Issue with xtdidregress implementation

[Kiran Maharana]

BANK	Year	PRE=0/POST=1	ROE(%)	Inflation Rate (%)	Gross Domestic Product growth rate (%)	Real Interest rate	Total assets	Bank Size (Natural log of Total Assets)
1	2012	0	20.64	7.934386	5.456389	2.473522	4,47,321.47	13.01103278
1	2013	0	15.07	6.186504	6.386106	3.865993	5,47,135.44	13.21245166
1	2014	0	13.36	3.331757	7.410228	6.695176	6,59,504.53	13.39924413
1	2015	0	8.96	2.279588	7.996254	7.556488	7,14,988.55	13.4800218
1	2016	0	-13.48	3.237975	8.256305	6.232711	6,71,376.48	13.41708533
1	2017	1	3.44	3.969258	6.795383	5.327609	6,94,875.42	13.45148786
1	2018	1	-5.81	3.88424	6.453851	5.361666	7,19,999.77	13.48700617
1	2019	1	0.97	2.405517	3.871437	6.894875	7,80,987.40	13.5683143
1	2020	1	0.84	4.746549	-5.83105	4.20391	11,57,915.52	13.96213198
1	2021	1	1.11	8.538525	9.050278	0.147236	11,55,364.77	13.95992667
1	2022	1	8.93	8.225869	7.239693	0.315334	12,77,999.83	14.06080678
2	2012	0	14.00	7.934386	5.456389	2.473522	3,84,535.47	12.85979132
2	2013	0	12.25	6.186504	6.386106	3.865993	4,52,602.72	13.02277002
2	2014	0	10.14	3.331757	7.410228	6.695176	5,73,190.20	13.25897288
2	2015	0	5.57	2.279588	7.996254	7.556488	6,18,697.76	13.33537216
2	2016	0	-19.50	3.237975	8.256305	6.232711	6,09,913.93	13.32107312
2	2017	1	-5.04	3.969258	6.795383	5.327609	6,26,309.27	13.34759956
2	2018	1	-18.23	3.88424	6.453851	5.361666	6,09,574.83	13.32051699
2	2019	1	-14.37	2.405517	3.871437	6.894875	6,25,222.84	13.34586342
2	2020	1	-6.92	4.746549	-5.83105	4.20391	6,56,995.48	13.39543242
2	2021	1	4.83	8.538525	9.050278	0.147236	7,25,856.45	13.49510754
2	2022	1	6.75	8.225869	7.239693	0.315334	7,34,614.01	13.50710049
3	2012	0	9.91	7.934386	5.456389	2.473522	91,137.39	11.42012338
3	2013	0	13.66	6.186504	6.386106	3.865993	1,16,952.81	11.66952576
3	2014	0	5.61	3.331757	7.410228	6.695176	1,36,320.05	11.82276072
3	2015	0	5.84	2.279588	7.996254	7.556488	1,46,018.79	11.89149061
3	2016	0	1.19	3.237975	8.256305	6.232711	1,60,957.32	11.98889454

This is my dataset similar to this there are total of 12 banks in the dataset.

I want to perform DID to understand effect of implementation of a law on banks which is defined in binary column PRE=0/POST=1. ROE is dependent, rest are independent.

* Set up the dataset with panel structure
xtset BANK Year

* Generate the treatment dummy variable
gen treat_dummy_var = (PRE0POST1 == 1)

* Generate the post-treatment dummy variable
gen post_time_dummy = (Year >= 2017)

* Generate the interaction term for DID
gen did = treat_dummy_var * post_time_dummy

* Run xtdidregress with fixed effects specified for the bank and time fixed effects
xtdidregress (ROE InflationRate GrossDomesticProductgrowthrate RealInterestrate BankSizeNaturallogofTotalAssets) (did), group(BANK) time(Year)


Performed these commands for the last command getting this, please help out.
xtdidregress (ROE InflationRate GrossDomesticProductgrowthrate RealInterestrate BankSizeNaturallogofTotalAssets) (did), group(BANK) time(Year)
invalid group specification
None of the groups defined by BANK is a control.
r(198);

Would be really grateful for any answers. Thank You!

[George Ford]

Are some banks treated and others not? Which variable marks the treated?

[Kiran Maharana]

Dear George Ford ,

Have treated all the banks

g treat = cond(pre0post1 == 1 & year >= 2017, 1, 0)

xtset bank year

xtdidregress ( banksizenaturallogoftotalassets ) (treat), group( bank) time (year)


These are the commands which I have used

[Kiran Maharana]

Sharing the dataset so you can review it

[Enrique Pinzon (StataCorp)]

Dear Kiran,

I think your issue might be that there is no bank for which all values of the variable -treat- is zero for every period in the estimation sample. This is what the error message is saying. It suggest that all banks eventually are subject to the law and none of them remain untreated, i.e. none of them are controls. There is an FAQ on the topic that might help you diagnose the issue.

https://www.stata.com/support/faqs/s...specification/

[Kiran Maharana]

Thank You for the reply, got the issue I have marked treated to all the banks in my dataset. As per the case in the dataset, there was a law which applied to all the banks after 2017 and due to this, I have marked all the banks as treated after 2017. And as per the response, I got to know maybe DID cannot work for this case.

[Jared Greathouse]

No control group=no DID