By Myrna Wooders
It's always difficult to respond to the question of how I do research since I'd rather do research. But I'll give it my best shot. How to start doing research may be the biggest question for students so let me begin there.
As a PhD student and research assistant at the University of Minnesota, I worked on a simulation model of the Twin Cities, Minneapolis/St.Paul. My major task was to evaluate the model. The Twin Cities were modelled by their neighbourhoods and measures of amenities provided in each neighbourhood, with some adjustment process describing the movements of individuals in response some change. A single change in the amenities could generate an unbounded number of changes in population distribution in subsequent periods. I tested the model by making changes in the parameters and then looking at the results in the long run. The outcomes seemed almost random. I threw up my hands in dismay -- and wondered how to model a city with multiple regions.
Sitting at my office window, looking out across the Mississippi River at downtown Minneapolis, it seemed to me that any model of a city should have people affecting other people. People make a city different than a farm. On a farm in a sparsely populated rural area, unless your neighbour hasn't cleaned his barn for a long time and the wind is blowing in the wrong direction, you're quite separate from your neighbours. Not so in the city! But while one may be adversely affected by the pollution of the traffic or the smells from garbage accumulating during a strike of city employees, the joy of city life is also the people. Streets in London are crowded with colourful characters; interesting shop windows abound. There are enough people so that entertainment -- and food -- can be diverse. Museums are handy. People together, bustling, talking, hurrying, have an energy and excitement that has no match. In general, depending on the activity and the costs, there may be too many or too few people but in a city, one can pick and choose. Most public swimming pools almost always have too many people for cautious and awkward swimmers like me. Most operas that I've attended have the right numbers -- not too many people to significantly interfere with enjoyment and enough people generating a lively environment and sharing the costs. I imagined an individual's utility may be an initially increasing, but eventually decreasing, function of the number of people in his community.
But how to start a model? Actually, that was almost done with the last line of the above paragraph! For a very simple sort of model, suppose all individuals are identical, assume all costs of public amenities are shared equally, and assume that the utility of a representative individual, as a function of the number of people in his community, has a maximum -- visualise a "utility hill". I had a few more ideas about this, expressed these as simple examples, and talked to other people about the problem. One person mentioned a paper by Paul Samuelson describing some results of M. Joseph (Margaret, I'm told) on multi-plant production. I turned the Joseph/Samuelson diagram "upside down. Their average cost curves became my utility hills. Constant costs for firms with many plants became constant per capita utilities in economies with local public goods (or clubs) and many people. Minimum average costs became maximum per capita payoffs. I worked at turning intuition into results and reading the literature to determine what I might be able to add. This led to my thesis, published in part in the Journal of Economic Theory with a subsequent paper in Econometrica. This line of research still occupies me today.
When one starts to ask questions of models, one thing just leads to another. Some results for economies where it is optimal to have multiple jurisdictions or clubs producing public goods seemed to depend on only a few features of the economies, in particular, the feature that all or almost all gains to collective activities can be realised with many relatively small clubs or jurisdictions. I wanted to share my intuition. How could the intuition be expressed and conveyed was the question. This led to a model of large games with small effective groups of players, many extensions and generalisations of that model, and results that show that small group effectiveness is a fairly fundamental property of competitive economies, including ones with indivisible commodities, clubs, local public goods, nonmonotonicities and so on.
In conversations with another colleague, we wondered about price systems. Need prices for local public goods with differentiated crowding -- where different sorts of people affect others differently -- depend on preferences? That was several papers ago. Others have now adopted our approach.
Last year, for a small course assessment, I proposed to a student that he examine the robustness of some results in a paper inthe American Economic Review to changes in the properties of the cost function.The student did a good job, read the relevant literature, and it's expected that the paper will eventually appear in a refereed journal. Much research, I believe, arises in precisely this way -- from reading the literature and from asking questions of the literature. What if the cost function exhibited decreasing costs? What if governments taxed capacity rather than output? My hope is that in Microeconomics B, students will learn how to ask such questions of models in the literature and how to start to answer them.
Sometimes research arises from looking at life and wondering about things. I wonder about how female survival strategies -- in any environment, including professional -- differ from male strategies and why. I had some conjectures, read some literature to help refine my own views -- whether I agree or disagree, for example. And now have one paper on this topic, as yet unpublished. (But I have hope.) Approaching research in this way is trickier and more time-consuming. I had to read much literature on primates (they now fascinate me). There was no obvious journal in which to publish this paper. And so on. In any case, I want to understand the question.
These days I do think about how my research has affected my career path. Other than that it is necessary to publish, thinking of this aspect of my research was something that I've been slow to do. My current research on tax competition must surely have been influenced by the fact that many others are working in this area and one is drawn into debates and scientific discussions. If you want to have many citations, then it helps to work in a "hot" area. But really I can't imagine deliberately choosing an area because it's hot. The fact of the matter is that doing research requires great interest in a topic; without that interest, it can be painful!
This is more, perhaps, than anyone who wants to do research wants to read about it and I have papers to write. So let me try to summarise. I start with a question. Then, typically I wonder about it, think about it, ask how something could be the way my intuition or observations tell me it is. I start to think about how one might develop a model to use to test my intuition, or how one might modify an existing model. (I use the word "one" deliberately here; I wonder about too many things to hope of actually working out models for some of them.) Talk to colleagues to see what they think about the perplexities and if anyone else has written on the topic or related topics and what they have do say. Eventually maybe I actually start to write down a model. Then, typically, if not before, I start to look at the literature. And so it goes.
But having an idea and testing the logic and assumptions behind it with a model do not, by themselves, constitute a contribution to science. So that the reader can fit the research into the context of the research already done, it's necessary to search for and understand related literature and include informative reference to that literature in your paper. Also, one might find that some basic questions and models have already been introduced -- my single utility hill was already in a 1965 paper due to James Buchanan. (It sometimes seems to me that in this era of league tables and citation indexes, the scientific purpose of citations is ignored. In addition to recognising the prior work of others, citations are to help the reader understand. They should also make clear your own contribution.) The research must be sufficiently well written so experts in the area can easily understand it. Simple clarity in writing is itself a big part of making a scientific contribution -- which, although it is neither necessary nor sufficient, typically boils down to getting the research published.