What Would the Optimal Tax System Look Like?

When we talk about taxes, the focus is usually on how much people pay. But there’s another equally important consideration: how exactly we structure the tax system. The way we design taxes can dramatically affect the economy—not just through revenue, but through behavior.

So what would an optimal tax system look like?

One key principle is that it would include fewer loopholes—special deductions, exemptions, and carve-outs for different types of income or spending.

Why? Because these tax breaks distort decisions. They encourage people to change how they work, save, invest, or spend in order to lower their tax bill.

For example, the mortgage interest deduction pushes people toward buying larger homes. The exclusion for employer-sponsored health insurance encourages overconsumption of health care—yes, there is such a thing—and locks workers into jobs. Preferential tax treatment for carried interest lets some investment managers pay lower tax rates than regular wage earners.

These provisions aren’t free. They reduce government revenue and shift the tax burden onto everyone else, all while adding complexity and often benefiting higher-income households the most.

Now, I get it—you may really like your favorite deduction. But just because it benefits you doesn’t mean it’s good policy for society as a whole. When we build a tax system around special breaks, we end up with something that’s less fair, less efficient, and harder to defend.

That said, not all tax breaks are bad.

Some can be justified—but only when there’s a clear market failure. That means the market, left on its own, would under-provide something that benefits society.

Other tax provisions may serve redistributive goals—helping those with lower incomes or fewer opportunities. But even then, they should be carefully designed to minimize economic distortions and avoid creating bad incentives.

For example, R&D tax credits aim to boost innovation, which has broad social benefits. The Earned Income Tax Credit supports low-income workers by increasing take-home pay without discouraging work—helping both fairness and labor market efficiency. It’s a good example of redistribution that’s structured to support, not weaken, economic participation.

But the bar should be high. A tax subsidy or credit should exist only if:

1. We clearly want more of that behavior or outcome

and

2. The private market won’t deliver it effectively on its own.

If we do implement such a policy, it should be designed to achieve its goals with as few unintended side effects as possible.

We’ve made progress in the past. The Tax Reform Act of 1986—one of the most ambitious US tax overhauls—eliminated dozens of tax shelters, simplified brackets, and broadened the tax base. It didn’t raise or lower total taxes much—it just made the system more efficient and less distortionary. I’m not claiming that every provision in that Act was a good one, but it did greatly improve efficiency.

That’s the goal of an optimal tax system: raise the revenue we need, with the fewest distortions of incentives and the least unnecessary complexity.