|
论文题目：Statistics and Econometrics的一个Assignment，要用统计学软件做的
论文语种：英文
您的研究方向：统计经济学
是否有数据处理要求：是
您的国家：英国
您的学校背景：university of exeter 英国13名
要求字数：在附件有
论文用途：本科课程论文 BA Assignment
是否需要盲审（博士或硕士生有这个需要）：
补充要求和说明： [email protected](1455780998)
BE2006-Statistics and Econometrics
Assignment - 2008/2009
Note: You are allowed to use any statistical software that you
nd convenient to solve this assignment. If you choose to use the software Time Series Modelling (TSM), you can
nd it installed in the computers of Rooms 116 and 206, Streatham Ct, and in the X
computer suite. You can also install TSM on your own PC if you wish. Visit for downloads and installation instructions. As well as the TSM installation program you will need two other components, Ox 4 Console and the Sun JREJava package. Ox is a free download for students. Java is also free, and is often installed on PCs by the manufacturer, but updating to the latest version is recommended.
You will also need a licence key, and you may use the time-limited licence issued to the university. After installing and starting the program, open Help / Regis-ter this Copy. . . and type in the codes that you can
nd on the course web page Assignment Length: The maximum length of the Assignment is 10 sides of A4(including Appendices).
Whenever conducting a test use a 5% signi
cance level. Also be sure to state null and alternative hypotheses, null distribution (with degrees of freedom), rejection criterion and your decision. Do not answer questions with a yesor noonly,but carefully justify your answers. You will be penalized if you present tables and results that are not subsequently discussed in your answers.
In this Assignment you are going to estimate a so-called gravity model. In applied international economics the gravity model has been a successful empirical model for explaining bilateral trade ows between countries. Its basic speci
cation relates trade ows between trading partners with their size and trade costs. Typically size is approximated by income, while trade costs are measured by geographical distance between trading partners.#p#分页标题#e#
The gravity model has been estimated for a wide range of countries and periods.Usually researchers
nd positive e¤ects for income, that is, large countries trade more with each other than small countries. Concerning distance, a negative e¤ect on trade ows has been found often as transaction costs are low when countries are neighbours(e.g. Spain and France or Canada and the US), but rise with distance. Hence, trade ows between countries depend negatively on distance.
In this Assignment you are going to analyze bilateral trade ows for a cross section of OECD countries. In total 221 observations are available. The data have been ex- tracted from a database used in Glick and Rose (Does a currency union a¤ect trade:
The time series evidence, European Economic Review, vol46, June 2002, 1125-1151). Both the paper and the database can be found at On the web page 10 dif- ferent EXCEL and TSM
les are available named grav88.xls, ..., grav97.xls and grav88.tsm, ..., grav97.tsm respectively. Each Excel and TSM
le contains data for a particular year. TSM
les also contains the settings for the Assignment.
The speci
c year you have to analyze is determined by the 9th digit of your student ID number:
0 1988 5 1993
1 1989 6 1994
2 1990 7 1995
3 1991 8 1996
4 1992 9 1997
For example, if the 9 digits of your student ID number are 555488567, you choose year 1995 and download grav95.xls or grav95.tsp (click on the right button of your mouse, choose Save Target As and choose a directory and click OK). If you choose to use the TSM software, click on the TSM
le. The 4 imported series are:
ltrade = logarithm of real bilateral trade
lrgdp = logarithm of real GDP
ldist = logarithm of distance
regional = 1 if two countries have regional trade agreement, 0 otherwise.
1. [5%] Briey describe the main features of the data. [Hint: Consider
rst descriptive statistics, for example, means, medians, variances/standard deviations, minima and maxima, graphs of variables, histograms, cross- plots of variables, comparisons of data across various groups and correlation matrices.]
2. [10%] Basic gravity models relate trade ows to GDP and distance. Estimate the model
ltrade = 0 + 1lrgdp + 2ldist + u
Given the speci
c functional form explain in words what the regression coe¢ cients 1 and 2 measure. Interpret the results obtained.
3. [5%] Suppose you are working on this assignment together with a fellow student.
Your colleague claims that the least squares estimators of the regression in question
2 are unbiased because the
t of the regression is satisfactory. More speci
cally, it is claimed that least squares estimators are unbiased because the R2 is large and all#p#分页标题#e#
regression coe¢ cients are signi
cant. Do you agree?
4. [10%] For the model in question 2 perform Ramseys RESET test. What do you conclude about the validity of the estimation results from question 2?
5. [10%] In the empirical literature on international trade an important issue is whether regional trade agreements have an impact on bilateral trade relations. An additional measure of trade costs is the presence of a regional trade agreement (re-gional). Estimate the following model and interpret the results
ltrade = 0 + 1lrgdp + 2ldist + 3regional + u:
6. [10%] Use the estimates of question 5 to test whether regional trade agreements have a signi
cant impact on trade.
7. [10%] Using your model estimated in 5, construct a 95% con
dence interval for E[ltradejlrgdp = lrgdp; ldist = ldist; regional = 1], where lrgdp and ldist denote the sample means of lrgdp and ldist respectively. Assuming normality of the error term u; estimate E[tradejlrgdp = lrgdp; ldist = ldist; regional = 1], where trade = real bilateral trade.
8. [10%] In the existing empirical literature the usual way of measuring a trade agreement e¤ect on trade ows is to incorporate the dummy variable regional as an intercept shifter as in question 5. It is possible, however, that the regression model as a whole is di¤erent for trading partners with and/without trade agreement. Estimate again the model of question 2, but now for two subsamples, i.e. both trading partners have a trade agreement or not. Use the Chow Test to test whether the regression coe¢ cients di¤er between the two subsamples.
9. [10%] An alternative way of testing on a structural break is by using the dummy variables. Construct a model using dummy variables that allows you to test the same
null hypothesis of the Chow Test in 8, that is of no structural break. Use the Lagrange Multiplier statistic to test the null hypothesis of no structural break.
10. [10%] In your model estimated in 8, test if the error term is homoskedastic.
What do you conclude about the validity of the tests performed in 8 and 9? In case you consider that these tests are invalid, test whether the regression coe¢ cients di¤er between the two subsamples using the appropriate test statistic.
11. [10%] For the model estimated in question 9 perform Ramseys RESET test.
What do you conclude? If you
nd evidence that the model is misspeci
ed, suggest an alternative model that does not have speci
cation problems and interpret the results.
|
