Cowpox, Smallpox

Unnoticed among the financial headlines last October was the scariest news story of 2008:

Carbon Is Building Up in Atmosphere Faster Than Predicted

The gist of the article is that carbon dioxide emissions in 2007 exceeded IPCC's worst-case forecasts, and that most of the growth in emissions came from the developing world. While we have always had a hard time modeling the effects of carbon emissions on climate, this demonstrates that we can't even solve the much simpler problem of forecasting our direct carbon input. It also means that any agreement to stabilize emissions (let alone begin to lower them, the only way to reduce the eventual severity of global warming) would require even more Draconian measures than we had anticipated.

It was interesting to see this bit of climate news juxtaposed against the fall financial crisis. The two crises have a lot in common. In both cases we have a systemic problem that can be forestalled only through immediate and very expensive government action, paid for by a public that has to take the diagnosis on faith (since even the domain experts do not understand the problem). The solution has to cross national boundaries, with the most advanced economies paying a disproportionate share of the price. And in the best case, the solution will give a negative result (lack of a depression, absence of catastrophic climate change) where it may be impossible to prove that the adopted policy was better than doing nothing.

These crises test the limits of modern democracy. Governments have to convince people to pay an enormous price today in order to lessen the impact of a future disaster whose timing and scope are not known. It is my belief that taking meaningful action to forestall an economic crisis lies just within the realm of the possible, while doing anything substantive about global warming is beyond our capacity, and that the fundamental reason for this is a difference in time scales.

One of the more arresting details of the financial collapse has been the complete discrediting of macroeconomics. This is a field in which Nobel Prizes are awarded, one that calls itself a science, and yet one in which only a tiny dissident minority gave any warning about structural problems that are now obvious in hindsight. If the same thing had happened in theoretical physics, the streets would be filled with newly unemployed CalTech PhDs. Yet the bearded heads who failed to anticipate the global collapse have proved more unsinkable than the companies whose health they were vouching for. The companies are gone, but they are still out there, offering television commentary and advising governments, as if nothing had happened.

Much of this has to do with dynamics of money and power that are outside the scope of this essay. But another important reason is that we recognize that the difficulty of analyzing complex, dynamic systems. Just as we do not fire a weatherman after he forecasts a nonexistent storm, we give economists and financiers wide latitude for failure because we haven't found a better alternative.

Both climatology and economic forecasting are difficult for the same reason: they are trying to model systems that interact with themselves. We know from mathematics that these kinds of interactions give rise to incredible complexity. The thing we call 'climate' is the net result of hundreds of interlocking feedback loops. An increase in ocean temperature, for example, causes more water to evaporate and condense into clouds. These clouds reflect sunlight back into space in the daytime, but also trap heat that would otherwise escape at night. At the same time, they affect patterns of precipitation and plant growth on the ground, which in turns affects the amount of sunlight absorbed by the earth (vegetation is darker than desert) and the amount of carbon that gets removed from the atmosphere.

Figuring out the net effect of these interconnections is not just a matter of describing them and then running a big computer simulation. The system as a whole is highly non-linear: a tiny nudge in one direction might result in an equally tiny outcome, or it might launch a cascade of side effects that culminates in a dramatic shift. Even with complete knowledge of how the system works, it might not be possible to make long-term predictions because of sensitivity to initial conditions.

Feedback loops in the financial world are even worse, since the entities being modeled are aware of their behavior - and aware of the models being used to study them. Investors form strategies based not just on market conditions, but on their perceptions of others' perceptions of market conditions, and so on in a hall of mirrors effect. Any algorithm that can reliably predict the behavior of a financial market will be used by participants in that market to earn money, altering the system in a way that leaves you right back where you started. In this sense our ability to model economics will always be worse than our understanding of the weather, since we don't have to worry about a raindrop anticipating that it will hit the ground before it even forms, and taking steps to change the outcome.

Just because the job is hard, though, doesn't absolve us from doing it. And so we have climatologists and economists. They have come up with simplified models that have at least limited predictive power, by validating them against historical data. The problem comes when we enter uncharted territory - say by doubling the amount of carbon dioxide in the atmosphere, or creating an enormous, unregulated derivatives market. At this point we have no historical data to compare against.

