ETC3410: Applied Econometrics

Assignment 1 (20%): 1 individual component, 1 group component with STATA, 1 project proposal for a research project. The individual and group components comprise of questions on the theoretical and practical part of the lectures and comp labs, and these are pretty standard actually albeit its length. The research project is designed to test students’ ability to design projects and use econometric techniques taught in the lectures to analyse data. Students can opt to choose 1 of 2 topics provided by the lecturer, or make up their own project topic, which was pretty cool. Our group did an entire project on global HIV/AIDS data. Marking was pretty harsh though.
Assignment 2 (20%): 1 individual component, 1 group component with STATA, 1 project report for a research project. Again, a similar structure, but the project report with the results section is handed in instead. Marking was quite harsh for this as well.
Final exam (60%): 3 long questions. The questions were fair, and there were definitely a few challenging bits here and there. Marking was quite harsh apparently, and students don’t do as well compared to other ETC units despite being an “applied” and “easy” unit. If one were to focus in the comp labs and do the assignments well it is quite easy to do well in the finals.

Thoroughly enjoyed the unit (partly biased as an econometrics enthusiast). This unit is divided into 5 topics:
1)A review of econometric topics (statistics and linear algebra, OLS from ETC2410, GLS and an array of hypothesis tests)
2)Binary choice models (logit, probit)
3)Instrumental variable estimation
4)Panel data estimation (fixed effects, random effects, etc.)
5)Project evaluation (not examined)
Even though this is technically an ‘applied’ unit, Jun does spend a great deal of time on developing the theoretical foundations of the models and econometrics techniques taught, while the applied side of the unit is mainly taught in the tutorials. STATA is used as the main software for this unit, and basic commands need to be recalled for exam purposes, which isn’t a very steep learning curve compared to R.
The only gripe I have with the unit is how econometric methods are used in a ‘black box’ manner without any form of confirmation through mathematical reasoning (i.e. serial correlation and what-not). Otherwise, it is a really good unit.

Jun Sung Kim. Great guy, super funny and approachable, gives really good advice for the research project.

1 sample exam with solutions

5 out of 5

Yes, with screen capture

Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. Textbooks are not required as lecture slides are more than sufficient.

2x 1 hour lectures, 1x 1.5 hour comp labs.

S1 2019.

94 HD

Two assignments worth 20% each, and 60% exam (2 hour). No group assignments.

Pretty decent unit. Decent amount of content without being too much. Doesn’t take too much work to stay on top of the content. Gives the impression that it is extremely useful if you plan to do any research in economics (such as honours) as you learn how limited OLS taught in ETC2410 is. Highly recommended to any considering research.
Highly recommend attending the tutorials, even though participation is not required (attendance is taken though). The lectures are not as necessary, and ended up going to about half of them. The lecture slides are very comprehensive, so no textbook is required. The lectures are not recorded.

Tutorials are split roughly 50/50 between tute work (maths on the whiteboard and theory discussion) and using STATA on the computers. The computer tutes are quite basic, but can be quite boring. All you do is replicate results by following instructions. John will talk for about 45 minutes going through how to get the results. Takes no more than 20 minutes to replicate the results. You will need to know how to use STATA as it is required for both assignments. The tutes can be ahead of the lectures at times, so you won’t be at much of a disadvantage by not attending lectures. In fact, as John is also the lecturer for half of the unit, I found that I learnt much more from the tutes than the lectures.

With regards to the assignments, they are quite straightforward. In my opinion, they are easier than the ones from ETC2410. Not too time consuming. The assignment from Jun for the first four topics is very easy, and shouldn’t take more than 6 hours to complete. John’s is a bit more challenging, but probably still easier than the second assignment from ETC2410 in my opinion.
In terms of the content covered, it is mainly going through the estimators used in econometrics. Gone is the days of OLS for a lot of the models used. Where ETC2410 mainly goes through how to calculate marginal effects, derivation of OLS, etc.., ETC3410 goes through the different estimators used in different cases.

Topic 1: Basic Theory and Revision (Asymptotic Theory)
Topic 2: Binary Choice Models (modelling dependent variables which are binary such as employment using linear probability, logit and probit models)
Topic 3: Endogeneity Bias (consequences of having an endogenous regressor)
Topic 4: IV Estimation (using instrumental variables when an endogenous regressor is present – SIV and GIV, 2SLS/two-stage least squares)
Topic 5: System of Equations (estimating multiple equations in one model)
Topic 6: Pooled OLS
Topic 7: Static Panel Data Models (Fixed Effects and Random Effects Estimators)
Topic 8: Dynamic Panel Data Models (Anderson-Hsiao, GMM, Arellano-Bond and Blundell-Bond)

Although the unit is called “Applied” Econometrics, the lectures cover a lot of the derivation and theory. After completing the exam, I can say that, while knowing the derivation and theory of how the estimators work is necessary, the bulk of the assessment is on how it would affect models when applied. So while it doesn’t seem very “applied” in the lectures, the assignments and the exam is mainly concentrated on hypothesis testing and interpretation of models (about 70/30 split between applied and theory).In terms of the maths skills required, ETC2440 is helpful (another of John's units), but isn't necessary.

Jun Sung Kim for the first half covering Binary Choice Models, Endogeneity and IV Estimation. Seems like a nice guy, but lectures can be a bit hard to follow at times.

John Stapleton for second half covering system of equations and panel data models (time series). He also doubles as the tutor for the whole semester. Excellent lecturer/tutor. Easy to understand for me, and less boring in general.

Yes, most available from Monash Library online. Sample questions for some topics also provided

4 out of 5

No

Don’t need one.

Two 1-hour lectures and a 2-hour comp lab/tute

2015, Semester 2

TBA