Wednesday, 16 March 2011
Modelling university behaviour and choices
Earlier this week I was reading a very interesting theoretical economics paper on the incentives and constraints that influence universities' choices between teaching and research. The paper is by John Beath, Joanna Poyago-Theotoky and David Ulph and it is entitled: 'University Funding Systems: Impact on Research and Teaching'. The paper is available as a Discussion Paper at the Economics e-Journal, and you can download it using this link.
You might imagine that theory has no place in thinking about our universities, or that if it does it should not be economics, but perhaps sociology, political science, anthropology, etc. But it seems to me that this is a big mistake, not because I believe economic theory has all the answers, but because I find that building an economic model, even quite a simple one as in this paper, often helps to illuminate important questions, and sometimes in quite unexpected ways. It is always easy to criticise a model by claiming that this or that assumption is unrealistic, and in a strict formal sense such criticism is nearly always right. But in my view, it is also nearly always irrelevant, because it misunderstands how models work, why we build them, and what we think we can learn from them.
The present paper illustrates these points quite nicely. The model makes a variety of assumptions, some of which clearly relate to core aspects of what universities do and how they function, while other assumptions are more technical and are made to ensure analytical tractability. After all, there isn't great value in a highly complicated model that is considered to be 'correct' in some sense if no one can solve it or derive any interesting conclusions from it. The first type of assumption includes the idea that universities engage in teaching and research, and they have preferences over these two 'outputs' or 'products'. Where the preferences come from remains unexamined, but different universities in the system can place different weights on research as against teaching. The ideas that universities need to produce teaching of at least some minimum quality level, and that quality is costly in terms of academic staff time, are other assumptions built into the model. Achieving research quality is also costly, institutions face a budget constraint implying that there are real tradeoffs between teaching and research, and individual academics have a time budget - if they spend more time on research, less is available for teaching, and conversely.
Most of this seems to me highly reasonable, at least as a useful starting point for analysing universities. Of course, one might want to extend it by modelling aspects of system-wide competition, or by having multiple disciplines (let's say, at least two, to start with the simplest case) in each institution competing for resources. But even the present paper, with reasonable values for the main parameters, reaches some intriguing conclusions. Most notably, it finds that universities are likely to divide into two groups, one of which opts for a very research intensive strategy (the research elite), the other consists of institutions placing greater weight on teaching - they may still do some research but it is not such a central part of their portfolios. Since this is not a terrible approximation to UK reality, I thought it was interesting to see it drop out of a model with such a nice, intuitively appealing and simple structure.