# Tobit regression

I am running a tobit regression to examine the effect of firm attributes on access to bank finance. specifically, I want to find out the effect of variouus firm attributes on the level of credit that is offered by banks. The depdent variable is the actual credit offered by the bank as a percentage of the amount of credit requested by the firm - it takes the value of 0 to 1. The independent variables are firm attributes such as location, sector, gender, income, collateral, age, size, year of education, etc. I have tried several times to write the tobit command for this but it always pops up on the screen at some point that there is an error. As a first time user of Shazam I gues I am not doing something right. Unfortunately I have a very limited time to get this results into a report. Can someone please help me to write the tobit format using the dependent and independent variables stated above.

edit retag close merge delete

Please post your data and script so we can assist further.

( 2013-01-13 11:27:31 +0000 )edit

Sort by ยป oldest newest most voted

In general the TOBIT capabilities in SHAZAM are relatively easy to use. Here is a rough approach based on the problem you have described.

* Set a sample size and read the data, the variable names are in the first row
* so use the NAMES option. N.B. Without a sample command SHAZAM will attempt to
* figure it out from the amount of data read in.
sample 1 100
read (mydata.dat) / names

* Perform the TOBIT regression. Use LIST to print the value of the index,
* the density, the cumulative probabilities for index I and the observed,
* expected and conditional values of the dependent variable.
tobit credit location sector gender income collateral age size edu / list

* Perform some hypothesis tests to derive the standard error, and associated
* asymptotic normal test statistics for the regression coefficients. Note that
* the coefficient on the dependent variable (credit) is 1/sigma
test location/credit
test collateral/credit
test constant/credit


There are further details within the SHAZAM manual and a number of examples contained within the \help\manuals, \textbook and \help\examples subfolders of the SHAZAM installation directory. Examples include censored (double censored, left censored etc), multiplicative heteroskedasticity and the calculating of marginal effects.

more

## Stats

Asked: 2013-01-04 22:03:48 +0000

Seen: 2,309 times

Last updated: Oct 20 '13