Back-of-the-envelope low-quality musings: #takes

Longform stuff I've put time and thought into is at #effortposts.

To hear from academics and policymakers, #interviews.

Compiled scribbles on what I read can be found in #notes.

Topic-related tags include #economics, #coronavirus and #miscellaneous.

# All that Glitters is Not Gold: It's Consumption

29 Jul 2020

I recently saw an Instagram post about billionaires - why they were inherently problematic, why they didn’t deserve it and just a litany of complaints about their wealth. And what it was gesturing at was the broad agenda we increasingly see of taxing income, corporations and wealth much more aggressively - which made we wonder, why do we worry so much about income or wealth instead of consumption?

So I decided to make a poll on Instagram, asking people if they would rather:

• Earn $100,000 a year, but only be able to spend$25,000 a year
• Earn $100,000 a year and have$1 million, but only be able to spend $25,000 a year • Earn$80,000 a year and be able to spend $80,000 a year To be clear, this assumes that you are sure the arrangement would occur in perpetuity, and that the$75,000 a year in the first case as well as the extra $1 million in the second case can never be used or given to others. What would you pick? People in my low-effort, low-quality poll overwhelmingly chose the third choice, and those who didn’t choose it after having the options explained - that’s because when we consider our living standards, we are interested in our consumption, rather than income or wealth. And as I will go on to explain, this has important implications for how we think about taxing the rich. What do we want from our tax system? When designing the tax code, a helpful starting point is to raise revenue with the minimum distortion to economic activity and growth. That means minimising the marginal rate, because the amount of tax you pay per dollar at the margin determines how hard you decide to work for that money. That means having a broad tax base without the excessive number of exemptions and deductions that exist in tax codes under the status quo. And that means taxing consumption. Why not personal income tax? To begin with, let’s consider Alice and Bob - they both earn$100,000 a year - Alice spends it all on wine, while Bob spends it all on beer. Because they are free to buy either drink, it seems unfair to tax one above the other.

Now suppose Alice spends it all on wine while Bob spends $60,000 on wine and invests$40,000 on a bond that doubles in 40 years. This allows Bob to buy $80,000 worth of wine 40 years later. Again, they had the same choice of spending now or saving and investing the money - and crucially, the present value of their lifetime consumption is equal. Because stuff in 40 years costs half as much (since you can double the amount of money you have), we cannot simply add up the spending in the two time periods as$60,000 + $80,000 =$140,000. Instead, we need to halve the future consumption if we want it in present value, which gives us a total of $100,000, which is the same as Alice’s. This is the problem with including capital income when considering inequality - it is a form of double counting. Suppose we did tax all personal income at 50% - Alice would be left with$50,000 to spend on wine today, while Bob would spend $20,000 now. The remaining$30,000 would double in 40 years, and with half of those capital gains being taxed, he is left with $45,000. That means in present terms, Alice is left with$50,000 while Bob is only left with \$42,500 - in other words, Alice only faces a 50% tax rate while Bob faces a 57.5% tax rate. So taxing capital gains is problematic, because it double-taxes and leads to an unequal tax treatment between two people who had the same choices - just as taxing beer over wine is unfair, this punishes the person who is thriftier over the person who spends more. By contrast, a payroll tax that only taxed labour income or a consumption tax would keep the equal treatment of Alice and Bob - with a consumption tax having a less distortionary effect on disincentivising work.

Why not corporate income tax?

Mitt Romney got a lot of flack in the 2012 election when he said that “corporations are people” - while that was undoubtedly a bad political choice, it is a useful way of thinking about what a corporation tax does. When corporations are taxed, they can react in three ways - by raising prices, lowering wages or lowering returns to their shareholders. The important question is what shifts the most and who is least able to avoid the tax. It seems plausible to believe that given a fairly mobile capital market, shareholders are most able to buy other assets - that is, their demand of shares is incredibly elastic and responsive to changes in returns. By contrast, workers are less able to change jobs when wages fall and consumers less able to adjust to buying other goods as prices increase - this means that they will be the ones bearing the burden of the tax.

But even if the incidence were on the shareholders, as soon as a corporate income tax is announced, the share price drops in order to maintain the same after-tax risk-adjusted rate of return as the rest of the market - that is to say, it only hurts existing shareholders at the point of the tax being announced, with no impact on future shareholders. Hence the Chamley-Judd conclusion that the optimal corporate income tax rate is 0.

So the corporate income tax is a very convoluted way of implementing a tax for which the incidence is at best on an arbitrary section of shareholders, but more likely on ordinary consumers and workers. That doesn’t even begin to recognise the fact that a significant proportion of stocks are owned by institutional investors - that is, people’s pension funds and such, making it an incredibly subpar way of targeting the rich.

What are the advantages of the consumption tax?

We’ve seen the issues with personal income and corporate income taxes - but there are other reasons to prefer a consumption tax. For one, it provides a disincentive against consumption, which encourages people to save. As I explain in Part I of my growth primer, savings and investments are crucial in providing the capital formation needed for economic growth. And this means that we no longer need the incredibly excessive number of exemptions in tax codes for various savings accounts, simplifying the system.

Secondly, there is no fundamental reason to tax and redistribute income or wealth. As evidenced by my incredibly non-rigorous poll, the most meaningful measure of well-being and inequality is by the things you can get with your money i.e. your consumption - and the best proxy for consumption is consumption, not income or wealth. People who are wealthy but don’t spend it are either donating or investing that money - that is a good thing.

Thirdly, a progressive consumption tax doesn’t hugely impact people’s welfare - not only for reasons of diminishing marginal utility as is true of all progressive taxes, but also because people’s psychological happiness is much more correlated with relative consumption at the top end.

Fourthly, to the extent to which we use fiscal policy in recessions, a temporary cut in consumption taxes is much more likely to be a useful stimulus, because it is only advantageous to spend more in that period.

Why not a wealth tax?

The mirror of the second benefit of a consumption tax is why a wealth tax similarly makes little sense - to the extent to which we are worried about rich people being able to afford extravagant lifestyles of marshal their resources in problematic ways, we are worried about consumption. Wealth by definition is sitting in investment funds or in businesses - that is, being spent to grow businesses. And it is incredibly unclear to me why wealth is the indicator we should therefore be worrying about.

What happens when we do impose a wealth tax? In the same way the income taxes caused problematic incentives, a wealth tax incentivises people not to save, but instead to spend. Why invest your money to grow businesses when that’s going to get taxed? Go on a cruise instead! And to the extent to which some people have far too much money to spend on private jets or lobbying politicians, why invest in risky but potentially high return businesses? A progressive wealth tax encourages investing in safer but lower return assets like blue chip stocks and government bonds, rather than the startups of the world. This is not to mention the immense complexities of tax evasion that a wealth tax incentivises - whether that is rich people owning intermediaries that own various assets or design shares in a convoluted way, all designed to make enforcement impossible.

How would a progressive consumption tax work?

There’s a lot of concern that a flat consumption tax like a VAT would be regressive - because the rich save a lot more and spend a smaller proportion of their income. The problem is that progressivity depends on what you measure it against - when lifetime consumption measured against lifetime earnings, flat consumption taxes are actually proportional. This is unsurprising, since income has to be consumed now or saved/invested such that it can be consumed later.

Nonetheless, I believe in the arbitrary nature of a birth lottery and the diminishing utility of money, so let’s make tax-and-transfer progressive. The easiest way to do so is to impose a flat consumption tax and then redistribute it towards those who are poorer - this makes the overall system progressive. If that isn’t enough, there are three ways to make the consumption tax itself progressive.

The first is simply by altering the VAT on different goods depending on who buys these goods - that is, the tax on yachts would be greater than on hamburgers. This is not a terribly effective way of targeting the rich. The second is by having households report their incomes and savings - they would be taxed based on consumption equalling income (excluding capital income) minus savings. Then it would simply be a matter of imposing a progressive tax schedule on consumption.

The third is to force everyone to pay the maximum VAT rate at the point of purchase, and as people turn in their receipts, they get a refund of some of the VAT paid - by altering that refund as the total amount of consumption increases, we have a progressive consumption tax. This mirrors the experience of tourists who buy things and turn in receipts at the airport in order to get a rebate on some of the VAT - and of course, this would be done electronically rather than sending paper receipts, which allows it to occur immediately rather than at the end of the year.

Because it is a refund, there is an incentive to declare all of one’s consumption, whereas the first method keeps the existing incentive to hide one’s income and make labour income appear as capital income. However, the practical difficulties of setting up an electronic system and making sure this doesn’t create cashflow problems for those who are less well off are significant. But nonetheless, it does seem that there are options for creating a tractable and progressive consumption tax.

Coda

Neither personal nor corporate income taxes make much sense. And you probably shouldn’t have one but not the other, because it creates a huge incentive for arbitrage - that is, you create an incentive for individuals to incorporate or vice versa. Wealth taxes don’t make much sense either, because they are taxes on saved income that produce growth. Consumption is the most meaningful yardstick for people’s quality of life and inequality - a progressive consumption tax not only creates an incentive to save and invest, it is also the best way to target inequality between people’s well-being levels. (Of course, there are valid critiques that well-being is about more than consumption. More specifically, wealth provides some form of insurance against bad things happening. So there is some reason to look at wealth - I just don’t think the reason I hear most people give has anything to do with this.)

Notice I haven’t touched upon the size of government revenue or the degree of redistribution - that’s because we can fiddle with the specific rates to meet those criterion. Although even a revenue-maximising consumption tax may not be large enough for the “eat the rich” crowd, I am not proposing abolishing all other taxes. But I am simply pointing out that there is a lot of revenue and redistribution available from first-best options like Pigouvian taxes and taxes on economic rents, as well as second-best options such as the consumption tax. And so we should be looking there, before we look at the more popular remedies of wealth taxes and income taxes.