How Reliably Do Empirical Tests Identify Tax Avoidance?
Bibliographic record
Abstract
ABSTRACT Research on the determinants of tax avoidance have relied on tests using GAAP and cash effective tax rates (ETRs) and total and permanent book‐tax differences. Two new proxies have emerged that overcome documented limitations of these proxies: one, developed by Henry and Sansing (2018), allows for more meaningful interpretation of results estimated in samples that include loss observations. The other, reserves for unrecognized tax benefits (UTB), provides new data on tax uncertainty. We offer empirical evidence on how well tests using these new proxies perform relative to those extensively used in prior research. The paper finds that tests using the proxy developed by Henry and Sansing (2018) have lower power relative to those using other proxies across all samples, including a sample that includes loss observations. In contrast, when firms accrue reserves for uncertain tax avoidance, tests using the current‐year addition to the UTB have the highest power across all proxies, samples, and levels of reserves. In the absence of reserves, tests using the GAAP ETR best detect uncertain tax avoidance, on average. This study contributes to the literature by using a controlled environment to provide the first large‐scale empirical evidence on how the power of tests varies with the use of relatively new proxies, the inclusion of loss observations, and the advent of FIN 48.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.125 | 0.615 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.008 | 0.009 |
| Science and technology studies | 0.001 | 0.009 |
| Scholarly communication | 0.008 | 0.011 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.004 | 0.004 |
| Insufficient payload (model declined to judge) | 0.007 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".