There are people who devise and employ elaborate schemes to evade or avoid (I never know which is which) their taxes whilst staying just within the law. There are bankers who stay technically within a literal interpretation of the banking regulations, whilst engaging in dubious practices which undermine the regulator’s intention. There are employers who try to evade workplace regulations by reclassifying their workers as independent contractors. There are states which harp on about the technical details of the laws of war as they happily murder civilians. Well sometimes we need to punish the technically innocent but morally culpable.
Rule-of-law fetishists, Hayekians, and the like tend to think this is just appalling. Legislators, regulators, sports administrators and similar, should devise watertight systems of rules within which people are entitled simply to go for it. But it is highly questionable that such watertight frameworks are possible, even in principle. So we need to give the enforcers some discretionary power to zap the bad people: people who knew perfectly well that what they were doing was at or over the moral and legal boundaries but didn’t care. (On the tax front, the UK’s plan to introduce a General Anti-Avoidance Rule is designed to punish just these characters.) Such power is, of course, liable to abuse. But that’s just the way things unavoidably are. The solution is not to pretend that we can make the rules work perfectly, but to make sure the enforcers are genuinely democratically accountable and removable.
I would probably count as one of those rule-of-law fetishists. Chris's formulation is bad in theory, and worse in practice. Is it true that no law can cover every conceivable case? Yes. Is it true that some people somewhere will abuse the corner cases? Yes. Is it true that some discretion is necessary? Certainly. Does this mean that we should stop focusing on creating clear simple laws and just give enforcers lots more power to make things right? No. He is throwing out something very important because of imperfections in tough cases without apparently realizing how it functions in the not so close cases.
His examples (banking regulations, workplace regulations and law of war cases) suggest that he is concerned about using legal discretionary techniques to hit rich people who "knew perfectly well that what they were doing was at or over the moral and legal boundaries but didn't care." In the comments Chris writes "The OP said nothing about what the general level of regulation should be. What it did say is that, at any level, you are going to have this kind of conflict and that the solution then is not to proliferate further rule and sub-rules and definitions (because that’s just an arms race) but to allow the enforcers a general power to capture the instances that fall outside the literal rules on mere technicalities, where the perps are obviously acting in bad faith (as the badminton players, corporate tax evaders etc are)." These quotes heavily frieght the issue with judgments that are quite murky in the real world: "knew perfectly well", "outside the literal rules", "mere technicalities", "obvious bad faith". He seems to want us to give this power to virtuous prosecutors and regulators in a much more open way so that they will be empowered to bring these nasty rich people in line.
Theoretically, I'd prefer clean rules with little need for interpretation. But I understand that from time to time the rules aren't clear enough and we have to let administrators act with some discrection. But for the purpose of citizen/government interactions, it would be much better to make it as easy as humanly possible for people to know ahead of time what the rules are so they can guide their actions accordingly.
Also in the comments, Phil makes a good point that brings the issue more into the real world: "The general trend is towards more regulators with more power. Endorsing it, on the basis that all we have to do is make them accountable, is a bit like voting for an Enabling Act on the grounds that we can always democratise the new regime later on."
We are already in a world with increasing regulatory power, and that power is not increasingly under democratic control. We need more control over regulators first, before we start thinking about giving them more authority to punish the people who are 'obviously acting in bad faith'. You can see interesting discussions of this problem at Volokh here and here.
But as a practical matter, Chris's focus is wrong. We have all sorts of practice with regulatory and prosecutorial discretion, and in reality it tends to be employed against the poor, politically weak, and/or disfavored minorities while it tends to be employed in favor of the rich, politically strong, and/or favored majorities. We give prosecutors discrection in charging, but they tend to wildly overcharge (for the purpose of getting a plea bargain) poor people far more often than they try to do so against rich people. We know that drug sniffing dogs are far more likely to give false alerts when searching Hispanics in Chicago than when searching white people--suggesting that police discrection in searching (and in interpreting what counts as a drug alert from the 'trained' dog) is employed to the detriment of Hispanics [see here]. We know that the discretion offered by vagrancy laws is more likely to be employed against black people. We know that cities will interpret disfavored groups as rioting while favored groups are demonstrating.
In practical reality, clear simple rules operate to level the field somewhat between the rich and the poor, the favored and the unfavored. They don't make it perfectly level, of course, but when ambiguity can be leveraged, it is the rich and favored who can leverage it, not the poor and unfavored. Ramping up the ambiguity is more likely to let the IRS hit 5,000 waitresses than 1 millionaire.