Welcome to another blog post in which we attempt to look at serious economic questions through the lens of dubious statistics and anecdotes.
Today's subject: this provocative editorial by Ryan Avent in the New York Times, "One Path to Better Jobs: More Density in Cities," the central thesis of which appears to be that American cities should become denser if the nation wants to create more jobs.
Really? I live in one of the world's densest cities and that conclusion strikes me as suspicious.
Avent writes: "Density isn't a magic elixir. One can't create wealth just by crowding people together; otherwise the super-dense metropolitan areas in emerging Asian countries would be richer than American cities."
Subtext: cities outside the US need not apply to be part of this editorial.
Okay, so he has included a qualifier. However, we at this blog (okay, it's just me) would like to open up Avent's argument for global consideration.
What are the world's densest regions? The United Nations tracks population density by region in its Demographic Yearbook. In 2008 (the most recent year I can find), the global average (population per square kilometer) was 50. North America had a density score of 16. Asia came in at 128, Africa at 33, Latin America at 28, Europe at 32 and Oceania at 4. These data aren't all that useful, mainly because the stats don't measure urban density, but they do suggest that any general discussion of population density shouldn't begin with the second-least dense region of the world.
If we're going to talk about the impact of density on jobs growth, we should take a holistic view of the full spectrum of urban density.
1. Some important world cities and their densities
Originally, I wanted to list the world's densest cities. Unfortunately, the UN's data on city density repeatedly returns a charming Page Not Found error. In fact, after spending over an hour Googling and reading, I've yet to find ANY reliable source that has calculated urban population density the way I want. So I've calculated density myself for a sample of cities from around the world.
Ranked in order of density these cities are:
1. Port Au Prince, Haiti, with a density of 47,169 people/sq. km
2. Dhaka, Bangladesh, with a density of 34,634
3. Phnom Penh, Cambodia, with a density of 33,522
4. Ho Chi Minh City, Vietnam, at 21,541.
5. Paris, France, at 20,246
6. Mumbai, India, at 19,865.
7. Delhi, India at 17,832.
8. Cairo, Egypt, at 14,919.
9. Buenos Aires, Argentina, at 14680.
10. Tokyo, Japan, at 13,671
11. Jakarta, Indonesia, at 11,919.
12. New York City, US, at 10,641
13. Moscow, Russia, at 9,704.
14. San Jose, Costa Rica, at 7,746.
15. Lahore, Pakistan, at 7,711.
16. Sao Paulo, Brazil, at 7,361.
17. Santiago, Chile, at 6,861.
18. Singapore, at 6,816.
19. San Francisco, US, at 6,686.
20. Hong Kong, at 6,320.
21. Mexico City, Mexico, at 5,960.
22. Amsterdam, Netherlands, at 4,500.
23. Brussels, Belgium, at 4,467.
24. Nairobi, Kenya, at 4,236.
25. Stockholm, Sweden, at 4,219.
26. Shanghai, China, at 4,079.
27. London, England, at 3,966.
28. Berlin, Germany, at 3,800.
29. Washington DC, US, at 3,745.
30. Havana, Cuba, at 2,985.
31. Manila, Philippines, at 2,575.
32. Rome, Italy, at 2,073.
33. Cape Town, South Africa, at 1,382.
34. Phoeniz, US, at 1,177.
35. Beijing, China, at 1,167.
36. Sydney, Australia, at 362.
(Here's how I did it: I took population and area data from the UN's Demographic Yearbook where available, otherwise from official city stats. If I couldn't find official city stats, I looked for at least three sources that confirmed the same stat.)
Remember, these are cities meant to represent major urban centres around the world, not the "world's densest cities." The only numbers that don't make a lot of sense are those for Beijing and Shanghai, but the Chinese government, in its most recent census, released provisional data not for urban centres but for entire municipalities, which probably includes areas outside of the heavily crowded urban centres. Or maybe China isn't as crowded as I always thought. I'm not gonna lie, there are elements here that seem a little wacky. Wikipedia, for example, lists the world's densest city as Manila, with a density of 43,079 people/sq. km. That's a LOT more than the density of 2,575 suggested by the UN Yearbook, which means either that Wikipedia is profoundly wrong or that Manila is misrepresenting itself to the UN. (What other option is there?)
2. Economic growth and density: not related?
But ok, we have a first set of numbers about density. Avent makes an interesting claim in his essay, which is that denser cities mean higher economic growth. Economic growth - defined as GDP growth - is an easy parameter to track for pretty much every country on the planet. So let's look at the rate of GDP growth for the above countries. Here they are, ranked in order of fastest-growing to slowest-growing. (The World Bank has a handy table with all this.)
