orgtheory.net

where’s the capital?

I am a little late to this, but last week a bit of a debate erupted about whether it is better or worse to split-up the political and commercial capitals of countries. It was sparked by an off-hand concluding remark by Luigi Zingales in an article for the City Journal on whether New York is likely to retain its title as the world’s financial capital:

The Founding Fathers wisely decided that the nation’s political capital should be separate from its financial capital (in both senses of the word). Now this splendid segregation has ended. 

Greg Mankiw picked up on that quote and spread it around the interwebs.

Then Matt Yglesias weighed in, countering with the boring vs. exciting capital hypothesis:

One model, seen in France and the UK, is of a single dominant city. Another model, seen in Italy, is where your capital is also your largest city (Rome), but the main financial and business center is elsewhere (Milan). Then you have your scenarios, seen in the US and Canada, where a capital is established someplace a bit random specifically to avoid choosing between major cities. This tends to lead to capital cities with a reputation as ‘boring.’

No one was taking this too seriously.  But they raise some interesting questions about the relationship between government and economic growth.  Zingales and Mankiw’s idea is that separating government from the financial sector is a good thing.  They don’t really expand on why, but I can guess at their reasoning: they think that a financial industry that is too close to government is likely to be over-regulated and hamstrung.  Yglesias’s idea is simply that stand alone government capitals may be boring (Springfield, Ottawa) or beautiful (Edinburgh, Rome), but whether the capital stands apart shouldn’t matter for much when it comes to actual growth.  I’d offer two contrasting alternative hypotheses for why co-locating government and industry might actually matter:

  1. The Full of Themselves Hypothesis:  This one has to do with status hierarchies.  Where the government capital is also in the largest city, your status as a government type competes with other status hierarchies (industry, media, education, etc).  But in places where the government is the most important game in town, there is really only one status hierarchy.  My idea is that legislators are more full of themselves in places where government is the big dog in town.  The prediction, then, is that that would be worse for economic performance as legislators go off willy-nilly coming up with big, but ill-conceived, ideas.
  2. The Social Capture Hypothesis:    This one has to do with networks.  Where the government capital is also in the largest city, you’d expect bureaucrats and legislators to hobnob fairly regularly with elites from industry and other sectors.  This should tend to make their decision-making more elitist (and in the case of finance, more sympathetic to the interests of bankers).  Where government stands alone, government should be (a) less beholden to elite interests and/or (b) more influenced by a wider spectrum of special interest group lobbying.

Thanks to the joys of Wikipedia, it’s fairly simple to pull down a few quick and dirty statistics.  Nothing definitive, but enough to whet the appetite…

First, some descriptive stats are in order:  What is the breakdown of countries and/or states where the largest city is also the capital?

All of the stats I’m going to quote are for 89 countries with more than 6 million residents and, for comparison sake, the 18 US states and 16 Indian states with more than 6 million residents.  I should also mention that I am using data on the metropolitan area of cities, not cities-proper (so, e.g., New Delhi is within the metro area of Delhi, not a separate city).  The data on cities are from the World Gazetteer for consistency-sake.  Everything else comes from Wikipedia.  And the numbers are:

Capital is Largest City

Countries (n=89)         70%

US States (n=18)          33%

Indian States (n=16)    44%

So, in general, most national capitals are located in the largest city in the country, but this isn’t true for sub-national states.  This raises a few other questions on what determines whether a government capital city is located in the largest city.  For instance, one might assume that places where one city really dominates the population of the country, it would be more likely to find the capital in that city.  To do this right, you’d want to look at the size of various cities in a country when the capital was established.  But it’s too much work to do that right now, so I’m just looking at current correlations.  Surprisingly, the correlation between the size of a country or state’s largest city and the rest of its population is negative among countries and Indian states, and practically zero among large US states.

Average Size of the Largest City as a Percent of Total Population
Capital is Largest Capital is not Largest
Countries 15.71% 18.82%
US States 43.04% 41.17%
Indian States 4.93% 17.70%

.

I would have expected it to be the opposite, but one probable explanation for this is the fact that the larger the country or state, the more likely it is that the capital is not in the largest city.  There are probably multiple reasons for that, but one of them is simply because the larger the country is, the smaller a percentage any one city can take up (e.g., Shanghai is a city of 21 million people, but that still accounts for just than 1% of China’s population while Athens, at 3 million, takes up 34% of Greece’s population).  Another explanation, of course, is that the city that becomes the capital can capture a lot of economic development and therefore becomes the largest over time.

