Does Optimal Tax Theory reward avoidance, evasion, and lax enforcement?

The standard optimal tax frameworks, particularly the Diamond-Saez model, which are focused on "efficiency," would suggest lowering top rates if the elasticity of taxable income (ETI) increases. Since the optimal rate formula is roughly τ/(1-τ) = (1/e) × a/(1-a), where e is the ETI and a is the Pareto parameter, a higher elasticity mechanically implies a lower optimal rate. This creates what you might call a "tax evasion veto." The wealthy can effectively vote for lower rates by finding better ways to avoid paying tax or by lobbying for weaker IRS enforcement. This is not, I think, what most people would expect.

This seems to reveal something deeply problematic about treating efficiency as the primary normative criterion. The OTT approach essentially takes behavioral responses to taxation as exogenous constraints rather than endogenous reactions that themselves have normative content. Consider the implications:

1. The perverse incentive: If aggressive tax avoidance lowers optimal rates, then the wealthy have an incentive to invest in avoidance technologies or methods and lobby for weaker enforcement, knowing this will eventually translate into lower statutory rates being deemed "optimal."
2. The enforcement paradox: Because weakening enforcement will lead to increases in measured ETI (because more income disappears), optimal tax theory would recommend lowering rates in response – rewarding the very behavior that reduced revenue collection. (What this means is that when Congress cuts the IRS enforcement budget, OTT might argue that they should also cut the top tax rates in anticipation of an increased ETI for high incomes.)
3. The moral content of elasticity: Not all behavioral responses are created equal. Real economic responses (e.g. reduced work effort or entrepreneurship) differ fundamentally from tax avoidance (e.g. accounting games and income shifting). Yet standard ETI measures seem to conflate these, treating them as equivalent efficiency losses.

I understand the desire to determine "efficient" tax rates, but I suspect that the average non-economist would complain that while efficiency has benefits, there should be more focus on defining a tax regime which is "fair." Of course, what it means to be fair would be very hard to define, but I suspect that few would argue that it is fair for a tax system to reward those who seek to thwart it.

Am I missing something important here? Does OTT reward avoidance and evasion of taxes as well as lax enforcement of tax laws?