The roots of liberal policy failure are intertwined with the dark side of scientific rationality.

During the 1980s and 1990s, experts working for the World Bank and development agencies persuaded African nations like Malawi to stop subsidizing fertilizer. Subsidies for fertilizer were extremely popular among Malawi's people -- roughly 90 percent are small farmers growing staples on depleted soils who cannot afford fertilizer at market prices. Malawi's political leaders resisted the expert advice for years. But donor nations are powerful in aid-dependent countries, and Malawi eventually acceded to their demands.

The results were disastrous. Malawians were not able to produce enough corn to feed themselves. By 2005, more than one out of three Malawians were dependent on foreign food aid, and the country was on the brink of famine. In an agriculture-dependent region where poor harvests can have devastating effects, Malawi's government changed tack, and later that year started subsidizing nitrogen fertilizer over the objections of its expert Western advisers, who predicted a worsening of the disaster.

Instead, the opposite happened. Over the next four years, corn production per hectare of land tripled. Acute child hunger rates plunged. Malawi was able to turn away powdered milk from Unicef and export food to its neighbors. Britain's Department for International Development evaluated the program and found that Malawi's $74 million annual corn fertilizer subsidy was worth $120 to $140 million annually -- a nearly 100 percent return on investment.

Malawi wasn't alone. In Rwanda, corn yields rose 75 percent after it started subsidizing the distribution of fertilizer. Zambia and Mali launched their own efforts to expand fertilizer use. Egg on its face, the World Bank was forced to admit, albeit in its crabbed bureaucratic vernacular, that subsidies "may be justifiable on a temporary basis to stimulate increased fertilizer use in the short term."

How did the experts get it so wrong -- and for so long? There's nothing to suggest mal intent. The experts in question have in many cases dedicated their lives to helping poor nations develop. All are highly-educated, not ignorant.

Writing in Breakthrough Journal, innovation expert Daniel Sarewitz blames something else: "scientific rationality unchecked by experience, empathy, and moral grounding." By "scientific rationality," Sarewitz, a geologist by training, doesn't mean science, per se, but rather the reductive chains of logic that end up conflating science with policy. It's not the rationality itself that's the problem, but rather its detachment from a larger social, moral, and historical context.

The logic of ending fertilizer subsidies was, for example, ironclad. The law of comparative advantage says that nations will grow faster if they focus on doing things they can do better or more cheaply than other nations. For countries like Malawi and Rwanda, that meant producing (and thus subsidizing the production of) more high-value crops for export, like tea and coffee, and less of low-value food crops for domestic consumption, like corn and wheat. Experts produced rigorous, scientific studies showing that eliminating fertilizer subsidies would result in faster economic growth.

But this rationality only worked in a vacuum. Cutting fertilizer subsidies required ignoring Malawi's history of depleted soils. It required ignoring the experience of other nations in Europe, Asia, North America, and Latin America, which have all massively boosted productivity through fertilizers and subsidies. And it required dismissing the cries of protest from Malawi's leaders and farmers.

Sarewitz offers climate policy as a case study in the ways that an overdependence upon scientific rationality can lead experts and policy makers to places they probably don't want to go. Over the last two decades, the range of policy options considered to be "science-based" has narrowed dramatically. Even after the collapse of international efforts to establish legally binding limits on carbon emissions, many policy experts and climate advocates continue to maintain that pricing carbon and hence raising the cost of fossil energy is the only "science-based" response to global warming.

Indeed, as the expert consensus on carbon pricing congealed, other policy responses were often characterized as unscientific, or even anti-science. "The main points from climate science," wrote Environmental Defense Fund economist Gernot Wagner recently, "are no longer up for debate: the planet is warming; humans are the cause of it; we need to limit emissions... carbon is a pollutant; we need make polluters pay, either through a cap or a price.... Once again, this one is not up for debate.... you can have very real economic debates around whether to tax or cap carbon.... But it's not up for debate whether 'carbon = pollutant' leads to the need to cap or price carbon. It does."

It is a remarkable chain of reasoning. If you accept that the planet is warming, then you must support an emissions cap or a price on carbon. To question the latter is to deny the former. Their devotion to a particularly narrow scientific rationality led many liberals to imagine carbon pricing to be the pure policy expression of unadulterated science.

But like ending fertilizer subsidies, raising the price of energy to combat climate change is perfectly rational only so long as the larger context is ignored. "If one were seeking a policy intervention that could simply and effectively erode economic and social equity worldwide," Sarewitz writes, "one could hardly do better than to increase the cost of energy.... This fact is so blindingly obvious that nearly every large developing country has treated the idea of a global agreement to raise energy prices as a joke of Swiftean character."

There is, of course, an important role for scientific rationality. Morality unchecked by scientific rationality can descend into dogmatism. Historical awareness unchecked by scientific rationality results in a commitment to traditionalism for its own sake.

But for scientific rationality to be moral, wise, and result in positive outcomes, Sarewitz suggests, it must incorporate, rather than dismiss, other ways of knowing. A rationality that relies on the natural sciences and economics to the exclusion of history, culture, and politics, is a threat and not an aid to liberal goals and values.

In the case of dealing with global warming, Sarewitz suggests updating a successful past model, the Tennessee Valley Authority. That New Deal-era project was focused on expanding access to cheap energy among the poor. It should be placed in a global context, Sarewitz argues, and used to accelerate both energy access and decarbonization through the demonstration and deployment of zero carbon technologies.

But Sarewitz is after prey far larger than climate policy. Over the last few years, many prominent progressives have come to blame liberal policy failures on a society that cares too little about science, rationality, and the truth. Sarewitz begs to differ. In many cases, he convincingly demonstrates, the roots of policy illiberalism lie not in popular irrationalism but rather in reductive, scientistic rationalism.