We know that our models sometimes show catastrophic behavior. Like a bowling ball being pushed to the edge of a staircase, sometimes an tiny additional stimulus causes a surprising - and irreversible - result. But which catastrophic events are artifacts of the model, and which are features of the real world? History shows us that both the financial system and the global climate have exhibited catastrophic behavior in the past, but how can we have any confidence that our models have got the triggers and the thresholds right?

The honest answer, both in climate change and the economic collapse, is that we are in the dark. Economists use words like 'meltdown' or 'panic' and draw parallels to the Great Depression, but leave unclear what this would mean to the average working person. Ten percent unemployment for half a year? Thirty percent for three years? The complete collapse of international trade? Opening canned goods with sharpened sticks?

The range of forecast outcomes for global warming is wider. One thing we do know is that there is a long lag time between cause and effect in the climate, because of the enormous inertia of the system. In the best case, some hidden moderating feedback kicks in to make the effects of global warming less severe than we now expect, and the result is limited desertification, sea level rise, and a far more livable Canada. Maybe the ocean mans up and starts absorbing its share of atmospheric carbon; maybe cloud patterns change in a way that reflects a greater amount of sunlight.

In the worst case, the Earth turns into Venus.

Given that all the plausible outcomes of climate change will have a profound effect on our economy, and some may threaten our survival as a species, why are we not putting at least the same efforts into adressing climate change as we have devoted to the current economic crisis?

The problem is a difference of time scale. Financial panics and depressions unfold over a scale of months. Severe depressions have happened in living memory, and we have learned to recognize some of their features. We also have a cultural memory of the effects of a global depression that make us willing to take drastic steps to avoid another one. And we have a political system that is capable of making sacrifices on the scale of a few years (think of the Second World War, or shock therapy in the former Eastern Bloc) in return for a credible promise of future prosperity.

Climate change takes place on a scale of decades. Though the earth has existed in a variety of metastable configurations during its history (snowball, hothouse, ice age) our only experience as a species has been with a very predictable, benign climate. Human beings have never before caused a radical shift in global climate (possibly because this is something a species only gets to do once), and prehistoric evidence of abrupt irreversible change remains too abstruse to convince a nontechnical public. If a third of your electorate doesn't even believe in geology, how do you convince them to shell out for climate change? Trying to prevent severe global warming in advance, even if it is still possible at all, would require massive expenditures on a time horizon of twenty years or more, with a strictly negative benefit (the absence of droughts, floods, and mass extinctions) if the policy succeeds. We would have to rebuild the entire basis of our economic life, and be content that nothing bad happened.

Modern, mass-media-driven representative democracies cannot address risks on this kind of time scale, especially when the risks are necessarily vague and couched in technical language. Domain experts who are unwilling to frame their case as a choice between stark opposites find themselves ignored in favor of those who are less squeamish. The subsequent polarization of the field removes what little light was previously being shed. And given the complexities and uncertainties in the underlying science, it becomes easy for well-funded interests to find an expert to back any point of view. This rewards opportunism by people unrestrained by intellectual honesty or notions of scale ("this latte is carbon-neutral", "home loans to poor people caused our banking system to fail").

I do not know if there is any way to structure society so that it can make wise decisions over a time scale of fifty years. Maybe someday a comet will be found heading for Earth, with advance warning several decades, and we will get our answer. In the meantime, it is clear to me that whatever cosmetic efforts we make at reducing emissions (such as cap-and-trade agreements), these will never involve the kinds of astronomical sums needed to make a difference.

At some point within the next few decades, we will begin to feel the first direct economic impact of global warming, and at that moment the range of politically possible solutions will expand. When disaster strikes, we pay what it takes to rescue people and worry about the cost later. Mitigation will cost vastly more than prevention, but the problem will at last be politically tractable.

The danger zone lies in the next few years, during the span of time when we will be increasing our contribution to the problem, but not yet feeling its effects. Since taking preventative steps is apparently beyond our power, the only hope is that some outside force will reduce our emissions for us.

Ironically, this outside force may already be here, in the form of the economic crisis itself. Already fossil fuel use has dropped significantly, and a prolonged worldwide depression might be able to achieve what no Kyoto protocol ever could - a net decrease in greenhouse gas output. Anyone who seriously believes in the threat of catastrophic climate change is therefore left in an uncomfortable position. We are facing an economic crisis that is within our capacity to solve, and an ecological crisis that we lack the political means to prevent. It's only by failing at the former that we might have a chance at surviving the latter.