The IRS recently received a great deal of funding through the Inflation Reduction Act, enabling it to hire more agents and cut down on backlogs. They also got permission to build a direct filing website, replacing expensive alternatives like Turbotax. These changes were fought, however, by Republicans. The proposed changes would very obviously make it cheaper and more convenient to pay taxes, so the cynical mind might think that Republicans opposed it for the same reasons they oppose any positive changes during a Democratic administration — any bad things which happen are blamed on the president. The cynical mind is correct, of course — my oldest political conviction is that Republican legislators are largely damned jackasses, and I don’t expect that to change any time soon — but there is a deeper, and perhaps better, reason for opposing it. Taxes vary in salience. Sales tax is far more visible than a value added tax, and property tax is more visible than withheld income tax. We should expect that the less salient a tax is, the higher it is. If you are concerned about the effects of taxation and government spending on the economy, as I am, making taxes maximally salient might lead to spending and tax cuts. But is this true?
Some excellent evidence on the political effects of salient taxes is provided by Cabral and Hoxby (2012). Most people pay property tax by writing one big check every year. It’s annoying. You can’t be unaware of it, like you can be unaware of the effect of corporate tax on prices. Some places, however, don’t have this. Some lenders offer tax escrow as part of their mortgage, so the property tax payment is broken apart into many small pieces, none of which are easily noticeable. They argue — to my mind, successfully — that the prevalence of tax escrow is uncorrelated with anything. The prospective buyer does not learn whether tax escrow is offered until well into negotiations, and if they were to change their mind and try to shop around for a better bargain, the deal would likely fall through. This requirement to reveal how property tax is handled only exists because people weren’t ever informed, in the recent past. They can plausibly control for all factors because lenders are assessing borrowers on the same basis. They show that whether escrow happens varies arbitrarily, and that is sufficient for me to believe their instrument.
Changing the salience of the tax has an enormous effect on tax rates. A 1 standard deviation increase in the number of people with tax escrow increases property tax rates by 1.2 standard deviations. The only plausible channel for this is through political action. This does not seem to come at the expense of government efficiency, although one cannot measure this precisely. They suggest math and reading scores after controlling for student characteristics as a measure of government quality, and find no improvement for having more government, but this is just a suggestive test. Voters believe that property taxes are the taxes which are spent most efficiently, yet it is the tax despised the most. If you want to keep the government small and efficient, then, one should make taxes as salient as possible.
The counterargument is that voters are stupid, and tax rates need to be kept higher to balance out the services they will inevitably demand. Keeping taxes salient would still be valuable in reducing distortions in consumer behavior. There are two main sources of evidence on this. Chetty, Looney, and Kroft (2010) measure the changes in consumer behavior induced by different tax structures. The first method is a field experiment in a grocery store. Rather than display the price without tax, and have it apply at the register, they post price tags which display the price with and without tax. They are not simply including the tax in the price, which might confuse some customers; instead, they are simply revealing more information. Reading the price with tax reduced customer demand by 8 percent! The second strategy is comparing states which include alcohol taxes in the price tag, versus those which apply it at the register. Places which apply it in the price tag, as you might now expect, discourage more alcohol consumption than those who apply it at the register.
Finkelstein (2009) is another study showing the distortion in consumer behavior from changes in salience. Turnpikes used to require cash tolls, leading to long backups at the toll booth. The changeover to EZ-Pass had two effects, then — reducing congestion, and reducing the salience of the toll. (The reduction in pollution, by the way, was used in an extraordinary study from Currie and Walker (2009) to measure the effect of pollution. Being within a kilometer of a toll booth led to lower birth weight and infant health than those 2-10 kilometer’s away). Finkelstein cannot use, then, the changeover itself to show the effects of salience. Instead, she shows that driving behavior varies less when tolls change after the change than before. These are not small changes — she estimates that the profit maximizing toll rose 20-40% after the adoption of EZ-Pass. The other piece of evidence is that toll rates became less sensitive to the local election calendar. Before, local politicians would cut toll rates before an election to try and gin up votes. Afterward, not so much.
I think taxes should be salient, but convenient. There is no reason to carve out a giveaway to tax preparation companies, but neither should we so hide our taxes that we have no idea what we are paying. In truth, I do not know what is best. It is so much nicer, isn’t it, to be positive rather than normative! I can say with certainty, though, that salience has a meaningful effect on consumer behavior, and political action.