Jonah Lehrer, “A physicist solves the city“, NY Times. This article is a few years old but still thought-provoking. It discusses a physicist’s analysis of the data on cities. The key finding is that resource consumption increases more slowly than city size, while economic activity and human interaction increases more rapidly. While I think the title overstates things – the work identifies a set of strong correlations but not a causal mechanism – it’s still important:

The correspondence was obvious to [Geoffrey] West: he saw the metropolis as a sprawling organism, similarly defined by its infrastructure. (The boulevard was like a blood vessel, the back alley a capillary.) This implied that the real purpose of cities, and the reason cities keep on growing, is their ability to create massive economies of scale, just as big animals do. After analyzing the first sets of city data — the physicists began with infrastructure and consumption statistics — they concluded that cities looked a lot like elephants. In city after city, the indicators of urban “metabolism,” like the number of gas stations or the total surface area of roads, showed that when a city doubles in size, it requires an increase in resources of only 85 percent

What Bettencourt and West failed to appreciate, at least at first, was that the value of modern cities has little to do with energy efficiency. As West puts it, “Nobody moves to New York to save money on their gas bill.” Why, then, do we put up with the indignities of the city? Why do we accept the failing schools and overpriced apartments, the bedbugs and the traffic?

In essence, they arrive at the sensible conclusion that cities are valuable because they facilitate human interactions, as people crammed into a few square miles exchange ideas and start collaborations. “If you ask people why they move to the city, they always give the same reasons,” West says. “They’ve come to get a job or follow their friends or to be at the center of a scene. That’s why we pay the high rent. Cities are all about the people, not the infrastructure.”

…The challenge for Bettencourt and West was finding a way to quantify urban interactions. As usual, they began with reams of statistics. The first data set they analyzed was on the economic productivity of American cities, and it quickly became clear that their working hypothesis — like elephants, cities become more efficient as they get bigger — was profoundly incomplete. According to the data, whenever a city doubles in size, every measure of economic activity, from construction spending to the amount of bank deposits, increases by approximately 15 percent per capita. It doesn’t matter how big the city is; the law remains the same. “This remarkable equation is why people move to the big city,” West says. “Because you can take the same person, and if you just move them to a city that’s twice as big, then all of a sudden they’ll do 15 percent more of everything that we can measure.”

Relatedly, here’s a map that Oxford University researcher Max Roser put together of land use changes in Europe over the last century. I noticed two key trends: first, there are a lot more cities (the red areas), and second, the forests have become denser. In other words, if you want to protect nature, stay out of it!

Over in the US, FiveThirtyEight’s Gary Rivlin takes a look at “Why the plan to shrink New Orleans failed“. It’s a good account of the difficulty of rebuilding or rethinking a city after a major disaster, especially when there are big demographic and racial divides:

“The government is going to spend, what, $100 billion or more to rebuild New Orleans — and for what?” Sharky asked a few months after Katrina. “If we don’t do things differently, it can happen again next year.”1

Money was very much on Canizaro’s mind in early 2006. He asked me to imagine the cost of providing police, fire protection, garbage pickup, and other city services to half-built neighborhoods in low-lying New Orleans. Every expert was telling him the same thing: The city couldn’t afford to let every last neighborhood come back. Fewer people meant less tax revenue and therefore less money to spend on everything from police and fire to street repairs.

Yet by that point, Canizaro was admitting that it was inevitable that he would duck the tough call and not prohibit low-lying neighborhoods from rebuilding. He was a white man leading the redevelopment of a city that was two-thirds black. Of the 353,000 people living in a part of the city that flooded, 265,000 were black.

“Unfortunately, a lot of poor African-Americans had everything they own destroyed here,” he said. “So we have to be careful about dictating what’s going to happen, especially me as a white man.”

One other thing that this brought home to me was that the US has a stupidly large amount of money to spend on infrastructure. New Orleans has a slightly smaller population than greater Christchurch, and yet various levels of government spent around US$100 billion (~NZ$160bn at today’s exchange rates) rebuilding it. By comparison, the Treasury’s latest estimate is that the NZ government will spend a total of NZ$7.6bn rebuilding infrastructure in Christchurch. The total cost of the rebuild, including insurance payouts to homeowners, is likely to be more like $40bn.

