Treatment Effects and Program Evaluation ECOM90023
Research Project
Details
Your Research Projects will be the basis for 70% of the assessment for this subject.
It is thus important that you work steadily on your Research Projects over the whole
semester.
Themain benefit of completing these Research Projects is to gain experience in conducting
credible micro-empirical research.
PROJECT OBJECTIVES
The main objective of your Research Project is to conduct new micro-empirical research
(defined below) which provides an arguably causal estimate of the effect of some important
“treatment” on an interesting outcome or outcomes.
A project that meets or exceeds the requirements for passing this assessment will ideally
have the following elements:
1. A clear research question involving a causal relationship that is potentially of interest
to economists, policy-makers or other economic actors.
This research question should have:
(a) a specific “treatment” or causal variable,
(b) a specific outcome or outcomes of interest, and
(c) clear intution for the existence of a causal relationship from the “treatment”
variable to the outcome(s).
2. A defensible estimation strategy for yielding causal estimates.
3. Appropriate use of micro data when estimating a causal effect that answers the
research question.
4. “Some” contribution to knowledge i.e. not simply a replication of prior research.
5. Careful implementation of your estimation strategy.
6. Appropriately detailed description of what you have done and why you estimated your
effects in this manner.
This includes describing clearly:
(a) the data employed (including its source), the construction of key variables and
details of data linkages and dataset cleaning / sample selection (if any),
1 of 6
Treatment Effects and Program Evaluation ECOM90023
(b) the estimation method – the main equation and key identifying assumptions,
(c) why this method is appropriate for estimating causal effects in your setting,
(d) the construction of standard errors and the use or otherwise of weights, and
(e) the statistical tests conducted to support the use of your chosen estimation
method in your specific setting.
7. Clear and appropriate interpretation of your results.
8. A concise discussion of how you may have constructed your estimates using an alternative
estimation method also discussed in this subject.
A list of estimation methods covered in this subject is provided below.
What is micro-empirical research?
It is econometric estimation of ideally causal relationships using micro data. Micro data has
individual observations for many “units”:
people, households, houses, small areas, states, firms, or even countries.
PROJECT DELIVERABLES
There are three main “milestones” throughout the semester related to your Research Projects.
1. Research Proposal and Plan (Assignment 1, 10%):
– due Friday 27 March at 4 pm (end of week 4).
2. Oral Presentations (10%):
– Monday 11 or Tuesday 12 May (start of week 10), exact timing TBA.
3. Final Reports (50%):
– due Sunday 31 May at 4 pm (the very end of week 12).
Further details regarding specific requirements for each of the three deliverables are provided
on the LMS under the “Assignments” menu item.
The remaining 30% of assessment in this subject is allocated to Assignment 2 and the mid-
semester test.
The requirements for Assignment 2 (Peer Review) are described in a separate document
available on the LMS under the “Assignments” menu item.
Details regarding the mid-semester test will be provided via the LMS, including its coverage,
a practice exam, and other requirements.
2 of 6
Treatment Effects and Program Evaluation ECOM90023
GROUP WORK
Students will complete these Research Projects in groups with a maximum of three student
members.
We will begin the process of forming these groups in the first tutorial for the subject, held
during week 1 of semester:
Tuesday 3 March at 3:00 pm or Wednesday 4 March at 5:00 pm.
Ideally these groups will be confirmed towards the beginning of week 2 of semester.
As far as practically possible, these groups should be formed among students within the
same enrolled courses:
ˆ Master of Applied Econometrics,
ˆ Master of Economics (including exchange Masters),
ˆ PhD in Economics (including exchange PhDs), and
ˆ other students, e.g. PhDs in Finance.
Note that group members do not need to be attending the same weekly tutorial.
Given the nature of the assessment for this subject, groups should meet at least weekly to
ensure adequate progress throughout the semester. It is also important that actual progress
on the Research Project occurs each week.
Students should undertake their group work using a collaboration tool. An appropriate
collaboration tool or app will allow for the storage of project files, allow group members
to work on the same files, provide space for communication with other group members and
allow weekly tracking of progress on project tasks.