Also somewhat surprising is the question of whether places with two relatively closely matched cities end up putting the capital somewhere in between.  My assumption (and Matt’s) was that countries/states where two cities are relatively well matched (think LA and San Francisco, Pittsburgh and Philadelphia) will have remote capitals (Sacramento and Harrisburg).  But the data actually show the opposite: the more dominant the largest city is, the more likely it is that the capital is located somewhere else, except in US states where, again, it’s a wash (true for California and Pennsylvania, but not true for Tennessee, which is evenly matched between Nashville and Memphis, and for New York which is decidedly mismatched between New York and Buffalo).

Average Ratio of the Largest City to the Second Largest City
Capital is Largest Capital is not Largest
Countries 1.83 3.32
US States 5.85 5.54
Indian States 1.94 5.11

.

My guess is that this is due to at least two factors: (1) a number of countries have consciously moved their capitals away from the major cities in order to both relieve overcrowding and provide a nicer lifestyle for government bureaucrats and (2) populist or dictatorial regimes sometimes move their capitals away from the major city intentionally as a means of maintaining control by buffering themselves from elites (see hypothesis 2).

Ok.  But the big question is whether there is any correlation between the economic success of a place and the status of its capital.  I looked at the annual growth of per capita GDP for countries and US states between 1990 and 2007, and for Indian States between 1993 and 2006.  The results are mixed.

Average Annual
Per Capita Growth of GDP
Capital is Largest Capital is not Largest
Countries 5.74% 6.92%
US States 2.02% 1.83%
Indian States 4.5% 4.6%

.

In general, countries where the capital is not located in the largest city seem to be growing faster than countries where the capital is in the largest city.  However, this is where the comparison within countries is important: holding constant the overall growth rate of the country, the results are a wash (or if anything, slightly tilted toward the alternative hypotheses in the case of US States).  This tells me that something else is going on…  One would want to do a more careful statistical analysis (easily done but beyond what I’m willing to put into it right now) to really test the hypotheses.  But I thought the preliminary results are interesting enough to warrant weighing in.

UPDATE: Here are the data, FYI.

