3.2. Successor-Predecessor File (SPF)
3.2.1. Overview
The firm identifier used in the LEHD infrastructure tables is a state-specific unemployment insurance (UI) account number used to administer the tax and benefits of the UI system. These account numbers, prepended by a state identifier and here called “SEIN,” can change for a number of reasons, including a simple change in legal form, a merger with another firm, or a divestiture into several firms. Typically, the separation of a worker from an employer is identified by a change in the SEIN on that workers’ UI wage records. If an employer changes SEINs but makes no other changes, the worker would appear to have left the original firm even though his or her employment status remains unchanged from an economic viewpoint. These spurious employer changes are known to induce biases in both employment and job flow statistics. For example, a simple change in account numbers would lead to the observation of a firm closing and all workers separating, even though all (or most) workers remain employed.
Successor-Predecessor File (SPF_ZZ)
Scope: State
Key: SEIN SEIN_SUCC QTIME
To identify and correct for such events, the Successor-Predecessor File (SPF) was created to track large worker movements between SEINs using quarterly UI wage records. Benedetto et al. [2007] used the SPF for an early analysis in one particular state of the impact of such an exercise. The SPF provides link characteristics based on the number of workers leaving an origin SEIN and entering a destination SEIN, presented in absolute terms and relative terms to the overall size of the firms. In addition to the UI-sourced information, predecessor and/or successor firms may be reported directly on the firm microdata from the Quarterly Census for Employment and Wages (QCEW or ES202), and these linkages are also recorded in the SPF. Note, links are identified using UI data for one state at a time. Cross-state firm relocations are not identified here, and such events would be interpreted as separations in the origin state and accessions in the destination state.
Research use of the SPF
The linkages identified in the SPF have already been incorporated into jobs on the JHF and QWI tables as described in Accounting for Successor-Predecessor Events. The JHF is the recommended input for analyses of job starts, tenure, separations, at the job level, and the QWI can be used for establishment-level calculations of job flows. The SPF itself may not be needed by most researchers.
Researchers who wish to explore an alternate method of accounting for SEIN changes (e.g., modifying the thresholds that identify firm linkages) may want to use the full set of flows presented in the SPF. Researchers interested in the evolution of firm structure across mergers and divestures may also benefit from linkages found in the SPF.
3.2.2. User Guidance
Identifying Transitions in UI Data
SPF processing starts by finding all job-level separations and accessions from the quarterly UI data. A separation in period t is defined for PIK i from firm j when i receives earnings from j in quarter t, but not in t+1. Similarly, an accession in period t is defined for PIK i into firm k when i receives earnings from k in quarter t but did not in t-1. Then for every PIK, separations in period t are linked with all accessions in periods t or t+1 to identify possible candidates for SEIN transitions at the job level. The timing in this linkage strategy recognizes that a transition could be reported to occur within a quarter or across quarters, or indeed a combination of both at the firm level.
These SEIN transitions found at the PIK level are then aggregated across PIKs to get the count of transitions from predecessor j (SEIN) to successor k (SEIN_SUCC) in period t (qtime), where t is the quarter of separation from i (accession to j in t or t+1). The count of PIKs making the transition is recorded in the NUM_LEFT variable. For informational purposes, the MATCH_PERIOD variable is created to report the share of jobs for which the accession occurred in t+1. The de minimis SEIN transitions are ignored by removing links where the NUM_LEFT is less than 6. The remaining records form the UI-sourced frame of the SPF.
Assessing Strength of SEIN Transitions in UI Data
Several factors are used to assess whether a linkage between two firms in a quarter constitutes a predecessor-successor relationship:
The share of employment at the predecessor that transits to the successor
The share of employment at the successor that transited from predecessor
Death of the predecessor firm, such that UI records are no longer reported after the transition
Birth of the successor firm, such that UI records were not reported prior to the transition
The shares of jobs are provided in the variables RATIO (predecessor) and SUCC_RATIO (successor). Because of timing issues employment patterns and transitions, the numerator (the count of jobs in the transition) is not always a strict subset of the employment count at the firm, so the ratios are considered estimates of the share of the firm employment that is part of the transition. The ratio is capped at 100%. A ratio of 80% is considered the threshold for representing a large share of the predecessor or successor.
