Scott Sumner  

Why I favor taxing subsidized goods

PRINT
Karl Smith's Weak Critique of ... Is Taxing One Item Because It'...

A number of people have asked me why I opposed the repeal of the medical devices tax. Taxes generally create deadweight losses, as they reduce output to a position below the socially optimal level (which is usually the free market level.) There are a few exceptions:

1. Taxes on externalities, such as pollution.

2. Taxes on goods that are already heavily subsidized.

Here's the standard graph showing the deadweight loss from a subsidy:

Screen Shot 2017-03-07 at 4.33.23 PM.png
Recall that a subsidy will tend to increase output and reduce government revenue. A tax imposed on this same good will reduce output and increase government revenue, essentially reversing the effect of the subsidy, and thus moving you back closer to the optimal level of output.

Of course you can overdo it, but the medical devices tax was far smaller than the subsidy on medical devices, which is enormous. (Subsidies include Medicare, Medicaid, tax deductibility of insurance, etc.) So the medical devices tax was a modest step in the right direction. And now it is to be abolished. I wouldn't object to its abolition if they had also abolished government subsidies such as the deductibility of health insurance. They should have abolished that tax deduction, and replaced it with a fixed tax credit for anyone with company provided health insurance plans. (Then they would no longer need the Cadillac Plan tax either. I seem to recall that John McCain once proposed this approach.)


Comments and Sharing






COMMENTS (11 to date)
pyroseed13 writes:

Scott, wouldn't this logic seem to imply that we should be imposing a whole host taxes on things in the health care sector? The medical device tax just seems like a poor way to address this problem, because medical devices are not directly subsidized by the tax code like health insurance is.

brad writes:

Does the fact that the Medical Device tax is an excise tax?

Thus the company pays taxes whether they are profitable or not. That would seem to put a big drag on innovation in this space as many new products are probably money losers at the beginning.

This is particularly concerning to me as the parent of a Type I diabetic where innovation in the medical device space (insulin pumps) have drastically improved the life of my daughter and innovations on the horizon (external pancreas) show the ability to improve it even more.

And I agree with pyroseed13 that this seems unusually targeted and thus distorting.

Thaomas writes:

Shifting the demand curve out for an item is not the same as a subsidy for producing it. There is no "subsidy" for producing medical devices so I see not reason to tax them on that second best ground.

brad writes:

Should have been: "Does the fact that the Medical Device tax is an excise tax change your opinion"?

RPLong writes:

Doesn't the subsidy make up for the reduction in supply caused my medical device market over-regulation?

I mean, where do we begin and end with this kind of analysis? The entire medical device and pharmaceutical markets are so tightly regulated, taxed, subsidized, and mandated that perhaps the best thing to do is to take out the distortions, one-by-one. This might leave the markets distorted in an unbalanced way at a given point in time, but if the goal is to let markets do what they do best, what other options do we have? If we wait for quid-pro-quo deregulation, it's never going to happen.

Jim Glass writes:

Is there a "medical device subsidy" I somehow missed along the way?

"the subsidy on medical devices, which is enormous. (Subsidies include Medicare, Medicaid, tax deductibility of insurance, etc.)"

Ah, those aren't subsidies for medical devices, they are subsidies for medical consumption, of all kinds.

Taxing medical devices as a way to offset the massive subsidies for total medical consumption seems to me illogical, probably harmful, and certainly penny-ante. To get over the penny-ante objection we could of course make a list of every product and service consumed for medical purposes and impose a separate tax on each at a different rate, with differing rules and collection mechanisms, etc, by the whims of the politics-of-the-item. The revenue could add up, but then the illogical and harmful points would become a lot more evident.

The very principle of "let's subsidize it and tax it to undo the subsidy" seems like something out of Lewis Carroll come to politics. But if we want to go that route, a much better thing ISTM would be to offset the subsidy to consumption by taxing consumption.

A politically proven and actually used method of doing this for over 30 years is including the subsidy in taxable income -- Social Security benefits were made subject to income tax in the 1983 reform. It's progressive, hitting the richest mostest, was needed to get the system back to fiscal solvency (for a while) and has been happily accepted by the voters, including AARP. Because this tax isn't indexed to inflation it has become quite a significant revenue-raiser for the SSA over 30+ years. (The tax collected is returned to the SSA instead of sent to general revenue -- a carefully disguised method of means testing.)

The obvious parallel is making Medicare benefits subject to income tax. (No, the medical expense deduction would come nowhere near to offsetting this.) In fact, small steps in this direction have already been put in the law. If one wants to tax subsidized consumption, it is time to push them further, Perhaps add Trumpcare benefits too.

It is precedented, proven politically viable as to entitlements -- and probably inevitable as a form of future means testing. And in my mind it is also a heck of a lot better than taxing specific items that medical consumers buy. (Imagine if the 1983 SS Reform instead of making benefits uniformly subject to income tax had instead tried to separately tax item-by-item the things that seniors spend upon. Eech.)


AlanG writes:

Jim Glass writes, "The obvious parallel is making Medicare benefits subject to income tax. (No, the medical expense deduction would come nowhere near to offsetting this.) In fact, small steps in this direction have already been put in the law. If one wants to tax subsidized consumption, it is time to push them further, "

Medicare benefits are indirectly taxed already as the providers, labs, and employees of hospitals all pay income tax. Medicare requires more providers and services because of the population size, hence income tax revenues are likely to be higher than in its absence.

Additionally, as anyone on Medicare well knows, Part B premiums are means tested. There are five tiers of premium payments depending on the Adjusted Gross Income (single of joint filing) and one's monthly premium ranges from $134 to $428 per month. You would be surprised at the number of people who believe Medicare is free to seniors!!

Scott Sumner writes:

Pyroseed, Yes, an excise tax on all forms of health care would be useful.

Brad, No, I assumed an excise tax, which is appropriate.

Thaomas, You said:

"Shifting the demand curve out for an item is not the same as a subsidy for producing it."

Yes it is, if the shift in demand is caused by a subsidy. One of the most basic tenets of public finance is that it makes no difference whether taxes or subsidies are applied to demanders or suppliers.

Everyone, I agree that it would be much, much better to undo the current government subsidies.

Thaomas writes:

Scott,

I assumed that you would understand that I meant not every policy that results in higher output of an item is a "subsidy" to producing it. As Jim Glass said explicitly "Ah, those aren't subsidies for medical devices, they are subsidies for medical consumption, of all kinds."

Scott Sumner writes:

Thaomas, I don't agree, they are subsidies for medical devices.

(It would be like saying "That's not a tax on 87 octane gasoline, it's a tax on all types of gasoline.")

RPLong writes:

Scott, the demand for medical devices looks very different from your graph. The demand curve is far more inelastic and, depending on the type of device we're talking about, may indeed be perfectly inelastic with exactly zero substitutes.

This changes matters in important ways. There are "orphan drug" programs in which the subsidy provided by the government raises the quantity supplied from zero to... anything more than zero. It's difficult to even talk about dead-weight loss in these scenarios.

In short, health care is nothing like 87 versus 85 octane. There is no aggregate supply curve for "all of health care" that can be addressed in meaningful ways by a product-specific tax.

Comments for this entry have been closed
Return to top