On a different note, Emily Washington at Market Urbanism takes a look at the scientific validity – or lack thereof – of common traffic engineering standards. She describes them as a case of “Engineering in the dark“:

In The High Cost of Free ParkingDonald Shoup explains the origin of municipal parking requirements. Municipal planning offices do not have the resources to study the amount of parking that businesses should provide. Even with more staff, it’s not clear that planners would be able to determine optimal parking requirements unless they allowed business owners themselves to experiment and choose the amount of parking on their own in a learning process of how to best serve their customers.

The Institute of Transportation Engineers is one of the only organizations that provides estimates of the number car trips that businesses generate. Given the lack of information planners have to determine parking requirements, they often rely on ITE’s information to set their parking requirements. However, ITE studies are often conducted at businesses that already provide ample free parking, ignoring the potential for businesses to manage demand for parking on their property through prices. Furthermore, ITE estimates of trip generation are typically based on a very small sample of locations, which are unlikely to be representative of businesses and cities in general.

In the example below, the ITE provides a recommendation for fast food parking requirements based on their floor area. Even though the chart includes a line of best fit for the plot of peak parking spot occupation and floor area, the ITE hasn’t demonstrated a correlation between these two variables. Shoup points out:

We cannot say much about how floor area affects either vehicle trips or parking demand at a fast food restaurant, because the 95-percent confidence interval around the floor-area coefficient includes zero . . . This is not to say that vehicle trips and parking demand are unrelated to a restaurant’s size, because common sense suggests some correlation. Nevertheless, factors other than the floor area explain most of the variation in vehicle trips and peak parking occupancy at these restaurants.

Parking R2

Even though this study fails to show much of anything about parking at fast food restaurants, the ITE parking requirement recommendations carry heavy weight with city planners. Shoup finds that “the median parking requirement for fast food restaurants in the US is 10 spaces per 1,000 square feet–almost identical to the ITE’s reported parking generation rate of 9.95 spaces per 1,000 square feet. After all, planners expect minimum parking requirements to meet the peak demand for free parking, and parking generation rates predict this demand precisely. When the ITE speaks, urban planners listen.”

A linear trendline through 18 non-randomly sampled data points with an R2 of 0.038 tells you precisely bugger all about how much parking a new restaurant is likely to need. Anybody who thinks this is a good basis for regulation ought to be sent to Statistics Re-Education Camp.

However, it turns out that regulators aren’t the only ones coming up with absurd requirements based on no data whatsoever. Reddit came up with many, many examples of crazy HOA (Home Owners’ Association) rules. HOA rules, which influence building design and maintenance in many American suburbs, are enforced through private covenants, but that doesn’t prevent them from being costly and unreasonable:

I worked in property management for a while and at one time I had a portfolio with 30 HOA’s. And guess what I go to do? …..Write rule violation notices for all of them, based on what the HOA board members told me to send notices for.

Sometimes I cringed while putting them in envelopes, feeling so bad for fining some sweet old lady $50 for planting purple flowers when purple is NOT on the approved color list, and demanding she dig up all her hard work.

Or fining someone $75 for their garbage can being “too visible from the street.”

Or demanding that someone plant a new tree in their park strip, then fining them and making them plant a NEW new tree because it wasn’t “between 60 and 72 inches.” It was 58. The board president measured it.

Fun fact: many states allow HOAs to foreclose on your house if you refuse to pay fines or membership fees!

I have a bunch, but the one that sticks out is when we got fined for having a 2×4 in our backyard. Our FENCED backyard. In a place you couldn’t see from public property.

I know someone where his HOA prevents you from taking the trash out the day before pick up. Like if taking your trash out on sunday night for trash day monday was worthy of a $50 or $100 fine. 12:01AM Monday is fine, but 11:59PM sunday wasn’t

I was once made to resod my front lawn. In the middle of summer with average daily temperatures over 100 degrees. During one of the worst droughts on record. While the whole city was under watering restrictions.