Region Country US State Indian State Political Largest Metro as Percent of Pop Largest City Pop Second City Second City Pop Region Population Largest Metro as Percent of Pop Ratio of Largest to Second Largest Large=Pol Regional Growth
Australia 1 Canberra Sydney 4,399,722 Melbourne 3,892,419 21,848,000 20% 1.130331 0 5.31%
Bolivia 1 Sucre La Paz 2,268,683 Santa Cruz 1,614,618 9,863,000 23% 1.40509 0 6.01%
Brazil 1 Brazillia Sao Paulo 20,534,112 Rio De Jenaro 12,058,824 191,519,000 11% 1.702829 0 4.58%
Canada 1 Ottawa Toronto 5,113,149 Montreal 3,635,571 33,716,000 15% 1.406423 0 4.35%
China 1 Beijing Shanghai 17,783,119 Beijing 12,230,029 1,331,860,000 1% 1.454054 0 12.37%
Cote d’Ivoire 1 Yamoussoukro Abidjan 5,878,609 21,075,000 28% 0 3.30%
Ecuador 1 Quito Guayaquil 2,248,800 Quito 1,621,817 14,014,947 16% 1.386593 0 8.41%
Germany 1 Berlin Rein-Ruhr 11,831,297 Berlin 4,055,548 82,062,200 14% 2.917312 0 3.72%
India 1 Delhi Mumbai 21,347,412 Delhi 18,639,762 1,166,430,000 2% 1.145262 0 5.71%
Israel 1 Jerusalem Tel Aviv 3,340,850 7,424,400 45% 0 6.35%
Malaysia 1 Putrajaya Kuala Lampur 7,635,858 Pinang 1,856,370 27,468,000 28% 4.113328 0 8.63%
Morocco 1 Rabat Casablanca 3,269,962 Rabat 1,787,307 31,491,578 10% 1.829547 0 5.65%
Myanmar 1 Naypyidaw Rangoon 4,994,082 50,020,000 10% 0 7.78%
Netherlands 1 split Amsterdam 1,443,258 Rotterdam 1,186,818 16,517,532 9% 1.216074 0 5.20%
Nigeria 1 Abuja Lagos 13,134,029 Ibadan 5,210,392 154,729,000 8% 2.520737 0 9.86%
Pakistan 1 Islamabad Karachi 12,827,927 Lahore 6,936,563 166,953,500 8% 1.84932 0 6.37%
Poland 1 Warsaw Katowice 2,710,397 Warsaw 2,660,406 38,130,300 7% 1.018791 0 11.64%
South Africa 1 Pretoria Johannesburg 7,864,994 Cape Town 4,899,329 48,697,000 16% 1.605321 0 5.60%
Sri Lanka 1 Sri Jayawardenapura Kotte Kolamba 2,618,658 20,238,000 13% 0 8.40%
Switzerland 1 Berne Zurich 1,110,478 Geneva 812,000 7,725,200 14% 1.367584 0 2.89%
Syria 1 Dimashq Halab 2,925,478 Dimashq 2,759,230 21,906,000 13% 1.060252 0 7.40%
Tanzania 1 Dodoma Dar es Salaam 3,092,430 43,739,000 7% 0 7.02%
Turkey 1 Ankara Istanbul 13,778,373 Ankara 3,791,642 71,517,100 19% 3.63388 0 5.48%
UAE 1 Abu Dhabi Dubai 2,353,215 Abu Dhabi 945,268 4,599,000 51% 2.489469 0 10.74%
US 1 Washington New York City 23,141,893 Los Angeles 17,881,561 306,924,000 8% 1.294176 0 4.17%
Vietnam 1 Hanoi Ho Chi Minh 5,318,392 Hanoi 2,717,562 88,069,000 6% 1.957045 0 15.15%
Bihar 1 Patna Panta 2,569,775 Gaya 499,874 82,998,509 3% 5.140845 0 4.4%
Chhattisgarh 1 Thiruvananthapuram Kochi 1,541,175 Visakhapatnam 1,047,560 31,841,374 5% 1.471205 0 7.8%
Gujarat 1 Gandhinagar Ahmedabad 5,334,314 Surat 3,196,799 50,671,017 11% 1.668642 0 5.1%
Jharkhand 1 Ranchi Jamshedpur 1,252,815 Dhanbad 1,210,235 26,945,829 5% 1.035183 0 5.3%
Kerala 1 Dispur Guwahati 1,038,071 26,655,528 4% 0 4.5%
Madhya Pradesh 1 Bhopal Indore 2,049,193 Bhopal 1,751,766 60,348,023 3% 1.169787 0 4.1%
Orissa 1 Chandigarh Bhubaneswar 1,666,429 Raurkela 550,668 24,358,999 7% 3.026195 0 2.2%
Punjab 1 Raipur Ludhiana 1,107,012 Amritsar 1,206,918 20,833,803 5% 0.917222 0 4.0%
Uttar Pradesh 1 Lucknow Kanpur 3,494,275 Lucknow 2,991,280 190,891,000 2% 1.168154 0 3.7%
California 1 Sacramento Los Angeles 15,250,000 San Francisco 4,274,531 36,756,666 41% 3.567643 0 1.68%
