[Michael Bond]

Am I right about this interpretation? the last interpretation

[Clyde Schechter]

No, it's not right. Again, you have to do the algebra:

At age A, the expected value of log expense is:
0.123*A  + 0.232*A + other terms

At age 1.10*A, the expected value of log expense is:
0.123*1.10*A  +0.232*1.10*A + the same other terms.

So the difference in log expense is:
0.123*0.10*A + 0.232*.10*A, or
0.3905*A

Notice that A does not drop out of this calculation. So the change associated with a 10% increase in A depends on A itself and cannot be expressed as a single number.

I should also point out that had you done the regression using factor-variable notation correctly in the first place, you would be able to get the various marginal effects you are looking for out of the -margins- command without having to think through the algebra (and risk getting it wrong). In any case, I've now shown you several example of how these calculations are done, and I expect you can do them on your own from this point.

[Michael Bond]

Clyde no more questions for you. thanks a lot, i will figure out myself. You have been such a saviour

[Michael Bond]

I hope my calculations are right, please check it for me
The expression exponential^ (.232+2.905)-1 *100 that you proposed would actually be the percentage difference in expenses between Plan 3 at age 1 and no plan at age 0.
Answer

As you said, it means, 1 percentage point or unit increae in age leads to 26% increase in expense in Plan 3

Another example, Whereas exponential^ (.232*.953+2.905*0.953)-1 *100= (0.221+2.768)= exponential^ (2.989)-1= 19.86-1= 18.86% percent. Hope these are right and I have improved my knowledge. log 1.1= 0.953