by Paul Collier
In the simple scenario in which all the gains of agglomeration accrue to landowners it is evident that in taxing their gains we do not change their behaviour in any way that retards the city. You may recall that we left them on a beach: once we tax them they may need to work like the rest of us. But even in the other scenarios, taxing the rents is efficient. The corporate lawyer in her bedsit will lose some of that spectacular excess of her income over her rent, but as long as we leave her better off than working in a small town she will keep working in the metropolis. Similarly, in our other scenario, we can tax the smart workers producing services in the metropolis without changing their behaviour as long as we leave them better off than working as farmers.
In terms of tax efficiency, finding economic rents are the equivalent of finding the Holy Grail: revenue without collateral damage. If this sounds too good to be true, brace yourself: it is about to get even better. For this we need another handy economic concept – rent-seeking.
Rent-seeking is a menace; here is an example. Suppose that a legislature passes a law that grants a monopoly to a group of producers. Why did the legislature do such a thing? Because the legislators were lobbied and coaxed with rewards. The regulation generated rents; the lobbying was rent-seeking. The distinguished economist Anne Krueger showed that lobbying, and other rent-seeking, will increase up to the point at which one extra dollar spent on it yields one extra dollar of rent. The resources devoted to rent-seeking are a total waste.
The gains from agglomeration are rents: so, do they attract rent-seeking? Economists have never posed that question and there is a simple reason for their neglect. If the Henry George Theorem is right and the gains accrue only to landowners, then there is no scope for rent-seeking. Land is in fixed supply and so not amenable to lobbying or any other action. But the Henry George Theorem is wrong. In a metropolis, most of the gains of agglomeration accrue to those people with high skills and little need for housing. Suddenly, many opportunities for rent-seeking open up. People elbow their way into jobs by lobbying well-connected relatives; they pay tutors for the extra study that gets them more credentials; they go to hundreds of interviews. Or they squeeze their need for housing by delaying marriage, or delaying children. Each of these is a form of rent-seeking. Behaviour gets distorted in the competition to capture the lucrative rents of agglomeration. The rent-seeking does not increase the overall size of the pie, it just inflicts a collective loss of well-being upon mid-career people scrambling against each other. Potentially, these losses from rent-seeking are massive.
By taxing the gains of agglomeration, we reduce the pressure for rent-seeking. Getting that job in the metropolis would still be worth doing, but since it would be less lucrative, people are less likely to be driven to extreme measures. Delaying having children in order to be able to remain in a pricey London or New York City apartment might become a sacrifice too much. The economic rents of agglomeration in our thriving big cities are currently staggeringly large. Not only is the scramble for them probably inflicting damage on the people who are scrambling, but its sheer momentum may blind people to the irreversible damage they can do to their own lives.
Putting it all together: how can the gains of agglomeration be taxed?
As a general idea, taxing economic rents is now being recognized as wise. The most influential recent advocate is Robert Solow, Nobel Laureate and the founder of the theory of economic growth, who has argued that economic rents have increased, and that taxation should be shifted to them and away from earned income. With this reassurance, I will now bring the two blocks of argument together. Taxing the gains from agglomeration is a smart policy on grounds both of ethics and efficiency. Each of these criteria matters, and there are few other taxes that satisfy them both.
On ethical grounds, the case for taxing the gains of metropolitan agglomeration is unusually powerful. Usually with a tax the best we can hope for is to say that the burden is fairly shared, but in this case, taxing the rents is necessary to better align gains with desert. Similarly, on efficiency grounds, usually the best we can hope for with a tax is that it does little collateral damage. Few enough taxes are able to satisfy even that modest-sounding condition, but taxing the gains of agglomeration may even increase efficiency, by curbing rent-seeking.
The pertinent question is how, in practical terms, can the gains be taxed? Recall that they are spread between city landowners and skilled city workers. Capturing these gains through taxation therefore requires differentially high taxation of these two groups.
A sensible starting point is to capture the appreciation in the value of land and property. This is best done as an annual charge as a percentage of land and property values.* The revenues from such taxes should accrue nationally: they will be needed to finance redistribution to other cities that have been hard hit by the same forces that have benefited the metropolis. Currently, instead of being taxed more heavily than other sources of income, the appreciation in metropolitan land values is taxed more lightly. In many countries, Britain being one of them, it is barely taxed at all. This is a mis-design of the tax system of monumental proportions. In the nineteenth century, politicians agonized over ‘the undeserving poor’. Politicians of the twenty-first century should be agonizing over the legacy of our policy negligence; we now have many thousands of ‘the undeserving rich’. Unfortunately, not a few of them are politicians. The right wants to protect the rich; the left wants to roast them. We need to discriminate among them. Some are hugely useful to society, others have merely captured the fruits of collective effort.