Florida 1 Talahassee Miami 5,414,772 Tampa 2,733,761 18,328,340 30% 1.980704 0 1.87%
Illinois 1 Springfield Chicago 9,569,624 Peoria 371,206 12,901,563 74% 25.77982 0 1.86%
Michigan 1 Lansing Detroit 4,425,110 Grand Rapids 776,742 10,003,422 44% 5.697014 0 1.60%
Missouri 1 Jefferson City St. Louis 2,816,710 Kansas City 2,002,047 5,911,605 48% 1.406915 0 1.57%
New York 1 Albany New York City 19,006,798 Buffalo 1,124,309 19,490,297 98% 16.90532 0 2.04%
North Carolina 1 Raleigh Charlotte 1,701,799 Raleigh 1,088,765 9,222,414 18% 1.563054 0 1.99%
Ohio 1 Columbus Cleveland 2,088,291 Cincinatti 2,009,632 11,485,910 18% 1.039141 0 1.72%
Pennsylvania 1 Harrisburg Philadelphia 5,838,471 Pittsburgh 2,351,192 12,448,279 47% 2.483196 0 1.91%
Texas 1 Austin Dallas 6,300,006 Houston 5,728,143 24,326,974 26% 1.099834 0 2.02%
Virginia 1 Richmond Virginia Beach 1,658,292 Richmond 1,225,626 7,769,089 21% 1.353016 0 1.90%
Washington 1 Olympia Seattle 3,344,813 Spokane 456,175 6,549,224 51% 7.332302 0 1.80%
Afghanistan 1 Kabul Kabul 2,413,032 28,150,000 9% 1 5.74%
Algeria 1 Algiers Algiers 6,488,795 34,895,000 19% 1 4.58%
Angola 1 Luanda Luanda 4,577,500 18,498,000 25% 1 10.03%
Argentina 1 Buenos Aires Buenos Aires 14,393,015 Córdoba 1,819,522 40,135,000 36% 7.910328 1 3.70%
Austria 1 Viena Viena 2,179,769 Linz 532,995 8,356,707 26% 4.089661 1 4.44%
Austria 1 Vienna Vienna 2,123,829 8,356,707 25% 1 4.92%
Azerbaijan 1 Baku Baku 2,255,268 8,629,900 26% 1 9.66%
Bangladesh 1 Dhaka Dhaka 13,778,179 Chāţţagām 3,761,337 162,221,000 8% 3.663107 1 4.83%
Belarus 1 Minsk Minsk 1,758,453 9,671,900 18% 1 5.23%
Belgium 1 Brussels Brussels 1,800,663 Antwerp 915,258 10,741,000 17% 1.967383 1 4.55%
Cameroon 1 Yaoundé Yaoundé 2,184,528 19,522,000 11% 1 3.20%
Chad 1 N’Djaména N’Djaména 1,554,578 11,206,000 14% 1 9.72%
Chile 1 Santiago Santiago 5,145,599 16,938,000 30% 1 9.79%
Colombia 1 Bogota Bogota 8,286,535 Medellin 3,337,722 45,006,200 18% 2.482692 1 7.70%
Congo 1 Kinshasa Kinshasa 10,076,099 Lubumbashi 1,713,852 66,020,000 15% 5.879212 1 0.16%
Cuba 1 Habana Habana 2,601,335 11,204,000 23% 1 3.14%
Egypt 1 Cairo Cairo 16,254,102 al-Iskandarīyah 4,545,218 76,855,037 21% 3.576089 1 7.45%
Ethiopia 1 Addis Ababa Addis Ababa 3,230,771 79,221,000 4% 1 2.55%
France 1 Paris Paris 11,089,124 Lyon 1,717,300 65,073,482 17% 6.457302 1 3.77%
Ghana 1 Accra Accra 4,092,613 Kumasi 1,852,449 23,837,000 17% 2.209299 1 5.38%
Greece 1 Athens Athens 3,853,451 11,262,500 34% 1 7.33%
Guatemala 1 Guatemala Guatemala 3,481,976 14,027,000 25% 1 9.80%
Guinea 1 Conakry Conakry 1,931,184 10,069,000 19% 1 2.42%
Haiti 1 Port-au-Prince Port-au-Prince 1,753,767 10,033,000 17% 1 4.87%
Honduras 1 Tegucigalpa Tegucigalpa 2,187,434 7,466,000 29% 1 7.44%
Hungary 1 Budapest Budapest 2,573,664 10,029,900 26% 1 8.10%
Indonesia 1 Jakarta Jakarta 18,924,470 Bandung 6,492,753 230,512,000 8% 2.914707 1 6.08%
Iran 1 Tehran Tehran 12,949,621 Esfahan 3,075,665 70,495,782 18% 4.210348 1 7.10%
Iraq 1 Bagdad Bagdad 11,206,300 Basra 4,094,523 30,747,000 36% 2.7369 1 8.59%
Italy 1 Rome Rome 3,457,690 Milan 3,076,643 60,090,400 6% 1.123852 1 3.45%
Japan 1 Tokyo Tokyo 37,468,203 Osaka 17,411,857 127,580,000 29% 2.151879 1 2.00%
