Econometrics postgraduate personal statement example

Personal statement example

My interest in econometrics grew out of frustration with how easily a regression table can look convincing. In my second year I took a module on applied microeconometrics and spent a weekend trying to reproduce a textbook wage equation using UK Labour Force Survey data. My coefficients did not match anything in the lecture slides, and working out why — sample restrictions, the treatment of missing values, my own careless handling of weights — taught me more than the lectures had. I began to see estimation as a set of decisions that have to be defended rather than a button to press, and that is the part of economics I want to study properly.

My degree in Economics with Statistics gave me a reasonable foundation: matrix algebra, probability, statistical inference, and two econometrics modules covering ordinary least squares, instrumental variables, panel methods and an introduction to time series. I also took an optional module in numerical methods, which was the first time I had to think about whether an estimator could actually be computed rather than only whether it existed. I finished with a 2:1, with my strongest marks in the quantitative modules, and I am aware that the gap between my econometric theory and my statistics is something the MSc would close.

For my dissertation I examined the relationship between commuting time and part-time working among women with dependent children, using several waves of a household panel study. I used a fixed effects specification to absorb time-invariant preferences, then spent a long time on the limitations: commuting time is partly chosen alongside hours, so my estimates are associations rather than causal effects. My supervisor pushed me to say this plainly rather than hedging, and the honest discussion of identification was the section I was most pleased with. I wrote all the cleaning and estimation code in R and kept it organised well enough that a friend on the same module could run it on her machine without my help — a small thing, but it made me care about reproducibility.

That friend and I also worked together on a shorter group project comparing two approaches to forecasting monthly retail sales. She handled the seasonal adjustment and I fitted the models; we disagreed about how to compare them and settled it by holding back the final year of data and looking at out-of-sample errors. It was a modest piece of work, but deciding in advance how we would judge the answer was a useful discipline.

Since graduating I have worked three days a week as a data assistant at a housing association, producing arrears and void reports from the tenancy management system. Much of it is routine SQL and spreadsheet work, but it has made me careful about what administrative data actually records: a repair marked complete is an entry in a database, not necessarily a repaired window. I have automated two monthly reports, which freed up time I now spend checking figures rather than assembling them. On Thursday evenings I tutor adults working towards a numeracy qualification, which has improved my ability to explain a method in more than one way.

I am reading Angrist and Pischke's Mostly Harmless Econometrics alongside Hayashi's Econometrics to get used to a more formal treatment. I want the MSc to give me rigorous asymptotic theory and panel and time-series methods, with a view to working as an applied economist in housing or labour policy, where I would rather be the person asking whether an estimate can be believed.

Review this personal statement

Latest reviews