Being a tax lawyer means sometimes advising clients that the clever tax scheme they’ve come up with doesn’t work. The tax authorities would challenge it; they would probably win; and there would be penalties for underreporting income to boot.
Occasionally, the client’s response would go something like this. “OK, understood, thanks. But just out of curiosity… how likely is it that the tax authority would find out about this arrangement, if we did it? A lot of it happens overseas, so maybe they just wouldn’t notice. Also, I hear their audit resources are very stretched at the moment. Perhaps we should go ahead anyway.” And we would respond “sure, fraud is always an option”. And then they would look a bit sheepish, and then we would try to find a structure that did work.
Why discourage fraud? Hardly for selfless reasons. Our goal was to get paid for clever structures that succeeded, under existing tax law, in reducing our client’s tax liability. If you were prepared to just stash money in a secretive jurisdiction and lie about it on your tax return, you didn’t need us. (You needed a different kind of financial professional.)
Being a tax lawyer means sometimes advising clients that the clever tax scheme they’ve come up with doesn’t work. The tax authorities would challenge it; they would probably win; and there would be penalties for underreporting income to boot.
Occasionally, the client’s response would go something like this. «OK, understood, thanks. But just out of curiosity… how likely is it that the tax authority would find out about this arrangement, if we did it? A lot of it happens overseas, so maybe they just wouldn’t notice. Also, I hear their audit resources are very stretched at the moment. Perhaps we should go ahead anyway.» And we would respond «sure, fraud is always an option». And then they would look a bit sheepish, and then we would try to find a structure that did work.
Why discourage fraud? Hardly for selfless reasons. Our goal was to get paid for clever structures that succeeded, under existing tax law, in reducing our client’s tax liability. If you were prepared to just stash money in a secretive jurisdiction and lie about it on your tax return, you didn’t need us. (You needed a different kind of financial professional.)
«We have complied with all existing tax laws,» say Amazon, Google, Apple.
Of course they have. They have paid a lot of money to make sure of it. They may have taken some… optimistic views about what precisely the tax laws say. But their interpretation will be defensible, and will be supported by at least some legal authority.
So the ethical analysis of tax avoidance starts with the idea that it complies with existing tax laws. Tax fraud (often called tax «evasion», because that word’s semantic and phonetic similarity to «avoidance» makes the distinction seem more subtle) is very bad, and severely punished, so it’s understandable that companies would want to deny they’ve engaged in it. But it’s rarely the charge levelled at big companies.
When is legal conduct unethical? This is a familiar problem to ethicists. The converse question – when is illegal conduct ethical – might get more attention in political contexts, since it involves interesting issues like necessity, civil disobedience, and revolution. But everyone agrees that the law doesn’t prohibit everything that one ought not to do. It’s not illegal to be ungrateful to your parents, or to fail to do your share of household chores, or to vote for fascists.
When it comes to tax avoidance, there is a tension between two normative principles. The first is social: one has a duty to make a fair contribution to one’s society, in accordance with the distribution of burdens agreed on by that society. (Provided the contributions are fair. We can assume that taxes are used for social benefit, at least in general, and that tax burdens are distributed justly. Let’s also assume that distributing the burdens according to income is just, since it correlates with ability to pay and with the diminishing marginal utility of money.) Looked at this way, tax avoidance is shirking: the tax-avoider is shying away from their obligations to their society, while still taking advantage of the benefits it offers.
The second principle is a political one. The rule of law requires that the state’s impositions on its citizens be stated clearly, so that they can be complied with, and that they not be subject to post-facto reinterpretation. Looked at this way, tax «avoidance» is perfectly legitimate: no one has an obligation to fork over money to the state except where explicitly demanded by the law. Requiring that people pay tax based on what the law should have said, but didn’t, is unjust.
To understand the moral status of tax avoidance, both principles have to be taken into account. The strict libertarian view (defended for example by Machan, 2010) is irresponsible. Taxation in a democracy is not simply the government taking our money against our will. Quite apart from our duties to our fellow citizens, it would be impossible to earn that income if it wasn’t for the services that taxation allows the state to provide. (For a good summary of this argument, see Goff, 2017.)
This is why, in my view, «we have complied with all existing laws» is not just insufficient, but obnoxious. In advancing that as a defence, companies are not merely failing to address the question of tax avoidance. They are implying that they have no other social obligations worth mentioning.
But simply claiming that we should all just pay «our fair share» ignores that taxation is fundamentally a legal obligation. It is a duty decided and imposed by the state, ultimately at gunpoint. If the tax law does not mean what it actually says, but rather something vaguer and more discretionary, that can be dangerous.
How to balance those two principles? Which transactions should be condemned as tax avoidance, and which respected under the rules as written? Common approaches to defining tax avoidance try to do it by reference to the nature of the transactions involved. They look for some technical feature that distinguishes legitimate transactions from illegitimate. I want to suggest this is the wrong way to go.
I’ll consider two examples: the «economic substance» approach, and the «intention of the tax law» approach.
The first approach looks for features of transactions that are there in order to create tax benefits, and disregards them. (The new UK general anti-avoidance rule, in section 206 of the Finance Act 2013, is an example.) The idea is that if we disregard what people have done for tax reasons, we get to the real nature of the transaction. That allows us to tax the participants’ genuine economic income.
