Dear Statalisters

This is my first post so please excuse if the question is not posed correctly.

While running a tobit regression (with censoring at 0), my dependent variable captures spending on education (hence there are many zeros), and hence I would like an estimate of proportional changes in expenditure with respect to the regressors. In a simple regression model, this is achieved by taking the log of the dependent variable, however, here the dependent variable has zeros and taking a log results in missing values since log of zero is not defined.

Could anyone suggest how to get around this issue? Will the margins command help here?