honestdid for Stata

[blanelewis]

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth

[jonathandroth]

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

[blanelewis]

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.***>

[jonathandroth]

Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.***> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

[blanelewis]

Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.***> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.***>

[blanelewis]

Hi Jon,

I hope you don’t mind me following up with you on this.

I guess you would have seen Clément’s reply to your question on Twitter. I’ve done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.*** Clément’s commands to retrieve the VCOV matrix:

@.***

Now, I’m not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with ‘bootstrap’ as the VCOV matrix (and various alternatives) with the following results.

@.***

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.>, blanelewis/quasi-experimental @.> Cc: Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.***> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.***>

[jonathandroth]

Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.***> wrote:

Hi Jon,

I hope you don’t mind me following up with you on this.

I guess you would have seen Clément’s reply to your question on Twitter. I’ve done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.*** Clément’s commands to retrieve the VCOV matrix:

@.***

Now, I’m not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with ‘bootstrap’ as the VCOV matrix (and various alternatives) with the following results.

@.***

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.>, blanelewis/quasi-experimental @.> Cc: Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.***> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

[blanelewis]

Hi Jon,

Sounds good. Thanks very much.

Regards, Blane

From: Jonathan Roth @.> Date: Tuesday, 15 November 2022 at 1:00 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.***> wrote:

Hi Jon,

I hope you don’t mind me following up with you on this.

I guess you would have seen Clément’s reply to your question on Twitter. I’ve done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.*** Clément’s commands to retrieve the VCOV matrix:

@.***

Now, I’m not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with ‘bootstrap’ as the VCOV matrix (and various alternatives) with the following results.

@.***

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.>, blanelewis/quasi-experimental @.> Cc: Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.***> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors ‘e(estimates)’ and ‘e(variances) respectively. It doesn’t store a variance-covariance matrix like the other diff-in-diff commands. Maybe that’s the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.***

From: Jonathan Roth @.> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>, Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.***> wrote:

I have been experimenting with using ‘honestdid’ across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, ‘eventstudyinteract’), Callaway and Sant’Anna (2021, ‘csdid’) and Borusyak et al (2022, ‘did_imputation’). But I have been unable to get it to work with de Chaisemartin and D’Haultfuille (2022,'did_multiplegt’). I suspect this is because their procedure does not save the “last estimates” in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that ‘honestdid’ is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1313754817&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7KpNvFRM5EehN3R1LdkRYgzxEyUju7EVedqtnHHRpu8%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRL2SF6CNWZNIHRJ74LWIJAWPANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=zAe8ESbruPGfnMUZ1aTQ%2FuDhltKF2mHaRxp1SVVS4e0%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.***>

[blanelewis]

Hi Jon, I see from my twitter timeline you've been keeping busy. (Your recent paper on the dangers of ln(y+1) and arcsinh(y), which I found there, is very interesting; congratulations.) Just thought I would check on this-any new developments? Thanks. Best regards, Blane

From: Jonathan Roth @.> Sent: Tuesday, 15 November 2022 1:00 AM To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>; Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1)

Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.<mailto:@.>> wrote:

Hi Jon,

I hope you don't mind me following up with you on this.

I guess you would have seen Clément's reply to your question on Twitter. I've done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.<mailto:@.> Clément's commands to retrieve the VCOV matrix:

@.<mailto:@.>

Now, I'm not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with 'bootstrap' as the VCOV matrix (and various alternatives) with the following results.

@.<mailto:@.>

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.<mailto:@.>>, blanelewis/quasi-experimental @.<mailto:@.>> Cc: Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1) Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.<mailto:@.>> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors 'e(estimates)' and 'e(variances) respectively. It doesn't store a variance-covariance matrix like the other diff-in-diff commands. Maybe that's the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.<mailto:@.>

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.<mailto:@.>> wrote:

I have been experimenting with using 'honestdid' across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, 'eventstudyinteract'), Callaway and Sant'Anna (2021, 'csdid') and Borusyak et al (2022, 'did_imputation'). But I have been unable to get it to work with de Chaisemartin and D'Haultfuille (2022,'did_multiplegt'). I suspect this is because their procedure does not save the "last estimates" in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that 'honestdid' is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1313754817&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7KpNvFRM5EehN3R1LdkRYgzxEyUju7EVedqtnHHRpu8%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRL2SF6CNWZNIHRJ74LWIJAWPANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=zAe8ESbruPGfnMUZ1aTQ%2FuDhltKF2mHaRxp1SVVS4e0%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.**@.>>

[jonathandroth]

Ah, sorry for the delay. We worked with Clement and team so that the latest versions of the two packages should now be compatible. I'd thought my RA had updated the Github README, but turns out I forgot to reply to the email approving the changes. It should be up soon, but in any event, here's the example that will be added to the README:

