Funding the US Government

As I explained earlier, in the US it takes about 32% of GDP to fund government seriously (I’m ignoring federal-state distinctions here – I’ll talk about them later). It’s a little higher than the current level of spending, but with somewhat different priorities – less defense and farm aid and health care waste, more welfare. In other countries, where welfare payments need to be higher relative to GDP, it’s a higher proportion, probably in the mid-to-high 30s.

Now there’s the question of how to raise that money. Strictly speaking the US needs more than 32%, mostly thanks to Reagan and both Bushes’ fiscal irresponsibility. Interest on the debt is about 3% more, so in fact the US needs to raise 35% to break even.

Now, the simplest way to raise money for that is to tax individual incomes flatly. Individual income is about two thirds of the USA’s national income, so that level would have to be 53%. That’s too high, especially at the bottom, so the easiest way to divert money away is by taxing corporate profits. The current tax, 35%, generates 2.5% of GDP in income, but is full of loopholes. Corporate profits are 10.7% of GDP and rising, so closing off loopholes and hiking corporate tax to 40% should increase revenue to 4%.

Another way to divert costs away from most people is to tax capital gains at the same rate as normal income. Capital gains taxes are heavily dependent on how the stock market’s doing, and even so in 2000 they only reached 1.5% of GDP, albeit at a reduced rate.

Finally, dividend tax is supposed to complement corporate tax. If corporations pay 40% and the top individual rate is 60%, then dividends should be taxed at 33%. In terms of revenue, there’s no difference between a 40%/33% tax and a 60%/0% tax, so the revenue from that can be viewed as 60% of corporate profits.

This leaves the big two – income and consumption taxes. Assuming a top rate of 60% for capital gains and corporate/dividend purposes, it yields maybe 8% of GDP in those taxes, i.e. 27% in income and sales (note that social security taxes are functionally identical to income taxes). Right off the bat it means an average of 40% on income, if there’s no sales tax.

The question of how to divide this 40% is heavily dependent on the degree of social engineering permitted. Strictly speaking, anything but payment based on need is social engineering. For example, anything other than a precise child credit and marriage penalty socially engineers birth rates, marriage status, and within-family division of labor.

On the other hand, taxation that encourages behavior that helps the economy may be acceptable; the gold standard is whether the economic gains outweigh the social engineering effect, in some sense. Sin taxes barely decrease public health spending, so they’re out; if you don’t believe me, look up non-smoking Sweden’s health expenditure versus smoking Japan’s. Taxes on environmental externalities encourage innovation in green technologies, and on top of that raise a lot more money, so they’re in.

Raising US gas prices to European levels, i.e. $7/gallon, will net $4.50/gallon * 581 gallons/year per car * 243 million cars = $630 billion, or a little over 5% of GDP. The 581/year average is for sedans only; for SUVs it’s 813. And assuming this will reduce gas consumption to French averages, i.e. 6.2 l/100 km = 38 mpg, it will cut gas consumption by 43% from the sedan-only average, which will reduce revenues to $360 billion. In fact there’s no way it will go up to 38 mpg immediately, because it will require people to immediately move closer to the city centers and buy cars that consume less fuel, and cities to spend more on public transportation. So for the first year, a 5% receipt is likely; subsequently it will go down, but so will debt interest payments.

Assuming no further sales taxes, it makes the average income tax burden 33%, which roughly means household income tax must raise 2 trillion dollars. Now I’m going to base calculations on the assumption that all households have the average number of people, 2.65. Since my data is stratified only by income, it’s equivalent to assuming that household size is independent of income (it’s not, but that’s the best I can do).

From a need-based perspective, it makes no sense to tax people on welfare. Since my minimum income proposal goes up to 150% of poverty, which for 2.65 people equals $23,000/year, it means no taxes on income under that amount. Income under that amount averages $20,000/year, which means that adding the total income of households making less than $23,000/year and $23,000/year for each household making over that threshold yields an average of $20,000/year. That’s a third of household income, so the average tax above that amount needs to be 50%.