Singapore 14.5
China 10.3
India 9.7
Argentina 9.2
Philippines 7.6
Brazil 7.5
Hong Kong 7.0
Vietnam 6.8
Indonesia 6.1
Bangladesh 5.8
Sweden 5.5
Mexico 5.5
Kenya 5.3
Egypt 5.2
Chile 5.2
Japan 5.1
Pakistan 4.4
Cuba 4.3 in 2008
Russia 4.0
Germany 3.6
Costa Rica 3.5
US 2.9
South Africa 2.8
Belgium 2.2
Netherlands 1.8
France 1.5
Italy 1.3
Australia 1.3 in 2009
England 1.3
Cambodia -1.9
Haiti -5.1
There are really not a lot of surprises here. Looking at these two side by side isn't as helpful as it could be (ideally we'd also look at what percent of the population lives in urban centres, which the World Bank does make available here).
A few takeaways: Port-au-Prince is the world's densest city, but Haiti's economic growth has plummeted over recent years (for the obvious reasons).
Our three fastest-growing economies - Singapore, China and India - appear scattered at random across the density spectrum.
3. Why American cities are not emblematic of a global norm
Before I go loosely linking together more numbers, here's what I think Avent's argument should have taken into account: American cities are a notorious exception to global rules. In terms of safety, they are as dangerous as cities located in warzones. In (highly unscientific) lists of the world's most dangerous urban areas by murder and violent crime rates, American cities often appear in the Top 10. Even if we were to claim that accounting for violent crimes in developing countries and warzones is pretty lax, that doesn't explain why no European cities (where murder rates are tracked as well as they are in the US) EVER appear on any of these lists.
That's not the case, as we can see when we list our representative countries based on the percentage of the urban population living in poverty. (The data aren't really complete, so I've listed the countries that had data available in the order of economic growth, highest to lowest. I've taken the percentage of the urban population living in poverty and divided it by the percentage of the rural population living in poverty.)
Here we sort of have the beginnings of what might be a trend - it seems like countries where the rates of urban and rural poverty more closely converge also contain less dense cities. In other words, in countries that contain denser cities, the city-dwellers do seem to be somewhat better off than their rural counterparts, at least when it comes to the number of each who live in poverty.
(Important caveat: we're talking about the percentage of the urban population below the urban poverty line and the percentage of the rural population below the rural poverty line - so we'll probably get some errors because we're looking at two different poverty lines. Unfortunately, I can't find the actual poverty lines for each of these countries. Of course, if the urban poverty line is significantly higher than the rural poverty line, then the trend in these results would be even more pronounced. We're also not accounting for the fact that the density of our chosen cities might not reflect the average density for all the cities in that particular country - an important "if.")
5. The other things we think should matter, like literacy
Cambodia, 0.34, 78% literate
At this point, if I had the software and time, I would graph economic growth and density together, adjusting for literacy, to see if that demonstrated some sort of pattern.
But here's the interesting thing right up front: density and literacy don't seem to be as connected as density and poverty.
Anyway. This has all ranged a bit far afield. The idea was to try and assess the relationship that Avent discusses - urban density leads to more jobs and economic growth - in a more international context.
And while he may have a point about US cities, and it looks like he may or may not have the beginnings of a point about density in general, I do think it's unwise to talk about a variable like "density" without acknowledging that urban density is a sword that slices both ways. A lack of density can be an economic problem - and countries like Canada and Australia do face economic challenges because of their very sparse populations - but the patchwork of traffic congestion and land grabs that characterize Indian cities certainly doesn't encourage growth either.
In fact, it seems like Avent's qualifier "Density isn't a magic elixir. One can't create wealth just by crowding people together" may be the meat of his argument. Cities shouldn't focus on "density" so much as "smart density," which isn't really all that different from "smart urban design" anyway.
Today's subject: this provocative editorial by Ryan Avent in the New York Times, "One Path to Better Jobs: More Density in Cities," the central thesis of which appears to be that American cities should become denser if the nation wants to create more jobs.
Really? I live in one of the world's densest cities and that conclusion strikes me as suspicious.
Avent writes: "Density isn't a magic elixir. One can't create wealth just by crowding people together; otherwise the super-dense metropolitan areas in emerging Asian countries would be richer than American cities."
Subtext: cities outside the US need not apply to be part of this editorial.
Okay, so he has included a qualifier. However, we at this blog (okay, it's just me) would like to open up Avent's argument for global consideration.
What are the world's densest regions? The United Nations tracks population density by region in its Demographic Yearbook. In 2008 (the most recent year I can find), the global average (population per square kilometer) was 50. North America had a density score of 16. Asia came in at 128, Africa at 33, Latin America at 28, Europe at 32 and Oceania at 4. These data aren't all that useful, mainly because the stats don't measure urban density, but they do suggest that any general discussion of population density shouldn't begin with the second-least dense region of the world.
If we're going to talk about the impact of density on jobs growth, we should take a holistic view of the full spectrum of urban density.