HonestDiD is also compatible with the estimator introduced in Chaisemartin and D'Haultfoeuille (2020), available for Stata via the did_multiplegt package.

local mixtape https://raw.githubusercontent.com/Mixtape-Sessions
use `mixtape'/Advanced-DID/main/Exercises/Data/ehec_data.dta, clear
gen byte D = (year >= yexp2) & !mi(yexp2)
did_multiplegt dins stfips year D, robust_dynamic dynamic(5) placebo(5)
breps(50) cluster(stfips)
honestdid, pre(7/11) post(1/6) vcov(didmgt_vcov) b(didmgt_results_no_avg)

On Tue, Dec 13, 2022 at 6:56 PM blanelewis @.***> wrote:

Hi Jon, I see from my twitter timeline you've been keeping busy. (Your recent paper on the dangers of ln(y+1) and arcsinh(y), which I found there, is very interesting; congratulations.) Just thought I would check on this-any new developments? Thanks. Best regards, Blane

From: Jonathan Roth @.> Sent: Tuesday, 15 November 2022 1:00 AM To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>; Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1)

Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.<mailto:@.>> wrote:

Hi Jon,

I hope you don't mind me following up with you on this.

I guess you would have seen Clément's reply to your question on Twitter. I've done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.<mailto:@.> Clément's commands to retrieve the VCOV matrix:

@.<mailto:@.>

Now, I'm not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with 'bootstrap' as the VCOV matrix (and various alternatives) with the following results.

@.<mailto:@.>

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.<mailto:@.>>, blanelewis/quasi-experimental @.<mailto:@.>> Cc: Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.<mailto:@.>> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors 'e(estimates)' and 'e(variances) respectively. It doesn't store a variance-covariance matrix like the other diff-in-diff commands. Maybe that's the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.<mailto:@.>

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.<mailto:@.>> wrote:

I have been experimenting with using 'honestdid' across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, 'eventstudyinteract'), Callaway and Sant'Anna (2021, 'csdid') and Borusyak et al (2022, 'did_imputation'). But I have been unable to get it to work with de Chaisemartin and D'Haultfuille (2022,'did_multiplegt'). I suspect this is because their procedure does not save the "last estimates" in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that 'honestdid' is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0

, or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub <

https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313

, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1313754817&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7KpNvFRM5EehN3R1LdkRYgzxEyUju7EVedqtnHHRpu8%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRL2SF6CNWZNIHRJ74LWIJAWPANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=zAe8ESbruPGfnMUZ1aTQ%2FuDhltKF2mHaRxp1SVVS4e0%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.**@.>>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1350100823, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFB4YDFRYCOYTZPC4UTWNEEL5ANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>

[blanelewis]

It works! Thanks very much.

From: Jonathan Roth @.> Sent: Wednesday, 14 December 2022 11:33 AM To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>; Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1)

Ah, sorry for the delay. We worked with Clement and team so that the latest versions of the two packages should now be compatible. I'd thought my RA had updated the Github README, but turns out I forgot to reply to the email approving the changes. It should be up soon, but in any event, here's the example that will be added to the README:

HonestDiD is also compatible with the estimator introduced in Chaisemartin and D'Haultfoeuille (2020), available for Stata via the did_multiplegt package.

local mixtape https://raw.githubusercontent.com/Mixtape-Sessions
use `mixtape'/Advanced-DID/main/Exercises/Data/ehec_data.dta, clear
gen byte D = (year >= yexp2) & !mi(yexp2)
did_multiplegt dins stfips year D, robust_dynamic dynamic(5) placebo(5)
breps(50) cluster(stfips)
honestdid, pre(7/11) post(1/6) vcov(didmgt_vcov) b(didmgt_results_no_avg)

On Tue, Dec 13, 2022 at 6:56 PM blanelewis @.<mailto:@.>> wrote:

Hi Jon, I see from my twitter timeline you've been keeping busy. (Your recent paper on the dangers of ln(y+1) and arcsinh(y), which I found there, is very interesting; congratulations.) Just thought I would check on this-any new developments? Thanks. Best regards, Blane

From: Jonathan Roth @.<mailto:@.>> Sent: Tuesday, 15 November 2022 1:00 AM To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>; Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue #1)

Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> wrote:

Hi Jon,

I hope you don't mind me following up with you on this.

I guess you would have seen Clément's reply to your question on Twitter. I've done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.<mailto:@.mailto:***@***.***%3cmailto:***@***.***> Clément's commands to retrieve the VCOV matrix:

@.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>

Now, I'm not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with 'bootstrap' as the VCOV matrix (and various alternatives) with the following results.

