Question-First Economic Research
Start with a real-world question, then choose the method that answers it
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
Question-First Economic Research reverses the temptation to begin with an elegant model and search for something it can explain. The researcher first states a consequential question about the real economy, describes the proposed mechanism intuitively, and identifies the evidence needed to distinguish among possible answers. Only then does the researcher choose equations, econometrics, or other methods suited to the problem. Empirical results are judged by whether they improve understanding of the world, not merely by technical sophistication. Clarida adopted this approach after spending several early-career years trying unsuccessfully to produce breakthroughs through abstract mathematical work. The method preserves rigor while making methodology subordinate to inquiry, and it offers a practical guard against reverse-engineering economic stories around fashionable tools.
Origin
Clarida credits Harvard economist Benjamin Friedman with teaching him that good economics begins with the questions asked rather than the methodology used. Clarida adopted the principle after reconsidering his own early focus on abstract mathematical economics.
Core principles
- 01Important questions outrank impressive techniques
- 02Mathematics is a tool rather than the purpose of inquiry
- 03Intuition should precede formal modeling
- 04Models must help explain the world as it is
How to run it
- 1
Frame the question
Write the real-world question in plain language and explain why answering it matters.
Pro tip Make the question understandable before introducing disciplinary jargon.
Watch out Do not disguise a preferred technique as a research question.
- 2
Build an intuitive mechanism
Describe how the relevant actors, incentives, and constraints could generate the observed outcome.
Pro tip Teach the intuitive version before presenting equations.
Watch out Intuition is a starting hypothesis, not a substitute for evidence.
- 3
Specify discriminating evidence
Determine what observations would support, weaken, or falsify the proposed mechanism.
Pro tip Identify competing explanations before examining the results.
Watch out Avoid evidence selected only because it is convenient or available.
- 4
Choose the methodology
Select the model, data, and analytical technique that best answer the stated question.
Pro tip Use the simplest method capable of resolving the uncertainty.
Watch out Technical novelty does not compensate for a poor fit between method and question.
- 5
Confront the messy world
Compare theoretical predictions with empirical evidence and investigate discrepancies rather than hiding them.
Pro tip Treat model failures as information about omitted mechanisms.
Watch out Do not force observations to fit assumptions made for mathematical convenience.
- 6
Refine and communicate
Revise the question, mechanism, or method as evidence demands, then explain the result intuitively before presenting technical details.
Pro tip Return explicitly to the original question in the conclusion.
Watch out A technically correct result can still fail if its practical meaning remains unclear.
In the wild
Clarida spent about five years expecting an abstract mathematical breakthrough. He then recognized that his motivation was understanding the actual economy, shifted toward empirical work, and made questions rather than methodology the organizing principle of his research.
→ His research became more closely connected to observable economic problems.
A researcher using the framework would first ask whether weak measured productivity reflects real underinvestment, measurement error, or both. The researcher would then select data and methods capable of separating those explanations instead of beginning with a favored model.
→ The analysis addresses a meaningful ambiguity rather than showcasing a technique.
Common mistakes
Starting with equations
Beginning with a preferred model encourages the researcher to reverse-engineer a question that fits it.
Equating rigor with complexity
A more mathematical method is not automatically more informative or appropriate.
Ignoring conflicting evidence
Forcing the real world into a model conceals the very discrepancies that could improve understanding.
Is it for you?
Best for
It is best for economists, analysts, and researchers choosing among quantitative methods.
Not ideal for
It is not ideal for purely mathematical research whose explicit purpose is advancing formal theory.
From the transcript
“Ben always thought good economics was about the questions that we asked, not about the methodology we use.”
“I always try to start with with an interesting question and then let the methodology work out as it will.”
“I really realized that what I got into economics was to actually be trying to understand the world the way it was, not the way…”
From the episode
Richard Clarida of Pimco on the New Neutral of Monetary Policy
Richard Clarida