Tuesday, July 30, 2013

In Which I Fix Everything

A while ago I started thinking about how I would go about designing a functional welfare state. As a free market guy I'm not a big fan of a welfare state, but if we accept that we've got to have one we can certainly do it better than what we've got, and fix a bunch of other problems while we're at it. I limited myself to preserving existing institutions in some form, since traditional libertarian shibboleths like abolishing the Federal Reserve in favor of a 100% reserve requirement gold standard or ending all welfare payments immediately don't really address the problems of institutions that we're basically stuck with whether we like it or not. So, if I had the power to reform political institutions but not replace them outright, I would approach the problem with the goal of eliminating the following problems:

I. Unnecessarily complex and divisive tax policy.
II. Obscure individual costs of government programs.
III. Welfare system is wasteful and creates disincentives to work.
IV. Ineffective, unfair monetary policy tools.

I'll discuss each of these in order.

Unnecessarily Complex and Divisive Tax Policy

So, first up is the tax system. Taxes cause deadweight loss that increases proportionate to the square of the tax rate. Economists across the political spectrum agree on this, but those who favor higher taxes generally view that the government is just so much better at spending money than the rest of us that it's still a good idea to have high taxes. Deadweight loss is also affected by the price elasticity of supply and price elasticity of demand with smaller deadweight losses suffered when elasticity of supply and demand are relatively low. To maximize revenue for a given deadweight loss, tax rates would be inversely proportionate to price elasticity.

However, actually figuring out price elasticity is very much like shooting in the dark at a moving target. Nobody really knows what the demand curve for anything actually looks like since consumer preferences are only made known through purchases at a given price point, and changing the price of a good to get another price point only gives us one more point with which to construct a curve that is by no means a straight line. Furthermore, the greater the time disparity between the collection of purchase data at different price points, the more likely it is that consumer preferences have changed in the meantime. Consumers also take note of frequent changes in price, and this affects their purchasing behavior. As a practical matter, actually constructing an accurate demand curve is a pipe dream.

Companies hire economists to try to figure this stuff out anyway, since taking an educated guess is a bit better than nothing, but if we were to try to base tax policy on it, we would introduce a lot of inefficiency-inducing political pressures at every point where a judgment call is necessary, and the accuracy of the results would be even worse. It would also run counter to the goal of avoiding unnecessary complexity in tax policy.

So, what I would do is replace all current taxes (income tax, payroll taxes, corporate income tax, etc) with a Value Added Tax. It's basically a sales tax that is collected at each stage of production instead of just at the end. To eliminate the possibility of carousel fraud and not distort relative prices of imports and domestically produced goods, the VAT would also apply to the value of imports. The IRS would perform a function similar to the UK's HM Revenue and Customs, although there would be no exceptions or variable VAT rates for different goods. This would minimize the potential for abuse and inefficiencies induced by political pressure, as well as prevent nonsense like the case of Jaffa Cakes being a cake or a biscuit for tax purposes.

Aside from reducing compliance costs with taxation, a VAT would give everyone a stake in government spending. There is a very strong tendency in a representative system of government to try to make other people pay for things that benefit you. As a general rule of thumb, politicians respond to people according to their political significance. Swing voters (and in presidential elections, swing states) get subsidies and lots of political attention, while voters whose vote is a foregone conclusion will be exploited for all they're worth by politicians on the other side. Non-voters are a source of revenue at best, but there is no political capital to be gained by doing anything for them.

The complexity of the tax system means that any individual's actual share of the tax burden is extremely difficult to figure out. A good example of this is the payroll tax, which is supposedly paid 50% by the employer and 50% by the employee, but in practice the actual burden of the tax shifts according to the relative elasticity of supply and demand for labor. Remember when I talked about how hard it is to plot out an accurate demand curve earlier? You'd need to be able to do this to determine what your actual share of the payroll tax burden is, as well as every other tax on any transaction you are involved with. Good luck!

Once you've assembled a team of specialists including (at minimum) a tax accountant, an economist, and a wizard (to verify the accuracy of the economist's assumptions), you could look at the share of the budget that each government program comprises, and from there you could calculate how much a particular program cost you personally that year.