@.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>, blanelewis/quasi-experimental @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Cc: Author @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Cc: Blane Lewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>, Author @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors 'e(estimates)' and 'e(variances) respectively. It doesn't store a variance-covariance matrix like the other diff-in-diff commands. Maybe that's the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>

From: Jonathan Roth @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Cc: Blane Lewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>, Author @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>> wrote:

I have been experimenting with using 'honestdid' across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, 'eventstudyinteract'), Callaway and Sant'Anna (2021, 'csdid') and Borusyak et al (2022, 'did_imputation'). But I have been unable to get it to work with de Chaisemartin and D'Haultfuille (2022,'did_multiplegt'). I suspect this is because their procedure does not save the "last estimates" in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that 'honestdid' is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0

, or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>

- Reply to this email directly, view it on GitHub <

https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313

, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>

- Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.mailto:***@***.***%3cmailto:***@***.***>>

- Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1313754817&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7KpNvFRM5EehN3R1LdkRYgzxEyUju7EVedqtnHHRpu8%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRL2SF6CNWZNIHRJ74LWIJAWPANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=zAe8ESbruPGfnMUZ1aTQ%2FuDhltKF2mHaRxp1SVVS4e0%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.**@.mailto:***@***.******@***.***>>

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1350100823, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFB4YDFRYCOYTZPC4UTWNEEL5ANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHubhttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1350174909&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C9148c5b1c96b4c84028108dadd6ab4ea%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638065747632322520%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=EFVk8YKFgHLCQT2pgZeuk0kk%2FyICqU6Qo2I%2Bcma3lsM%3D&reserved=0, or unsubscribehttps://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKFH7RCJMEY5IYNKN3WNEISLANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C9148c5b1c96b4c84028108dadd6ab4ea%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638065747632322520%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Xn5O9qtCoGMfYeR6CW4vzi7NPLFj2GFnC6IlB72thKk%3D&reserved=0. You are receiving this because you authored the thread.Message ID: @.**@.>>

[jonathandroth]

Fantastic! Thanks for using our tools :)

On Tue, Dec 13, 2022 at 7:32 PM Jonathan Roth @.***> wrote:

Ah, sorry for the delay. We worked with Clement and team so that the latest versions of the two packages should now be compatible. I'd thought my RA had updated the Github README, but turns out I forgot to reply to the email approving the changes. It should be up soon, but in any event, here's the example that will be added to the README:

HonestDiD is also compatible with the estimator introduced in Chaisemartin and D'Haultfoeuille (2020), available for Stata via the did_multiplegt package.

local mixtape https://raw.githubusercontent.com/Mixtape-Sessions
use `mixtape'/Advanced-DID/main/Exercises/Data/ehec_data.dta, clear
gen byte D = (year >= yexp2) & !mi(yexp2)
did_multiplegt dins stfips year D, robust_dynamic dynamic(5) placebo(5)
breps(50) cluster(stfips)
honestdid, pre(7/11) post(1/6) vcov(didmgt_vcov) b(didmgt_results_no_avg)

On Tue, Dec 13, 2022 at 6:56 PM blanelewis @.***> wrote:

Hi Jon, I see from my twitter timeline you've been keeping busy. (Your recent paper on the dangers of ln(y+1) and arcsinh(y), which I found there, is very interesting; congratulations.) Just thought I would check on this-any new developments? Thanks. Best regards, Blane

From: Jonathan Roth @.> Sent: Tuesday, 15 November 2022 1:00 AM To: blanelewis/quasi-experimental @.> Cc: Blane Lewis @.>; Author @.> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Hi Blane -- let me look into this and try to get back to you later this week.

Jon

On Fri, Nov 11, 2022 at 4:44 PM blanelewis @.<mailto:@.>> wrote:

Hi Jon,

I hope you don't mind me following up with you on this.

I guess you would have seen Clément's reply to your question on Twitter. I've done what he suggests but am not sure how to make it all work with honestdid. (It would help if I were better at programming but unfortunately that is not the case.)

Anyway, here is what I get.

The basic did_multiplegt output:

@.<mailto:@.> Clément's commands to retrieve the VCOV matrix:

@.<mailto:@.>

Now, I'm not quite sure what to do with that in the context of implementing honestdid.

I tried running your command with 'bootstrap' as the VCOV matrix (and various alternatives) with the following results.

@.<mailto:@.>

Do you have any suggestions? Thanks very much!

Best regards, Blane

From: Blane Lewis @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:30 am To: blanelewis/quasi-experimental @.<mailto:@.>>, blanelewis/quasi-experimental @.<mailto:@.>> Cc: Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Excellent! Thanks Jon. Cheers, Blane

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 7:25 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Hi Blane,

Thanks very much! Yes, we need a joint variance-covariance matrix for the pre-treatment and post-treatment event study coefficients. I wrote to Xavier and Clement to see whether it is possible to get this from the did_multiplegt package. I'll let you know if I hear anything back from them. Thanks!

