Halfway down the Danube

Farm to Factory, Robert Allen

Okay, my secret is out. I love nothing better than to find a new book on some incredibly boring arcane subject and worry at it until it makes sense. So when an online friend recently brought up Robert Allen's book on Soviet industrialization, Farm to Factory, I thought I would give it a look-see.

It's an interesting but extremely frustrating book. Interesting in that Allen tries to analyze Soviet industrial development from the point of view of its decision makers, but using modern econometric techniques, in the same way a psychologist or a novelist might analyze a bad relationship from the perspective of its hapless participants: how it made sense at the time to the people involved.

On the other hand, it's extremely frustrating when Allen makes comparisons that just don't make sense to me, the reader.

For instance, Allen repeatedly makes the claim that Russia without Communism would most likely look a lot more like India demographically than it would any western European state; it would have an extremely large and poor population of perhaps a billion people. At first I thought this was rhetorical hyperbole, the first few times I encountered it. Nope; it forms the basis for chapter six. The possibility of a non-Communist demographic transition is where?

But what really made me wince was his comparison of Soviet agricultural and climatic zones to American ones. Now, I'm from Wisconsin, where we know both climate and agriculture pretty well. (Wisconsin is climatically like Belarus, but with more lakefront, and has dairy like Denmark. The closest American analog to Romania would probably be Iowa or Missouri.) So when Allen compared Russian grain yields to North Dakota's -- compared to Wisconsin, a barren dessicated wasteland, though its inhabitants are very sweet -- and used this to claim that the Soviet Union had somehow caught up to American grain yields before Stalin, I knew something was a little weird.

Here's Allen on his methodology:

To reach a better assessment of Russian performance, the comparison should be made with a region of similar climate and soil. In this chapter, I compare Russian productivity in 1913 with productivity on the Great Plains of North America -- a region that includes the Canadian prairie provinces of Manitoba, Saskatchewan, and Alberta, as well as the American states of North Dakota, South Dakota, Montana, and Wyoming. Johnson and Brooks (1983) undertook a very careful assessment of Soviet agriculture in the post-World War II period and compared it with the same region studied here. As in Russia, the climate was cold and dry.

Yes, Johnson and Brooks did make a very careful and detailed comparison between Soviet and North American agricultural regions. Here it is, from Table 6.1 in Johnson and Brooks, Prospects for Soviet Agriculture in the 1980s:

Ukraine and Moldavia, Spring grains: South Dakota

Ukraine and Moldavia, Winter wheat, Area I: Nebraska, SE
Ukraine and Moldavia, Winter wheat, Area II: Nebraska, NW
Ukraine and Moldavia, Winter wheat, Area III: Nebraska, SW
Ukraine and Moldavia, Winter wheat, Area IV: Nebraska, Cen.
Ukraine and Moldavia, Winter wheat, Area V: Nebraska, NW
Ukraine and Moldavia, Winter wheat, Area VI: Montana, EC or NC, Colorado, NW

Ukraine and Moldavia, Winter rye : US, Canada

Central Black Earth: Manitoba
Middle Volga: North Dakota
Belorussia: Minnesota
Baltic Republics and Kaliningrad: Minnesota, EC, WC
European West: Minnesota, EC, WC
European Northwest: Minnesota, EC, WC
Central Industrial: Minnesota, NW
Upper Volga : North Dakota

Krasnodar: Nebraska, NE, East
Stavropol: South Dakota (spring), Nebraska (fall)
Kamensk: Wyoming, SE
Remainder, North Caucasus: Nebraska
Lower Volga: New Mexico

Molotov and Sverdlovsk: Alberta, EC
Bashkir and Udmurt: Montana
Chkalov and Cheliabinsk: South Dakota, West
Western Siberia: Manitoba, Saskatchewan
Eastern Siberia: Saskatchewan
Far East: North Dakota
Kazakh, except Alma Ata: Wyoming, Montana, South Dakota, Saskatchewan

(Apologies for the lack of a map.)

The Soviet wheat growing regions comparable to Allen's chosen North American analogs leave out most of Ukraine! Huh? Meanwhile, they include such great Russian breadbasket regions like the Amur river valley.

It gets worse. In fact, Allen compares these North American regions to European Russia in 1913. But that reduces the overlap even further, to a thin strip starting at the Black Sea, and widening upwards to include the upper and middle Volga before it hits the Urals.

Going by Johnson and Brooks's figures, had Allen used the five states and one province of Manitoba, Minnesota, Nebraska, North Dakota, New Mexico, and South Dakota as his comparisons, he would have incorporated 90% of European Russia's grain-growing areas. (In a footnote, Allen considers and rejects using Nebraska, since it produces so much maize. I don't know what that has to do with comparing yield.) This combination is warmer, wetter, and significantly more agriculturally productive than the four states and three provinces Allen did use.

Why is this important? Allen wanted to show that Russian, and thus later Soviet agriculture, was not as inefficient as sometimes is claimed. Unfortunately, noting that yields on rich Russian land were comparable to yields on more marginal North American land is probably not the way to do it.

So you see what I mean. Extremely frustrating.

Posted by coyu at September 9, 2004 08:55 PM