Hello Stata community,
First, I wonder if I posted in the right section since my question is not strictly related to the technical aspects of using the software.
Getting to the point, I am using panel data to estimate the behavior of Italian banks' non-performing loans (NPLs) with respect to a given set of variables.
This is the first time I am working with panel data.
Since my dependent variable represents the stock of a bank's nonperforming loans (i.e., those loans for which the bank estimates it will not be able to obtain the amount originally lent and therefore will become impaired, nonperforming to be exact), it seems to me that the explanatory variables I want to use actually should be used as lags. This is because I imagine that a certain kind of time is needed to generate an effect on NPLs. For example, a change in a bank's operating costs needs some time to produce an effect on nonperforming loans. Since in my view, this reasoning (i.e., that it takes time for certain effects to occur and influence NPLs) is almost common to all variables, I am left wondering whether I should also use the variables as non-lags.
I understand that this is not a question strictly related to the use of the software, but I have also been advised to reach the community for points of concern like this one.
I thank all those who will find time to help me.
Best regards.
First, I wonder if I posted in the right section since my question is not strictly related to the technical aspects of using the software.
Getting to the point, I am using panel data to estimate the behavior of Italian banks' non-performing loans (NPLs) with respect to a given set of variables.
This is the first time I am working with panel data.
Since my dependent variable represents the stock of a bank's nonperforming loans (i.e., those loans for which the bank estimates it will not be able to obtain the amount originally lent and therefore will become impaired, nonperforming to be exact), it seems to me that the explanatory variables I want to use actually should be used as lags. This is because I imagine that a certain kind of time is needed to generate an effect on NPLs. For example, a change in a bank's operating costs needs some time to produce an effect on nonperforming loans. Since in my view, this reasoning (i.e., that it takes time for certain effects to occur and influence NPLs) is almost common to all variables, I am left wondering whether I should also use the variables as non-lags.
I understand that this is not a question strictly related to the use of the software, but I have also been advised to reach the community for points of concern like this one.
I thank all those who will find time to help me.
Best regards.
Comment