Cumulative inflation

[Jean Jacques]

Hi! I have a panel data with prices (p) item and time(t). I would like to compute the cumulative inflation. I.e.:

t	p	item	p/p[t-1]
1	10	apple	.
2	12	apple	1.2
3	10	apple	0.833
4	15	apple	1.5
1	20	orange	.
3	21	orange	1.05
4	20	orange	0.952
5	20	orange	1
6	20	orange	1

that implies for apples the cumulative inflation was: 1.2 for t=2, 1 for t=3 and 1.5 for t=4 and for oranges: 1.05 for t=3 and then 1 until the end of the period. How can I compute that with stata.

For the last column (the monthly inflation) I have:

gen price_change_per = price/price[_n-1] if item==item[_n-1]

but I cannot come up with the doce for the cumulative one? I'm concerned not only for the final value of the cumulative inflation (1.5 for apples and 1 for oranges) but also for the evolution it has.

Best and thanks a lot!

[Joro Kolev]

Is that what you want?

. sort item t

. by item: gen ratio = p/p[_n-1]
(2 missing values generated)

. list, sep(0)

     +-----------------------------------+
     | t    p     item   ppt1      ratio |
     |-----------------------------------|
  1. | 1   10    apple      .          . |
  2. | 2   12    apple    1.2        1.2 |
  3. | 3   10    apple   .833   .8333333 |
  4. | 4   15    apple    1.5        1.5 |
  5. | 1   20   orange      .          . |
  6. | 3   21   orange   1.05       1.05 |
  7. | 4   20   orange   .952    .952381 |
  8. | 5   20   orange      1          1 |
  9. | 6   20   orange      1          1 |
     +-----------------------------------+