Tax complexity is a subsidy for lobbyists, and the code shows how it works

If you have ever sat at a kitchen table in March, surrounded by shoeboxes, with a calculator and the expression of a man defusing a bomb, I want you to know something. Nobody did this to you by accident. Your confusion was manufactured, installed in the code line by line, and paid for. Somebody had to ask for each of those worksheets. And somebody got something in return.

Today's subject is the Internal Revenue Code as a product, and the question is the oldest one in politics: who benefits? The answer is not the person with the shoeboxes.

What a "tax expenditure" is

Start with a term that sounds like jargon and is actually the key to the whole racket. A tax expenditure is a provision of the tax code that reduces someone's tax bill to achieve a goal other than raising revenue: a deduction, an exclusion, a credit, a special rate, a deferral. The Joint Committee on Taxation and the Treasury each publish an annual list of them with estimated costs. The list is long.

The point of the label is that these provisions work exactly like spending. Here is a labelled hypothetical to make the equivalence plain.

Suppose Congress wants to encourage families to buy a particular gadget, a made-up one. It has two ways to do it.

• Method A, a direct grant. The Treasury sends each buyer a check for $1,000. The cost to the Treasury is $1,000 per buyer. It shows up in the budget as spending, it gets a hearing, and an appropriations committee can cut it.
• Method B, a tax credit. The buyer subtracts $1,000 from his tax bill. The cost to the Treasury is $1,000 per buyer, in revenue not collected. It shows up in no appropriations bill. It is on autopilot, reviewed by nobody, and continues until Congress votes to end it.

Same money, same effect on the buyer, same hole in the budget. The only differences are that Method B is harder to see, harder to cut, and delivered through a code that is already thick. Politicians love it for those reasons. A program that does not appear on the spending side does not draw the same scrutiny, and the lobbyist who wins a credit has won something that outlasts several Congresses.

A deduction works the same way but with a twist that exposes the fairness problem. A $1,000 deduction lowers the income you are taxed on. In a hypothetical 10 percent bracket, it saves $100. In a hypothetical 35 percent bracket, it saves $350. The same deduction is worth three and a half times as much to the taxpayer in the higher bracket. If you ever hear that a deduction "helps middle-class families," ask what bracket the family is in, and then check the arithmetic.

The business model

Now the part that gives this post its title. Think about what a lobbyist sells. He does not sell policy; he sells a change in a rule that is worth more to his client than the cost of the lobbyist. A tax provision that saves a company or an industry tens of millions of dollars a year is worth a large annual retainer. And the better the provision lasts, the better the return.

Count the participants in the market.

1. The industry or interest group that wants the provision. It pays for the lobbying and the lawyers.
2. The lobbyists and lawyers who draft the language and shepherd it through committees.
3. The members of Congress and staff who get a hearing, a donation, a grateful constituency, or the satisfaction of having helped.
4. The accountants and tax-prep industry that gains from complexity because they are the ones who interpret it. A simple tax is a threat to a business that sells interpretation.
5. The ordinary filer, who pays the cost through a more complicated return, a higher rate than might otherwise be needed to raise the same revenue, and a code that nobody fully understands.

Four of those five parties benefit. The fifth is everybody else. Economists call this a concentrated-benefit, diffuse-cost problem, and it has a long record. The benefit to each beneficiary is large and obvious and worth fighting for. The cost to each taxpayer is small, spread thin, and hidden inside a form. No one forms a lobby to defend the interests of everyone who has to do the paperwork.

This is James Madison's insight about faction in Federalist No. 10, applied to the tax code. A faction is a group united by a common interest adverse to the rights of other citizens or the permanent interests of the community. Madison worried that small, organized groups could capture the legislature. A tax code made of thousands of special provisions is what that capture looks like on paper.

I want to be fair here, because a good-faith reader might object that not every provision is a shakedown. Many credits and deductions began with a plausible public purpose, and many people who support them sincerely believe in them. Granted. The problem is structural. A provision does not have to be corrupt to be costly, and an honest idea written into a complex code still adds to the pile.

What the National Taxpayer Advocate says

You do not have to take my word for the scale of the burden. The National Taxpayer Advocate is an official inside the Internal Revenue Service, an ombudsman for taxpayers, whose office publishes an annual report to Congress. The report lists what it considers the most serious problems facing taxpayers, and over many years complexity of the tax code has appeared at or near the top of that list. The Advocate's reports describe taxpayers spending large amounts of time and money on compliance, often paying for professional help, and having trouble understanding what they owe. Read the latest report for the figures on the hours and costs. It is available on the Advocate's website, and the numbers are the agency's own.

What I want you to hear is that the officer paid to represent the taxpayer inside the tax agency keeps telling Congress the same thing: the code is too complex for the people it governs to follow without help. And Congress, which wrote the code, keeps adding to it.

The 1986 experiment

There is a hopeful chapter and a cautionary one, and they belong together.

The Tax Reform Act of 1986, signed by President Reagan on October 22, 1986, was a deliberate attempt to go the other way. It was a bipartisan effort. It cut the top individual rate sharply, from 50 percent to a 28 percent headline rate (with a narrow 33 percent band for some incomes), collapsed the bracket structure into a small number of rates, and, in exchange, eliminated or narrowed a long list of shelters and preferences. The bargain was precisely the one conservatives should want: lower rates paid for by a broader base, fewer special favors, a simpler return.

