What (and who) in the world are we trying to tax when we tax corporate income?

Inside the wild world of corporate tax theory

A major topic of discussion around election season is the corporate income tax. With this being election season, I thought it was appropriate for me to jump in on the fun. However, unlike the politicians or economists who tell you what they believe is the optimal corporate income tax rate, I want to take a step back and ask a different question. What is the corporate income tax? Or, more directly, what does it actually tax?

On its face, it seems like the answer to both of these questions is easy. The corporate income tax is a tax on the income made by corporations. Okay, easy enough. But that answer is superficial. Let’s dig into the details.

In a standard optimizing model of the firm, a corporate income tax increases the user cost of capital and therefore reduces the demand for capital. Thus, if you want a simple answer, you might say that the corporate income tax is a tax on capital – although that is also not a complete answer because it ignores the issue of tax incidence. However, let’s stick with calling it a tax on capital and see where that takes us.

Suppose that all firms are corporations and capital is homogeneous. Under that scenario, it seems permissible to refer to the corporate income tax as a tax on capital. Alternatively, suppose that there are two types of capital and two industries. One type of capital is used by one industry and the other type of capital is used by the other. If the corporate income tax is levied on one industry, then this is a tax on the capital of that industry.

However, if we assume that some firms are corporations and some are not and that both types of firms produce the same product, then calling this a tax on capital is not correct. It is not correct to call it a tax on capital because the capital is only taxed if the firm calls itself a corporation. It seems reasonable to me to think that both corporations and non-corporations produce the same goods and therefore use the same types of capital. Only one firm is subject to the tax.

What this framing reveals is that when both corporate and non-corporate firms produce the same good, a tax on corporate income is not really a tax on capital, but rather a tax on incorporation. Okay, now we are getting somewhere.

We know that if you tax something, you get less of it. If we tax corporate income, we should see fewer firms that call themselves corporations relative to the counterfactual of no corporate income tax.

To understand what this means, we first have to know what constitutes a corporation. According to the IRS, a corporation is a firm that has filed articles of incorporation with the state. (Thanks IRS!)

In general, corporations have some specific characteristics, such as shareholders with easily transferable shares and limited liability. Thus, if we think of a corporation as a firm that issues marketable ownership shares, then a tax on corporate income is actually a tax on the decision to issue marketable ownership shares.

This raises the question as to why a firm might sell marketable ownership shares and the corresponding implications of making it more expensive to do so. On this point, Laurence Kotlikoff and Jianjun Miao develop a framework for thinking about this. Their crucial point is this. The decision for the marginal entrepreneur about whether or not to incorporate depends on risk, the fixed cost of issuing ownership shares, and the cost of the corporate income tax. On the one hand, an entrepreneur would like to incorporate because adding shareholders allows for risk-sharing. However, to get the benefit of risk-sharing, the firm will have to pay a tax that it would not have to pay if the entrepreneur operated the firm on his/her own. Ultimately, whether an entrepreneur will decide to incorporate depends on whether the marginal benefits of incorporation are greater than or equal to the marginal costs. What this tells us is that there is some threshold for the skill or productivity of entrepreneurs. Those above the threshold will find it beneficial to incorporate. Those below it will not. Thus, when the government increases the corporate income tax, they are increasing the marginal cost of incorporating. As a result, the marginal benefit of incorporation must be higher than it was before the tax increase. This means that the skill or productivity threshold that the entrepreneur must cross to want to incorporate increases. In other words, fewer entrepreneurs will choose to incorporate.

In comparison to a world with no corporate income tax, the world with a corporate income tax will have fewer entrepreneurs who decide to incorporate. Since the benefit of incorporating is that it brings about risk-sharing, this also means there is less risk-sharing. Recall that when you tax something, you get less of it. What their model shows is that when you tax corporate income, you get less risk-sharing. Thus, it seems like the corporate income tax is a tax on risk-sharing.

To this point, I have focused on determining what is being taxed by the corporate income tax. We also need to know who is paying the tax. Politicians often complain when companies like Amazon do not “pay their fair share.” Of course, fairness is subjective. Nonetheless, what this highlights is that politicians often frame the tax as a tax on corporations, but corporations don’t pay taxes. Capital doesn’t pay taxes. People pay taxes. Do the owners of capital pay the tax through lower profits? Do firms pass along the cost of the tax to workers through lower wages and/or less employment? Who pays the tax depends on tax incidence. Of course, who pays the tax is related to what is being taxed. As it turns out, the assumptions made about what distinguishes corporate and non-corporate entities as well as some ancillary assumptions about the supply of capital have important implications for determining who pays the tax.

Maybe what politicians, like those I described, have in mind is that the owners of the firm (i.e., the shareholders) pay most of the tax. Arnold Harberger’s pioneering paper suggests that capital owners pay the tax — even those in the non-corporate sector. However, Harberger assumes that corporate firms produce a different good than non-corporate firms. In addition, Harberger assumes that factor inputs, like capital, are in fixed supply. The assumption that corporate and non-corporate firms produce different goods seems to violate casual empirical observation. In addition, since it assumes that the factor supplies are fixed, the model is unable to examine the effects of the tax on capital accumulation.

If we drop the fixed factor supply assumption, we get a much different result. Suppose that all firms are corporations and use the same capital. In this case, the corporate income tax is just a tax on capital. Here, the neoclassical model will suffice. The neoclassical model has a horizontal supply curve for capital. As a result, the equilibrium allocation of capital is entirely demand-determined. In the neoclassical model, since the supply curve is horizontal, a tax on capital is entirely paid by workers.

Of course, as I said before, there might be characteristics of corporations that make them different from non-corporations. For example, we could assume that corporations and non-corporations produce different goods and use different types of capital. We can now modify the neoclassical model to have multiple types of capital and assume that the corporate income tax is a tax on a particular type of capital. In the new textbook, Chicago Price Theory, the authors have a nice example of a corporate income tax as a tax on a particular type of capital. The conclusion that workers pay the entirety of the tax holds, but the cost of the tax to workers is higher than a tax on capital itself because of the additional distortion created by having different taxes on different types of capital. (Kevin Murphy provides an excellent discussion of this here.)

If we drop the assumption that corporate and non-corporate firms produce different goods, we can turn to Jane Gravell and Laurence Kotlikoff. As they show, this assumption results in a much larger deadweight loss than what Harberger finds because it distorts production within a sector that consists of both corporate and non-corporate firms producing the same good. In addition, capital owners are worse off from the tax at the expense of non-corporate entrepreneurs. Whether workers are better off or worse off depends on how the labor demand of entrepreneurs adjusts.

Finally, we could just abstract from capital completely like Kotlikoff and Miao do in the paper referenced above and examine the corporate income tax from the perspective of choices about being a worker, an entrepreneur, and incorporation. They show that the corporate income tax is a tax on high-skilled entrepreneurs who incorporate and workers who see lower wages due to the decline in the demand for labor as a result of the tax.

Are you following all of this?

If not, that is okay. Maybe the fundamental lesson here is that what and who is paying the corporate income tax and the size of the deadweight loss are issues of considerable disagreement because it depends on assumptions that we make about the differences between corporate and non-corporate entities as well as other ancillary assumptions. This disagreement is something to remember the next time your favorite politician (or economist) speaks confidently about the precise corporate income tax rate we should choose.