Sachs Debate, Day Eight. Surprised that Sachs did not understand the point about Occam's Razor -- that a theory should be as simple as possible but no simpler. African poverty is complex, but our theories about it should not have so many complex Buts, Ors, and Excepts that they are impossible to disprove. Ignoring Occam's Razor is how astrologists stay in business. An astrologer might say: "Watch out for strangers, especially those that are short, or dark, or fat." You are likely to have a number of bad encounters with strangers (especially in New York!), and being short or dark or fat covers such a large share of the population that you will likely encounter one such bad stranger sooner or later.
Explaining poverty with a flexible theory of geography when you already know the outcome is similarly easy. First, you notice that Africa had the worst poverty in the world. Second, you notice that Africa also was the only region that had a particular mosquito species, had a lot of landlocked countries, had a particular soil type, had a large share of continent in the tropics, and did not have snowmelt-irrigated agriculture. Third, you define the worst geography as consisting of exactly these mosquitoes, landlocked locations, soil types, tropical locations, and non-snowmelt-irrigated fields. Fourth, presto, you have proven that the worst geography causes the worst poverty!
(Some of the above are bad, but there are many bad things in the world, and only such a four-step exercise guarantees that you will predict rightly in hindsight. Try this at home to impress your friends! If they object that you are just using hindsight, tell them they don't understand "complex systems.")
The more general argument why geography is not destiny is what economists call comparative advantage -- you can export what your geography gives you an advantage at, and import what you geography makes you bad at. Rainy countries (Thailand) could export rice (a thirsty crop) and import minerals, and desert countries could export their minerals (Chad) or maybe tourism (Dubai), and import rice. Landlocked countries can always export high value to weight items by air (Swiss watches, Botswana's diamonds). Other geographic problems have their own human adaptations. Sachs contradicts his own geographic determinism by arguing how easy it would be to solve these problems (like bed nets and medicines for malaria). When such adaptation to a geographic problem fails, there is usually some kind of social barrier to problem-solving.
One such social barrier is a bad government -- like a government that fails to deliver the bed nets and malaria medicines -- a factor that Sachs still refuses to do justice. Sachs' admission that Zimbabwe has a bad government is not exactly revolutionary. This continues his longtime reluctance to admit there is bad government in any except a handful of extreme despots. In his UN Millennium Project in 2005, he only named four bad governments: Belarus, Myanmar, North Korea, and Zimbabwe. At the time, he listed 63 poor countries that are "potentially well governed" (a phrase equivalent to calling Paris Hilton "potentially a virgin.") Sachs' list included 5 out of the 7 countries singled out by Transparency International at the time as the most corrupt in the world, and fifteen governments that Freedom House classifies as "not free." Even a despot like the late Saparmurat Niyazov of Turkmenistan, who so terrorized his country that he renamed the months of the year after himself and his mother, couldn't get into Sachs' bad government club at the time. Meles Zenawi of Ethiopia is inexplicably a Sachs darling despite rigging elections, jailing opposition politicians, shooting demonstrators, fielding an army accused of atrocities in the Ogaden, and fomenting corrupt practices that give Ethiopia a ranking on Transparency International of 138th out of 179 countries.
So Sachs' re-admission that Zimbabwe has a bad government does not get us very far. Unlike Sachs' flexible theory of bad geography, the bad government theory is inflexible enough that it is not rigged to pass in advance -- but it does pass this test according to studies by three different and independent groups of economists published in prominent economics journals. All science is essentially about testing one thing -- like bad government -- at a time, but nobody thinks that such testing implies only one thing matters.
The bottom line remains the same -- bad government is a serious obstacle to development, and ignoring bad government is a formula for the same kind of bad aid policy that gave us decades of failure. Isn't it time for a change? Even cash transfers directly to poor people are demonstrably better than cash transfers to poor governments.
Follow William Easterly on Twitter: www.twitter.com/bill_easterly