Composite flags LINK_UI and SUCC_LINK_UI are then created to combine the strength of the link with information on firm birth and death. The values of the flags are as follows (description for SUCC_LINK_UI in parentheses):
1 |
Firm death (birth), flow constitutes a large percentage of predecessor (successor) employment |
2 |
Firm death (birth), flow does not constitute a large percentage of predecessor (successor) employment. |
4 |
Not firm death (birth), flow constitutes a large percentage of predecessor (successor) employment |
5 |
Not firm death (birth), flow does not constitute a large percentage of predecessor (successor) employment. |
Directly Reported Links from Firm Data
The UI sourced flows on the SPF are supplemented with links that are provided on the firm microdata (ES202). In each quarter, the firm can be reported with having a predecessor, successor, or both. These links are collected and merged onto the UI-sourced links by quarter. The SOURCE flag is created to identify which source provided the SEIN transition in a quarter, with values as follows:
UI |
Wage record flow observed in UI data, only |
ES |
Transition reported in ES202 data, only |
UIES |
Wage record flow observed in UI and reported in ES202 data |
Because the timing of the report on the ES202 and the observation of the event in UI data do not always coincide precisely, an ES202 transition reported between two firms in a proximate quarter will be flagged as happening at the time the UI transition is observed.
Identifying Successor-Predecessor Relationships
For LEHD data products, only the strongest links based on the UI wage record flows are used to filter out spurious employer identifier changes. A link between two SEINs is also used if it was directly reported on the input ES202 data, even if it does not meet the UI requirements. The following are the precise business rules used:
80% of jobs at the predecessor are linked to accessions at the successor, and the predecessor firm dies after the transition.
LINK_UI=1
80% of jobs at the successor are linked to separations from the predecessor, and the successor firm is born in the transition.
SUCC_LINK_UI=1
The separations in the flow represent 80% of employment at the successor, and the accessions in the flow represent 80% of employment at the successor.
LINK_UI=4 AND SUCC_LINK_UI=4
A flow of wage records is identified between firms in a quarter, and the predecessor and/or the successor SEIN is reported on the ES202.
SOURCE=”UIES”
Successor-predecessor relationships that meet at least one of these requirements are used to link job spells in the Job History File (JHF). The process of job linkage is further described in the documentation for the JHF.
3.2.3. Codebook
Table Metadata for Successor-Predecessor File (SPF_ZZ)
| State Approval Required | IRS Approval Required | SSA Approval Required | |
|---|---|---|---|
| Access Requirements | X |
- Description
Flows of separations and hires between employers (SEINs) used to identify successor-predecessor relationships.
- Scope
State
- Key
SEIN SEIN_SUCC QTIME
- Sort Order
SEIN SEIN_SUCC QTIME
- File Format
SAS Data Table
- Download Codebook
Variable Information
| Variable Name | Type | Length | Description |
|---|---|---|---|
| SEIN | char | 12 | SEIN - predecessor |
| num | 8 | Quarter of separation, 1985Q1=1 (See details below) |
|
| SEIN_SUCC | char | 12 | SEIN - successor |
| MATCH_PERIOD | num | 8 | Percent of transitions where separation precedes quarter of accession |
| NUM_LEFT | num | 8 | Number of jobs transitioning between firms |
| ACTIVE_BEG_QTR_A | num | 3 | First quarter predecessor is active on UI |
| ACTIVE_END_QTR_A | num | 3 | Last quarter predecessor is active on UI |
| ACTIVE_BEG_QTR_B | num | 3 | First quarter successor is active on UI |
| ACTIVE_END_QTR_B | num | 3 | Last quarter successor is active on UI |
| num | 3 | Type of link for predecessor firm (See details below) |
|
| num | 3 | Type of link for successor firm (See details below) |
|
| BPEMP_MASTER | num | 8 | Pred UI B Employment, max of last 3 quarters |
| RATIO | num | 8 | Percent of jobs at predecessor transitioning to successor (estimated) |