1. Some important world cities and their densities
Originally, I wanted to list the world's densest cities. Unfortunately, the UN's data on city density repeatedly returns a charming Page Not Found error. In fact, after spending over an hour Googling and reading, I've yet to find ANY reliable source that has calculated urban population density the way I want. So I've calculated density myself for a sample of cities from around the world.
Ranked in order of density these cities are:
1. Port Au Prince, Haiti, with a density of 47,169 people/sq. km
2. Dhaka, Bangladesh, with a density of 34,634
3. Phnom Penh, Cambodia, with a density of 33,522
4. Ho Chi Minh City, Vietnam, at 21,541.
5. Paris, France, at 20,246
6. Mumbai, India, at 19,865.
7. Delhi, India at 17,832.
8. Cairo, Egypt, at 14,919.
9. Buenos Aires, Argentina, at 14680.
10. Tokyo, Japan, at 13,671
11. Jakarta, Indonesia, at 11,919.
12. New York City, US, at 10,641
13. Moscow, Russia, at 9,704.
14. San Jose, Costa Rica, at 7,746.
15. Lahore, Pakistan, at 7,711.
16. Sao Paulo, Brazil, at 7,361.
17. Santiago, Chile, at 6,861.
18. Singapore, at 6,816.
19. San Francisco, US, at 6,686.
20. Hong Kong, at 6,320.
21. Mexico City, Mexico, at 5,960.
22. Amsterdam, Netherlands, at 4,500.
23. Brussels, Belgium, at 4,467.
24. Nairobi, Kenya, at 4,236.
25. Stockholm, Sweden, at 4,219.
26. Shanghai, China, at 4,079.
27. London, England, at 3,966.
28. Berlin, Germany, at 3,800.
29. Washington DC, US, at 3,745.
30. Havana, Cuba, at 2,985.
31. Manila, Philippines, at 2,575.
32. Rome, Italy, at 2,073.
33. Cape Town, South Africa, at 1,382.
34. Phoeniz, US, at 1,177.
35. Beijing, China, at 1,167.
36. Sydney, Australia, at 362.
(Here's how I did it: I took population and area data from the UN's Demographic Yearbook where available, otherwise from official city stats. If I couldn't find official city stats, I looked for at least three sources that confirmed the same stat.)
Remember, these are cities meant to represent major urban centres around the world, not the "world's densest cities." The only numbers that don't make a lot of sense are those for Beijing and Shanghai, but the Chinese government, in its most recent census, released provisional data not for urban centres but for entire municipalities, which probably includes areas outside of the heavily crowded urban centres. Or maybe China isn't as crowded as I always thought. I'm not gonna lie, there are elements here that seem a little wacky. Wikipedia, for example, lists the world's densest city as Manila, with a density of 43,079 people/sq. km. That's a LOT more than the density of 2,575 suggested by the UN Yearbook, which means either that Wikipedia is profoundly wrong or that Manila is misrepresenting itself to the UN. (What other option is there?)
2. Economic growth and density: not related?
But ok, we have a first set of numbers about density. Avent makes an interesting claim in his essay, which is that denser cities mean higher economic growth. Economic growth - defined as GDP growth - is an easy parameter to track for pretty much every country on the planet. So let's look at the rate of GDP growth for the above countries. Here they are, ranked in order of fastest-growing to slowest-growing. (The World Bank has a handy table with all this.)
Singapore 14.5
China 10.3
India 9.7
Argentina 9.2
Philippines 7.6
Brazil 7.5
Hong Kong 7.0
Vietnam 6.8
Indonesia 6.1
Bangladesh 5.8
Sweden 5.5
Mexico 5.5
Kenya 5.3
Egypt 5.2
Chile 5.2
Japan 5.1
Pakistan 4.4
Cuba 4.3 in 2008
Russia 4.0
Germany 3.6
Costa Rica 3.5
US 2.9
South Africa 2.8
Belgium 2.2
Netherlands 1.8
France 1.5
Italy 1.3
Australia 1.3 in 2009
England 1.3
Cambodia -1.9
Haiti -5.1
There are really not a lot of surprises here. Looking at these two side by side isn't as helpful as it could be (ideally we'd also look at what percent of the population lives in urban centres, which the World Bank does make available here).
A few takeaways: Port-au-Prince is the world's densest city, but Haiti's economic growth has plummeted over recent years (for the obvious reasons).
Our three fastest-growing economies - Singapore, China and India - appear scattered at random across the density spectrum.
3. Why American cities are not emblematic of a global norm
Before I go loosely linking together more numbers, here's what I think Avent's argument should have taken into account: American cities are a notorious exception to global rules. In terms of safety, they are as dangerous as cities located in warzones. In (highly unscientific) lists of the world's most dangerous urban areas by murder and violent crime rates, American cities often appear in the Top 10. Even if we were to claim that accounting for violent crimes in developing countries and warzones is pretty lax, that doesn't explain why no European cities (where murder rates are tracked as well as they are in the US) EVER appear on any of these lists.
