by Paul Collier
Faced with that proposal, the ideologues will each be salivating. The right will pontificate about the disincentive effects of high taxation, while mumbling against turning the provinces into a giant Benefit Street full of scroungers: ‘shackled to a corpse’. The left may overreach in its enthusiasm to fleece the City, inadvertently triggering an exodus of alarmed companies that unravels the economies of agglomeration.
Both have just enough truth on their side to convince their adherents, but not enough to be right. The truth perceived by the right is that turning provincial cities into Benefit Street cannot be the goal. Well-being depends upon dignity and purpose, not just on how much you can afford to consume. A strategy of supplementing unrewarding jobs with public benefits is not a substitute for creating jobs that require skills that a worker can take pride in having mastered. So, the goal is productive jobs, not public supplements to the earnings from unproductive jobs. The truth perceived by the left is that the strutting affluence of those milking the highly paid metropolitan specialisms is ethically offensive. These people think they earn their incomes; I am about to show that they don’t.
The strategy I propose separates naturally into two halves: taxing the metropolis, and restoring the provincial cities. Each half depends upon a distinct analysis.
TAXATION AND THE METROPOLIS: ‘WE’VE EARNED IT’?
Taxation should be guided by ethics and efficiency. Ethics matter both because of their intrinsic value, and because an unethical tax will meet resistance and evasion. Efficiency matters because taxes drive wedges between prices; for example, the price a consumer pays for a product becomes higher than the amount the producer gets. Such tax wedges distort the allocation of resources and so reduce efficiency.
What ideologies of left and right think they know about taxation has polarized and poisoned our politics. A dose of pragmatism is liberating: smart new taxes could beat existing taxes on both ethical and efficiency criteria.
The ethical rationale for a tax is probably more important for tax design than its efficiency. Tax administration depends critically upon voluntary compliance. The standard philosophical method for the analysis of ethical propositions is practical reasoning. Despite its centrality to tax policy, practical reasoning has not been part of conventional economic methodology. In consequence, economists have largely ignored ethical aspects of taxation. As advisors to ministries of finance, they quite often propose taxes that breach promises that they consider to have been foolish (and quite probably they are right in this judgement). Indeed, economists appear to think that they have addressed ethical issues merely by considering income inequality, which is analysed through the standard Utilitarian calculus.* As Jonathan Haidt has found, for most people fairness means proportionality and desert, rather than equality. Yet they have been ignored.4 Forget desert: if the idle have less money than the hard worker, a transfer raises ‘utility’. Forget entitlement: if someone who has built a pension retires with more money than someone who has spent their life on a beach, a transfer raises ‘utility’. Forget obligation: by now you will get the picture. Utilitarian economists might caution that some transfers might have disincentive effects and so be inefficient, but they would not recognize them as unethical. This blindness to wider ethical considerations is an instance of the larger phenomenon: these people are WEIRD.
Once we accept that issues of desert should loom large in tax design it has powerful implications for the gains from agglomeration. The first person to spot this was Henry George, a nineteenth-century American journalist and political-economist. Once he explained his idea it became a sensation.
Henry George’s big idea
George made an ethical case for the distinctive taxation of the gains from agglomeration. He saw why they were ethically distinctive and concluded that the appropriate policy was to tax the appreciating value of urban land.
You can grasp his insight by posing a sequence of questions. Start with ‘who gets the gains from agglomeration?’ To think it through, here is a stylized version of the industrial revolution. Initially everyone is a farmer. Industry starts up in a new city, and people move to it to work in factories. As the cluster of factories grows, people become more productive than they were in farming: this extra productivity is what is meant by ‘the gains of agglomeration’. The extra productivity is reflected in wages because firms compete with each other for workers. But in order to work in the factories, people have to live near them, and so they need to rent land from whoever owns the land on which the city is forming. So, the gains from moving to the city are the higher wage minus this rent.* As long as this rent is less than the productivity difference between farming and industry, more people will move to the city. But as they do so the rents get bid up. This process keeps on until rents rise to eat up the entire productivity difference. At that point, there is no further incentive to move; in economic jargon, we have reached equilibrium. But, more excitingly, we have reached a powerful punchline that answers our question: all the gains from agglomeration accrue as rent to landlords. For those at the rightward end of the political spectrum who may be growing a little uneasy, let me reassure you that this is not Marxism: George was not a socialist. But he was a smart economist; many years after he had died, two economists proved his conclusion as a theorem. They had the decency to name it ‘The Henry George Theorem’.5
Henry George then went on to pose a second question, incomprehensible within a conventional economic framework: ‘Do the landlords deserve to get these gains?’ Although incomprehensible to economists, this is perfectly comprehensible to everyone else. To answer it we need no theorem: what we need is practical reasoning. To see whether someone deserves an income we trace back to find one of their actions that generated the income which has accrued to them. But when we trace the gains from agglomeration, the actions that generated the gains were done by everyone who works in the city. By working in the city, each person has contributed to the overall increase in the productivity. The gains from agglomeration are generated by interactions between masses of people, and so they are a collective achievement that benefits everyone. This is something that economists refer to as a public good. So, what part have the landowners played in this process? For all that they have done, they might as well have been lying on a beach. Indeed, quite possibly that is how they have spent their time. They have received the income because they owned land where people happened to cluster. Their activity played no part in generating the gains from agglomeration. In the confusing vocabulary of economics, it is classified as an ‘economic rent’.
