How to implement government policies
Tax breaks are opaque and someone else’s problem
Democracies love to fund policies via tax incentives. They love it because it is seemingly free money. It’s paying for things by giving back money. Magical! You want more people to be homeowners? Make mortgage interest payments tax-deductible. More people to get married? Tax break. More people to go to college? Tax break. Saving for retirement? Tax break. Charitable donations? Tax break. More kids? You got it. You name it, there’s probably a tax break for it.
Tax breaks can come in two forms: tax deductions (something you subtract from your gross income, lowering the taxable income), and tax credits (something you subtract directly from your tax bill). They can be designed with or without expiration, although more often than not they are designed without.
Tax breaks are usually put in place by the government in order to increase the supply of some good or feature by incentivizing either the citizenry or other entities like corporations to supply more of that “good”. That “good” can be children, savings, homeowners, investments, etc. For instance, by permanently taxing capital income at a lower tax rate, the government wants to incentivize investment. It is a form of tax credit. Tax breaks that target corporations are typically meant to incentivize them to provide goods that benefit the citizenry but that the government deems itself too incompetent to provide.
So how is this a bad thing? When I say it’s bad, I’m not talking about the good targeted by the tax break, I’m talking about the method of using a tax break to increase the supply of that good. It is bad because not only do tax breaks create deficits, they do so with unpredictable and permanent effect. They are highly desirable for a politician because they don’t require a budget here and now. They are unpredictable in their effect because it is hard to calculate the present value of all future tax breaks. And they are permanent in their effect because tax law is typically not systematically and regularly reviewed, unlike an annual budget. Voters unfortunately are under-indexed on opaque funding via future discounts on taxes and over-indexed on transparent subsidies.
How does one compute the cost of a tax break? For the duration of the tax break (and remember, they often have no expiration), every year the people or entities eligible will get the tax break. Therefore to compute the value to the beneficiaries, or the cost to the government, one needs to sum those benefits across all future years the tax break will be in place, using a discount factor to make the value of the benefit today comparable to the value of the benefit in, say, ten years (you need to hold a degree in economics to figure out the financial impact). Quickly assumptions on the future size of the eligible population, the discount factor, and other things, increase the uncertainty of the value of the benefit. The cost is therefore not only opaque, it’s also someone else’s problem. Perhaps one of the most egregious examples of this in recent memory is the terrible Foxconn deal in Wisconsin, with tax incentives amounting to north of $240,000 per job created (and this is the optimistic estimate) and about 43 years for the government to maybe recoup their investment.
Unfortunately, the financial abstruseness of tax incentives is precisely what makes them popular amongst politicians, because, while the balance sheet impact may be illegible, the headline impact (a picture of a politician shoveling dirt for the inauguration of some brand new industrial plant where cousin Jerry might get hired) is highly legible, and is known to swing undecided voters.
Use subsidies instead
As a rule of thumb, you should incentivize behaviors via subsidies, not tax breaks. Increase the demand for the things you want to see more in your society (homeowners, married people, savers, etc.) by directly giving money to those deciding to pursue those things. Contrary to tax breaks, subsidies are predictable, require funding, and are guaranteed to be periodically reviewed.
Now how should subsidies be provided? They should as much as possible be a continuous function f of some feature x of the beneficiaries. Most of the time that feature is income. (But it doesn’t have to be. If you want to incentivize fertility, x would be the number of kids.) If f is non-continuous, i.e. for instance brutally stops for some level of the feature, boundary effects appear. People right below the boundary after which the benefit drops will be incentivized to not increase whatever x the subsidy is conditioned on. Similarly, people right above the boundary that puts them outside the scope of the benefit will tend to reduce their supply of the level of the feature the subsidy is conditioned on. Distortionary behavior will happen at the boundary. Typical examples of boundary effects include hiding income or disincentivizing work. An example of this is subsidized housing in some cities: they are frequently only valid for people earning less than some arbitrary number. Depending on the value of that subsidy to the beneficiary, the beneficiary will be incentivized to not increase their income. Despite such undesirable boundary effects, such subsidies targeting citizens in a specific bracket of whatever metric (typically income) are far from uncommon.
Finally, subsidies should follow a not only continuous but also sliding scale. More precisely, you want the subsidy function to be logarithmic in x because any good has decreasing returns as you increase its supply.