But the crux of our analysis is that much of the rents do not accrue to landowners, they accrue to skilled metropolitan workers. Capturing these rents requires a tax innovation: tax rates need to be differentiated not just by income, as at present, but by the combination of high income and metropolitan location.
Metropolitan workers with only modest skills do not capture any of the rents of agglomeration. A large majority of those with modest skills work in the provinces, and so the pay of a modest-skilled worker in London who makes the morning coffee for the lawyer is going to be set in the provinces, plus the additional amount needed to cover the extra rent payable on a London bedsit over that of a provincial bedsit. So, the basic rate of tax that is levied nationally on those with modest incomes is equally appropriate for those working in the metropolis. But the high-earning corporate lawyer in her bedsit does capture rents of agglomeration that should be shared with others. So, she should pay a higher tax rate than were she working in the provinces, where she would not get the rents. This is not cranky: if she worked in New York City, she would already pay an extra 8 per cent income tax than if she earned the same amount in a smaller city. She pays it by dint of working there, even if she lives outside the city boundaries. If she works in London she doesn’t – but she could. At modest rates of taxation of economic rents, few employment decisions would be changed and so the tax would be much less damaging than current taxes. The challenge, which is entirely feasible to solve with modern techniques of fiscal analysis, would be to work out how high the supplementary tax on high incomes of metropolitan workers should be before the efficiency costs are comparable to current taxes. The difference between what New York does already and this proposal is only on where the tax accrues. In New York City, the revenues from that 8 per cent income tax accrue to New York City; in my proposal, they would accrue to the nation to benefit the recovery of cities like Detroit and Sheffield.
What all this means is that the basic rate of tax, which is the only one paid by most people, would continue to be applied nationally. But each tax rate applicable to higher incomes would carry a metropolitan supplement that would target the rents of agglomeration captured by that skill group. Since the gains of agglomeration are far greater for the most highly skilled, the supplements would be progressively larger at higher levels of income.
Since tax administrations know where people live and work, this is in practical terms surprisingly straigh
tforward: indeed, as in the New York example, many taxes are already geographically differentiated.* The most likely obstacle is the disproportionate political influence of wealthy city-dwellers, not least through being heavily over-represented in legislatures. Despite their high estimate of their moral self-worth, this proposal for an ethically just and economically efficient tax is likely to be greeted with self-righteous outrage. But recall, since we are taxing economic rents, the predictable arguments about disincentives and desert are self-serving: prepare for an avalanche of ‘motivated reasoning’. Taxation is not only analytically warranted: it is a fitting response to the new urban arrogance.
REGENERATING PROVINCIAL CITIES: ‘SHACKLED TO A CORPSE’?
How can cities like Sheffield, Detroit and Stoke be revived? The purpose of taxing the metropolis is not to finance welfare benefits for the inhabitants of these places, but to meet the costs of restoring them as clusters of productive work. As we have seen, the market will not replace a broken cluster with a new one; instead, the city fills up piecemeal with low-productivity activities. But why can’t market forces generate a new cluster, and, if markets can’t do it, why should we think that government can?
A successful cluster is the common location of many different companies, some of which are in competition with each other. Being clustered together enables them to reap common economies of scale and so they all benefit from lower costs. Once a cluster is formed, market forces maintain it: no firm wants to leave because it knows that tomorrow the other firms will still be there, not somewhere else. But forming a new cluster is far more demanding. Precisely because companies are interdependent, a single company will be much more inclined to move to a new location if it expects many others to make the same decision. But how can the company know whether others will do so? If the pioneer goes ahead, yet another firm may join it as the second firm in the cluster, and if that happened yet another firm might decide to become the third. But there is no market mechanism for generating and revealing these decisions. The formation of clusters faces a co-ordination problem and so it needs a co-ordinator. Silicon Valley co-ordinated around Stanford University; what might work in less-favoured places?
Private-sector solutions to co-ordination
The co-ordination problem arises because the decision of each firm depends on each other firm. In economics, these effects are known as externalities; because they affect other firms rather than the firm itself, they are not taken into account in its decisions. But there are market solutions to this interdependence: think local, or think big.
Think local . . .
One sector of the economy has a natural role in the co-ordination of companies: finance. At its best, the financial sector hoovers up information about firms and allocates capital with a view to future opportunities. A bank whose business was legally restricted to a particular city would understand that its own future depended upon the success of the local economy. The bank would itself internalize the effects that were external to each of the firms that it was financing. In order for this not to be suicidal for the bank, it would need to learn a lot about the opportunities and interdependencies, firm-by-firm. It would thus be very different from the financial sector institutions described in Chapter 4. Are such banks a fantasy? On the contrary, prior to a legislative change in 1994, they were the norm in the USA. In Britain, we have to go further back, but names such as The Midland Bank and The Yorkshire Bank are testimony to a localized past, and local banks are still common in Germany. Potentially, the policy change to global banks could have enhanced the finance potential for cities needing a new industry, by giving access to a wider pool of capital. But in practice, global banks have little incentive to invest in local knowledge. When a city starts to contract, their local branches are instructed to reduce credit and the recovered money is moved to other cities. A return to localization would give the financial sector the incentive to perform its socially useful role: generating and judging information about the real economy.
