The Power and the Money

Was the French Empire predatory?

There is a long debate over colonialism, particularly the late-19th-century variant. Most of the literature focuses on long-term effects. So it often loses sight of a simple question: did imperial institutions function basically the same as capitalist institutions in non-colonized places or did they serve to transfer resources to the metropole? That is, did colonialism simply bring areas into the modern world economy (for better or for worse) or did it directly suck resources out of the colonies?

This is not an easy question to answer. In a classic work, for example, Davis and Huttenback showed that British colonial enterprises were less profitable than similar British enterprises at home, in the Dominions, or in the Americas. But profitability tells us little. Imagine that British colonial enterprises had been phenomenally profitable. Would that have meant that they were bad for the colonies? Clearly not! Some other measure is needed. 

So here comes Federico Tadei (Universitat de Barcelona) with an obvious measure: did colonial commodity producers receive the same price for their goods as producers elsewhere, accounting for differences in transportation and insurance costs? 

In a bit more detail, here is what he does:

1. Gather prices at French ports for each commodity between 1898 and 1959. Also gather producer prices (what African farmers received at the farm gate) at African ports. Tadei can do the latter because French African statistics broke port out trade costs from producer prices; 
2. Construct estimates of trading costs including Atlantic shipping, insurance, inland transportation, port charges, and export taxes;
3. Calculate (1) − (2) to get what producer prices should have been under non-predatory institutions;
4. Compare (3) to the actual prices received by African farmers;
5. Check the results by comparing (1) against producer prices for U.S. cotton and wheat exporters or European settlers in Africa to see if American producers selling to Britain were as disadvantaged as French African producers selling to France;
6. Check the results by seeing how much of an increase in world prices is passed on to French African producer prices, plus a few more abstruse checks.

Here is the headline result, as a % of prices:

And here is French Africa compared to the United States for cotton:

Moreover, there is a plausible mechanism for extraction: monoposonistic trading companies and compulsory harvest quotas. The effect of the former is obvious; the effect of the latter was to prompt African producers to grow more than they would have in a competitive market. In addition, poll taxes might have induced African workers to enter the labor market, driving down production costs for African farmers and thereby prompting them to grow more. (Forced labor may have had the same effect, inasmuch as it also drove down costs for African producers.)

What I would like to see is the same method applied to Latin American and Asian producers. Brazil, Ecuador and various Caribbean countries exported cocoa; China exported peanuts; Brazil exported cotton. As for palm oil … well, forget that. (United Fruit introduced it into Guatemala, Panama and Honduras in the 1920s without much success. I was involved in a failed effort to introduce it to Chiapas in 1999.) But it is a way of stress-testing the hypothesis. If producers in Latin America and China faced similar extraction rates, then formal imperialism might not have been the problem.

Thoughts?

NAFTA and agriculture

There was some interest in this topic but not a whole lot. So I will put up a short post, which will be pretty straightforward. I will discuss U.S. exports today.

The key thing to keep in mind is that agriculture is a global market. So if third-countries displace American agricultural products in the Mexican market, then American producers will be able to export their products to the places the third-country used to sell to. But if Mexico just reduces its consumption of all agricultural imports, then U.S. producers will lose unambiguously.

Mexico takes 8% ($1.8 bn) of American soybean exports. It takes 21% ($2.4 bn) of U.S. corn exports. It takes 11% ($0.9 bn) of our wheat exports.  It turns out that 36% of U.S. corn travelled to Mexico by sea in 2007-10. See page 10. The share is even higher for other products: 55% for wheat and 38% for soybeans; see pages 20 and 23. I used USDA data to calculate costs per ton-mile for soybeans: subtracting out $230,000 per ship Panama Canal tolls for cargoes from the U.S. costs gives you an average cost per ton-mile of 49₵ from the U.S., 43₵ from Argentina, and 39₵ from Brazil; fixed costs come to $5.70 from the U.S., $7.20 from Argentina, and, inexplicably, only $2.70 from Brazil.

With that data in hand, let’s look at those three markets and imagine that Mexico slaps MFN tariffs on the U.S. while opening to the South Americans. 

Soybeans: Mexico could easily supply itself from Argentina and Brazil. Landed costs to Mexico are already cheaper for the Latin American nations, even with higher transport costs; only South American export taxes keep us competitive.

Slap a 4% tariff on American exports (while removing them on the South Americans) and Argentina and Brazil would drive us out of the market. But remember that soybeans are sold in a world market. So more South American exports to Mexico means less South American exports to somewhere else. American production will fill up the gap. When it all shakes out, you would have a small loss for the Americans, a small gain for the Argentines and Brazilians, and about a wash for the Mexicans. 