The problem is that tax effects can be taken account in perfectly legitimate ways. In deciding to make an investment via debt rather than equity, for example, the parties will inevitably consider the tax effects – in particular, that interest is deductible. Does that mean the investment is really equity? If you run your own business, you can choose whether to act in your own name or to incorporate a company. Incorporating a company sometimes has tax benefits; does that mean the company should be disregarded? We still need to distinguish between acceptable and unacceptable tax planning, and that was the problem we came in with. More generally, as long as the tax law distinguishes between types of income (interest versus dividends; current income versus capital gains; share sales versus asset sales) there will be alternative ways to carry out transactions that have different tax consequences. There is no general reason to allow the tax authority to impose the most expensive. As one British judge put it (before the general anti-avoidance rule): «No man in this country is under the smallest obligation, moral or other, so as to arrange his legal relations to his business or to his property as to enable the Inland Revenue to put the largest possible shovel into his stores.» (Ayrshire Pullman v CIR (1929) 14 TC 754.)
In response, some anti-avoidance regimes look for artificial elements of the transaction, ones that lack any economic effect at all. But it is easy enough to build in some economic consequences for the artificial elements. The expected cost of that consequence can then be weighed against the tax benefits. The result is that tax-avoidance becomes slightly more expensive, but not much less common.
A second approach looks to frustration of the intent of the tax law. This gives voice to the intuitive sense that tax avoidance exploits «loopholes»: infelicities of drafting that allow for consequences very different from those the legislators intended. The US IRS takes this approach to so-called step transactions. (See IRS 2017, p.8) These are arrangements where multiple steps are used to achieve an end state that could be achieved in a single step, but where the longer route has better tax consequences. For example, instead of selling a business, you might incorporate it, and sell the shares in the new company, to a buyer who promptly liquidates it and runs the business directly. The result is exactly as if the business had been sold directly, but in some cases the tax treatment will be much better. (You know the children’s trick where one person in a queue lets a friend cut in in front of them, after which the friend lets the first person go in front of them? That’s why step transactions are bad.)
The obvious problem is deciding what the intention of the tax law is. At one level of detail, it’s: to raise tax! At the opposite extreme, it’s: to impose tax on the transactions it describes as subject to tax. Finding the «real» goal of tax legislation raises very similar problems to the «real» nature of a transaction. But there’s also a major type of tax avoidance that this approach handles very poorly: international tax arbitrage. This form of avoidance exploits the differences between two tax regimes, in a way that complies with the letter and spirit of both, but which in combination reduces tax.
As an example, consider hybrid debt/equity instruments, whose terms are a mixture of debt and equity. They have some features of debt, such as a fixed return; but other features are share-like, such as subordination to the company’s other obligations, or an indefinite term. Is such an investment debt or equity?
Philosophers faced with such categorisation questions might say something about «vague predicates» or «prototype-based concepts», and refuse to give a definite answer. But the law has to decide. If necessary, it will draw an arbitrary line. Inevitably, different jurisdictions will draw their arbitrary lines in different places – and hybrid instruments exploit those differences. In the classic case, the investment is treated as debt on the borrower’s side of the Atlantic, and as equity on the other.
That means the borrower can treat the yield on the investment as interest, and deduct it from their income as it accrues. The investor, on the other hand, can treat the return as dividends. They might be exempt from tax altogether, or they might go untaxed until the dividends are actually paid out – perhaps at the end of the investment.
Whose intention does that frustrate? No one’s. From the point of view of each jurisdiction, the instrument is treated precisely according to the policy it has adopted. But tax-avoidance it surely is.
The same thing goes for many of the strategies adopted by the likes of Amazon, Google and Apple. Intermediary shell companies, with just enough substance, are set up to exploit differences in the treatment of royalties, dividends, or trading income, or even of the existence of those companies themselves. The notorious US “check the box” rules have allowed US companies to treat wholly-owned European subsidiaries as non-existent for US tax purposes, even as their European homes treated them as separate taxpayers. (For a useful, if somewhat technical, discussion of hybrid tax strategies, see OECD, 2012.)
The problems with these approaches are not accidental. Both offer technical answers to an ethical problem. Tax avoidance, I suggest, is what philosophers call a «thick» concept – one that has both evaluative elements and descriptive, value-neutral, ones. (See Dancy, 1995.) A more familiar example is «bravery»: to be brave is to act despite the presence of serious risk – but only when doing so is praiseworthy. Ignoring risks in a blameworthy way is not brave, but reckless or foolhardy.
Similarly, tax avoidance is reducing your tax liability in a way that shirks your moral obligation to pay your fair share. It has a descriptive element – your tax liability needs to be reduced relative to the alternatives – but it has an irreducibly normative element besides. And that means that it can’t be defined in strictly value-neutral technical terms. Trying to do so leads down rabbit-holes that are very familiar to both moral philosophers and analytically-minded lawyers. There will always be another counter-example to your technical definition, however refined and baroque you make it.
To say all that is not, of course, to answer the question of when tax planning is ethically acceptable and when it is immoral. It’s just to ask it in a more transparent way.
What it suggests, though, is why legislating against tax avoidance as such, as the UK and other general anti-avoidance rules try to do, will ultimately be unsuccessful. While such laws have some deterrent effect, tax planners are still fully employed. What tends to result is a body of pseudo-ethical laws, in which judges’ verdicts on whether particular transactions are «reasonable» are fossilised into precedent. Given time, they become just another set of technical rules for clever tax planners to work around.
The better approach is to address the structural flaws that large-scale tax avoidance appeals to. For example, by addressing jurisdiction-shifting through international cooperation, such as formulary apportionment. And in doing so, to be guided by the principle that companies – especially ones that can afford expensive tax planning – ought to contribute their fair share to the societies they operate in. That principle is ill-suited for determining how much tax a company ought to pay in any specific case; but it should be an uncontroversial guide for tax policy.