Over the past several years, a series of leaks related to offshore tax avoidance and evasion (SwissLeaks, LuxLeaks, the Panama Papers, Bahama Leaks, and Paradise Papers, to name a few) has fueled calls for tax transparency. To date, most discussion of the leaks has been policy-oriented (leaks: good or bad?) and largely anecdotal (based on some truly outrageous revelations). It was not until very recently, however, that a small group of researches started delving into the data exposed by these leaks to make statistically significant empirical findings. Alstadsæter, Johannesen & Zucman’s (AJZ) paper is an excellent example of such paper, which combines methodological sophistication, public data, and leaked data, to make important new contributions to the voluminous literature on the offshore tax world.
Matching leaked data with data from random audits in Scandinavian countries, public wealth records in those countries, and data from voluntary disclosure programs, AJZ find that offshore tax evasion (meaning, the act concealing income from tax authorities in offshore accounts), is not evenly distributed across wealth groups. Rather, they demonstrate that “the probability to hide assets offshore rises sharply with wealth, including within the very top groups of the wealth.”
AJZ findings stand in sharp contrast to what is known (or at least assumed) about tax evasion from stratified random audits, which—per AJZ— is “the key source used so far in rich countries” to study tax evasion. AJZ argue that random audits miss large part of tax evasion, “because it assumes that detected and undetected forms of tax evasion are similarly distributed across the income spectrum”. They suggest, however, that “sophisticated forms of evasion involving legal and financial intermediaries—that are only accessible to wealthy taxpayers—are unlikely to be uncovered in random audits.”
AJZ find, however, that “[Scandinavian] households who own around $10-12 million in net wealth are twice more likely to conceal assets abroad than households with around $5-6 million; households with more than $45 million are four times more likely.” Consider another example: random audits suggest that the average evasion rate in Scandinavian countries is 3% (meaning, about 3% of all individual taxes are evaded in Scandinavia). AJZ find, however, that the top 0.01% of Scandinavian households (in terms of net wealth) evade about 25% of their income through offshore accounts. This stands in sharp contrast to the average 3% benchmark.
These findings are profound in many important ways. It means, for example, that models for assessing the tax gap (meaning, the amount of uncollected taxes) probably understate the tax gap, if such models usually assume even distribution of detected and undetected tax evasion among income groups. If undetected evaded income is concentrated at the top, it also means it should have been taxed at higher rates than previously assumed for purposes of calculating the gap. The findings also mean that standard measures of inequality severely understate the level of inequality, because such models rely on reported data, which by definition does not include unreported income, which—we now know—is concentrated within top-earning households.
In their paper, AJZ address the effects of their findings on inequality and the tax gap at length. But several other important themes—some more implicit—emerge. First, consider the fact that in Scandinavian countries tax compliance is a cultural trait. What might be the outcomes in economies where tax avoidance (and evasion) have a more favorable public view, such as the United States? AJZ indeed note that because other economies “own much more wealth offshore than Norway, the results found in Scandinavia are likely to be lower bounds for these countries.” Second, another important theme relates to the data on which we rely in tax policy making. Proposed legislation, enforcement actions, and administrative guidance all rely on modeling, which in turn rely on reported tax data. The AJZ paper at least raises a plausible argument that the entire tax-policy making process relies on severely faulty (if not incorrect) data. Finally, the AJZ paper emphasizes how leaks may affect our tax system, and point to the role of civil protest in tax. For years, policy makers and taxpayers invested significant effort in protecting taxpayers’ privacy, sometime legitimizing the offshore system as a protector of legitimate privacy concerns. These arguments make some sense in theory. But when 95% of the accounts identified in the leaks are not reported to tax authorities (as AJZ find when trying to match the leaked accounts with required reports on offshore accounts), the legitimacy of the privacy argument is diminished. It really seems that, for the most part, intermediaries in secrecy jurisdictions main function is to facilitate tax evasion and other illicit activities.
In a utopian world, we would not need to rely on leaks to provide the information needed to prescribe successful tax policies. But in this world, the AJZ paper suggests that leaks may play a helpful role in pushing us towards better tax policy. It is the leaks (and the data exposed in papers such as AJZ’s paper) that have driven recent changes in international tax law (such as the adoption of the common reporting standards and the automatic exchange of information by multiple countries). It is difficult to understate how profound has been the effect of leaks, and the following public outcry, on real tax policy. It is truly a bottom-up legal change.
There is a lot more to like in the paper, and I cannot hope to discuss it all in the confined space of this review. I will just note one of my favorite aspects: AJZ develop a model showing that after disclosing tax evasion in an amnesty program, tax evaders do not then find legal ways to avoid taxes. This suggests that discovered tax evasion is not substituted by “legal” tax avoidance, which means that anti-evasion enforcement action are efficient.
To summarize, AJZ’s paper is an early one among several others that use recent leaked data to try to empirically assess what we didn’t know we didn’t know about offshore tax evasion. The implications are profound and should guide data-driven tax policy making for the foreseeable future.