The supply of good x is increased by a subsidy either because the beneficiaries directly produce that good in order to receive the benefit (e.g., being paid for giving birth to a child), or because the subsidy comes in the form or earmarked money, or vouchers, only spendable on good x, thereby increasing the demand for x. An important condition for the success of subsidies increasing the demand for x is the lifting of restrictions on the supply of that good. If restrictions are in place, markets will not be able to do their job and the result of the subsidy will simply be the increase in price of good x. For instance, if the government were to provide people vouchers for housing based on their income, but at the same time do not make the real estate market of cities more flexible, by maintaining highly questionable yet extremely common restrictions like floor-area ratios (FAR), or a poor public mobility system, this won’t solve anything.
Minimum standards: virtuous only in appearance
Democracies also love to regulate by just telling people what they should do. It’s no longer even about incentivizing behavior (be it via tax breaks or subsidies), it’s about telling you what’s better for you. Such policy-making strategy can be captured by the term “quality floors”, or “minimum standards”. For instance, many cities impose minimum housing consumption, as defined by a level on an index scale where the index is a mix of floor space, access to jobs, water and electricity infrastructure, etc. The regulation simply says “no one can live in this city if their housing doesn’t meet some arbitrary minimum standard”.
How is this a bad thing, you might ask? It sounds pretty virtuous. Don’t we want the best for everyone? Of course. But dictating those minimum standards does absolutely nothing to help the people who can’t afford them. They won’t just disappear because of such regulations. Well, they will actually leave the city. In fact such consumption mandate only exacerbates the duality and chasm of our current economy between city and non-city, a major problem of our times. It’s again a hidden cost. Rather than mandating it, it’s infinitely better to directly subsidize housing consumption: complement the salaries of everyone on a sliding scale in such a manner that anyone can afford a socially-desirable level of housing consumption. But again this has to be done while simultaneously breaking the shackles of many zoning laws that aim to decide top-down what kind of activity should happen where.
Perhaps more importantly and insidiously, policy-making via minimum standards can also create perverse incentives for the bureaucrat in charge. In practice, minimum standards artificially limit rights (e.g., the rights of a developer to build cheap housing by building high towers resulting in high FAR, or by building small dwellings). And when the bureaucrat limits rights, it’s only one step before they start trading those rights off for “favors”. This in fact happens all the time in urban policy: a bureaucrat will give a developer FAR exceptions for some project, if the developer agrees to dedicate a share of some other project to affordable housing, or if they agree to fund a public plaza. This in effect means the government thinks it can do better at providing for low-income people than the market.
This is probably how civilization dies: at the hands of a bureaucrat parceling out licenses. In practice, the bureaucrat has created a new currency, namely the rights artificially restricted, and effectively “bribes” entrepreneurs with that currency. It also creates major market friction. Markets, to thrive, need transparency: all sides to a transaction should ideally have access to the same information. If artificial costs to access that information are created (the difficulty of regulation, the unpredictability of zoning categories in which a building will fall, which will impact its long term value), then the market will lose efficiency. This uncertainty, this information asymmetry, reduces the level of trade, and only creates opportunities for value-less middlemen like specialized zoning lawyers.
I’ve focused on cities so much in this post because city planning exhibits some of the worst examples of bad regulation (to learn more about cities and urban policy, you should read Alain Bertaud’s amazing book). Cities are beacons of high productivity: agglomerating humans creates positive externalities that yield prosperity. We should have as many people as possible participate in the urban economy. People are attracted to cities because people want access to large, fluid labor markets. However, many big cities in the US and around the world are plagued by high housing prices, which are either due to ineffective transportation (the effective size of a city or labor market is a function of the effectiveness of its transportation system), or an undersupply of housing. There is nothing “natural” or inevitable about rising house prices. In truth, many policies in place are meant to keep people out of cities, most likely not intentionally, but in effect that’s what’s happening. Policies that a first-order analysis would describe as virtuous, like minimum housing consumption, or pollution reduction by only allowing odd-numbered license plates on the road, or even deliberately limiting the geographic expansion of the city, are in fact, on closer inspection, when you take second-order effects into account, tremendously bad.
Most people agree on the broad values that a democratic government should uphold. But there are good and bad ways to go about it. Let’s be more thoughtful about that and recognize how often hell is indeed paved with good intentions.
