Your task in this assignment is construct a forecasting model of Melbourne’s house prices, conditional on other variables, and then to undertake a scenario forecast examining what is likely to happen to house prices if there is a major change in the economy. This assignment will count for 10% of your final mark.
The assignment must be electronically submitted by 5pm Sunday October 15th in PDF format. It needs to be submitted online via Moodle by a member of the group. However, all group members must click the “Submit Assignment” button on Moodle and accept the University’s submission statement. This is essential so please make sure to do this.
There is also a peer evaluation component to the assignment similar to what was done for the first assignment. This time you will receive a single survey request from TEAMMATES to be submitted after the assignment is completed. This is to determine group members’ contribution to the assignment. It will be used in determining your final mark. In particular, if we find that certain group members have not actively participated in the assignment, as indicated by the feedback from other group members, then these students will have their grade adjusted downward to reflect this. This survey must be completed by 5pm Wednesday October 18th. The completion of this survey is required.
The objective of this assignment is to assist you in developing model building, model evaluation and forecasting skills. Tasks that many of you will be involved with on a daily basis once you hit the work force. The specific objective is to build a forecasting model for Melbourne house prices and examine a certain future scenario. The final output of this assignment will be a short report.
The data set to use in this assignment will be provided. The variable which you are required to model is “HPI_MEL”. A range of other variables are included which can be used to model and forecast this variable. The Excel file includes a “KEY” sheet which provides some explanation of the variables.
The references section below lists some resources that you may find useful in this assignment. These are mostly papers that have looked at modelling Australian house prices. Please note that some of these go well beyond what is expected of you in this unit. However, they may provide some background context and discuss ideas that are useful to you. We outline instructions for the assignment below. Please follow them closely. Note the presentation of your results must not be more than 1,500 words and must not be more than 5 pages (this excludes the reference section but the page limit includes ample space for tables and figures, no appendices are needed). Markers will focus on the clarity and conciseness of your presentation, overly wordy or detailed submissions will be penalised.
When constructing your report present it as if you were engaged to undertake the work as a professional consultant. Here the focus is on the methodology, the results and their interpretation. For example, if you do a hypothesis test explain the test you did, why you did it and what it means but you don’t need to outline every step of the test (we will save this exercise for the exam where you will most certainly need to outline every step of a hypothesis test to get full marks!). Also, make sure to present results tidily. For example, if you are presenting regression results put the results that you think are important in a table. Do not simply copy-and-paste Eviews output into the document.
(1). Introduction: provide a short outline of the task and compile some descriptive statistics of the variables that you will use in your modelling. Discuss any relevant findings from your examination of the data. This may include, for example, discussion of: mean, median, maximum, minimum of certain variables or their changes, skewed or symmetric distribution, trends, SACF, SPACF, seasonality. Focus on issues that are important for the forecasting exercise. Make sure to transform any nonstationary data accordingly for modelling purposes.
(2). A Forecasting Model for Melbourne House Prices: develop a model of Melbourne house prices and discuss the process you used in model development. This may include: choice of variables, significance tests of variables, lags chosen, functional form chosen, estimation sample used, forecast accuracy. (HINT: The task of constructing a forecasting model is challenging even for experts. This is because there are lots of potential models and we simply cannot evaluate them all. Here is some advice. First, make use of the model selection approaches you have discussed in lectures. Second, make sure the model is intuitively sensible. A variable may be ruled out from inclusion in the model on the grounds that we can’t think of any reason why it should affect the dependent variable. For example, would we ever include the rainfall in Peru in a model of Melbourne house prices even if it had a statistically significant effect? No (at least not unless someone can come up with a plausible explanation why one effects the other, I certainly cannot think of one). Third, construct a model that makes sense in terms of what you are asked to do in section 3 below. This requires at least one variable in the model that is affected by the state of the economy so you can explore the impact of an economic shock. You should also consider whether you find the direction/sign of this effect plausible).
(3). Conclusion: in your conclusion you need to do two things. First, construct a baseline forecast for the period 2017Q2 to 2018Q1 for Melbourne house prices using your chosen model. Second, we would like you to explore how your baseline forecast would alter were there to be a major change (shock) to the economy. There are at least two examples of future scenarios that would be interesting to explore. Please choose one of these or make up your own: (1) a recession (i.e. a fall in state final demand, rise in the unemployment rate, fall in employment and so forth), (2) a significant tightening in monetary policy by the Reserve Bank of Australia (i.e. higher interest rates). Construct the future paths of the exogenous variables in your model so that they are consistent with the scenario you are envisaging. Explain the approach you took to constructing these future paths and then discuss the scenario forecast and compare it with the baseline forecast.
(4). References: please include any references you use in this section.
In the Appendix we have outlined the marking rubric that will inform the grading of your assignments. Please look at this when putting together your assignment to try and best address the marking criteria.