Home Economist Interview questions
Economist interview questions
The questions people actually get asked for this role, and what each one is really testing.
What to expect
Economist interviews mix technical modelling questions with tests of how well you can communicate results to a non-specialist audience. Expect a panel format, often with someone from a policy or research team and someone from a more general management background.
- Technical modelling: Questions probing your grasp of econometric methods, model choice and the tools (R, Python, Stata) you use to build and test them.
- Policy scenario: Hypothetical policy or regulatory changes where you're asked to reason through likely economic effects on the spot.
- Behavioural: Questions about past research or analysis projects, used to check how you handle ambiguity, deadlines and disagreement over findings.
- Communication: Questions testing whether you can explain a technical result to a non-technical stakeholder without losing accuracy.
Interviews typically open with background and motivation questions, move into a technical or scenario-based segment (sometimes with a short written or take-home exercise beforehand), and close with behavioural questions and a chance for you to ask about the team's current research program.
- 1
Walk me through how you'd build a forecasting model for a given economic indicator, from data to final output.
Why they ask: Tests whether you understand the practical steps of economic modelling, not just the theory, and whether you can explain your process clearly.
How to structure your answer: Walk-through: describe the sequence from data sourcing and cleaning, through model specification and validation, to presenting the forecast and its uncertainty.
Example answer
“I'd start by pulling the relevant series, often from the ABS or RBA, and checking for structural breaks or seasonality issues before doing anything else. Then I'd specify a model, usually starting simple and adding complexity only if the data supports it, and test it against a holdout period. In Stata or R, I'd run diagnostics on residuals to check the model isn't missing something obvious. Once I'm confident in it, I'd present the forecast with a clear range around it rather than a single point figure, because decision-makers need to understand the uncertainty, not just the headline number.”
- 2
The government is considering raising a particular tax rate. How would you assess the likely economic impact?
Why they ask: A policy scenario question that checks your ability to reason through second-order effects under time pressure, a core part of the job.
How to structure your answer: Judgement under pressure: state your initial framework, identify the main channels of impact, flag key assumptions and data gaps, then give a provisional view.
Example answer
“First I'd separate the direct revenue effect from behavioural responses: how much does the tax change alter consumption, investment or labour supply. I'd look for comparable historical changes or interstate variation to estimate elasticities, and flag that the answer depends heavily on which group bears the incidence. I'd be upfront that without proper modelling I can only give a provisional view, but my instinct is the effect would depend more on the size of the change and the substitutes available than on the tax rate alone.”
- 3
Tell me about a time your analysis contradicted what stakeholders expected or wanted to hear.
Why they ask: Behavioural question checking whether you can hold a technically sound position under pushback, which matters when advising decision-makers on economic risk.
How to structure your answer: STAR: situation, task, action, result, focusing on how you communicated the disagreement.
Example answer
“I was evaluating a proposed policy change and my modelling showed the expected benefit was much smaller than the sponsoring team had assumed. My task was to present that finding without it being dismissed as a technical quibble. I walked them through the assumptions driving my result and ran a couple of alternative scenarios so they could see the finding wasn't an artefact of one modelling choice. The team adjusted their expectations and the final policy brief included a more conservative estimate, which held up better once the change was actually implemented.”
- 4
How would you explain a regression result to a policy officer with no statistics background?
Why they ask: Communication is one of the biggest gaps between technically strong economists and effective ones, so interviewers test it directly.
How to structure your answer: Direct explanation: give a plain-language example of how you'd translate a specific technical result, avoiding jargon.
Example answer
“I'd drop the coefficients and confidence intervals and talk in terms they can use: 'for every extra year of experience, wages tend to rise by about this much, and we're fairly confident that's a real pattern rather than noise.' I'd use a chart rather than a table where I can, and I'd always state what the result doesn't tell us, like whether the relationship is causal, so they don't overinterpret it.”
- 5
What software do you use for statistical work, and how do you decide between them?
Why they ask: Checks practical fluency with the actual tools (R, Python, Stata, Excel) the role uses day to day, not just theoretical knowledge.
How to structure your answer: Direct comparison: name the tools, give a concrete reason for choosing one over another in a real task.
Example answer
“I use Stata for standard econometric work because the syntax is quick for panel and time series models, and R or Python when I need more flexibility, like building a custom simulation or automating a data pipeline. Excel still gets used for quick checks or when sharing a working file with someone who isn't going to open R. I don't have a strong preference beyond matching the tool to the task and the audience who needs to use the output afterwards.”
- 6
Where do you see economic research heading given how much routine data analysis can now be automated?
Why they ask: With the role's moderate AI exposure, interviewers want to see you've thought about how automation changes the value you add, rather than treating it as a threat to dismiss.
How to structure your answer: Reflective: acknowledge the change directly, then explain where you think human judgement still matters.
Example answer
“A lot of the routine data cleaning and even first-pass modelling can be automated now, and I think that's mostly a good thing, it frees up time for the harder parts of the job: deciding which questions are worth asking, checking whether a model's assumptions actually hold in a messy real-world situation, and making the judgement call on how to present uncertainty to people who have to act on it. I see my job shifting more toward that judgement and communication layer over time, rather than away from it.”