Think big . . .
The need for co-ordination can be surmounted by means of a mega-firm: a firm like Amazon that is so big that it reaps sufficient scale economies of a cluster entirely with its own operations to justify being a pioneer. The firm is a cluster in itself, and its location will suck in an entourage of supporting suppliers. In most industries, being that big is not beautiful: the cluster efficiencies are liable to be offset by the difficulties of managing an elephant. So being big enough to be your own cluster is rare. There are far fewer such firms than there are broken cities whose mayors would like a mega-firm to come to them. The problem of which broken cities succeed in attracting the mega-firms also has a market solution, but not a pretty one. A smart mega-firm seeking a new location will organize an auction in which cities bid against each other for the prize of getting the firm. The value of the prize is the gains from agglomeration that will accrue to the city from the new cluster. New research that compares cities that win these auctions with those that lose confirms that these gains are real.7 Auction theory tells us how much the winning bid will be: it will be equal to the prize.* So, the market ‘solves’ the co-ordination problem facing a broken city by handing all the gains of a new cluster to the mega-firm that creates it. As I write, Amazon is conducting an auction among American cities for a new headquarters location. The company is sufficiently big to revive a broken city, and sufficiently ruthless to extract these benefits for itself.
Public-sector solutions to co-ordination
The government as the co-ordinator of business decisions sends shudders down the spines of market fundamentalists. But I am writing this section in Singapore, and from my desk I have a panoramic view of an outstandingly prosperous city that has been achieved by public planning. When I first visited the city in 1980 it had just raised the minimum wage in order to drive out what the government recognized was a doomed sector – garments. The strategy met with scathing critiques from the market fundamentalists: minimum wages would just create high unemployment. In America and Europe, the government as co-ordinator indeed has an embarrassing history of politically skewed interventions, but East Asia is a valuable corrective: co-ordination can work. Singapore’s founder, Lee Kwan Yew, also understood both the economics and the ethics of agglomeration. His policies reflected this: ‘I saw no reason why private landowners should profit from an increase in land value brought about by economic development and the infrastructure paid for with public funds.’8
Here is an approach that superficially seems the least distorting. If the metropolis is to be subject to supplementary taxation, then why not use the revenues to finance correspondingly reduced taxation of firms in broken cities, thereafter leaving it to the market to determine which firms move where? This, however, does not address the co-ordination problem, and it fails for the same reason that the market works to maintain clusters once they have formed, but not to establish them. Knowledge that firms going to a broken city would pay reduced taxes does not help a pioneer firm know which firms will move, where they will move, or when they will move. Mayors would still have no option beyond bidding for mega-firms. But the mega-firm auction would now have an added twist. Since all broken cities would have this fiscal advantage, they would still have the same incentive to bid against each other to win the auction. As before, the mega-firm would capture a payment equal to the value to the city of the prize, but now it would also get the tax subsidy as a bonus. So, what might work?
Compensate pioneers
Broken cities need to attract firms that are dynamic enough to start a new cluster in their wake. Such pioneering firms are scarce, however, because, unless other firms follow them, they are likely to go bankrupt. Even if other firms do follow, the pioneer will still be at a disadvantage relative to these later entrants. When pioneer firms look for the skilled workers that they need, they are unlikely to find them. How can local workers have such skills when there have been no firms that use them? So, the pioneer will have to bring skilled workers fro
m elsewhere so that they can gradually train its local employees, which is likely to be expensive. But if a second firm decides to set up shop in the same city, it will have an easier time recruiting the skilled workers it needs – it can poach some of the workers that the pioneer has trained. As a result, the second firm’s set-up costs will be lower than the pioneer’s, enabling it to have a higher return on capital.
In other words, pioneers of clusters face what is known as a first mover disadvantage. This is distinctive: more commonly, pioneers enjoy first mover advantage, but this applies to pioneers of new markets, and new technologies. Being the first in a market entrenches the firm ahead of subsequent entrants because it builds brand loyalty – think of Hoover; being the first to a technology allows the firm to patent it – think of Apple. But if the firm is pioneering a new cluster that will sell what it produces in an established market using an established technology, the pioneer bears the costs that later firms avoid.
For a broken city, however, a cluster pioneer is socially valuable. So, what can be done about this? Since the pioneer generates externalities, this public benefit should be compensated with public money. As a principle this is straightforward, but to implement it requires competent specialist public agencies. How is this best administered?
