Hal Varian’s Tips on Building an Economic Model
Look for your ideas outside the academic journals — in newspapers, in magazines, in TV and radio programs. A shallower analysis may be more stimulating: there’s nothing like a fallacious argument to stimulate research.
Origin
In a 1997 chapter titled How to Build an Economic Model in Your Spare Time, Hal Varian distilled decades of craft knowledge into practical advice for economists. The recommendations travel well beyond economics — any discipline that builds formal models of messy reality can borrow from them.
What it says
Varian offers three counterintuitive suggestions that run against standard research training.
Get ideas from outside the literature. The conventional view says journal papers are the best source of original ideas. Varian argues the opposite: they give you someone else’s ideas. Newspapers, magazines, and even shallow social media arguments are better fuel because they present problems in raw form, unconstrained by disciplinary framing. A bad argument can be more useful than a polished one.
Postpone the literature review. Most researchers rush to survey what has been written. Varian advises waiting a few weeks. Working through the problem first builds modelling muscle, increases the odds of a genuinely different approach, and lets your ideas incubate before they are colonised by established methods.
Talk about it early. Giving a seminar forces you to refine the idea. An audience will not tolerate turgid prose, complex notation, or tedious detail — and neither will readers. The seminar is a filter: use it to get the clutter out before it hardens into a draft.
Applied
Indian policy research often inverts Varian’s advice. Researchers begin with exhaustive literature reviews and end with models that replicate what is already known. Policymakers, meanwhile, form mental models from anecdote and ideology without ever testing them formally. Varian’s middle path — start with the messy real-world observation, build your own representation, then check the literature — would improve both camps.
The seminar rule is especially valuable in institutions where feedback is ritualised and gentle. A model that cannot survive a sceptical audience will not survive implementation either.
When it falls short
Varian’s advice assumes the researcher already has some modelling facility. For beginners, delaying the literature review can mean reinventing wheels badly. And not every policy question deserves a formal model. Building a utility function to justify a ban on child labour would be grotesque. The advice is for problems where simplification genuinely clarifies.
Further reading
- Varian, H. R. (1997). How to Build an Economic Model in Your Spare Time.
Originally explored in A Framework a Week: Hal Varian’s Tips on Building an Economic Model on Anticipating the Unintended.