Kenya 1 Nairobi Nairobi 4,112,322 39,802,000 10% 1 5.96%
Madagascar 1 Antananarivo Antananarivo 1,740,322 19,625,000 9% 1 5.31%
Mali 1 Bamako Bamako 2,209,225 13,010,000 17% 1 6.07%
Mexico 1 Mexico City Mexico City 23,293,783 Guadalajara 4,205,153 107,550,697 22% 5.539343 1 5.97%
Mozambique 1 Maputo Maputo 2,058,054 22,894,000 9% 1 5.69%
Nepal 1 Katmandu Katmandu 1,687,102 29,331,000 6% 1 6.43%
North Korea 1 Pyŏngyang Pyŏngyang 3,198,937 24,051,706 13% 1 0.02%
Peru 1 Lima Lima 7,902,851 29,165,000 27% 1 8.00%
Phillipines 1 Manila Manila 19,888,419 92,226,600 22% 1 7.18%
Portugal 1 Lisboa Lisboa 2,638,211 10,631,800 25% 1 6.60%
Romania 1 Bucuresti Bucuresti 2,177,669 21,496,700 10% 1 8.79%
Russia 1 Moscow Moscow 14,837,510 St. Petersburg 4,798,526 141,868,000 10% 3.092097 1 5.18%
Saudi Arabia 1 Riyadh Riyadh 5,150,000 Jeddah 3,600,000 25,721,000 20% 1.430556 1 4.54%
Senegal 1 Dakar Dakar 2,535,431 12,534,000 20% 1 3.56%
Serbia 1 Beograd Beograd 1,774,990 9,850,000 18% 1 0.99%
Somalia 1 Muqdisho Muqdisho 1,663,223 9,133,000 18% 1 5.65%
South Korea 1 Seoul Seoul 20,550,000 Busan 3,650,000 48,333,000 43% 5.630137 1 7.13%
Spain 1 Madrid Madrid 5,804,829 Barcelona 4,233,638 45,828,172 13% 1.371121 1 5.34%
Sudan 1 Khartum Khartum 9,514,261 42,272,000 23% 1 7.63%
Sweden 1 Stockholm Stockholm 1,860,872 Göteborg 873,335 9,283,722 20% 2.130765 1 3.33%
Taiwan 1 Taipei Taipei 8,422,488 Kaohsiung 2,718,589 23,027,672 37% 3.09811 1
Thailand 1 Bangkok Bangkok 9,972,337 63,389,730 16% 1 6.41%
Uganda 1 Kampala Kampala 1,876,485 32,710,000 6% 1 7.47%
UK 1 London London 13,219,506 Manchester-Liverpool 5,163,631 82,062,200 16% 2.560118 1 5.82%
Ukraine 1 Kiev Kiev 2,987,132 Donets’k 1,623,194 46,143,700 6% 1.84028 1 2.67%
Uzbekistan 1 Toshkent Toshkent 3,235,029 27,488,000 12% 1 1.60%
Venezuela 1 Caracas Caracas 4,368,552 Maracaibo 2,484,277 28,359,313 15% 1.75848 1 9.97%
Yemen 1 Şan’ā Şan’ā 1,976,081 23,580,000 8% 1 11.39%
Zambia 1 Lusaka Lusaka 2,851,411 12,935,000 22% 1 6.89%
Zimbabwe 1 Harare Harare 3,124,683 12,523,000 25% 1 -8.00%
Andhra Pradesh 1 Hyderbad Hyderbad 6,290,397 Visakhapatnam 1,511,687 76,210,007 8% 4.161177 1 5.1%
Assam 1 Bhubaneswar Bhubaneswar 1,666,429 Raurkela 550,668 36,804,660 5% 3.026195 1 2.4%
Karnataka 1 Bangalore Bangalore 6,466,271 Maisuru 1,230,039 52,850,562 12% 5.256964 1 5.6%
Maharashtra 1 Mumbai Mumbai 21,347,412 Pune 5,273,211 96,878,627 22% 4.048276 1 4.6%
Rajasthan 1 Jaipur Jaipur 3,102,808 Jodhpur 971,407 56,507,188 5% 3.194138 1 3.8%
Tamil Nadu 1 Chennai Chennai 7,305,169 Coimbatore 1,644,224 62,405,679 12% 4.442928 1 5.1%
West Bengal 1 Kolkata Kolkata 15,414,859 Asansol 1,317,369 80,176,197 19% 11.70125 1 4.8%
Arizona 1 Phoenix Phoenix 4,281,899 Tucson 1,012,018 6,500,180 66% 4.23105 1 2.46%
Georgia 1 Atlanta Atlanta 5,376,285 Savanna 329,329 9,685,744 56% 16.32497 1 1.63%
Indiana 1 Indianapolis Indianapolis 1,715,459 Ft. Wayne 390,156 6,376,792 27% 4.396854 1 2.06%
Massachusetts 1 Boston Boston 4,522,858 Worcester 781,352 6,497,967 70% 5.788502 1 2.24%
New Jersey 1 Trenton Trenton 365,449 Atlantic City 270,644 8,682,661 4% 1.350294 1 1.58%
Tennessee 1 Nashville Nashville 1,550,733 Memphis 1,285,732 6,214,888 25% 1.206109 1 2.17%
About these ads