Although a 50% tax on income above $23,000/year sounds scary, it’s not so bad for the middle class as it seems. The median household will pay a 25% income tax, while under the current system, federal income tax starts at 10%, FICA is 14.2% of what it costs the employer to employ a person, and state income tax typically starts around 1% and rises rapidly. A head of household making $46,000/year pays 14.9% of his gross income in taxes, or 13.8% of the cost to his employer. To take one state, in Wisconsin he’ll also pay 6.1% of gross income, or 5.6% of the cost to the employer. So in fact, his income taxes will be cut to 25% from 33.6%.

Obviously, this system becomes more and more of a tax increase as income goes up. Marginal tax in Virginia for a couple filing jointly with each person making $25,000 is 13.9% federal, 14.2% FICA, and 5.3% state, or 33.4% (the couple’s total liability under my system will be 27%, down from 31.6% right now). But even at $92,000, it’s 37.5% in my system compared with 35.6% in the current one.

If raising taxes on households making above $85,000-90,000 is too onerous, there’s a way out. Note that the assumption about corporate/dividend and capital gains taxes is that individual income is taxes at 60% in the highest bracket. The top quintile’s average income is just short of $150,000/year, so hiking its tax burden by 10% nets the government $6,000 per top quintile household, or $1,200 per household, or 2% of all taxes.

2% seems low, but it allows any of a) getting rid of the debt faster, b) spending more on social security, c) having a few deductions, or d) reducing every taxpaying household’s tax burden by $1,600, or equivalently raising the income tax threshold to $26,250. The tax hike at the top makes the break-even point the same as in a pure 0/50% system, but it cuts the taxes of people in the middle three quintiles more.

In either case, although the marginal tax hike seems huge, in fact it’s a tax cut for 80% of the population, which, incidentally, makes it more politically feasible than the spending increases. In fact if the FICA cap is lifted first, the plan will be an income tax cut on almost everyone; going by Massachusetts tax rates, which are actually relatively low at the top, lifting the FICA cap will increase the rich’s marginal income tax to 51.3% of employment cost.

Given that, you may ask yourself how this plan is revenue-positive. The answer is that in the real world, the rich pay the lower of the two rates available to them, income and corporate/dividend. Big businesses take care of their own, and small business owners can choose how to compensate themselves. Not coincidentally, in the 1950s, when corporate taxes were higher, the US collected 5-6% of its GDP in corporate tax revenues. In addition, the gas tax raises 5% of GDP, which is the same as a sales tax of about 7.5%, higher than most states’ sales taxes.

More interestingly, the bracket nature of the current US income tax is such that after eliminating the FICA cap, my plan will be a tax cut on people making less than about $85-90,000, depending on the state, and on people in the current 35% bracket, which starts at $336,550; on the rest it will be a small tax increase.

But the most interesting thing about this exercise is that it tends to highlight how tax brackets are an unnecessary complication. To someone making $50,000 a year, it doesn’t matter if there’s a $20,000 exemption and 50% on everything about it, or an $15,000 exemption, a starting rate of 25% between $15,000 and $25,000, and 50% thereafter. The starting rate will raise more money from people making $15,000 to $25,000, and only from them, netting about $9 billion dollars, or 0.075% of GDP.

Also, the actual revenues are heavily dependent on the distribution of household size, which I ignored entirely. In this system the easiest thing to do is to determine a poverty rate for each size of household and then let the tax be 50% on anything above 150% of that rate, possibly with an increase to 60% at a fixed level, say $100,000/year.

It’s possible to have a more individual system, but that requires having a more complex welfare payment system. The simplest thing to do is to have a hybrid system, in which welfare and tax payments are determined based on household size, but are then split equally among members. For welfare, it means each parent gets half the check in a family household; for taxes, it means the exemption is split equally among working members. This procedure helps increase women’s financial independence without egregious social engineering.

On one last note, I haven’t said anything about deductions. That’s mostly because most deductions cause more problems than they solve. For example, the mortgage deduction is a subsidy to the middle class at the lower and upper classes’ expense, which doesn’t even increase home ownership; the USA’s home ownership rate is 69%, compared with 67% in Canada, which has no mortgage deduction.