A Brief Overview of Good Taxation
So, I’ve started planning out various expansions of the US welfare state. They’re very expensive, and they’ll involve lots of tax increases. So, let’s discuss how to have good taxes!
The most important value in regards to tax revenue is elasticity. Elasticity is a unitless measure comparing the percent change in two different variables. For example, the price elasticity of demand refers to how much demand changes in response to increases in the price of a good. For our purposes, what we care about is the elasticity of various things we’ll be taxing (consumption, labor income, capital gains, land, etc.) relative to the net-of-tax rate (which is one minus the tax rate, AKA the proportion you keep). This will basically determine how much money a tax increase will raise. Let’s take an extreme example: say you’re taxing folks making over $100 million a year at 90%. If you decide to increase that to 95%, they will lose half of their take-home pay. This will obviously be a big disincentive for them to continue making that much money. Meanwhile, moving from a 90% rate to a 95% doesn’t remotely double revenue. So, you’ll probably actually bring in LESS revenue by raising the rate to 95%. If you’ve ever heard of Laffer curves, they depict this phenomena, and look something like this:
A result of this is that the first 5% tax raises more than the next 5% tax, and so on until the maximum. Thus, I can’t simply combine two estimates of raising the payroll tax by 10% and 7% or whatever. So we’ll need our own estimates, which I’ll discuss later.
But good taxation isn’t just about rates. It’s also about the structure of the tax system, especially the broadness of the tax base. For example, a value-added tax with lower rates for many items and exceptions for others would have a narrow base. We want to keep tax bases as broad as possible, for several reasons. First, “a broader tax base reduces the ETI [elasticity of taxable income to the net-of-tax rate], making it clear that a broader tax base and higher rates effectively complement one another”.² Because we’ll need lots of money to end poverty, we want to keep those revenue-maximizing rates high to… well, get more revenue! I specifically want to point out the concept of income shifting: basically, reclassifying income to save on taxes. A famous example is all the companies who say they’re located in Delaware to get lower tax rates. But people may also reclassify what is really labor income as capital income too. This is a common strategy for CEO’s: they get paid mostly in stock options, rather than in a salary. This is a type of tax avoidance, and should be minimized by taxing income from different sources at about the same rate.
Second, a narrow tax base is often regressive. In the US, we currently have a hodgepodge of different tax expenditures: tax deductions, credits, and exemptions, along with lower rates on capital gains as compared to normal income. A deduction, for example, lowers taxable income. So, if you make $50k a year and give $1k to something that’s tax-deductible, you’ll only have to pay taxes on $49k. These are inherently regressive, because the rich are taxed at a higher rate than the poor. So, a $1 deduction will save a rich person more than it does a poor person. Meanwhile, the rich get proportionally more of their money from capital gains, and they have the resources to find and exploit tax loopholes. So it’s no surprise that “the top 20 percent account for 75.2 percent of the benefits of the state and local tax deduction, 79.3 percent of the benefits of the mortgage interest deduction, 90.2 percent of the benefits of the pass-through income deduction, 91.5 percent of the benefits of the charitable deduction, and 93.2 percent of the benefits of the lower rates on capital income.”²
Third, expenditures are simply a roundabout way of doing things.³
“Many tax expenditures do not differ fundamentally from spending. The distinction between tax breaks and spending is often artificial and without economic basis. Education is one example. On the spending side of the budget, the federal government provides Pell Grants to help low- and moderate-income students afford college. On the tax side of the budget, so-called 529 accounts help parents pay for college by providing tax subsidies that are most generous for upper-income households. Both of these policies are government subsidies to promote higher education; the tax/spending distinction is not meaningful here.”
Not only are many expenditures similar to spending, they are also inefficient. To see why, let’s take the example of the charitable deduction. The idea of this deduction is to incentivize giving: if you don’t have to pay taxes on money you give to charities, it basically costs less to give to them, meaning you will probably give more. The measure of how much more you’ll give is the price elasticity of giving, which is estimated to be about -0.4⁴. What this means is that if giving money to charity became 10% cheaper, people would give only 4% more. In other words, the choice is between $10 of tax revenue and $4 of charity. You might think government is inefficient, but it’s generally not THAT inefficient. This is especially true when looking at what the top 10 charities cover⁷: healthcare, building homes, and disaster relief. All of these are prime areas for government involvement!
This is all, of course, leaving aside other critiques of how the rich use “charity” to stash their money away and keep power.⁵
So, exactly how much revenue could we raise by eliminating expenditures? An estimated 5.9% of GDP in 2021 if we kept the EITC (Earned Income Tax Credit)², or $1.357 trillion. Getting rid of the EITC would add another $70 billion a year to our revenue⁶, for a total of $1.427 trillion. This will be an excellent source of revenue for our future spending programs! And don’t worry, we’ll make sure no one who is poor is worse off from these changes.
So, we now have a framework to use when thinking about optimal taxation. I’ll also be writing a post soon on how to use elasticities to calculate the extra revenue from tax increases. But given that that will be very mathy, I didn’t want to expose folks to it who don’t care. After that, we can use some estimates of the size of different tax bases and their elasticities relative to the net-of-tax rate and make our own revenue estimates!
- https://www.cbpp.org/research/tax-expenditure-reform-an-essential-ingredient-of-needed-deficit-reduction (ignore the title, we don’t need to reduce the deficit)