Jeffrey Pfeffer has written an interesting post summarising research by Boris Groysberg of Harvard (found via Felix Salmon). The research looks at the impact of banks hiring away star performers from competitors. Here is the description of the research:
Grosyberg studied 1,052 stock analysts who worked for 28 U.S. investment banks over the period 1988 through 1996. He found that when a company hires a star away from another firm, the star’s performance falls (46 percent of the research analysts did poorly in the year they switched jobs and their performance remained lower even after five years), there is a decline in the performance of the group the star joins, the market value of the company hiring the star falls, and the star doesn’t stay with the new employer for very long.
Pfeffer and Salmon go on to discuss this research in terms of what it means for salaries in finance. Pfeffer suggests that the best strategy is to grow your own talent, and Salmon says it means that high salaries in finance are unjustified. I can see the logic in both of their arguments, but I draw a completely different conclusion, based on a related piece of research by Groysberg titled “The Effect of Colleague Quality on Top Performance” (co-written with Linda-Eling Lee). The summary of this piece is:
We show that top performers do not own their performance, even in the knowledge-intensive work performed in this professional business services context. While an individual’s past performance does indicate future performance, the quality of colleagues in one’s organization also significantly affects top performers’ ability to maintain their performance. Specifically, top performers in professional business services rely on high-quality colleagues both to improve the quality of their own work and to deliver it effectively to clients.
In other words, performance depends in large part on the network in which you’re embedded. And I contend that while this is true for financial analysts, it generalises to all economic ideas. The success of an innovation depends in large part on where and how it is embedded within the economy. Why wasn’t Charles Babbage’s Difference Engine the first working computer? In large part because there wasn’t any good way to embed it within the rest of the economy. In particular, there were insufficient machining skills available to make a workable version of Babbage’s plans. The idea and designs were good – we know that because the London Science Museum was able to manufacture a working Difference Engine in 1991, and it did what it was supposed to.
When you have an innovative idea, one of the most important questions to ask is ‘how does this fit into the economy?’ This requires you to think much more broadly – you can’t only consider the technological challenges involved with your idea. The innovations that spread are the ones that build an effective network around themselves.
“Specifically, top performers in professional business services rely on high-quality colleagues both to improve the quality of their own work and to deliver it effectively to clients.”
Hi Tim, this reminds me of other findings suggesting that disruption in one’s social network is detrimental to performance more generally (e.g. Srikanth Paruchuri’s paper below). It would be interesting to try and reconcile these findings with those of the Uzzi and Spiro type that argue it’s about balance (in diversity; turnover etc). Surely some degree of star turnover and hiring of outside starts is a good thing? I think the missing component is the types of problems they’re meant to be solving and the subsequent knowledge and social structure required to do this. I can think of a few examples from the research we’re currently doing where the same degree of turnover would have very different impacts depending on the group it occurred in.
Paruchuri, S., A. Nerkar and D.C. Hambrick. 2006. ‘Acquisition integration and productivity losses in the technical core: Disruption of inventors in acquired companies,’ Organization Science, 17(5): 545-562. pdf
I think that’s exactly correct Sam. Both Pfeffer & Salmon use that research to push their own specific barrows. I guess I did too, but I like mine better!
Tim,
The question that came to my mind is whether the kind of economy star operate in makes a difference to these findings. In particular, whether it only applies to economies of scarity rather than economies of abundance.
For example, research papers are economies of abundance because papers become more valuable as they get more disseminated and cited (regardless of whether they are good or bad). In a department where everyone is a star or a quasi-star, a pretty good player’s research output will be “crowded out” by the stars’ performance. Stars and quasi-star will tend to repulse each other (by going to separate departments or specialise in some research sub-discipline) rather than build on the agglomeration of talent.
Of course, stars will do their bit promoting and enhancing the quality of the papers of their lesser-known colleagues (they already have well established networks to disseminate their work). on the other hand, stars don’t need their lesser-known colleagues to promote their work (although they might use them to refine their papers — sometimes!).
But overall, it seems to me that the effect is less clear cut in such economies and mitigated by psychological factors.
This is an interesting point Marco,
“The question that came to my mind is whether the kind of economy star operate in makes a difference to these findings. In particular, whether it only applies to economies of scarity rather than economies of abundance.”
I’ve often wonder about whether this type of effect had an impact on Mark Granovetter’s findings about the strength of weak ties when searching for a job. Would the results have been different during the depression? hmm