The important point is that, by reasonable ethical standards, landowners are less deserving of the gain from the appreciation in the value of the land they own than if they had worked for it, or if it reflected a return on the capital they had accumulated from saving. This is not to say they have no claim whatsoever. As the legal owners of the land they have a claim to the gains of agglomeration based on entitlement. But this collides with the collective claim of all the workers in the city to those gains, which is based on desert. Where there is such a clash of reasonable criteria, pragmatism invites us to compromise rather than to retreat onto a pedestal of dogma. And that is where taxation comes in. Suppose that the society agrees on some tax rate for those incomes for which desert and entitlement coincide: the farmer produces output that he both deserves due to his work and to which he is entitled due to his ownership of the farm. Suppose the agreed tax rate is 30 per cent. Then in deciding on a tax rate on income from the appreciation in land values that reflects the gains from agglomeration, we should set that tax rate significantly higher than 30 per cent. This would reflect the fact that the claim of the landowner to this income is significantly weaker than the claim of the farmer to his income. Moreover, only by taxing the gains of agglomeration, and using the revenues to benefit the entire city, can the workforce that has generated the gain receive some share of it – which on the above reasoning they deserve.
Henry George’s idea was an ear
ly application of practical reasoning, resting on the distinction in desert between rent and other forms of income. He was careful to distinguish the rent generated by land appreciation from the income on capital, which he argued was ethically legitimate: his proposition was neither Marxist nor populist.
Were his views eccentric? On the contrary, his ethical common sense resonated: Progress and Poverty became the best-selling American book of the entire nineteenth century.
Unfortunately . . .
Henry George established a powerful ethical case for the heavy taxation of the appreciation in the value of urban land. Despite resonating with the public, his policies were never properly implemented. The people who were making fortunes from the ownership of land in the centre of big cities opposed taxing it. Rather than coming up with countervailing ethical arguments, their approach was to use some of their exploding wealth to buy political influence. In Britain, the man who owned much of central London, the Duke of Westminster, was conveniently sitting in the House of Lords: he became the richest man in the country. In the USA, a man whose core business was New York land deals is currently President.
It is never too late to introduce such a tax. The electorate is far better educated than it was in Henry George’s day and so it should be easier to build a political coalition that overcomes the resistance of vested interests. Further, since the 1980s there has been a surge in metropolitan growth, reflecting the large increase in the gains from agglomeration. Recall that this results from the leap in complexity and its concomitant deepening of skill differentiation. Thus, there are now far larger gains of agglomeration available to be taxed than in Henry George’s day, and so it has become ever more ridiculous that public policies have not done anything about it. Instead, we have been trapped in the gridlock of the old ideology-driven tax disputes.
But the ‘unfortunately’ that heads this section is not a lament about the current deficiencies in public policy. It is that the same rise in complexity that has powered the new metropolitan gains from agglomeration has also invalidated Henry George’s Theorem. His proposition that we can capture these gains through taxing land is no longer correct. The case for taxing the gains remains very powerful, but to do so requires a smart redesign of taxation. The analysis underlying the last two sentences is new: my colleague Tony Venables and I stumbled upon it as a result of working on something seemingly unrelated (which happens surprisingly often with academic discoveries).6 I will try to give you a sense of the excitement of making a new discovery. The ideas can be presented quite simply: indeed, that is how we stumbled upon them. You can reach the frontiers of economic thought on this topic by thinking through two simple scenarios:
Scenario 1: A metropolis in which workers have differing skills and differing needs for housing
The first scenario is a variant of our farmers-and-industry story, except that this time it is people with differing skills and housing needs who each decide whether to move to a metropolis. The high connectivity provided by the metropolis makes skills more productive: the more highly skilled you are, the more that being in the metropolis raises your productivity. But as people move to the city, the rents get bid up as previously. So, who moves and who stays put? Pretty clearly, the people who gain most by moving are single people with very high skills. So, the specialist corporate lawyer who works long hours in the office and spends her free evenings out on the town before returning to her bedsit is hugely more productive than if she worked in a small town, and she will not be spending much of her correspondingly spectacular income on rent. It is often useful in economics to look for the people who are indifferent between two choices, in this case between moving to the metropolis and staying in a small town. We know that for them the gain in productivity will be precisely offset by the extra rent they need to pay, but who will they be? Some will only be semi-skilled: they are single and only need a bedsit, but their earnings are not much higher than in a small town. Others may be highly skilled, but because they have a large family they need a lot of housing and the rent eats up their extra earnings. These people are important for the analysis (in economics they are termed marginal) because they are only just willing to live in the metropolis; if landlords were to charge a higher rent these people would leave and the landlords would be short of tenants. These ‘marginal’ people determine the rents that landlords can charge. That corporate lawyer will be paying the same rent for her bedsit as the semi-skilled singleton who rents the neighbouring one. We have arrived at the punchline: that corporate lawyer has been able to capture some of the gains of agglomeration.
