Monday, 4 March 2013

Tax in a Time of Robots


Ashok Rao has a post arguing for a wealth tax, pointing to the diminishing share of income that goes to labour and questioning the financing of government in a future where capital income is likely to be an even larger proportion of GDP than it is today.

I remember Adam Ozimek (I think?) saying that the left's policies would make a lot more sense if today's world was a lot more like the capital-dominated world we might end up in in the future – the world bloggers are talking about when they talk about ‘robots’. (A good list of links to this discussion is here) A wealth tax seems to make quite a bit more sense in this future, where income taxes aren't reliable and most people's wealth is inherited (because fewer people are earning wages and saving, relative to the size of the capital stock).

However, I don't think Ashok does enough to address the argument from economic theory that there are a lot of strong reasons to oppose a large wealth tax. Similarly to the points Evan Soltas raised in his post:
  • Capital taxes distort intertemporal consumption preferences. I don't think this is as big a problem as is often made out - how much more can a superrich person really consume if tax rates go up? - but it's definitely a concern.
  • Is it moral to disproportionately hurt savers? Yes, big savers are mostly the uberrich – but should we be encouraging the rich to spend all their money on yachts and Batcaves, rather than encouraging them to finance investment, or endow charities?
  • Capital taxes foster avoidance. It's really hard to collect all that money. People will stash their assets abroad, they'll hide cash, and at the very least you'll create a lot of jobs for tax lawyers. I'm pretty sure the literature suggests that the revenue-maximising top rate of capital taxation is a lot lower than for income taxes.
  • Not only will it be very unpopular, but even if you did succeed, there'd be all sorts of exemptions and caveats. Some of these would be desirable. Many would not be. All would increase complexity and implementation costs.
  • Capital taxes have other weird effects. They can distort the debt-equity tradeoff. They can bias rich people against inflation if the tax is not indexed to inflation, and if the experience of the last five years has taught us anything, it’s that rich people being biased against inflation is a very bad thing. How will the tax treat declining asset values? What will the effects of the tax be on the relative demand for different risk classes of assets?

I think we need to think about the fundamental problem, which is that ‘wages are falling as a share of income, therefore taxes are harder to collect,’ in a different way. If people save more, then that will mean a bigger allocation of the economic pie to capital, and less to labour. But that isn’t necessarily a bad thing.

The main problematic implication of this would be that it would become harder for governments to collect taxes on labour, because more of total income is interest on past labour, and less is the (taxable) current labour – i.e. wages.

But it’s a fundamental point that a wage tax raises the exact same amount of revenue as the equivalent consumption tax.

Imagine two doctors, A and B. Both of them earn £100000.

A pays a 50% wage tax. The government takes £50000 from him when he earns the money. He spends all of the remaining money straight away.

B pays a 100% consumption tax. He spends all of the money straight away. The government takes £50000 when he spends the money.

The two taxes are equivalent in this case.

Now imagine the case in which both A and B choose to invest all of their money at a risk-adjusted interest rate of 100% per decade. (Yes, that’s implausibly high, to make the maths easier).

Usually this sort of scenario is used to show that, because the net present values of £100000 and £200000 in ten years is the same, capital taxes are distortionary and hurt savers. Scott Sumner did so here.

I’m going to look at it a little bit differently, however.

A gives his £50000 to the government straight away, and saves the other half of his wages. He earns a return of 100% over the next decade and, ten years later, buys £100000 worth of goods.

B invests his whole £100000, and gets £200000 back in ten years. He spends it all, pays £100000 to the government in consumption taxes, and gets his £100000 worth of goods.
In Scenario A, the government gets £50000 in the first period. In Scenario B, the government gets £100000 after ten years.  The crucial point to remember here is that, because the risk-adjusted real interest rate is 100% per decade, the two are functionally equivalent. This point should also hold across time horizons.

If the state taxes wages, and the wage share of income is falling, it has too much income now, and not enough income later. But this is essentially a problem of consumption smoothing – and if anyone can do consumption smoothing, the government can.

So the government has two choices.

Either it can tax away wages now, invest that money, and later use it to pay the bills. This is what Miles Kimball has been plugging in his calls for a sovereign wealth fund and in our twitter discussion about capital taxes.

Or it can implement a consumption tax that allows people to save the money now, and the government will tax it away when they later spend the money. The USA would implement a VAT. European countries which already have one would raise theirs. A lot has been said about the various economic benefits of (progressive) consumption taxation, but this is an interesting angle on the idea that I haven’t seen discussed before.

I think consumption taxes generally trump the sovereign wealth fund idea (although they’re not necessarily incompatible – you could do a mixture of both). There are a few reasons for this.

Sovereign wealth funds may not be managed effectively. This is partly a problem of size. Norway’s sovereign wealth fund owns 1% of the global stock market. If the USA had a comparable fund, it would be enormous. Would it be able to find good places to invest that money? Even if there were good investment opportunities, it might be difficult for state-managed corporations to find them. There would be a powerful temptation to divert money to popular causes, such as ‘green’ industries or flagging auto companies.

Sovereign wealth funds would be difficult to accumulate. Government saving now to pay for spending later is extremely politically unpalatable. It would require huge government surpluses, perhaps of 10% of GDP or more, to finance future spending. Politicians – not to mention voters – would be quite aggrieved about this, when we have major problems today that we might be attending to. There would be the temptation to divert money to current spending at the expense of future spending, which is already a major problem in politics generally.

Consumption taxes are better than wage taxes in some other ways. Because you don’t have to have a distinction between wage and capital income – income is not taxed at all – there isn’t the incentive to reclassify wage income as capital income. This gets rid of a whole can of worms. Goodbye, carried-interest loophole.

Instead of saving money itself, the government tells people to take their money and invest it wherever they want, and spend it whenever they want – just with the knowledge that when they do spend that money the tax man will take his cut. This kind of implicit government saving solves quite a lot of the political problems associated with, for example, a sovereign wealth fund.

Moreover, it would appease Ashok and people like him by imposing a de facto wealth tax. People’s current wealth, accumulated under the old income tax system, would have incurred the old high rates of wage taxation when it was earned. When they spent it, it would also incur the new high rates of consumption taxation. So the rich get soaked.

But, crucially, they don’t get soaked in as distortionary a way as they would with a full-on wealth tax. Because the effective wealth tax is levied as part of tax reform, there is a credible commitment not to go back and impose another wealth levy. You can’t rebalance your system in favour of consumption taxes twice. It wouldn’t require assessment of people’s wealth, and it would be impossible to avoid. Further, because the wealth tax isn’t explicit, it would be easier to sell politically. And news of an impending rise in consumption taxation would stimulate spending now, hopefully jumpstarting the economy.

Ultimately, the government is going to need enough money to finance its operations and do some redistribution as well – especially in a future where inequality is potentially worse than it is now. I think wealth taxes are clearly unfeasible, and there are numerous practical objections to a sovereign wealth fund. A shift to greater consumption taxation would be more economically efficient – and it would also solve this problem.

No comments:

Post a Comment