Fixed effect for a multiple treatment analysis

[Jere RG]

Dear Statalist,

I'd like to study how the inflation of a country changes the years before and after it experiences a war using non conflict countries as a control group.

I have a panel with a column "Event" that indicates the three years before and after a war, being event=0 when the war started. I also have the inflation for every country every year and a dummy variable "conflict" that is 1 if the country experienced a war and 0 if not.

The data looks like this:

 
country
year
event
inflation
conflict

A
1990
.
8
1

A
1991
-3
4
1

A
1992
-2
9
1

A
1993
-1
5
1

A
1994
0
2
1

A
1995
1
4
1

A
1996
2
7
1

A
1997
3
2
1

A
1998
.
8
1

A
1999
.
10
1

B
1990
.
3
0

B
1991
.
6
0

B
1992
.
5
0

B
1993
.
3
0

B
1994
.
2
0

B
1995
.
7
0

B
1996
.
6
0

B
1997
.
9
0

B
1998
.
3
0

B
1999
.
6
0

C
1990
.
9
1

C
1991
.
10
1

C
1992
.
3
1

C
1993
-3
1
1

C
1994
-2
1
1

C
1995
-1
6
1

C
1996
0
7
1

C
1997
1
8
1

C
1998
2
3
1

C
1999
3
1
1

I would like to use the non conflict countries as a control group to determine the coefficients, confidence interval and p-values that the inflation has for the conflict countries the three years before and after the event. My ultimate goal is to have something that looks like this:

 
Event
dy/dx
Std. Err.
z
P>z
[95% Conf.
Interval]









-3
0.418232
4.333876
0.1
0.923
-8.07601
8.912474

-2
0.280138
4.248896
0.07
0.947
-8.04755
8.607821

-1
-1.1674
4.148458
-0.28
0.778
-9.29823
6.963426

0
0.002299
3.980347
0
1
-7.79904
7.803635

1
4.025524
3.844254
1.05
0.295
-3.50908
11.56012

2
7.296422
3.757677
1.94
0.052
-0.06849
14.66133

3
12.33721
3.721222
3.32
0.001
5.043751
19.63067

One colleague told me that maybe I should try with using dummy variables for the years and the conflict like this:

 xtreg inflation conflict 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999, fe

however, I don't see how this can give me information about the years I want to analyze (-3, -2, -1, 0, 1, 2, 3).

Any ideas to improve the regression?
Thanks!!