The University supports the use of the Microsoft 365 / Teams environment for collabora-
tion. All students should have access to the Microsoft 365 suite through the University of
Melbourne. Resources for students to make a start with Teams – including a user guide – is
available at the following link:
https://lms.unimelb.edu.au/learning-technologies/microsoft-teams
Students may, if they wish, choose to use alternative tools for collaboration such as Google
Drive, Google Docs, GitHub, Overleaf, GitLab, Chats, et cetera.
Group work can be a great experience, but sometimes they face challenges. Developing the
skills to work in groups efficiently and effectively is important for most jobs. This involves
dealing with issues related to: coordination of activities (meetings, task assignment), free-
riding, time-wasting social activities and conflict.
Students are encouraged to develop and agree to a group contract at the start of semester.
This contract would generally define and describe what is expected of all group members,
and may include details of:
3 of 6
Treatment Effects and Program Evaluation ECOM90023
1. who the team members are (contact details, key individual skills);
2. what you want to accomplish (main goals, both short and long term);
3. how the team will be organised (roles, task assignment, contingencies);
4. how the team will operate (meetings, communication, decision-making); and
5. how you will help each other to reach your goals (keeping to the contract).
Students should contact the subject coordinator – Mick Coelli – for a consultation if they
have any concerns with the operation of their group and the involvement of other members.
It is best to sort out any issues early on.
It is important for all members to contribute equally to the group’s research. At project
completion, students will be asked individually for their honest assessment of the contri-
bution of all group members to the project. If there are disagreements regarding member
contributions, these may be resolved by looking back at the interactions and contributions
on the group’s collaboration tool.
ESTIMATION METHODS
Below is a list of the econometric estimation methods that we will cover in this subject:
1. Ordinary Least Squares (OLS);
2. Instrumental Variables (IV);
3. Panel Estimators (such as Fixed Effects or FE);
4. Difference in Differences (Diff-in-diff or DinD),
with binary or continuous treatments;
5. Regression Discontinuity (RD) designs;
6. Quantile Treatment Effects (QTE);
7. Synthetic Control (SC) methods; and
8. Matching.
All Research Projects should ideally use one (or more) of these methods.
As noted above, your Final Reports should also briefly describe how you may have credibly
estimated effects in your setting using at least one alternative method that also is taken from
the above list.
4 of 6
Treatment Effects and Program Evaluation ECOM90023
POTENTIAL PROJECTS
These Research Projects are expected to meet the over-arching objectives described above.
One key element for successful completion of these Projects is quickly settling on a research
question that is viable. That is, choosing a question that can be answered using available
micro-data and an appropriate estimation method within three months.
Here is a potential starting point for your research that you may choose to use for your
projects.
Outcome of interest
House / unit prices or rents.
Data source for outcome
Australian Property Monitors (APM) held panel data on house price measures over time
(monthly) by geographic location / region across Australia.
This data was made available in the past by the Australian Urban Research Infrastructure
Network (AURIN) at various levels of geographic detail. Sadly AURIN are no longer able
to provide the data to researchers for free as it is now “owned” by for-profit firms such as
Corelogic and Domain Insights.
However, I have median house and unit price sales data from 1995 to 2021 at the local
geographical levels of Statistical Area 2 (or SA2, smaller than suburbs) and SA3 (a bit larger
than standard Australian suburbs). If students wish to use this data for their projects, they
can email me (mcoelli@unimelb.edu.au).
In addition, the Victorian government provides median house and unit prices for suburbs in
Victoria from 2014 to 2024:
https://www.land.vic.gov.au/valuations/resources-and-reports/property-sales-statistics
Potential “treatments”
Here are some potential drivers of home prices / rents that you may consider for developing
an interesting research question that requires estimation of a causal effect of “treatment”.
ˆ Government housing policies such as stamp duty, zoning, social housing, borrowing
subsidies.
ˆ School quality / availability of infrastructure.
ˆ Climate change / potential flooding risk changes.
ˆ Local labour market shocks – due to, for example, import competition from China,
installation of industrial robots or closures of large local firms.
5 of 6
Treatment Effects and Program Evaluation ECOM90023
Information on potential “treatments” will need to be sourced by students themselves, and
then merged with the housing price data using the geographical identifiers provided.
NOTE:
Students are not required to base their Research Projects on house prices.
Most groups chose to find their own data for their projects in 2023, 2024 and 2025.
This is a suggestion only for those who may find it difficult to come up with their own viable
project in a short amount of time.
6 of 6

学霸联盟