Maybe some economist or statisticiansstatistician here could give a theoretical answer, but I would tackle the problem empirically, by fitting both the models and checking how they perform in terms of residuals, metrics such as AIC/BIC, and using validation procedures such as cross-validation.