ONLY HIGHLIGHTED AREAS NEED TO BE DONE
Introductory Econometrics Assignment 2 Due: At or before 11:55pm on 7 June 2020 The following instructions supersede any instruction that is in the Unit Guide: The assignment must be electronically submitted by 11:55 pm on Sunday 7 June 2020 in PDF format. Your file needs to be uploaded by only one member of each 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. Each group member will also be required to complete an anonymous peer evaluation survey. The survey will be done via the TeamMates app which will email you a unique link to the survey (to your Monash student email address). You will be asked to rate your group members’participation and effort with the same aim as in Assignment 1. These surveys will also be used to adjust your assignment marks in the following manner: Let’s consider a hypothetical student called Johnny: n0 = number of (D) votes that Johnny receives from his teammates. A (D) means that Johnny has contributed nothing. n1 = number of (C) votes that Johnny receives from his teammates. A (C) vote means that Johnny contributed less than what was agreed by the group. GROUPMARK = Johnny’s group submission mark If n0+n1 >=2, then Johnny′s mark for the assignment = max{0, (1− 0.4× n0− 0.15× n1)} ×GROUPMARK and Johnny’s mark for the assignment = GROUPMARK otherwise. If you do not complete this survey before the deadline, we will assume that you have given everyone else in your group a (B), but have given yourself a (D). Failure to complete the survey will result in a loss of marks. It is important to communicate well with your group members and make sure that everyone understands what is expected from them. Remember to obtain full marks, you need to report your results in equation form, with standard errors provided in parentheses below parameter estimates, and when testing a hypothesis, you should state the null and alternative hypotheses for the test that you perform, the test statistics and its distribution under the null, any extra regression that you run to obtain the value of the test statistic, the critical value of the test, and finally your conclusion. Altogether, your report should not be more than 10 pages including its reference list and paragraphs clearly separated (use Times New Roman fontsize 12). Important graphs and equations (in equation form, not screen shots of EViews or other software output) must be included within the report, and the corresponding EViews or other software output should be included in an appendix (the appendix is not included in the page count, and it will not be marked. It will be used only to check if the results included within the text are reported correctly). Any work cited must be properly referenced and any information that is quoted verbatim should be in quotation marks with a reference to the source. You do not need to include the textbook or lecture slides in your reference list. 1 Estimating demand for Australian exports Hendrik Houthakker and Stephen Magee wrote an applied econometric paper in 1969, which has become a classic paper in the field of international trade. In this paper they estimate income and price elasticities for import and export for a list of advanced countries, including Australia. The aim of this assignment is to update the computation of these elasticities for Australian exports only. One of the main contributions of Houthakker and Magee was to construct trade weighted GDP of major trade partners and also relative export and import prices for each country. Given the time frame of the assignment, we are not going to do that. In fact, given that quarterly real GDP for China, Australia’s current largest export market, does not go back far, it would be diffi cult to do so. Hence, we are going to modify the task considerably, and use the US GDP as a proxy for income of Australia’s export destinations. Also, instead of constructing a relative price variable for Australian exports relative to the price of Australia’s competitors in the export market, we are going to use Australian export prices and the Australian trade weighted exchange rate, hoping that these will be suffi cient to capture how expensive Australian exports are for foreign consumers. Task 1. Downloading data: You will download data on Australian exports of goods and services and Aus- tralian export price index from the Australian Macro Database website http://ausmacrodata.org/. The easiest way is to go to Categories, and from the drop down menu under the “national ac- counts”choose “exports”. You will get many series with the title “Exports of goods and services”, but you need to pay attention to the details below the title and choose “Chain volume measure” (which means real as opposed to nominal value of goods and services) and “Original”, as opposed to “seasonally adjusted”or “Trend”. In the same group, also look for the chain price indexes, again “Original”. The export price index is only available from 09/1985. Download these two series. Then, download http://www.rba.gov.au/statistics/tables/xls/f15hist.xls, which is on the RBA (the Reserve Bank of Australia) website and contains the real export-weighted exchange rate. You then need to go to the Federal Reserve Economic Database (FRED) https://fred.stlouisfed.org/, which is similar to ausmacrodata.org but for the US data, and search for GDP, and download the Real Gross Domestic Product, Billions