Trace 03 · Research paper
Abraham and colleagues: survey versus administrative traces of nontraditional earnings
Katharine G. Abraham, John C. Haltiwanger, Kristin Sandusky and James R. Spletzer have published a body of work that compares household survey reports of work and earnings with administrative tax and employment records. Shift Ledger highlights this literature because it speaks directly to undercount risk when earnings arrive outside a single main-job narrative—including forms of nontraditional and side earnings that surveys may miss or misclassify.
The core methodological insight is simple to state and hard to implement: survey questions and tax forms observe different objects. A respondent may omit intermittent self-employment that still generates an information return. Conversely, administrative traces can include payments that respondents would not describe as a “job.” Abraham and coauthors map those mismatches rather than declaring one source universally superior.
Nontraditional work in this research program is a measurement category built from legal and economic signals—self-employment indicators, multiple job holding, contract status and related flags—not a lifestyle label. That precision matters for editorial use. Summaries that turn the findings into advice about how households “should” earn side income leave the evidence base and enter presupposition.
The survey–administrative gap literature also clarifies why “platform work” is difficult to time-stamp in national series. Platform payouts may appear in tax data with a lag, under entity names that hide the platform, or split across forms. Household surveys may capture the activity only when question wording cues respondents to include it. Both channels can be incomplete; their disagreement is data, not noise to hide.
Limits belong beside results. Linked datasets cover specific years, jurisdictions and match rates. Imperfect linkage can bias estimates of underreporting. Changes in tax law or form design alter administrative coverage. Abraham and colleagues discuss such constraints; Shift Ledger repeats them so that a citation cannot be mistaken for a universal constant.
For an editorial desk focused on uneven earnings, the practical takeaway is epistemic humility. Instruments that look definitive in a chart often rest on classification choices. Publishing those choices—who is counted, which form is used, which period is observed—is the difference between research coverage and marketing copy.
This note does not estimate any reader’s earnings, does not recommend platforms, and does not claim results from reading. It points to a research program that documents how surveys and administrative systems diverge when work is nontraditional. Primary papers and accompanying appendices remain the authoritative source for magnitudes and methods.