Hello, Everybody!
I wil be grateful for any help. I have a problem with frontier function. I want to make a model of life expectancy in 11 countries depending on some factors (consumption of alcohol, school years, gross national income per capita and etc.) using frontier method for panel data (from 2000 to 2013 year) to calculate efficiency of national care system at the end. I've made a model and then used the command
predict te, te
for each country but the meanings are very small and almost the same. Due to the fact I took into consideration different countries (from Australia to Nigeria) it is impossible that efficiency of national care system are equal. Besides, I've calculated u and tried to calculate e^(-u), but what is the difference between u and te? I've spent a lot of time to sort out this, but overall I don't understand clearly:(
Thanks!
Alternative Methods to Examine
Hospital Efficiency:
Data Envelopment Analysis and
Stochastic Frontier Analysis
HEALTH CARE EFFICIENCY
ACROSS
COUNTRIES
: A STOCHASTIC FRONTIER ANALYSIS
Efficiency estimates
of health care systems
Hi,
Are you using Stata? If yes, are you using the "xtfrontier" command which suits to calculate SF in panel data?
The stochastic frontier model is used largely in many studies of production, cost, revenue, profit and other models of goal attainment such as efficiency estimates and others. Since your objective is to estimate efficiency of health care system, I think your choice of stochastic frontier model is appropriate. What I will advise is that try to consult literature thoroughly particularly the original one developed by Aigner, Lovell, and Schmidt (1977). Note that the stochastic frontier framework involves two steps. In the first, the frontier model is estimated, usually by maximum likelihood. In the second, the estimated model is used to construct measures of inefficiency or efficiency. Individual specific estimates are computed that provide the basis of comparison. Good luck!
Hello Kseniia!
I am not sure in SFA, but using DEA it is very important the relation of the number of DMUs that you have and the number of inputs and outputs. If I understand, do you have 11 DMUs by year (which is a little bit). How many and which are the inputs and outputs of your model? Did you calculate the efficiency year by year?
Best regards,