Applied Econometrics: Empirical Project 2

Transcription

1 Applied Econometrics: Empirical Project 2 Introduction As part of the course, you are required to write two individual projects which demonstrate your ability to use econometric techniques in a practical context. Each project will count for 50% of the course grade. This instruction sheet relates to the second empirical project which is due on 27 April. You will be expected to analyze one of three data sets provided and write up an empirical project on the basis of this. Below you will nd a description of the three data sets and a brief discussion of the relevant research questions. You must choose one of these three data sets. The instruction sheet for Empirical Project 1 provides assignment submission instructions, grading criteria as well as complete details of how you should organize and format your project. Instructions for Empirical Project 2 are exactly the same as Empirical Project 1 and so I will not repeat this material here, but will refer you to the instruction sheet for the rst project (available on the course website or on Myplace). 1

2 Project Topics Topic 1: Demand for Medical Care Background: In the lectures I went through an empirical example involving the demand for medical care where the dependent variable was the number of hospital stays. The data I used was taken from the paper: Demand for medical care by the elderly: A nite mixture approach by P. Deb and P. Trivedi, published in the Journal of Applied Econometrics in The le medical_demand.xls is the complete data set used in this paper. In addition to the dependent variable I used in the lecture (number of hospital stays, which is one particular measure of demand for medical care), this data set contains several other measures of medical demand. It also contains several potential explanatory variables beyond the ones I used in the lectures. Table 1 on page 324 of the Deb and Trivedi paper (which is available on the course website) contains a list of all the variables in the data set. In this project you are expected to investigate research questions such as: Do the determinants of medical demand di er depending on which measure of medical demand is used as a dependent variable?. Of particular research interest in this eld is the role played by medical insurance in determining medical demand. To address such research questions you should estimate (and motivate why you choose) appropriate econometric models for the various measures of medical demand and interpret the results. 2

3 Topic 2: Estimating a Production Function for Spanish Dairy Farms Background: File dairy.xls contains data on 247 Spanish dairy farms for 6 years ( ). The variables in the data set are: MILK = number of litres of milk produced LABOUR = number of workers on the farm (full time equivalent) FEED = amount spent on animal feed (in pesetas) COWS = average number of cows on farm LAND = hectares of land available for cattle grazing FARM = identi cation number for farm YEAR = year In this project you are expected to estimate a production function for Spanish dairy farms. You will probably know what a production function is from your microeconomics course, but I o er a brief summary here. A production function speci es that a farm s output (Y ), depends on the inputs used (X 1 ; ::; X k ) through a production function: Y = f (X 1 ; ::; X k ) In this data set, the output would be MILK and LABOUR, FEED, COWS are LAND are inputs. When empirically estimating a production function, f (X 1 ; ::; X k ), there are two common functional forms that researchers use called the Cobb-Douglas and translog production functions. These both involve taking logarithms of variables. Let small letters denote logged variables (e.g. y = ln (Y ) ; x 1 = ln (X 1 ), etc.). The Cobb-Douglas production function says: y = x 1 + :: + k x k The translog production function takes a similar form, but adds squares and cross products. Let me illustrate assuming k = 2: y = x x x x x 1 x 2 : To do this project, you should estimate and interpret a production function for this data set using appropriate econometric methods for panel data (you must justify why your preferred model is an appropriate one). You should consider questions such as: Is a Cobb-Douglas or translog production function appropriate for this data set?, which explanatory variables are important in the production function? and what are the elasticities of output relative to the di erent inputs?. 1 1 By this is which for the Cobb-Douglas is simple to obtain, but is a bit i complicated for the translog. 3

4 You may assume that instrumental variable methods are not required for this project. 4

5 Topic 3: Volatility in International Stock Markets Data set Stock_indices.xls contains stock price indices from markets in Amsterdam (AMSTEOE), Frankfurt (DAXINMDX), Hong Kong (HNGKNGI), London (FTSE100), New York (SPCOMP), Singapore (SNGALLS) and Tokyo (JAPDOWA). This is daily data (not including weekends and holidays) for the stock indices from January 6,1986 until December 31,1997. Note that these are simply the values of the stock indexes and have not been transformed in any way. For this project, the research question relates to the volatility of these international stock markets. You should build (and justify) appropriate econometric models for each of the stock indices and investigate the patterns of volatility that exist in these international markets (e.g. Is there evidence that volatility is changing over time? If so, how?, Can you detect common patterns in volatility across the various cities?, etc.). 5