The new lawn (which I had spent several hundred dollars on) promptly died and they tried to make me replace it again, but apparently enough people had complained by that point that before I did they agreed not to make us replace our lawns until the water restrictions were lifted.

Confession bear time: I had to get up early for work (3am), so as I drove through the neighborhood I looked for people watering their lawns in the middle of the night on violation of restrictions. Most people were just trying to avoid being hassled by the HOA, I know, so I left them alone. But when I saw members of the HOA board doing it, I reported them to the city.

I suspect that, in the long run, HOA covenants may be worse for growth than traditional zoning rules. This is because zoning can be amended at the city-wide level to accommodate growth somewhere. Growing cities can undertake re-planning exercises where they identify where there are opportunities to develop more intensively. Under covenants, though, there is no mechanism for balancing out urban growth requirements. If any individual HOA doesn’t want apartment blocks, it can simply ban them, without having to consider the trade-offs of doing so.

On a completely different note, the Economist has taken a look at how UK’s regional governments are trying to “win back control of their bus networks”. The system put in place under Thatcher, which privatised both operations and service planning, has apparently been failing many regions quite badly. (New Zealand is a bit ahead of the curve in making similar much-needed changes – which I previously discussed here.)

Outside booming London, bus passenger journeys have fallen by 37% over the past three decades. Critics believe that deregulation has played a part in the decline: in 1986 Margaret Thatcher privatised the then publicly run bus networks outside the capital. Several commercial bus companies have come to dominate parts of England and Wales, and their fares have increased by at least 35% more than inflation between 1995 and 2013. “There can be few business sectors where profits continue to rise while customer numbers fall so significantly,” says Nick Forbes, leader of Newcastle city council.

So he, and others, are trying to re-regulate their regional networks. The aim is not renationalisation but taking control of the franchising of routes run by commercial operators. Recent commitments by the government to regional devolution have given the moves momentum. The bus companies are resisting strongly…

The aim, says Mr Forbes, is for Nexus to regulate the services and enable seamless switching between buses and other modes of transport, as with London’s electronic Oyster card. Under the proposed new regime, bus companies would bid for contracts to operate routes, rather than just operating the ones they like. Nexus would collect fares, pay the operators and invest some of the money (which the companies currently keep as profit) back into the system to boost subsidies. It says operators would still make a healthy margin.

Also on the topic of buses, it turns out that embracing low-cost policies enabling better bus systems and safe cycling can be an economic winner. Alex Davies, “The world could save trillions with buses and bikes“, Wired:

The argument that embracing a low-carbon future is a road map to economic ruin is bunk, say a band of economists who argue that investing in more efficient transportation, buildings and waste management could save cities worldwide at least $17 trillion. One way to unlock that savings is to promote bikes and buses.

The savings come from stimulating economic activity, decreasing health care costs, reducing poverty, and cutting the costs associated with urban sprawl, like time and productivity lost to traffic congestion. That’s according to a report, Accelerating Low‐Carbon Development in the World’s Cities, released today by New Climate Economy, a group of economists formed to examine the costs and benefits of addressing climate change.

“For too long, there’s been the same old argument used to prevent bold action on climate change, which is there’s some sort of tradeoff between economic prosperity and climate action,” says Nick Godfrey, an author of the report and the organization’s head of policy and urban development. “In cities, that is a false choice. Actually, there is a significant confluence between promoting economic growth and prosperity, and climate action.”

On a different note, some clever map-makers have come up with a great illustration of how Mercator projections can mislead us about the size of places in the extreme north or south. (More info here.) It turns out that Greenland is not the world-bestriding colossus that it is made out to be:

And to finish up, here are two good pieces discussing economic geography in the present day and the distant past.