of Chained 2012 Dollars, Quarterly, Seasonally Adjusted Annual Rate (note that we want this one to be seasonally adjusted, not original), which should be one of the first series that your search returns. Download that series from FRED. Place all of these series in one spreadsheet, only keeping data from 1985Q3 to 2019Q4. Make sure you do not make any mistakes in aligning the dates. Name your the real export series “export”, the export price index series “exprice”, the export-weighted exchange rate “exrate”, and the US GDP, “usgdp”. If you have done everything correctly, the first few lines of your data set should look like this: Task 2. Preliminary data analysis: Since the goal is to estimate elasticities, we will need to model the relationship between the logarithms of these variables. So, generate the logarithms of each of these series and name the series “l_export”, “l_exprice”, “l_exrate”and “l_usgdp”. Then, you should assign one series to each group member (groups which have 3 members should only consider the first 3 variables), and each group member must provide a plot and write a few sentences documenting the important features of the series assigned to them. This paragraph should include features like seasonality, clear trend in one direction, or high persistence but no consistent trend in one direction. Make sure that the reported features or statistics are relevant. For example, if a variable has an 2 Faisal Wani Faisal Wani obvious trend or does not seem to be mean reverting, then the sample average of that variable does not convey any relevant information. [The first two tasks have 20 marks, 4 marks for correct data, and 4 marks each for description of variables. Groups of 3 will earn 5 marks for correct data, and 5 marks for each of the 3 variables that they analyse.] Task 3. Developing a univariate model for l_export: Consider the goal of providing a 95% prediction interval for real exports of goods and services in 2020Q1. Develop a simple univariate model for l_export with a time trend variable T which starts at 1 in 1985Q3 and increases by 1 every period, and seasonal dummy variables for different quarters to provide a 95% prediction interval for l_export in 2020Q1. Then convert that to a 95% prediction interval for exports by simply exponentiating its bounds. In all sections of this assignment except for Task 4.e., please ignore any evidence of serial correlation in the errors of your estimated equations. This is to allow you to start working on the assignment before you learn about serial correlations in weeks 9 and 10. [No one will be surprised if the actual 2020Q1 exports fall below the lower bound of your prediction interval! - Econometric models cannot predict what happens in situations the likes of which never happened in the sample period.] Your report must explain the important insights that one can learn from your simple model (how good it fits, what does it reveal about trend and seasonality in exports, etc.), not just a report of numerical results. [10 marks] Task 4. Developing a simple export demand model: Consider the goal of developing an economic model of demand for Australian exports in order to estimate the price, exchange rate and income elasticities of demand for Australian exports (income here is the income of trade partners who buy Australian exports). Use the information in the data set to do this (again, ignore the serial correlation in the errors in 4.a, 4.b, 4.c and 4.d). Your report must include, (a) [For groups with 4 team members only] A test that after controlling for trend and seasonality, the economic variables are jointly significant at the 1% significance level. [5 marks] (b) A report of estimated elasticities, their interpretation and their respective 95% confidence intervals and a discussion of how similar or different they are from those reported for Australia in Table 1 of Houthakker and Magee (1969). [15 marks (+5 for groups with less than 4 members)] (c) A discussion around the following point: The demand for Australian exports should depend on export price in foreign currency, but the price index we have is export prices in the Australian currency. As a result, we should expect the price and exchange rate elasticities to be equal (i.e. to foreign buyers of Australian goods it should not matter if the price has gone up because price in Australian dollars has gone up or because Australian dollar has become more expensive), and examine this hypothesis against the alternative that they price and exchange rate elasticities are different, at the 5% level of significance. [10 marks (+5 for groups with less than 4 members)] (d) [For groups with 4 team members only] A test to check if the assumption of homoskedas- ticity for the errors of your equation is not rejected by the data, and why is it important to check this. Since the sample size is not too large, it would be wise to use the special form of the White test and perform the test at the 5% level of significance. If you reject homoskedasticity, only comment on the implications of this finding for your previous analysis, but do not re-do your analysis. [10 marks] (e) Only after finalising Task 3 and Tasks 4.a. to 4.d., include a discussion of the importance of checking for lack of serial correlation in errors when we have time series data, and then test this in your export demand model at the 5% level of significance