Tuesday, April 04, 2017

Tax policy colloquium, my 299th session (?): Kathleen Delaney Thomas, Taxing the Gig Economy

Yesterday, Kathleen Delaney Thomas presented Taxing the Gig Economy. Before getting to the very interesting topic, it has occurred to me recently to ask myself how many colloquium sessions I have done at NYU. I think the math goes as follows. This is Year 22.  14 sessions per year would mean 308 by the end of the semester and 294 before it began. But over the years I've missed one session due to illness, one due to a funeral, and one due to travel,. In addition, at least 2 were canceled due to weather (for papers by Yair Listokin and David Kamin - while a Saul Levmore session was also formally canceled due to snow, we got to hold it anyway).

If I'm not leaving out any missed sessions (such as further snow dates), that would mean I started this year at 289. Yesterday was session 10 for this year, so I suppose a milestone awaits. But it also means I've had a whole lot of micro-immersions over the years that give me a kind of backlog I find helpful. E.g., yesterday's paper was on a topic we've discussed at the colloquium previously, e.g., when Shu-Yi Oei presented a paper on what was then called the sharing economy, but is now (in a term I find more apt) more commonly called the gig economy.

The paper presents two interesting proposals regarding the taxation of certain workers who are not classified as employees for tax purposes. Instead, they are deemed to be independent contractors and thus, in effect, self-employed small business owners, which can present them with tax compliance challenges that they may be ill-equipped to meet.

Who is in this group? To start with, people who earn payments via platform companies such as Uber, Lyft, Task Rabbit, Airbnb, Etsy, etcetera. We may think of them as workers, but there's often property involved, such as the Uber driver's car, the Airbnb host's residence, and the Etsy seller's braided jute throw rug. (Also, Seamless? Stubhub? Craigslist? Square Card Reader? See below.) But one of the issues still open for analysis is to what extent the proposal might apply to non-employee service renderers and other small business owners who don't happen to connect with customers via platform companies.

The paper's two proposals are (1) non-employee withholding for certain gig workers and (2) providing a "standard business deduction" (SBD) that such individuals could claim in lieu of detailing their actual business expenses. But before discussing the details, a bit of useful background might pertain to how we want to tax independent contractors' work effort in the gig economy relative to other work effort. This might prove useful in evaluating the tradeoffs that the proposals inevitably present.

An obvious benchmark would be to try to tax gig economy work just the same as other work (i.e., the standard neutrality norm). But there are complications that might tilt the analysis in either direction.

Tax gig economy work less than other work? Suppose that, say, the Uber driver who does a few hours per week, perhaps as a second job, is doing this in lieu of leisure, rather than in lieu of other market work. Gig economy work that was not a substitute for other work, and that was more tax-elastic than such other work, in principle ought to be taxed at a lower effective rate, under the "Ramsey taxation" inverse elasticity principle. Likewise, suppose that gig economy work, which the government can trace and tax because there are electronic records of it, is a substitute for cash transactions that escape being taxed. Once again, there would be a case for taxing gig economy work more lightly than things that are not, to the same degree, substitutes for using untraceable cash.

Tax gig economy work more than other work? - Various regulatory regimes are run through regular, and in particular full-time, employment - for example, various employee benefits and labor protections. While sometimes these get tax subsidies (e.g., employer-provided health insurance), the regimes are also often regarded as costly, at least by employers, who may therefore want to avoid them. This both discourages the use of regular employment relationships that might otherwise be optimal to the parties, and undermines the social policy goals that underlie the legal regime for such employment. Hence, one might want to tax-penalize work relationships that lie outside it (a la the rationale for the mandate/penalty in the ACA).

Higher compliance costs per dollar of tax revenue - Figuring out your taxable income as, say, an Uber driver is more complicated than doing so with respect to, say, cash salary, given the issue of what expenses are allowable. So if we taxed the two "the same" in terms of equal accuracy - or even, perhaps lack of net bias in either direction - the tax burden on the Uber drivers would be higher, discouraging that choice. While this is a key analytical point, its implications are potentially complicated (there is work, e.g., by Kaplow and by Weisbach concerning the value of accuracy that might help in analyzing it). Clearly central to the issues, however.