In the New York Times, Adam Davidson takes a look at economic life as it was almost 4000 years ago: “The V.C.s of B.C.“:

In general, we know few details about economic life before roughly 1000 A.D. But during one 30-year period — between 1890 and 1860 B.C. — for one community in the town of Kanesh, we know a great deal. Through a series of incredibly unlikely events, archaeologists have uncovered the comprehensive written archive of a few hundred traders who left their hometown Assur, in what is now Iraq, to set up importing businesses in Kanesh, which sat roughly at the center of present-day Turkey and functioned as the hub of a massive global trading system that stretched from Central Asia to Europe. Kanesh’s traders sent letters back and forth with their business partners, carefully written on clay tablets and stored at home in special vaults. Tens of thousands of these records remain. One economist recently told me that he would love to have as much candid information about businesses today as we have about the dealings — and in particular, about the trading practices — of this 4,000-year-old community. […]

In 1962 A.D., as our modern era of globalization was just beginning, the economist Jan Tinbergen — who would later share the first Nobel in economic science — noted something curious: Trade within and between countries followed a mathematical formula. He called it the Gravity Model, sort of an E=mc2 for global business. It comes with an imposing formula: Fij = G(Mi x Mj)/Dij. Which, simplified, means that trade between two markets will equal the size of the two markets multiplied together and then divided by their distance. (The model gets its name from its mathematical similarity to the equation in physics that describes gravitational pull.) […]

Economists were drawn to the Kanesh archive because it offered an unprecedented chance to see how well the Gravity Model applied in an economy entirely unlike our own. This was trade conducted via donkey, through a land of independent city-states whose legal and cultural systems were totally dissimilar to any we know. But still, the model held up: Ali Hortacsu, a University of Chicago economist on the Kanesh team, says that the trade figures between Assur and Kanesh matched the formula almost perfectly. ‘‘It was a very nice surprise,’’ he told me.

Over at Medium, Matthew Romaine explains how proximity still matters for the success or failure of startups today: “Tokyo vs the Bay Area“:

Tokyo’s density allows for 2–3x the number of meetings per day.

In Tokyo, you can have a string of 1-hour meetings with 30min. between each. Start with a power-breakfast at 8AM, your next meeting at 9:30AM, then 11:00AM, 12:30PM, 2PM, 3:30PM, 5PM, 6:30PM, 8PM (dinner meeting). That’s 9 meetings. Yes it’s possible — I’ve done it before. I don’t recommend repeating this daily, but if you’re fundraising or hustling in sales, at least it’s possible. In the Bay Area there’s a lot of driving, even when many of the VCs are on Sandhill Road. When we were fundraising our seed round back in 2010, we maxed out at 4 meetings / day. You could probably do 5–6 meetings in San Francisco city proper, but you’ll likely be frazzled and sweating towards the end. More recently thanks to Uber you don’t have the stress of finding parking every time, but the traffic can still suck and be unpredictable. Tokyo’s impeccably reliable by-the-minute public transportation means you can pack-in a string of in-person meetings like no other. And nothing replaces f2f meetings when large sums of money are involved.

Share this


  1. I was amused by the comment comparing meetings per day in Tokyo and the Bay Area.

    My colleagues, who are based in the Bay Area, have to attend meetings with various media companies in L.A. They complain how inefficient it is compared to the Bay Area, and that only two customer meetings per day is a realistic schedule. It is guaranteed somebody will be at least half an hour late due to “traffic”.

    There is probably a lesson in that for Auckland!

  2. “New Orleans has a slightly smaller population than greater Christchurch” – um, not really; the metropolitan area is more than twice as large as greater Chch (and was larger before Katrina). More than 800,000 people were displaced there by the hurricane.

    I think that US Govt expenditure of $100+bn is also for the damage incurred from the hurricane across the entire affected region – including at least four different states. That includes some fairly major investment to repair the levees, flood walls, etc – about US$15bn alone.

    1. Fair point – I was writing in a hurry and didn’t check on metro area population. That said, the article notes that repairing the levees around New Orleans proper cost US$14.5bn. That’s at least twice the total government expenditure on repairing/replacing infrastructure in Christchurch.

  3. I can’t see what is wrong with the fast food graph. It shows the actual results they have, fits a line and notes that it isn’t very good as a model. What is wrong with that? Are you suggesting that someone has actually used1.95 per 1000feet plus 20? Who? They appear to have used an overall mean, so given a lack of data WTF does anyone expect them to do? Not use the mean?

    1. I think what is expected is for them not to regulate minimum car parking spaces based on floor area, as it is not very predictive of demand.

