create new column where values in each row based on previous row value

[aj_goodnews]

Hello all, I have the data looks like this:

data have;
input id year xrd;
datalines;
1 1991 10
1 1992 95
1 1993 100
1 1994 20
2 1994 115
2 1995 120
2 1996 100
2 1997 110
3 1992 950
3 1993 105
3 1994 112
3 1995 120
;
run;

I want to create a new variable X=0.85*Lag(X)+ xrd, and the initial X for each id is set equal the xrd in the first year divided by 0.15, for example, X for id=1 in 1991(first year for id=1) is calculated as 10/0.15=66.66667, X for id=1 in 1992(second year for id=1) is calculated as 66.66667*0.85+ 95= 151.6667;

similarly, X for id=2 in 1994(first year for id=2) is calculated as 115/0.15=766.66667, X for id=2 in 1995(second year for id=2) is calculated as 766.66667*0.85 + 120= 771.667.

so eventually I want to have data look like this:

id	year	xrd	X
1	1991	10	66.66667
1	1992	95	151.6667
1	1993	100	228.9167
1	1994	20	214.5792
2	1994	115	766.6667
2	1995	120	771.6667
2	1996	100	755.9167
2	1997	110	752.5292
3	1992	950	6333.333
3	1993	105	5488.333
3	1994	112	4777.083
3	1995	120	4180.521

Thanks!

[novinosrin · Accepted solution]

HI @aj_goodnews 

data have;
input id year xrd;
datalines;
1 1991 10
1 1992 95
1 1993 100
1 1994 20
2 1994 115
2 1995 120
2 1996 100
2 1997 110
3 1992 950
3 1993 105
3 1994 112
3 1995 120
;
run;


data want;
 set have;
 by id;
 if first.id then x=xrd/0.15;
 else X=0.85*_iorc_+ xrd;
 _iorc_=x;
run;

[aj_goodnews]

Thank you!