Hi Statalist members,

I'm trying to figure out if there is a two-way relationship between stock market development (stmcap_n) and venture capital investment (gVC). I think using SEM would be more beneficial than using the two separate OLS models. The model work for one way from gVC->stmcap_n but not for the other way, which has the error "initial value not feasible". Is there any problem with my model or data? Is it because my model is unidentified? How to determine the degree of freedom in my model. Actually, I only have 2 hypotheses: one is gVC affect stmcap_n and the other for the other way around. Do I need to create more hypotheses?
Can you please recommend any comprehensive document for a new starter of SEM? Or if for what I want to test, is there any other model more suitable?
Array