However, under a VAT this would be simplified somewhat. While people would still try to vote themselves disproportionate benefits out of the public treasury, under a VAT they at least wouldn't be able to use political influence to excuse themselves from taxation and push the costs of government programs onto others, which brings me to the next problem to fix: obscure costs of government programs.

Obscure Individual Costs of Government Programs

With a VAT, the tax burden is quite a lot easier to figure out. You wouldn't even need to know what your income is; you'd just need to know what the VAT rate is, which is about as hard as knowing your local sales tax rate. As an example, the cost of Medicare, Medicaid, and Social Security together comprise about 40% of the federal budget. Under a VAT with a tax rate of 30%, you could say that the cost of those programs was 12 cents on the dollar, for everybody.

Having the cost of government programs framed in such a way makes them comprehensible to everyone. I've got a strong suspicion that on some level, voters think that a billion is about ten million, and a trillion is about ten billion. The human brain sucks at actually comprehending large numbers.

A good demonstration of this was the Occupy Wall Street protests. A poll showed that 94% believed that the government spent more on the military than it did on education or healthcare and pensions. In reality, the government spends more than twice as much on healthcare and pensions as it does on the military. While it's tempting to say that liberal protesters are just particularly awful with large numbers, I think that ignorance of the costs of government is probably a little more widespread than that.

Making it so that the cost of government programs is easily comprehensible to the average voter can only improve political outcomes. At the very least, the political discourse would be a little harder to overwhelm with misinformation, and a VAT would make it harder to divide voters on tax issues by pushing the costs onto some other group.

Welfare System is Wasteful and Creates Disincentives to Work

This is one of those issues that suffers disproportionately from the difficulty in calculating individual tax burden. How much does a given person pay in taxes for welfare? How much does a household below the poverty line actually collect in benefits? The amount of money going into the system compared to the amount of money going out in benefits would give us a good idea of how much money is being lost to administrative costs or government waste.

Last year, the Senate Budget Committee found that combined state and federal welfare spending combined amounted to about $60,000 per household below the poverty line. This isn't to say that every household below the poverty line collected this much in benefits; some of these benefits accrue to households above the poverty line. But still, that's 20% higher than the median household income in the United States, so what gives?

Well, that figure includes Social Security, which accounts for an awful lot. There are also administrative costs for each of the wide variety of programs aimed at poverty. There are subsidized housing programs, food stamps, medicare/medicaid, and so on. The total cost of welfare spending is another one of those things that most people remain oblivious to. To the Left it's not enough and to the Right it's too much, but despite how often people bicker about this it's pretty rare for anyone to actually bring up hard numbers and break down where they all come from. To make matters worse, poverty statistics in the United States are based on income, but do not include government programs in that figure. A family of four with a cash income of $20,000 per year (which is below the poverty line of $23,550 for that family in 2013) and government benefits totaling an additional $30,000 would be living on median income, but would still count as impoverished just because we don't count the government benefits that account for 60% of their income.

I did some thinking about how I would design a social safety net that minimizes administrative costs and does not produce disincentives to work for those at the low end of the income spectrum. The current system has benefits that phase out in stages as income increases, so additional income is partially offset by decreasing benefits. It's also split into countless different programs. Nobody pays taxes on welfare benefits, which makes for a fundamental divide between those who collect benefits and those who pay taxes. From the perspective of the zero-income welfare recipient, they don't care what tax rates are required for a particular government program because their share of it is zero. Their cost-benefit analysis doesn't include costs, and they will consistently advocate greater spending on anything that they like.

What I would do is replace the entire array of social welfare spending (including social security, medicare, food stamps, subsidized housing, etc) with a universal basic income. It's the simplest possible solution: give every adult citizen a regular payment. When combined with a VAT, this would mean that there is a social safety net, but also that everyone would have a stake in the cost of government. It would also resolve one of the major objections to replacing the income tax with a consumption tax like a sales tax or a VAT, specifically the non-progressive nature of those forms of taxation. Some people really like the idea that a person with more income pays a higher percentage of their income, and oppose consumption taxes on the basis that they are not progressive. They occasionally claim that sales taxes are regressive, but I think it's just beyond stupid for anyone to make such a claim. Everyone facing the same tax rate is by definition neither regressive nor progressive.

Milton Friedman supported a negative income tax; he supported subsidies for people making less than a set amount, and taxes paid on income above that amount. That system still features a skewed cost-benefit analysis for people making less than the tax free amount, since they don't face any of the costs of any program that they support. Otherwise, there are similarities.