Historically, American cities lacked density because they were miserably managed and horribly governed. Could American cities afford to be more dense? Probably. But that doesn't mean that density, even carefully circumscribed, is really a goal for urban centres to strive towards.
4. Density and poverty - with an important caveat
4. Density and poverty - with an important caveat
Here's another interesting data point. The World Bank also lists the percentage of a country's urban population that lives in poverty. If Avent is right and a dense city should breed more professional opportunity and economic growth, then those living in denser cities should enjoy a more uniformly elevated quality of life.
That's not the case, as we can see when we list our representative countries based on the percentage of the urban population living in poverty. (The data aren't really complete, so I've listed the countries that had data available in the order of economic growth, highest to lowest. I've taken the percentage of the urban population living in poverty and divided it by the percentage of the rural population living in poverty.)
Vietnam 0.17
Indonesia 0.6
Mexico 0.65
Egypt 0.35
Chile 1.20
Pakistan 0.49
Russia 0.35
Costa Rica 0.9
Cambodia 0.34
Indonesia 0.6
Mexico 0.65
Egypt 0.35
Chile 1.20
Pakistan 0.49
Russia 0.35
Costa Rica 0.9
Cambodia 0.34
Ok what matters here is the comparison. As you can tell, there's really not much of a relationship between how fast a country is growing economically and this ratio. In most of these countries, the rural population is much poorer than the urban population, but that's not what we care about. Let's rearrange this list to reflect density. Here are the countries ranked by the density of their chosen city.
Cambodia 0.34
Vietnam 0.17
Egypt 0.35
Indonesia 0.6
Russia 0.35
Costa Rica 0.9
Pakistan 0.49
Chile 1.20
Chile 1.20
Mexico 0.65
Here we sort of have the beginnings of what might be a trend - it seems like countries where the rates of urban and rural poverty more closely converge also contain less dense cities. In other words, in countries that contain denser cities, the city-dwellers do seem to be somewhat better off than their rural counterparts, at least when it comes to the number of each who live in poverty.
(Important caveat: we're talking about the percentage of the urban population below the urban poverty line and the percentage of the rural population below the rural poverty line - so we'll probably get some errors because we're looking at two different poverty lines. Unfortunately, I can't find the actual poverty lines for each of these countries. Of course, if the urban poverty line is significantly higher than the rural poverty line, then the trend in these results would be even more pronounced. We're also not accounting for the fact that the density of our chosen cities might not reflect the average density for all the cities in that particular country - an important "if.")
5. The other things we think should matter, like literacy
Finally, you're going to tell me that I haven't adjusted for those other variables that Avent talks about - literacy, infrastructure, etc - as well as the important variables he doesn't talk about (political strife, etc).
So let's make a final list (this is not the same thing as statistically adjusting for something, I know) of those urban poor countries, in order of density, with their urban-poor ratio and the national literacy rate.
Cambodia, 0.34, 78% literate
Vietnam 0.17, 93% literate
Egypt, 0.35. 66% literate
Indonesia,0.6, 92% literate
Russia, 0.35, 100% literate
Costa Rica, 0.9, 96% literate
Pakistan, 0.49, 56% literate
Chile 1.20, 99% literate
Mexico, 0.65, 93% literate
At this point, if I had the software and time, I would graph economic growth and density together, adjusting for literacy, to see if that demonstrated some sort of pattern.
But here's the interesting thing right up front: density and literacy don't seem to be as connected as density and poverty.
This very fast, very handy little graph shows literacy rate on the X and urban-poverty on the Y. Even if you omit that topmost dot (Mexico) as some kind of crazy outlier, this looks a lot more like a plane than a best-fit line.
6. "Density" vs. "Smart Density"
6. "Density" vs. "Smart Density"
Anyway. This has all ranged a bit far afield. The idea was to try and assess the relationship that Avent discusses - urban density leads to more jobs and economic growth - in a more international context.
And while he may have a point about US cities, and it looks like he may or may not have the beginnings of a point about density in general, I do think it's unwise to talk about a variable like "density" without acknowledging that urban density is a sword that slices both ways. A lack of density can be an economic problem - and countries like Canada and Australia do face economic challenges because of their very sparse populations - but the patchwork of traffic congestion and land grabs that characterize Indian cities certainly doesn't encourage growth either.
In fact, it seems like Avent's qualifier "Density isn't a magic elixir. One can't create wealth just by crowding people together" may be the meat of his argument. Cities shouldn't focus on "density" so much as "smart density," which isn't really all that different from "smart urban design" anyway.
No comments:
Post a Comment