No, The Optimal Corporate Income Tax Isn't (Necessarily) Zero

It’s a relatively common sentiment among the economically literate to advocate for the complete abolition of corporate income taxes. I sympathize with and understand why some hold this view.

Many papers like the classic Atkinson and Stiglitz (1976), this 1999 paper, and these two from the 80s (Chamley and Judd), suggest that the optimal tax on capital income is approximately zero in the long run.

It only seems natural that one should apply the same logic to corporate income taxes. After all, a corporate income tax (CIT) is basically a tax on capital returns.

Add to that the host of legal and political problems that corporate income taxes bring to the table (e.g. tax avoidance and offshoring), and you’ve got yourself a pretty good case for zero.

But I see some problems in the reasoning of those who espouse the zero CIT mantra. Many of the same people who advocate for no CIT advocate for higher capital gains and estate taxes, completely forgetting the theoretical basis for why a corporate income tax should be abolished: the idea that taxing capital income is a bad idea altogether.

Interestingly enough, repealing the CIT is not necessary for there to be no taxes on capital income It's entirely possible to have both a positive corporate income tax and no taxes on capital income. The X-tax almost does exactly that. (Scott Sumner offers a similar proposal.)

That's ignoring the fact that there have been results suggesting a positive capital income tax is optimal. And let's not forget that some models suggest a high tax on the initial capital stock is desirable.

The Intuitions Against Capital Income Taxes

We can understand what capital income taxes do by looking at a model with simple assumptions.

Assume that we have a consumer who earns a wage in time period 1 and can choose to spend all their after-tax wages on consumption in the initial time period or instead save and invest all their after-tax wages for consumption in the next time period.

---

Further, we want to define some terms:

W := wage earned in the first time period

r := return on investment

w_t := wage tax rate

c_t := capital income tax rate

---

If there is a tax on wages but not on capital income, the consumer can either spend W(1-w_t) on consumption in period one or spend W(1-w_t)(1+r) on consumption in period two. One can see how this is equivalent to a tax on consumption.

But if capital income were to be taxed in addition to wages, then the consumer would face a dilemma between consuming W(1-w_t) in period one or consuming W(1-w_t)(1+r[1-c_t]) in period two.

If there were an infinite number of time periods, the implicit marginal tax rate on future consumption would (due to compounding) approach infinity 100%, and the consumer would have less incentive to invest.

Simply put, a tax on capital income causes a higher implicit marginal tax rate on future consumption relative to present consumption. Since the previously mentioned Atkinson-Stiglitz theorem roughly implies that taxes on consumption should be neutral with respect to time, the optimal tax on capital income is approximately zero.

And if we were to truly take the idea of infinite time periods seriously (as the Chamley and Judd results do), then the growing “tax ‘wedge’ between current and future consumption” as a consequence of time would create some serious Laffer curve (don’t you mean Rolle’s theorem?) problems. That strengthens the case for zero capital income taxation even more.

Corporate Income Taxes and Zero Capital Income Taxes are Compatible

Okay, so let’s embrace zero capital income taxes for now. It seems obvious that policymakers should work to repeal the CIT, right? Not necessarily!

Let’s suppose that the papers arguing against capital income taxes shift our preferences from taxes on production to those on the final consumption of goods. A VAT does exactly that, but are there better options?

VAT taxes run into the issue of them being proportional concerning consumption and regressive concerning income. Governments can offset this with cash transfers, but it needn’t be the case.

To address concerns about equity while taxing consumption rather than income, we can have a system where:

(1) Labor income is taxed progressively

(2) VATs are charged to firms directly

(3) Firms receive investment credits for labor costs so double taxation is avoided

Notice that this proposal covers the same tax base a VAT would, but firstly, it’s a lot more progressive, and secondly, it’s much closer to most tax systems you’ll see around the world.

If one wants a similar scheme that is nearer to the current system of the United States, here are some changes the US can make:

(1) Make the corporate income tax territorial

(2) Legalize full expensing

(3) Remove caps for certain tax-advantaged savings accounts

Those changes make the US tax system essentially the same as the previous proposal, and they’re a lot easier to sell.

Imagine you’re a well-informed politician who wants your country to shift from a system of income taxation to consumption taxes. Would you rather propose some minor shifts in the corporate income tax system and increased limits of tax-advantaged accounts, or an almost complete replacement of the current tax system with a flat VAT that looks incredibly regressive concerning income?

Both options are nearly impossible to politically implement, but a conversation with the median voter will tell you what looks more palatable. In any case, so long as a system with a CIT is economically equivalent to one without taxes on capital income, it is not immediately obvious that said CIT should be abolished.

The Optimal Capital Income Tax Ain’t Necessarily Zero Either

For several reasons, a positive tax on capital income may be seen as desirable, e.g.

(1) A tax on the initial capital stock imposes little to no deadweight loss

(2) Provided that investment is subsidized, taxing capital income not only allows for more progressive schemes of taxation, it could also improve welfare for “second best” reasons

(3) Assuming labor income is taxed, a lack of capital income taxes can reduce neutrality between investments in human capital relative to other capital

(4) Capital income taxes diminish incentives for one to disguise labor income as capital income,

etc.

All of these reasons add up to an argument for capital income taxation, and one that is not to be taken lightly. But that’s not all there is folks! Here’s a recent paper revisiting Chamley-Judd that contradicts the 1980s conclusion using the original model itself. And here’s another paper that goes against the Atksinson-Stiglitz “consensus.” (Stiglitz himself supports taxing capital and corporate income.) If that weren’t enough, I would like to comfort the Federal Reserve Bank of Minneapolis for this burn (just read the titles!).

And, oh yeah, take a look at these IGM surveys from the US and Europe. I don’t see much of a consensus for the abolition of capital income taxes, let alone corporate income taxes. Heck, add this podcast indicating a positive corporate income tax is optimal, and it looks like taxing capital income isn’t such a bad idea after all!

Optimal Taxation Goes Beyond Positive Economics

If you really want to get in-depth with the optimal taxation literature, Mankiw and Auerbach offer great places to start. Still, there remains a core problem with the idea that one can derive “the optimal tax rate” from positive rather than normative analysis (I’m not referring to the tax rate that maximizes some social welfare function), never mind the idea that this issue has great consensus among economists.

Like it or not, economics is a science (physicist_crying.jpeg), and the corporate tax incidence on labor being 80%, 40%, or whatever is still insufficient to tell us what policymakers ought to be doing.

If you’re a hardcore right-libertarian, maybe the corporate income tax should be zero, ditto with the consumption tax. If you’re sympathetic to socialism, it might be the opposite. It all depends on your political and moral leanings.

Just as theory doesn’t always translate well into practice, the same can be said of bad economics and bad politics. Economics, like any science, informs us about the way the world is, not what it should be.