Matching process.
This table presents the matching process employed to arrive at our final sample.
| Panel A: Matching Process for the Unionization Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union contract information from 2002 to 2021 (because ASSET4 data are available after 2002) | 352,750 |
| Manually match union contract records to Compustat | 38,122 |
| Aggregate establishment-year level observations to firm–year sample | 15,650 |
| Only keep firm–year observations with non-missing control variables | 5,934 |
| Match unionization sample to the Thomson Reuters ASSET4 ESG database | 2,120 |
| Add control sample through propensity score matching | 4,240 |
| Panel A: Matching Process for the Unionization Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union contract information from 2002 to 2021 (because ASSET4 data are available after 2002) | 352,750 |
| Manually match union contract records to Compustat | 38,122 |
| Aggregate establishment-year level observations to firm–year sample | 15,650 |
| Only keep firm–year observations with non-missing control variables | 5,934 |
| Match unionization sample to the Thomson Reuters ASSET4 ESG database | 2,120 |
| Add control sample through propensity score matching | 4,240 |
| Panel B: Matching Process for the Union Election Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union election records from 2002 to 2021 (because ASSET4 data are available after 2002) | 34,883 |
| Drop elections not used to certify unions (only keep Proof of Representative petitions) | 29,190 |
| Manually match remaining union election records to NETS | 17,581 |
| Use parent company information in NETS to link union election records to Compustat | 4,123 |
| Only keep the first union election if a firm has multiple union elections in one year | 2,289 |
| Match firms’ union election records to the Thomson Reuters ASSET4 ESG database | 663 |
| Limit sample to close elections where unions won or lost by a less than 20 percent margin of victory | 312 |
| Panel B: Matching Process for the Union Election Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union election records from 2002 to 2021 (because ASSET4 data are available after 2002) | 34,883 |
| Drop elections not used to certify unions (only keep Proof of Representative petitions) | 29,190 |
| Manually match remaining union election records to NETS | 17,581 |
| Use parent company information in NETS to link union election records to Compustat | 4,123 |
| Only keep the first union election if a firm has multiple union elections in one year | 2,289 |
| Match firms’ union election records to the Thomson Reuters ASSET4 ESG database | 663 |
| Limit sample to close elections where unions won or lost by a less than 20 percent margin of victory | 312 |
Matching process.
This table presents the matching process employed to arrive at our final sample.
| Panel A: Matching Process for the Unionization Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union contract information from 2002 to 2021 (because ASSET4 data are available after 2002) | 352,750 |
| Manually match union contract records to Compustat | 38,122 |
| Aggregate establishment-year level observations to firm–year sample | 15,650 |
| Only keep firm–year observations with non-missing control variables | 5,934 |
| Match unionization sample to the Thomson Reuters ASSET4 ESG database | 2,120 |
| Add control sample through propensity score matching | 4,240 |
| Panel A: Matching Process for the Unionization Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union contract information from 2002 to 2021 (because ASSET4 data are available after 2002) | 352,750 |
| Manually match union contract records to Compustat | 38,122 |
| Aggregate establishment-year level observations to firm–year sample | 15,650 |
| Only keep firm–year observations with non-missing control variables | 5,934 |
| Match unionization sample to the Thomson Reuters ASSET4 ESG database | 2,120 |
| Add control sample through propensity score matching | 4,240 |
| Panel B: Matching Process for the Union Election Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union election records from 2002 to 2021 (because ASSET4 data are available after 2002) | 34,883 |
| Drop elections not used to certify unions (only keep Proof of Representative petitions) | 29,190 |
| Manually match remaining union election records to NETS | 17,581 |
| Use parent company information in NETS to link union election records to Compustat | 4,123 |
| Only keep the first union election if a firm has multiple union elections in one year | 2,289 |
| Match firms’ union election records to the Thomson Reuters ASSET4 ESG database | 663 |
| Limit sample to close elections where unions won or lost by a less than 20 percent margin of victory | 312 |
| Panel B: Matching Process for the Union Election Sample . | |
|---|---|
| Operation . | Observations . |
| Compile union election records from 2002 to 2021 (because ASSET4 data are available after 2002) | 34,883 |
| Drop elections not used to certify unions (only keep Proof of Representative petitions) | 29,190 |
| Manually match remaining union election records to NETS | 17,581 |
| Use parent company information in NETS to link union election records to Compustat | 4,123 |
| Only keep the first union election if a firm has multiple union elections in one year | 2,289 |
| Match firms’ union election records to the Thomson Reuters ASSET4 ESG database | 663 |
| Limit sample to close elections where unions won or lost by a less than 20 percent margin of victory | 312 |
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
