration.
Question 4. You mention that one of the strengths of the data is that billable hours in the law firm are “obviously free of social desirability bias and related concerns that arise from survey based network data.” Have you found empirical or anecdotal evidence that social desirability plays a role with whom partners decide to share their work?
Social desirability undoubtedly plays a role in shaping inter-firm sharing networks among lawyers. Law firms are rife with status, reputation, power, and other social dynamics, as Emanuel Lazega and others have well documented.
Our point here was about the network measurement process: in this case it is not susceptible to social desirability bias. On social network surveys (the mainstay of much networks research), people may say they are friends with popular colleagues (even when such a friendship tie is not reported in the network surveys of others in the organization, including the popular colleague herself). Those surveys are also susceptible to other forms of bias related to salience, recall, cognitive limits, and so on. By contrast, our network data is objective and complete to the extent that it reflects mandatory reporting data entered into the firm’s internal system. (Of course, partners anecdotally also play games with client billings reporting data for various legitimate and illegitimate reasons, and that could be the subject of a separate fascinating study if one could somehow devise a way of detecting this kind of behavior.)
Question 5. One of the key tensions in your paper is that interunit client sharing increases revenue generation but decreases human capital development. This tension however was not predicted (H5 predicts an increase in human capital development). Was this strategic on your part to frame these hypotheses so as to find the opposite and create narrative tension? Or did you believe that the predicted argument, though ultimately not supported, was actually stronger?
Great question. Truth be told, we don’t quite remember what we initially believed at the start of the study (that was a while ago!). But we do think that an expectation of increases in human capital development would be most strongly supported by prior research, based on the common finding in the networks and social capital literature that exposure to new opportunities is critical to building your own human and social capital.
Though as we mentioned above, on talking with lawyers in the field, the logic of the opposite (surprise) finding was readily apparent.
Thanks for this chance to reflect on our project!