| SUCC_RATIO | num | 8 | Percent of jobs at successor transitioning from predecessor (estimated) |
| num | 8 | Size class of predecessor based on UI data (B employment in qtime) (See details below) |
|
| num | 8 | Size class of successor based on UI data (M employment in qtime+1) (See details below) |
|
| EMP_ES | num | 8 | Pred ES202 Month 1 Employment, max of last three quarters |
| ES_QTRS_OFF | num | 3 | Number of quarters removed ES202 event date is from UI flow |
| char | 4 | Data source of link between firms (See details below) |
Details for variable QTIME on SPF_ZZ
- Description
Quarter of separation, 1985Q1=1
- Codebook
Value Label 1 1985Q1 2 1985Q2 3 1985Q3 4 1985Q4 5 1986Q1 6 1986Q2 7 1986Q3 8 1986Q4 9 1987Q1 10 1987Q2 11 1987Q3 12 1987Q4 13 1988Q1 14 1988Q2 15 1988Q3 16 1988Q4 17 1989Q1 18 1989Q2 19 1989Q3 20 1989Q4 21 1990Q1 22 1990Q2 23 1990Q3 24 1990Q4 25 1991Q1 26 1991Q2 27 1991Q3 28 1991Q4 29 1992Q1 30 1992Q2 31 1992Q3 32 1992Q4 33 1993Q1 34 1993Q2 35 1993Q3 36 1993Q4 37 1994Q1 38 1994Q2 39 1994Q3 40 1994Q4 41 1995Q1 42 1995Q2 43 1995Q3 44 1995Q4 45 1996Q1 46 1996Q2 47 1996Q3 48 1996Q4 49 1997Q1 50 1997Q2 51 1997Q3 52 1997Q4 53 1998Q1 54 1998Q2 55 1998Q3 56 1998Q4 57 1999Q1 58 1999Q2 59 1999Q3 60 1999Q4 61 2000Q1 62 2000Q2 63 2000Q3 64 2000Q4 65 2001Q1 66 2001Q2 67 2001Q3 68 2001Q4 69 2002Q1 70 2002Q2 71 2002Q3 72 2002Q4 73 2003Q1 74 2003Q2 75 2003Q3 76 2003Q4 77 2004Q1 78 2004Q2 79 2004Q3 80 2004Q4 81 2005Q1 82 2005Q2 83 2005Q3 84 2005Q4 85 2006Q1 86 2006Q2 87 2006Q3 88 2006Q4 89 2007Q1 90 2007Q2 91 2007Q3 92 2007Q4 93 2008Q1 94 2008Q2 95 2008Q3 96 2008Q4 97 2009Q1 98 2009Q2 99 2009Q3 100 2009Q4 101 2010Q1 102 2010Q2 103 2010Q3 104 2010Q4 105 2011Q1 106 2011Q2 107 2011Q3 108 2011Q4 109 2012Q1 110 2012Q2 111 2012Q3 112 2012Q4 113 2013Q1 114 2013Q2 115 2013Q3 116 2013Q4 117 2014Q1 118 2014Q2 119 2014Q3 120 2014Q4 121 2015Q1 122 2015Q2 123 2015Q3 124 2015Q4 125 2016Q1 126 2016Q2 127 2016Q3 128 2016Q4 129 2017Q1 130 2017Q2 131 2017Q3 132 2017Q4 133 2018Q1 134 2018Q2 135 2018Q3 136 2018Q4 137 2019Q1 138 2019Q2 139 2019Q3 140 2019Q4 141 2020Q1 142 2020Q2 143 2020Q3 144 2020Q4 145 2021Q1 146 2021Q2 147 2021Q3 148 2021Q4 149 2022Q1 150 2022Q2 151 2022Q3 152 2022Q4 153 2023Q1 154 2023Q2 155 2023Q3 156 2023Q4
Details for variable LINK_UI on SPF_ZZ
- Description
Type of link for predecessor firm
- Codebook
Value Label 1 Firm death, flow constitutes a large percentage of predecessor employment 2 Firm death, flow does not constitute a large percentage of predecessor employment 4 Not firm death, flow constitutes a large percentage of predecessor employment 5 Not firm death, flow does not constitute a large percentage of predecessor employment
Details for variable SUCC_LINK_UI on SPF_ZZ
- Description
Type of link for successor firm
- Codebook
Value Label 1 Firm birth, flow constitutes a large percentage of successor employment 2 Firm birth, flow does not constitute a large percentage of successor employment 4 Not firm birth, flow constitutes a large percentage of successor employment 5 Not firm birth, flow does not constitute a large percentage of successor employment
Details for variable PRED_SIZE_CLASS_UI on SPF_ZZ
- Description
Size class of predecessor based on UI data (B employment in qtime)
- Codebook
Value Label 2 Predecessor <5 employment on UI 2 Predecessor 5-19 employment on UI 3 Predecessor 20-49 employment on UI 4 Predecessor 50-99 employment on UI 5 Predecessor 100-249 employment on UI 6 Predecessor 250-499 employment on UI 7 Predecessor 500+ employment on UI
Details for variable SUCC_SIZE_CLASS_UI on SPF_ZZ
- Description
Size class of successor based on UI data (M employment in qtime+1)
- Codebook
Value Label 1 Successor <5 employment on UI 2 Successor 5-19 employment on UI 3 Successor 20-49 employment on UI 4 Successor 50-99 employment on UI 5 Successor 100-249 employment on UI 6 Successor 250-499 employment on UI 7 Successor 500+ employment on UI
Details for variable SOURCE on SPF_ZZ
- Description
Data source of link between firms
- Codebook
Value Label ES Link provided on QCEW UI Sufficient flow found in UI earnings data UIES Sufficient flow found in UI earnings data and QCEW link