Generalizing, because of the differences in skills and housing needs, many of the gains from agglomeration no longer accrue to landlords, but stick with those highly skilled singles who don’t need much housing. When Tony Venables and I simulated what might happen in a metropolis like London or New York, we found that around half of all the gains from agglomeration end up with such people rather than with landlords. Once we allowed for a further layer of differences, this time among smaller cities, the share captured by landlords fell even further. The key implication is that no matter how heavily the government taxes landlords, it cannot get hold of most of the gains of agglomeration.
This is bad news, because the ethical argument for taxation remains powerful. To see this, I am going to sketch the second scenario.
Scenario 2: A metropolis that needs the rule of law
This scenario has a few more steps towards realism, and it packs a more powerful punchline. There are two products, food and services, and many countries. Food can be produced anywhere, but services can only be produced in those countries that have the rule of law. You can think of this as a proxy for many other aspects of good governance. In turn, the rule of law depends upon ordinary citizens co-operating and working together to support it. If each citizen just sits back and leaves it to others, that is, if everyone free-rides, the public good of rule of law is absent. In this scenario, in most societies people free-ride and the rule of law is rare. In consequence, only the few societies that have the rule of law are able to produce services; in the rest, everyone just produces food.
The gains from agglomeration apply to services but not to food, so in the few societies that have the rule of law there is going to be a metropolis in which these services are produced. Because not many countries are able to make services they are going to sell at a premium over food in world markets, so the service-exporting countries are going to be more prosperous than the food-exporting countries.
Next, we explore who in the service-exporting countries benefits from this prosperity. Suppose that in all countries there are two types of workers: those who are atypically smart, and everyone else. Suppose also that being smart is no help in farming. In contrast, being smart is potentially valuable in producing services, but it depends upon how many smart people are clustered together: an isolated smart worker in services is no more productive than a farmer, but the more that the smart people congregate in the metropolis, the more productive they all become. Finally, we add the usual story about rents: as smart people crowd into the metropolis, the rents go up.
So, who gets the gains from agglomeration, and do they deserve them? As in the previous scenario, the gains are shared between the workers living in the metropolis and the landlords. We could tease out quite what this division will be, but for present purposes it doesn’t matter. The punchline is that in this scenario only one group unambiguously deserves to get them, because only they are responsible for the actions that have been essential for the generation of the gains: namely, the ordinary citizens across the society who collectively sustain the rule of law. But they do not get any of the gains. Some of the gains accrue to the smart workers in the services sector, and the rest accrue to the landlords. Since the group that has an unambiguous ethical claim to a share of the gains gets nothing, there is a strong case for taxation. But, as in the previous scenario, land taxes alone will not hit the gains that accrue to the smart metropolitan
workforce.
These two scenarios have a significant feature in common, that the smart workers who capture the gains of agglomeration sincerely believe that they deserve them. Their belief is anchored on the fact that their earnings are high because their productivity is high. In turn, they believe that their productivity is high because they have developed a highly skilled specialism (scenario 1), or because they are atypically smart (scenario 2). There is indeed enough truth in these propositions that, given how convenient they are to such people, it is understandable that they believe them. But they are not the whole truth. The productivity of the metropolis depends upon public goods that have been provided by the entire nation, such as the rule of law and past investments in the infrastructure for connectivity. These provide some benefits for everyone, but they disproportionately benefit skilled metropolitan workers. More fundamentally, the gains of agglomeration are, by their nature, collectively produced. They are the result of interactions between millions of workers, not merely the outcome of the individual effort of each highly paid worker. The super-skilled deserve to retain a share of their high productivity. But they do not deserve all of it. Nor do they deserve as large a share as someone who is not based in the metropolis and whose productivity is not so augmented by others.
The efficiency case for taxing the gains of agglomeration
So far, I have only considered the ethics of taxing the gains from agglomeration. But there is another aspect of taxation that excites economists: efficiency. Economists are right to be excited about it, and on the taxation of the gains from agglomeration the profession at last has some valuable insights to offer.
The key insight is the concept of economic rents. They are any payments that accrue to someone for doing something that exceed what would have induced them to do it. On our previous criterion of ethics, the concept is irrelevant. Just because a star tennis player would be willing to play for less than the tournament prize money she wins does not delegitimize her keeping it. The star player earns economic rents on her exceptional talent, but since that talent is hers, so is the income arising from it. But when we switch from ethics to efficiency the concept of economic rents becomes really useful. By definition, taxing the rent does not affect the decision to work, and so the revenues do not come at the cost of inefficiency. The gains of agglomeration are economic rents: on the criterion of efficiency, they are the ideal target to tax.