With this as general background that admittedly fails to point us in a definite direction, herewith some aspects of the paper's two (still tentative) proposals:

1) Non-employee withholding - A company like Uber is extremely well-positioned to do federal income and self-employment tax withholding on behalf of its drivers. Indeed, conceivably it would already offer it on behalf of its drivers if not for concern that this would undermine its legal position with respect to other issues that turn on employee vs. independent contractor status. So it might be appealing, at a minimum to induce or require certain types of "platform companies" to withhold on behalf of their workers, or to extend the requirement more broadly.

     a) Why might one favor broader withholding? - One reason would be to protect the government, e.g., against the risk that people will disappear or become insolvent. A second reason would be to benefit the workers subjectively, if they were glad not to have to handle estimated taxes or large amounts due (perhaps with penalties) on April 15. A third reason would be to benefit the workers paternalistically, e.g., by giving them forced saving, albeit at zero interest, via the prospect of an April 15 refund. The mix between reasons could affect the scope of the policy, e.g., with regard to whether withholding is elective or mandatory.

     b) Who might be required to withhold for non-employees? - Maximum range of the proposal might be all businesses that are making payments of at least $600 to a given individual who is not an employee during the year. Or it might apply more narrowly to platform companies, which can be defined as those that make money by facilitating exchanges between two or more interdependent groups, such as consumers and producers. But this might only apply to those that get the money as an intermediary, e.g., not Craigslist, and would also exclude mere payment intermediaries (credit card companies, Square Card Reader, etc.) that don't really have a work relationship with the payee. Fleshing this out is a key area in which the paper can make a contribution.

     c) How much to withhold - This part is tricky, due both to the possibility that the workers have other income and their having expenses (e.g., an Uber driver's gas, insurance, depreciation, etc. while on the job). Due to the deduction issue, it overlaps to a degree with the second proposal, pertaining to the proposed standard business deduction.

2) Standard business deduction (SBD) - The paper discusses offering the relevant group of people (to be defined) an election to deduct, say, 60 percent of their gross income (or gross receipts?) from gig work. (Alternatively, there might be a flat dollar amount SBD, or floor on the SBD.) For people who were fully covered, this would eliminate the need to determine which expenses are allowable and to keep track of them (although one might nonetheless do so if one anticipated that they'd exceed the SBD). The paper notes that 60% expenses appears to be a rough average for certain broad categories under IRS data.

     a) Who would get the SBD? - The SBD might be limited to people below a specified threshold of gross income (or gross receipts - the difference here is whether one deducts cost of goods sold), and perhaps also to people below a specified level of adjusted gross income (AGI). Thus, to quote the paper at p. 37, n. 178, the SBD would not be available to "high-income employees (e.g., law professors) who perform low-cost consulting services on the side." But are gross income and AGI limits enough? One might also have to consider cutting out types of independent contractor work in which allowable expenses seem likely to be very low.

     b) Cliff and phaseout issues - Suppose one got a 60% SBD so long as one's independent contractor income was under, say, $50,000. If one's actual expenses were much lower, there would be a huge cliff or notch with respect to the $50,000th dollar that one earned, as this might increase taxable income by as much as $30,000. Various coordination methods might ease the problem, although none is perfect. E.g., slower phaseout of the SBD, although this adds complexity and creates a broader range over which there are high effective marginal tax rates. Or, let everyone deduct 60% vs. the first $50,000 of independent contractor income (at least, of the types that aren't screened out under (a) above), although this extends benefits to the top of the scale. In the latter case, one might also want to apportion allowable expenses, e.g., by disallowing a % that equals the % of independent contractor income that gets the SBD.

These are good issues to develop, and I look forward to subsequent drafts of the article that will tackle them in detail.

No comments: