Mr Medlock and the Classics: Part I

Tax positivity!

The basic idea behind social democratic taxation is to ignore the conventional wisdom of solely focusing on efficiency and progressivity - instead, we should aim for broad-based and large taxes. Broad taxes are efficient as they reduce the elasticity of taxable income to the tax rate, allowing a higher revenue-maximising rate. They also mean that the system is not dependent upon the existence of billionaires, avoiding the trilemma of a large welfare state, low middle class taxes and a small share of market income going to the rich not being able to co-exist. The sense of fairness arising from broad-based taxes that everyone pays also means people tolerate higher taxes rates.

More importantly, efficiency is not the be all and end all. By any account, France is a country with a huge amount of taxes and with very high tax rates. They have incredible amounts of regulation and lots of industrial action. And yet, France has a GDP per hour of work that ranks as one of the highest in the world, in line with the USA ad much higher than the UK or Australia etc. If all these cumbersome government interventions really distort markets, it’s not terribly visible. Indeed, research has suggested that the supply of inventions is much more affected by childhood exposure to innovation than changes in taxes. As James Tobin noted, “it takes a heap of Harberger triangles to fill an Okun’s gap”. So we can dispense with the notion that taxes inherently reduce growth and worry less about inefficiency.

Taxation should also be focused on being large rather than necessarily progressive - this is because all you need to reduce inequality is to be less pro-rich than the market, not necessarily to be pro-poor. Consider two people - one has £30,000 in income and the other has £300,000. If we tax at 50% and 30% and redistribute at 40% and 60% respectively, we get net incomes of £57,000 and £273,000. In other words, we see a Simpson paradox-esque situation, where a regressive tax system and regressive transfer system produce a progressive outcome, just by the sheer volume of redistribution. This is reflected in real life - some of the most social democratic countries like the Nordic ones have the least progressive taxes and transfers among developed countries. In fact, progressivity metrics tell us little - a country raising a significant chunk of its taxes from the top 10% could reflect more pre-tax inequality just as it could reflect higher rates of taxation on the wealthy. This is why we should look at the absolute amount of redistribution rather than these flawed measures of progressivity.

Obviously by the principle of MOAR we’ve established, we’re taxing lots of things and taxing lots of people. But what taxes do we want exactly? As I’ve established elsewhere, the first-best option (ignoring Pigouvian and land value taxes) is a tax on consumption - that’s because consumption is the most meaningful measure of inequality in people’s wellbeing. However, what we want in particular is a value-added tax, rather than a sales tax. This means that rather than having a tax paid at the retail sale of the final product, there is a tax levied at every stage of production. Via the credit-invoice method, this means that customer is told the VAT they have to pay, while firm will receive a credit on the VAT already paid on their inputs.

This provides a mechanism for self-enforcement, since there is cross-reporting i.e. the seller knows that the buyer is reporting the transaction in order to get the credit, creating an incentive for them to pay the tax. Another advantage is that it avoids tax pyramiding, which would occur if a sales tax were levied on every sale. That is, taxing sales at every stage of production would punish industries with lots of intermediate steps, while a VAT avoids this by letting firms claim credit on VAT paid on the value-added before their stage of production. As such, there will not be an undue incentive to vertically integrate.

Another lovely tax is the employer-side payroll tax. By being withheld from everyone’s paychecks, it is not that salient, is difficult to avoid and has a broad base. And when coupled with a minimum wage, it means that the incidence of the tax is not passed on to low wage workers due to a binding floor on wages.

The third prominent tax we ought to consider is wealth taxes. We can see the benefits in Norway, where a wealth tax seems not to have had a negative effect on employment. Instead, it has actually incentivised individuals to invest in employment within family-owned businesses to avoid the tax. Furthermore, it increases aggregate productivity by forcing the burden of the tax onto less productive entrepreneurs who don’t efficiently allocate their capital, making it advantageous compared to a capital income tax. To the extent to which this might incentivise a reduction in the capital stock, this seems not to hold up in practice in Norway, where the income effects dominate the substitution effects. But even if it did, that’s why we have a VAT to tax that spending. One example of a wealth tax would be an inheritance tax, which is also helpfully a Pigouvian tax on death. Something about “less dying, more gooder” seems relevant here.

All of this isn’t to reject capital income taxes - although the canonical Chamley-Judd results suggest capital income taxes should be zero, this is based on a very specific set of assumptions that likely don’t hold in real life. It takes a standard Ramsey growth model with constant returns to scale in its production function and it concludes that we shouldn’t tax capital but should tax labour, because capital can accumulate. Of course, this ignores the accumulation of human capital. And if people remember Mankiw, Romer and Weil, human capital is pretty damn important. Coupled with the many other assumptions which may not hold up, we get a world where capital income is very much taxable.

We can also look towards taxing various finance-related things. The most obvious is to tax capital gains on a mark-to-market basis. This involves taxing the capital gains as they accrue annually, rather than at the point of sale - this ensures there is no disincentive to hold on to assets for longer just to avoid the tax. Notice that there is no good reason to privilege capital gains over other possible things to tax, because doing so doesn’t incentivise investment as much as it incentivises the structuring of investments to avoid taxes on income. So that’s an option on the table.

Another would be to tax financial transactions, because there are gains from curbing speculative behaviour that outweigh the reduced level of liquidity. If you believe that the zero-sum nature of jobs in industries like finance are not things we should be encouraging due to their negative externalities, higher income tax rates might even grow the economy by changing the allocation of talent. And the amount we can tax is reasonably high here too, with some estimates going as high as 83% for the top 1% of earners.

Finally, there are also administrative and institutional things we can do. For example, we can reduce the salience of property taxes. Or we can get prepaid taxes which make taxes less burdensome for households and lead to significant cost savings. Or we can actually fund tax collection agencies, which is a policy that basically pays for itself. Or we can create a broader base that lowers the ETI by minimising tax evasion via third party reporting, while expanding the labour supply with good spending choices.