Corn: The U.S. and Argentina have more or less the same corn production costs. But transport costs from Argentina are quite a bit higher. If Mexico imposed MFN tariffs on corn and suspended them on South American exports, the landed cost of U.S. maize would still be about 13% less than Argentine corn. Americans would lose sales via a general cutback in Mexican purchases, not a switch to South America. 

Lose-lose, although more for Mexican consumers, since U.S. producers would presumably find alternative markets. That said, Mexican corn farmers would gain, and rural voters are a major constituency, about 30% of the electorate. (See page 10. Rural voters punch slightly above their weight in turnout terms: see page 9.)

Wheat: See soybeans. 

Upshot: ending NAFTA would not be catastrophic for American soy and wheat exporters, but both would have to discover new markets right quick or suffer some unpleasant transition costs. It would be a clear boon for Argentine and Brazilian farmers, who could capture a whole new market. Mexican corn farmers will gain; Mexican consumers will lose.

Still, NAFTA probably won’t end.

The great stagnation on the farm?

There has been a lot of discussion lately about the Great Stagnation: the apparent slowdown in productivity growth in rich countries since 1970. The phenomenon is indisputable, but mysterious, since it parallels the rise of computers and the internet, both of which should have saved a lot of resources in the production and distribution of goods and services. Maybe the robots will usher in a huge productivity increase, even if it has not quite happened yet. 

Oddly, there has been a running debate over whether agricultural productivity has suffered a similar decline. Well, now Julian Alston (U.C. Davis), Matt Andersen (Wyoming) and Phil Pardey (Minnesota) have taken a closer look! Now, calculating multifactor productivity (MFP) in agriculture is not as simple as it sounds, especially across a century, but they have done it. Here are their results:

Eyeballing the log scale gives you their main result: continuing productivity growth, but a clear acceleration in the 1930s and a slowdown around 1980. 

So what happened? They offer two hypotheses. First, stagnation in R&D spending:

As Julian Alston has pointed out, agriculture faces a Red Queen’s Race: the co-evolution of pests and diseases means that you have to keep running as fast as you can just to stay in the same place. Moreover, private spending does not necessarily compensate for public for the simple reason that the time horizons involved in developing agricultural technologies are very long. As a result, very high ultimate cost-benefit ratios (21 for state efforts and 18 for federal ones) translate into much more modest internal rates of return (19%). That is still a high return, no doubt, Panama Canal high ... but nothing like what you would expect from the size of the long-term payoffs. 

The second (related) hypothesis is that the big acceleration in productivity was a one-off from a few mid-century inventions:

The implication, of course, being that GMOs don’t hold a candle to hybrid corn (which is quite reasonable) ... or maybe that we haven’t seen nothing yet. In fact, all we can really say is that the argument Robert Gordon has advanced in his seminal work may apply to agriculture as much as manufacturing and services. 

What I really want to know, though, is what the hell was up with rural telephones?

Did drought abet the start of the Syrian civil war?

I have no idea. But there was a bad drought in Syria in 2006-11. (Hat tip: Wonkblog.) It is not unreasonable to think that there might be a connection.

Have we historical evidence?

Maybe. Many historians believe that the Mexican drought of 1906-10 played an important role in setting off the Mexican Revolution. Melissa Dell set out to test that hypothesis, combining one dataset of pre-1910 precipitation and another one of Revolutionary violence. She found that drought-affected counties were more likely to experience violent mobilization during the revolution.

I think she is right ... but I am not sure that her data is good enough to show it. The data on revolutionary mobilization were not the best. They came from the Encyclopedia of Mexican Counties, taking on a value of zero or one depending on whether the authors of the Encyclopedia noted spontaneous violent mobilizations. The data are summarized in this map:

The counties singled out on the map are  few of the ones where the coding is unclear. Tlahualilo, for example, produced few local leaders or organized revolt, but there were not a few complaints of banditry in 1911 and when Pancho Villa did roll in, the locals supported him. In Camargo, the Loyal Camargo Brigades formed an important part of the Villista army. (Inasmuch as they were no violent mobilizations in Camargo, it appears to be because the Revolution took over the state quickly enough to let Villa use it as a place for serious R&R as early as 1913.) If you go through all the possible switchers (and note that there might be something special about Yucatán) the results are weaker. Contagion models could weaken it further.

This does not mean that the argument is wrong! Just that more data is necessary.

Which brings us back to Syria. The report from the Center for American Progress (see page 25) is damnably vague. But data on precipitation, internal displacement, and the pace of revolutionary mobilization in the country should exist. (Doug? Carlos?) If they do, then it should be possible to see if climactic stress is linked to unrest or violence, at least in the modern Levant.