For a time, it worked as advertised. Rates were lower and the number of special provisions fell. Then the code filled back up. In the years that followed, Congress raised rates again in 1990 and later, and added new credits, new deductions, new phase-outs and new special rules, one at a time, each with a sponsor and a constituency. The base narrowed again. It was not a conspiracy. It was a market. The demand for favors did not disappear in 1986. It waited. Over time the demand found suppliers.

The lesson is the one the Founders understood about power: a reform that depends on the ongoing restraint of those who profit from its failure will not last. The remedy is a structural one, such as a rule that makes new tax expenditures hard to add, a sunset on each provision so that it must be justified again, or a flatter tax with so little room for exceptions that there is nothing left to lobby for. We look at some of those designs in A flat tax, a national sales tax and a VAT, compared for one family.

The strongest objection: targeted credits help families

The best answer to everything I have said comes from people who like these provisions, and it should be stated at its strongest.

A tax credit is the best way to help families, because it reaches people through a system they already use, without a new bureaucracy. A child tax credit gets money to parents without a caseworker. An earned income credit rewards work. A deduction for mortgage interest or for charitable gifts encourages behavior that benefits society. Using the tax code for these purposes is efficient, it respects individual choice, and it lets families keep more of their own money rather than sending it to Washington first.

That is a respectable argument, and parts of it are true. Here is my answer, point by point.

1. Efficient for the government is not the same as efficient for the family. A credit saves the Treasury administrative cost by pushing the cost onto the filer. The family must keep records, fill out schedules, and prove eligibility. The Taxpayer Advocate's reports, and the error rates in some of the largest credits, show how often the family gets it wrong, through no fault of its own. Complexity does not vanish; it moves to the kitchen table. See The EITC began as a conservative idea and now pays out billions in error for a credit that illustrates it.

1. "Letting families keep their own money" has a better form. If you want families to keep more of what they earn, lower the rate or raise the standard deduction. That helps everyone, through a rule everyone can understand. A targeted credit says instead: you may keep your money if you behave in the way Congress prefers. That is not letting families keep their own money. It is a permission slip.

1. Targeted provisions grow. A credit with a purpose invites a second credit for a neighboring purpose, and a third. Federal tax support for ethanol that began with a 1978 exemption ran for decades, and its blenders' credit expired at the end of 2011, an example of how a provision created for one reason survives long after the reason has faded. Each one is a monument to a lobbyist who succeeded.

1. The tax code is the wrong place for social policy. A nation's priorities belong in the budget, where they compete with one another in public, and not in a code where they are invisible. When Congress spends through the tax code it escapes the discipline of the appropriations process. We see the same dynamic in baseline budgeting, discussed in baseline budgeting, where a smaller increase gets reported as a cut, where a smaller increase gets reported as a cut. A spending program disguised as a tax break is the mirror image of that trick.

1. A simpler code is itself a form of fairness. A tax that a filer can understand is a tax he can judge. A tax that requires an accountant to interpret favors those who can afford one. The rule of law requires that the law be knowable. Madison wrote in Federalist No. 62 about laws "so voluminous that they cannot be read, or so incoherent that they cannot be understood." The tax code is the modern example.

What to do about it

I will offer a few reforms that do not require a revolution.

• A sunset on every tax expenditure. Each provision should expire after a set number of years unless Congress votes to renew it, and the vote should be recorded and public.
• A cost line in the budget. Treat each credit and deduction as the spending it is, listed in the budget alongside programs, so that voters can compare a tax break to a direct appropriation.
• A bigger standard deduction in place of narrow breaks. Fewer people itemize, fewer people hire preparers, and the benefit is available to all.
• Rate cuts paid for by base broadening. The 1986 formula was right. The lesson from 1986 is to pair it with structural protections so the base does not erode again.
• Return-free filing by those who want it. Some countries let the tax agency calculate simple returns. That has its own risks, since a government that fills out your return has an interest in the answer, and I would want strong safeguards. It is worth a debate.

What the Constitution says about all this

Article I, Section 8 gives Congress the power to lay and collect taxes "to pay the Debts and provide for the common Defence and general Welfare of the United States." The purpose clause matters. Taxes are for funding the government's legitimate functions. They are not a general-purpose tool for steering the private choices of hundreds of millions of people. A code that has become a sprawling instrument of social engineering has drifted far from that purpose. Read the clause on the Constitution page and decide for yourself.

The case in brief

1. A tax expenditure, whether a deduction, credit or special rate, works like spending but escapes the appropriations process, so it is hard to see, hard to cut and easy for lobbyists to defend.
2. A deduction is worth more to a taxpayer in a higher bracket, so a $1,000 deduction saves $100 at a hypothetical 10 percent and $350 at 35 percent.
3. Complexity has a business model: concentrated benefits go to interest groups, lobbyists and preparers, while the cost is spread across every filer, which is Madison's problem of faction in tax form.
4. The 1986 reform lowered rates and cut shelters, and the code filled back up, showing that reform without structural protection does not hold.
5. Targeted credits are not the best way to help families, since a higher standard deduction or lower rates help everyone through a rule anyone can read.

What you can do: ask your representative's office whether they will support a sunset date on every new tax credit or deduction, and read the National Taxpayer Advocate's latest annual report to Congress.