The idea of a government-funded basic income or social
dividend has been around for at least a couple of centuries. It has been
supported by some prominent advocates of individual liberty as well as by
collectivists. For example, it was proposed as an alternative to existing
welfare systems by Milton Friedman in the 1960s (as a negative income tax) and by
Charles Murray (as an unconditional basic income for all adults) in In Our Hands, published in 2006. More
recently Elon Musk among others, has suggested a government-provided
unconditional and universal basic income (UBI) as a solution to the hypothetical
problem of ensuring that people have adequate incomes when their jobs are
displaced by automation.
That problem is hypothetical because it seems reasonable to
expect - at a national level and over the longer term - that jobs displaced by
automation will be replaced by more highly paid jobs. That is what happened with
jobs displaced by mechanisation during the 19th and 20th
centuries. No persuasive evidence has emerged to support the view that the
effects of automation will differ in that respect. Nevertheless, UBIs might appear
to be an attractive social/political insurance policy, just in case automation
does result in widespread loss of income-earning opportunities.
The idea that one day most of the population will depend on UBIs
as their main source of income strikes me as inherently unappealing. Historically,
individual human flourishing has been closely related to the self-respect that
comes from earning a living, which is absent when people are able to live on “sit-down
money” – an appropirate term used by some Australian aborigines to describe
welfare benefits. Robert Colvile has
provided references to research relating to disincentive impacts of UBIs in a recent FEE article.
I want to focus here on a question of practicability: Is
there some easy way for a government raise sufficient additional revenue to
fund a UBI to reinforce expectations that the benefits of future economic
growth will be widely shared? How could substantial additional revenue be
raised without stifling the economic growth process? As I contemplated those
questions the thought crossed my mind that if I was back working in the
Australian public service (heaven forbid!) and was asked to recommend a way to
raise more tax revenue, I might suggest more reliance on taxes on the unimproved value of land, as proposed in Australia's Henry report, and as suggested much earlier by Henry George in Progress
and Poverty (first published in 1879). Land taxes get a fair amount of
support among economists, including some who write for The Economist.
At some point it occurred to me that I should actually read Progress and Poverty – or at least, the 2006 version, edited and abridged by Bob Drake – rather than rely on second hand
reports. As I read about Henry George’s theory of wages and interest it became
clearer to me why he was viewed as a crack-pot by some of the people who taught
me economics. For example, by rearranging the identity, Production = Rent +
Wages + Interest, he concludes: “wages and interest do not depend on what
labour and capital produce – they depend on what is left after rent is taken
out”. Of course, if you rearrange the terms another way, rent would appear as
the residual after payment of wages and interest. Modern economists should not
be overly critical, however, because George wrote Progress and Poverty before John Bates Clark had made his contribution
to the marginal productivity theory of distribution - and Clark apparently attributed his conception of the marginal
productivity of labour to George’s theory of rent.
Henry George provides an interesting discussion of the way
site rent rises with economic development. He asks readers to imagine a vast
unbounded savanna. Every acre seems as good as any other for the first family to
arrive, so they make a home somewhere, anywhere. When other families arrive,
one location is clearly better than the others, that is close to the family
that has already settled. Having a neighbour provides opportunities for the
families to help each other. As more people arrive, a village is established to
enable people to obtain advantages from local specialization and trade. As the
village grows into a town and then into a city, the productivity of the
original land increases. As a consequence: “Rent – which measures the
difference between this increased productivity and that of the least productive
land in use – has increased accordingly”. The original owners of the land
become rich “not from anything they have done, but from the increase in
population”.
George recognised that advances in technology, improvements
in manners and morals and government policy reforms (e.g. free trade) also
increase the productivity of land, and increase rents.
Following David Ricardo and John Stuart Mill, George argued
that a tax on rent would fall wholly on land owners. He went further, however,
in suggesting that all rent could be taxed away for the benefit of society
without ill-effect. He suggested that returns to labour would thereby be
enhanced:
“When all rent is taken by taxation for the needs of the community,
equality will be attained. No citizen will have an advantage over any other,
except through personal industry, skill, and intelligence. People will gain
what they fairly earn. Only then, and not until then, will labor get its full
reward, and capital its natural return”.
Henry George was correct to argue that, from an economic
efficiency perspective, rent taxes are superior to most other taxes because
they have a smaller impact on productive effort and investment. However, it is
hard to see how a large increase in land taxes could be viewed as providing an
equitable sharing of tax burden. Consider two people who have equal wealth, the
wealth of A is in entirely in land and the wealth of the B is entirely in
shares in companies that do not own land. Would you view it to be equitable for
a government to introduce a tax that would take away a large slice of the
wealth of A, while leaving the wealth of B unaffected?
Perhaps that inequity could be overcome by announcing that
the new land tax will only apply to future increases in land values. However,
the deadweight costs of a tax on future increases in land values would not be
negligible. For example, consider a firm that is planning to build a very fast
train and considering whether a stopping point along the route should be at City
X or City Y. The firm is buying land along the route because it needs to
capture some of the expected appreciation in land values to make its investment
worthwhile. The firm’s investment appraisal suggests that City X would be the
best location. However, it subsequently learns that City X is contemplating a
substantial tax on future increases in land values, while City Y has no such
plans. That information obviously has potential to tip the balance in favour of
City Y, resulting in a less efficient allocation of investment.
The potential deadweight costs of land taxes have been explored
in more depth by others, including Bryan Caplan and Zachary Gochenour.
My bottom line: Land taxes are better than many existing
taxes (much better than taxes on land transfers) but they don’t offer a
costless way to fund the substantial additional revenue that would be required
to fund an unconditional basic income sufficient to meet reasonable
expectations of a widely-shared dividend from future economic growth. If land
taxes can’t do it, I doubt whether any tax-transfer proposal can achieve that
objective. One way or another, even when robots do most of the work currently
done by humans, humans will still need to earn the bulk of the incomes they
live on - including by inventing and improving robots, servicing and managing
them, and owning them.