Need help with creating first difference operator for regression - econometrics

[azad mehrpoor]

Hi There Everyone.

I hope you're all well.
I had a question,
I'm really lost - i need to create a first difference operator in stata, I've been struggling to understand how to do this.

My data looks like this:

observation	Year	quarter	r	bill rate	pi
1	1950	1	0.51	11.53832	0.5
2	1950	2	0.51	11.53832	4.06
3	1950	3	0.55	11.56103	5.5
4	1950	4	0.623	11.60105	9.85
5	1951	1	0.693	11.61927	14
6	1951	2	0.75	11.60116	11.67
7	1951	3	0.803	11.5962	8
8	1951	4	0.913	11.59763	3.26
9	1952	1	0.907	11.64858	3.39
10	1952	2	1.013	11.67504	-2.93
11	1952	3	1.11	11.68812	-2.58

this is the regression i have to run

basically Δr_t=r_t - r_t-1
Any help and guidance would be greatly appreciated please.

[Marcos Almeida]

tsset observation year

now you can use the ‘D.’ operator.

[Clyde Schechter]

It looks to me like the data in #1 are quarterly, not yearly. And observation does not appear to be a panel identifier. This looks more like just a single time series to me. So I think it's just a little more complicated:

gen qdate = yq(Year, quarter)
format qdate %tq
tsset qdate

Then they can use the D. operator, which they can read about in -help tsvarlist-.

[azad mehrpoor]

Hi there thank you so much,
i had another issue,

. gen y1 = D1.GDPlog
(1 missing value generated)

. gen pi1 = D1.pi
(1 missing value generated)

. drop pi1

. gen pi1 = L1.pi
(1 missing value generated)

. gen dr1 = D1.r
(1 missing value generated)

. gen dr2 = D2.r
(2 missing values generated)

. gen dr = D.r
(1 missing value generated)

. regress dr pi1 y1 dr2 dr1

Source | SS df MS Number of obs = 186
-------------+------------------------------ F( 4, 181) = .
Model | 165.132359 4 41.2830897 Prob > F = .
Residual | 0 181 0 R-squared = 1.0000
-------------+------------------------------ Adj R-squared = 1.0000
Total | 165.132359 185 .892607344 Root MSE = 0

------------------------------------------------------------------------------
dr | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
pi1 | -9.47e-18 . . . . .
y1 | -1.74e-15 . . . . .
dr2 | -2.87e-16 . . . . .
dr1 | 1 . . . . .
_cons | 4.86e-17 . . . . .
------------------------------------------------------------------------------

i dont know what i'm doing wrong, can you specify what i'm doing wrong?

[Clyde Schechter]

You have

gen dr1 = D1.r
gen dr = D.r

D.r and D1.r are the same thing. So your dependent variable dr, is exactly equal to dr1. Consequently the correct regression has the coefficient of dr1 = 1, and the constant term and all other coefficients equal to 0. In fact, to within some very tiny rounding errors, that's what you got.

I'm pretty sure you didn't intend for dr and dr1 to be the same variable. But I wouldn't know what you actually need to do instead.