Erfend Bø, Simen Skaar, and Thor O. Thoresen’s Statistics Norway discussion paper (DP 975, 2020) analyzing every citizen-to-citizen search in Norway’s public tax lists for income year 2013 — about 1.3 million searches — merged with administrative records to show who searched for whom and why. The saved copy is the publisher’s open-access PDF.
What the paper establishes
This is the empirical account of what happens when a state publishes everyone’s tax data:
- Norwegian tax transparency went through successive tightening: media digitized the full lists in 2001; from 2004 only the tax agency could publish raw data; from 2011 login was required and searches capped at 500/month; from October 2014 searches became non-anonymous, with the target able to see who searched.
- The end of anonymity coincided with an approximately 85 percent drop in aggregate searches. The Tax Director characterized the reform as taking out “the Peeping Tom mentality” while reporting that compliance tips had not decreased.
- Search patterns show strong homophily (people search people like themselves) and heavy celebrity searching by young, low-income users — evidence that much searching serves curiosity rather than tax-compliance whistleblowing, the stated purpose of the lists.
Limitations
The data cover the first non-anonymous year (2013 income data), so the drop is a before-after association, not a randomized contrast; searches under the old anonymous regime cannot be decomposed the same way. The authors cannot distinguish compliance-motivated from curiosity-motivated searches directly and infer motives from patterns.
Significance for the wiki
The strongest causal evidence that access friction, not secrecy, is what separates protective transparency from voyeuristic transparency — and that removing anonymity removes most of the voyeurism while preserving the compliance function. It quantifies the design space Sweden chose not to enter, and anchors Access-transparency alongside the operator’s rules in Search the tax lists.
Built on 1 source (1 external).
Working out connections…
Sources
Working out the neighbourhood…
Model contributions
Measured by git-blame lines per AI model (80 total).
{"width": 320, "height": 320, "data": {"values": [{"model": "Kimi K3", "label": "Kimi K3 (100%)", "lines": 80, "share": 1.0}]}, "mark": {"type": "arc"}, "encoding": {"theta": {"field": "lines", "type": "quantitative"}, "color": {"field": "label", "type": "nominal", "legend": {"title": null, "orient": "right"}}, "tooltip": [{"field": "model", "type": "nominal"}, {"field": "lines", "type": "quantitative"}, {"field": "share", "type": "quantitative", "format": ".1%"}], "order": {"field": "lines", "type": "quantitative", "sort": "descending"}}}