ETC2410-3440: Introductory Econometrics

I liked this a lot. You basically look into OLS. The underlying assumptions, when it is unbiased, what happens when the assumptions fail etc.

You start off looking at deriving the formula for B₀_hat and B₁_hat under OLS, and then go into proving their unbiasedness using summation notation. The Gauss-markov assumptions are introduced, which when they hold show the OLS is BLUE (best linear un-biased estimator). You then move onto omitted variable bias, functional form, inference and hypothesis testing, prediction intervals, dummy variables and interaction terms, and heteroskedasticity. Some of these topics can take a little while to get your head around, but the first 2 assignments help a lot with learning the material. Phil covered these topics. His lectures did seem a bit rushed, but he knew his stuff and I found them helpful. That being said, the textbook and notes could suffice.

David took over from week 8, and went into time series regression for a couple of weeks. You get a few new assumptions, and go into FDL, AR and ARDL models, and the interpretation of coefficients using lag multipliers. Pretty interesting stuff. David's lectures are much more structured, and he taught the stuff using an example, checking assumptions as we go. He finishes off by basically re-teaching the first part of the course, but using matrices rather than summation notation. Very good revision for the exam.

The assignments can be tough, but if you can work through them you'll be set for the exam, as the exam is pretty well the assignment regressions and questions with a few more in-depth theory questions attached. Make sure you are very clear with all of your explanations, and perform a billion hypothesis tests. There are no word limits so go nuts if you want to get a high mark. They're not too difficult to pass though, but to get 80+ on the assignments you have to put in a fair bit of time.

Tutes are very helpful. You learn how to use eviews here (essential for the assignments) and basically run through a few questions, while the tutor teaches some of the important parts of the theory. While there are no tute marks, I do suggest going as the questions that you run through in the tutes are similar to the assignments. Eviews is very simple to use and shouldn't take too long to pick up.

As mentioned before, the exam is primarily interpreting regression outputs, performing tests for heteroskedasticity and various hypothesis tests with some OLS theory chucked on. About 20% is theory based, while the rest is mainly interpreting and practical. The 3440 students have 1 extra question worth 5 marks on the theory of OLS, otherwise exactly the same.

This is required for an econometrics major, and can count towards economics (2410) and finance (3440) as well as a fair few others. ETC1010 used to be a pre-requisite but no longer is. I did 1010 at the same time as 2410, and found that 1010 didn't help a great deal with 2410, but 2410 helps a lot with 1010. ETC1000 (or equivalent intro stats unit) is all you really need.
Overall it is a fairly challenging unit with a fair amount of content to cover. Unless you're very strong at maths/stats already it will take a fair chunk of study to understand the topics completely. I spent about as much time on this unit as my other 3 commerce units combined and still felt I could have been more prepared and spent more time on it. While the maths behind it isn't too tough (methods is all you need to start off with), knowing summation notation and matrix algebra prior would help a lot with the proofs. You could just wrote learn them, but better to learn the small amount of maths so that you properly understand them. I did like the content though, and I recommend it. Pretty interesting and useful stuff on the whole.

Phil Edwards - cross sectional data

David Harris - Time series + regression with matrices

Yes, about 4 were given out with solutions, but there are quite a few on the library database without solutions.

4.5 out of 5

Yes

Introductory econometrics 4th edition - Wooldridge. A pretty good textbook, the course follows it very closely, but you are given pretty comprehensive notes so its not entirely necessary. I bought it with the eviews software package, which you'll need if you want to do the assignments at home. I personally found buying it worthwhile, others didn't use it at all. Wait a week or two and see how you go.

Two 1-hour lectures and one 2-hour tutorial/computer lab = 4 hours

2013 semester 2

86 HD