Written by seansafford

July 21, 2009 at 4:32 pm

7 Responses

Subscribe to comments with RSS.

  1. Karl Smith comments on the issue here.

    Like

    TGGP

    July 21, 2009 at 5:34 pm

  2. Given the prior work of both, I find it much more likely that both Mankiw and Zingales’ reasons for preferring separation are due to the capture hypothesis – not the idea that regulators with more contact with financiers are likely to be more stringent in their regulation.

    Like

    Charlie

    July 21, 2009 at 7:52 pm

  3. The list shows 23 countries with separate political and commercial capitals (in Bolivia the President and Legislature sit in La Paz, which is also the largest city so Sucre is not really the capital, and in Sri Lanka the administrative capital is a suburb of Colombo). Of those, by my count five (Cote d’Ivoire, Malaysia, Burma, Nigeria, and Tanzania) are places where the capital has been moved in the last 10 years or so, so I don’t think you can attribute much historic economic growth to the choice of capital city.

    As for the rest, it’s hard to see why it matters. In the US and in India, the capital was originally in the biggest and most important city: Philadelphia and Delhi. But in one, the capital was moved for political purposes while in the other it wasn’t. But in both cases a new city is now the financial and business capital. Would America’s economy have grown more slowly if the capital had stayed in Philadelphia when presumably that would have done nothing to dent New York’s advantages as a trading center? Would India have grown faster if the capital and its corresponding massive bureaucracy had been moved Brasilia-style to the middle of nowhere? In both cases it’s hard to see why.

    Like

    T-Rock

    July 22, 2009 at 9:29 pm

  4. Admittedly, there are lots of complications. I’m certainly not claiming a whole lot, particularly from the data I threw up (and thanks for the clarification on Sri Lanka and La Paz).

    I was actually hoping to use the changes as some kind of an identification strategy: i.e., growth before and after. I still may at some point if I feel like this little idea has legs. Which it probably does not: there are simply too many complications.

    And you are also probably right that ‘growth’ may not be the best approach here. Maybe better would be some kind of specific legislation.

    Like

    Sean Safford

    July 22, 2009 at 10:20 pm

  5. Zingales and Mankiw also both need to learn a little more American history – the Founding Fathers put the capital in Philadelphia, which was the biggest city and financial capital of the US at the time. Washington only became the capital when the Constitution was written.

    Germany might be a natural experiment: Berlin was the political and financial capital of Germany until 1945; after the war Frankfurt became the financial capital and Bonn the political capital. But I don’t know how you compare governance in the Weimar Republic to governance in the Federal Republic.

    Like

    T-Rock

    July 23, 2009 at 10:49 pm

  6. I do think the natural experiment route is the way to go on this… Though, its actually hard to find examples of countries/states that shift a capital from a backwater to the major city.

    But more generally, I keep thinking that this might be a useful instrumental variable for something. Time will tell.

    Like

    Sean Safford

    July 24, 2009 at 2:27 pm

  7. [...] Journal that it was wise to have financial and political capitals in different cities, but that is not at all obvious. He may also undersell his case for Jackson’s refusal to renew the Second Bank of the United [...]

    Like


Comments are closed.

Follow

Get every new post delivered to your Inbox.

Join 1,191 other followers

%d bloggers like this: