Issue with adding "lackfit" statment

[mpiotro2]

I'm having an issuing adding a "lackfit" statement to a proc regression model... I keep getting a "0" answer for my Pure Error Sum of Squares value, which I don't think is correct. If my data set has 3 X variables, is the lackfit statement still functional? 

Obs	CITY	Y	X1	X2	X3
1	1	11.2	587	16.5	6.2
2	2	13.4	643	20.5	6.4
3	3	40.7	635	26.3	9.3
4	4	5.3	692	16.5	5.3
5	5	24.8	1248	19.2	7.3
6	6	12.7	643	16.5	5.9
7	7	20.9	1964	20.2	6.4
8	8	35.7	1531	21.3	7.6
9	9	8.7	713	17.2	4.9
10	10	9.6	749	14.3	6.4
11	11	14.5	7895	18.1	6
12	12	26.9	762	23.1	7.4
13	13	15.7	2793	19.1	5.8
14	14	36.2	741	24.7	8.6
15	15	18.1	625	18.6	6.5
16	16	28.9	854	24.9	8.3
17	17	14.9	716	17.9	6.7
18	18	25.8	921	22.4	8.6
19	19	21.7	595	20.2	8.4
20	20	25.7	3353	16.9	6.7

[PaigeMiller]

Lack-of-fit/pure error can only be computed if you have one or more exact replicates of your points (among the x-variables). You have zero exact replicates of your data points.

So SAS is giving the correct output.