Modelling Tourism Expenditure - Interpreting Coeff
Posted: Sat Aug 18, 2012 2:10 pm
Hi, would appreciate any thoughts on whether i'm specifying the model below correctly, if i am interpreting the coefficients accurately, if there is another way to do it more elegantly or if i'm committing any unpardonable econometrics error.
Problem: Per Capita Tourism Expenditure (PCE) = Length of stay (LOS) X Per Diem Expenditure (PDE)
I want to differentiate the effect of visitor characteristics on LOS and PDE, e.g. if visitors are business visitors (as opposed to leisure visitors), i want to know what impact that will have on length of stay, and per diem expenditure separately.
Proposed Model: PCE = B0 + B1 LOS + B2 LOS^2 + B3 Business.LOS +B4 Business.LOS^2 + B5 Business + u
Interpretation: According to the model, business visitors will:
1) Increase the length of stay by B3 (impact of business visitors on length of stay), and
2) Increase per diem expenditure by B4 (impact of business visitors on per diem expenditure) given that PCE / LOS = PDE,
3) Increase the minimum expenditure by B5 regardless of how many days visitor stays (i.e. shift y-intercept of PCE when LOS = 0 up by B5)
Would appreciate any thoughts on this, thanks!
Problem: Per Capita Tourism Expenditure (PCE) = Length of stay (LOS) X Per Diem Expenditure (PDE)
I want to differentiate the effect of visitor characteristics on LOS and PDE, e.g. if visitors are business visitors (as opposed to leisure visitors), i want to know what impact that will have on length of stay, and per diem expenditure separately.
Proposed Model: PCE = B0 + B1 LOS + B2 LOS^2 + B3 Business.LOS +B4 Business.LOS^2 + B5 Business + u
Interpretation: According to the model, business visitors will:
1) Increase the length of stay by B3 (impact of business visitors on length of stay), and
2) Increase per diem expenditure by B4 (impact of business visitors on per diem expenditure) given that PCE / LOS = PDE,
3) Increase the minimum expenditure by B5 regardless of how many days visitor stays (i.e. shift y-intercept of PCE when LOS = 0 up by B5)
Would appreciate any thoughts on this, thanks!