      1. Ok so what other independent variable do you base it on if the politicians set a minimum rule? It needs to be based on something that can’t be changed without a consent – any suggestions?

      2. “What’s wrong with that” = it’s absolutely indefensible scientific practice. Indefensible. The sorts of work that would (should) see a student fail statistics 101.

        And your question is stated imprecisely: Minimum parking requirements do not use the mean. They (typically) use the *conditional mean* for parking demand with respect to floor area. In this context, there is absolutely zero evidence to support the use of a conditional mean, i.e. we have no evidence to support the hypothesis that parking demands are related to floor area. Zero, zilch, nada. Continuing to use such a model, in the face of this evidence, rests not on scientific evidence (verifiable facts) but normative judgements (personal preferences).

        Given the lack of scientific evidence for minimum parking requirements, both in this context and many others, does it not beg the question as to whether there might be other methods for predicting parking demand which are more accurate than minimums? After all that is the point of the whole flaming exercise don’t you think?!? Imagine (if you will) a hypothetical alternative situation where developers provided as much parking as they thought was necessary. Would this lead to a higher R-squared? That is, would they use methods which had higher predictive power? I think it’s highly likely.

        In which case should we not prefer the developers’ method of prediction to minimum parking requirements? The fact that the engineering profession has not even researched this idea, i.e. whether developers can accurately predict their own travel demands, shows just how much the profession has its head in the sand on this issue. Personally, I suspect the traffic engineering profession’s blind confidence that their methods are always best, when confronted with evidence to the contrary, is a key reason why the profession is losing relevance.

        And *even if* it’s not more accurate, the implementation of policies (such as minimum parking requirements) should be expected to do better than the “do nothing”, i.e. outperform the don’t regulate situation. If there’s no evidence of a link between parking demands and floor area, then I find it hard to imagine of a situation where no regulation would have less predictive power! In which case we probably shouldn’t regulate unless there is clear evidence of externalities arising from the “do nothing”. And there’s not clear evidence, in fact there’s no economic evidence to support the contention that minimums reduce externalities. I encourage you to research the issue!

        Given what I know about what you think about these issues, I suspect I’m unlikely to convince you otherwise in this forum. However, to adapt fine words of Joe Lycett, minimum parking requirements “doth butter no parsnips” (Source:

        It’s also quite a funny parking related video which you might enjoy!

        1. Stu FFS! They are not promoting a model with a low R-squared they are presenting what little data they have and giving a warning about using it in the absence of anything else. They have fitted a regression line to the points they have because that is the first thing someone will do. Then they have reported it has a low R-squared. That is the correct thing to do. You can’t seem to understand there is a difference between reporting facts and telling people what you think is best for them. They have done the first and you are accusing them of the second. I hope in your current studies you go back and revise the difference between positive economics and normative economics. The graph above is not normative it just presents the data they have available with a warning. As for failing stats 101 don’t be a twit! A good stats question would be fit a line to this data, report the R-quared and comment on whether you think it is a good predictive model. They have done that P=1.95X +20, R-squared =0.038 (which means no correlation at all) and no they don’t think it is a good model to use. So nobody actually uses it. What is unscientific about finding your results inconclusive of anything?

          You are confusing your personal views on whether minimums are a worthy idea with some simple incomplete facts stated above.

  4. Tree density comes with a cost: greater fire hazard. This has been a major talking point in casual conversations on large US bush fires in the last couple of years, although I imagine in Europe it’s not as much of a concern. Still, something important to consider if one wants to generalise to say Australia or New Zealand.

  5. The Mercator projection certainly exaggerates it but……Greenland is one hell of a size; in fact the largest island in the world. (Oz is a continent, not an island).
    And as for parking minimums – why don’t we just get Joyce to decide? It’s how everything else is done.

  6. You can’t stress too strongly that talking about a ‘requirement’ for parking is intellectually incoherent if you don’t also discuss and justify your assumptions and scenarios about the price.

    If spaghetti was free, the ‘requirement’ for spaghetti would be – something. That imaginary figure is of no use whatever to the people involved in satisfying the real-world demand for spaghetti.

Leave a Reply