Here are some sample incomes to show what this would look like for various income levels with a VAT of 50% (meaning that 33% of the total cost is in taxes) and a universal basic income of $15,000. The break even point using these values is $30,000, where a person making less than this effectively gets a subsidy and a person making more than this is paying more than they are getting.

Income of $0 plus $15k minus 33% = $10,000 ($10,000 subsidy)
Income of $15k plus $15k minus 33% = $20,000 ($5,000 subsidy)
Income of $30k plus $15k minus 33% = $30,000 (No subsidy)
Income of $45k plus $15k minus 33% = $40,000 ($5,000 tax burden)
Income of $60k plus $15k minus 33% = $50,000 ($10,000 tax burden)

As income increases, the real tax rate approaches 33%. Higher tax rates or lower payouts push the break-even point lower. A VAT combined with a universal basic income ends up producing a progressive tax system without actually charging anyone different rates or granting exemptions. It gives everyone a stake in government spending, clarifies the cost of government, simplifies the tax code, and removes one of the most divisive tendencies in representative systems of government. Not only that, but it provides a perfect framework to reform our banking system, in particular the role of the central bank.

Ineffective, Unfair Monetary Policy Tools

Here is an overview of monetary policy tools available to the Federal Reserve. The biggest one is open market operations, which is sort of like a steering wheel that is connected to a few different cars at once. They try to influence the Fed funds rate and the monetary base using the same tool. If they want to influence one but not the other, that's not possible. If they want to increase the availability of funds to the public as a macroeconomic policy move, they use open market operations to expand the monetary base, making those funds available to banks who then lend money to the public.

There are two problems evident with this. The first is that it's unfair. Increasing the money supply diminishes the real value of debt, so there is a transfer of wealth from creditors to debtors. The devaluation of dollar-denominated assets such as savings accounts is what most people talk about when they complain about how unfair inflation is. However, if the change in the money supply is anticipated, it will be reflected in interest rates to account for this. Half of the Federal Reserve's job is just bluffing banks into doing what the Fed wants them to do.

That's not the really unfair part, though. If the increase in the money supply is known well enough in advance that interest rates can account for it, then it's a wash. The really unfair part is related to the distorting effects of monetary injections. When the money supply expands, prices rise, but they don't all rise at the same time. The people who get the new money first get to spend it before prices rise, and as that money circulates through the economy other prices rise as the new money reaches them. The Federal Reserve expands the money supply through commercial banks, meaning they consistently get the money before anybody else. This gives banks an opportunity to buy assets secure in the knowledge that they are paying a lower price than they'd have to pay for those same assets after their money has had a chance to slosh around the economy for a bit.

The ineffective part is something we've been seeing recently. If you're trying to add money to the economy through commercial banks, that's rather contingent upon the banks actually lending that money out instead of sitting on it. Which is exactly what they've been doing. The less money that banks lend, the more desperate the Fed is to give them money in the hope that they will lend it out. I think banks have caught on to this, and now have nearly two trillion dollars in excess reserves while the Fed scrambles to do anything it can to throw more money at them.

Coincidentally, if banks are holding more than the Fed's reserve requirement, then changes in the reserve requirements become entirely useless as a means of enacting macroeconomic policy unless you raise the requirements high enough that banks have to cut back on lending even with the extra money they were already sitting on. Lowering the reserve requirement won't have any effect no matter if banks don't do any extra lending with the extra slack you've given them. It's like pushing a rope.

So, what would I do to fix this? I would use the universal basic income as the outlet for monetary creation. The Fed would still retain control over monetary policy to distance it from political pressures to increase the payout, but changes in the money supply would come in the form of changes in the universal basic income. It's a blending of monetary policy and fiscal policy, but as long as the decision is made by the central bank rather than politicians, it's more insulated from the political pressures that produced awful monetary policy back when monetary authority was vested in Congress.

Rather than having interest rates derived from the fed funds target rate which is aimed at through open market operations, interest rates would be derived from the supply and demand for loanable funds, which is inherently counter-cyclical. In any case, it'd be a lot more reliable than just throwing money at the economy while eyeballing the Consumer Price Index in the hope that it'll tell them when they're doing too much.