Jon

On Mon, Nov 7, 2022 at 1:22 PM blanelewis @.<mailto:@.>> wrote:

Hi Jonathon,

Thanks very much for getting back to me. The did_multiplegt code, estimation output, and honestdid command, along with the error are shown below. did_multiplegt stores the matrices for coefficient estimates and variances in vectors 'e(estimates)' and 'e(variances) respectively. It doesn't store a variance-covariance matrix like the other diff-in-diff commands. Maybe that's the issue? Anyway, any advice you can provide would be appreciated. Please let me know if you need any additional information. Thanks again.

Best regards, Blane

@.<mailto:@.>

From: Jonathan Roth @.<mailto:@.>> Date: Tuesday, 8 November 2022 at 1:06 am To: blanelewis/quasi-experimental @.<mailto:@.>> Cc: Blane Lewis @.<mailto:@.>>, Author @.<mailto:@.>> Subject: Re: [blanelewis/quasi-experimental] honestdid for Stata (Issue

1)

Thanks, Blane! I do not personally have experience using did_multiplegt in Stata (I'm an R user myself). Can you please provide a minimal working example of how you're trying to use did_multiplegt and the error you got, and then we can go from there?

On Sat, Nov 5, 2022 at 4:56 PM blanelewis @.<mailto:@.>> wrote:

I have been experimenting with using 'honestdid' across some of the new diff-in-diff methods. It works well with the procedures developed by Sun and Abraham (2021, 'eventstudyinteract'), Callaway and Sant'Anna (2021, 'csdid') and Borusyak et al (2022, 'did_imputation'). But I have been unable to get it to work with de Chaisemartin and D'Haultfuille (2022,'did_multiplegt'). I suspect this is because their procedure does not save the "last estimates" in the same way that the other methods do. One can save those estimates to another file, but they are saved in a manner that 'honestdid' is apparently unable to read. Can you please offer me some guidance on how to implement it? Thank you. @jonathandroth https://github.com/jonathandroth

- Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDK6QGYWO2ZFMKNR33WG3CYFANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1305665635&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=dUDlwtPSG476yS81weiq4LACEU%2FxdyiQw%2FRQyYnSo2Q%3D&reserved=0

, or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRJ4WZWT3ZLYB73KXSLWHEEE3ANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc0b19ceb491f4172881d08dac0c937a8%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034267707518148%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=3DlVXiSXM5svtrAo97KSXFKluxO%2BtmgeMY9aMsf9tJ4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub <

https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1306013313

, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AE6EXFDOQ4NXSLN6FW36UKTWHFCGXANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1306145487&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wi9T98UbYRrcsygtwY6EoXTkT1wHsGd2mcNYPPyRcTE%3D&reserved=0 , or unsubscribe<

https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRKREMB7GAKXODLIZ7TWHFQTFANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7C621594ba63c346ceac9508dac0fe3411%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638034495275778577%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OuY9hW2A9h7FfLHOFJRouRnwCCuoga8Jq4V%2F89baWj4%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub < https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1312214400 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AE6EXFG7L3OWZX7N562M6Y3WH245LANCNFSM6AAAAAARYC4NS4

. You are receiving this because you were mentioned.Message ID: @.<mailto:@.>>

- Reply to this email directly, view it on GitHub< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fblanelewis%2Fquasi-experimental%2Fissues%2F1%23issuecomment-1313754817&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7KpNvFRM5EehN3R1LdkRYgzxEyUju7EVedqtnHHRpu8%3D&reserved=0>, or unsubscribe< https://aus01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA4AZFRL2SF6CNWZNIHRJ74LWIJAWPANCNFSM6AAAAAARYC4NS4&data=05%7C01%7Cblane.lewis%40anu.edu.au%7Cc548e5b7dfc54e5f5c0d08dac6488af7%7Ce37d725cab5c46249ae5f0533e486437%7C0%7C0%7C638040312125563331%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=zAe8ESbruPGfnMUZ1aTQ%2FuDhltKF2mHaRxp1SVVS4e0%3D&reserved=0

. You are receiving this because you authored the thread.Message ID: @.**@.>>

— Reply to this email directly, view it on GitHub https://github.com/blanelewis/quasi-experimental/issues/1#issuecomment-1350100823, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFB4YDFRYCOYTZPC4UTWNEEL5ANCNFSM6AAAAAARYC4NS4 . You are receiving this because you were mentioned.Message ID: @.***>