April 20, 2015

How to Think about Climate Risk



Editor’s note—our guest blogger today is Bob Litterman, a former partner at Goldman Sachs and presently the chairman of the risk committee at Kepos Capital. While at Goldman, Bob oversaw the Quantitative Investment Strategies Group, which at its peak managed over $185 billion in assets, and Global Investment Strategies, an institutional investment research group.  He is the co-developer of the Black-Litterman Global Asset Allocation Model, a key tool in investment management, and has co-authored books including The Practice of Risk Management and Modern Investment Management: An Equilibrium Approach. We are proud to have Bob on the Niskanen Center board of directors.

Twenty years ago the partners of Goldman Sachs decided that they needed to create a new position, head of risk management, and they asked me to join the partnership and to take on this job. The partners taught me many lessons about risk management that had been incorporated into the Goldman culture over the long history of the firm and which informed how I was expected to do my job.

These realities of risk management apply to the risks created by greenhouse gases in the atmosphere and, in particular, they imply that those risks should be priced immediately.

The key reality is that when confronting an uncertain set of outcomes, you cannot focus only on the expected outcome, you have to consider the worst case scenarios.

Another reality is that risk management is always an urgent priority.

The third reality is that the job of the risk manager is not to manage risk. At Goldman Sachs that was the job of the other partners. The job of the risk manager is to identify and quantify risks.

The final reality is that the purpose of risk management is not to minimize risk. The purpose of risk management is to make sure that risks are priced appropriately.

At Goldman Sachs we prided ourselves on our ability to make profits from smart risk taking. My job was not to get in the way of that activity; rather it was to inform it. Our job, as partners, was to make sure that we understood and managed all of the risks that we faced. Market risks, which were my responsibility, were among the easiest to manage.

The risks that we expected to get paid for were sized appropriately. Other risks were hedged. If, as risk manager, I identified a risk that was not intentional and was not being hedged, my job was to make sure that the responsible trader hedged that risk, or that the exposure was immediately escalated to senior management. Unintended risks, particularly those with catastrophic worst case scenarios, were an urgent problem, and in my career at Goldman they were never ignored.

The same principles apply to investing, and the same principles apply to managing climate risk. Having recognized the implications of these realities, years ago I became laser focused on appropriate pricing of emissions.

What grabbed my attention was some scar tissue in my brain.

When I was a graduate student in economics at the University of Minnesota in the late 1970’s, a very shrewd transportation professor, Herb Mohring, tripped me up. I was a macro student, but he put together the micro prelim that I was taking. And, thinking like a macroeconomist, I answered the question: “Provide a defense of smoothing oil prices” with a complicated solution to an optimal control problem, the key objective of which was to minimize squared changes in oil prices in the context of an economy with real shocks and a strategic oil reserve, the flow of oil into or out of which was in the control of the social planner.

For that answer, solution of which was technically fine, I got zero credit and flunked the prelim. I was in shock, and it meant I had to take the exam again, a very painful prospect. But the lesson—that prices do not belong in utility functions—was burned into my mind.

And so when I apply risk management to the question of greenhouse gases, the implication is clear. The price of emissions should immediately jump to the appropriate level. Incentives to reduce emissions should reflect the economic externality, the risk that is created by those emissions, immediately.

And so the question, “what is the appropriate incentive?” becomes the key question.

I was suspicious that the economic models then being used to answer this question—models that overwhelmingly suggested a low price for emissions—were not adequately considering worst case scenarios. And sadly, what I found was that this was indeed the case.

The standard models then in use suggested a low initial price of emissions followed by a slow increase in price over the course of many decades. It was an “ease-on-the-brakes” scenario. It also suggested that it doesn’t really matter too much when you start to price emissions. The hard work of emissions reductions would be done far in the future.

Unfortunately, those models did not adequately consider worst case scenarios.

Think of the earth’s atmosphere as a reservoir that is being filled up. There is no reason to assume a smooth function relates the level of greenhouse gases to future damages. It is perhaps better to think about boundary crossings followed by positive feedbacks, like a reservoir that when it overflows erodes the dam and suddenly releases a torrent—call it the “Johnstown flood” scenario.

Scientists worry about many different boundaries and positive feedback pathways. Methane release from the arctic regions as they warm creating a positive feedback is one such scenario. Deforestation of the amazon leading to reduced rainfall and drought is another. There are many others that have been considered, and no doubt there are many other scenarios that haven’t yet been considered.

How likely is such a catastrophic event at 400 ppm of carbon dioxide in the atmosphere (the current level)? How likely at 800 ppm? What level is safe? Given the uncertainty of what will happen, emissions of greenhouse gases create increased risk of catastrophic outcomes. This risk will, no doubt, be recognized and priced by society very soon. When that happens, the price will probably not follow the “ease on the brakes” scenario that most people today seem to expect.

You can understand that I am particularly sensitive to price smoothing—when I see prices being smoothed I hear a silent warning: “there is no defense of smoothing prices, prices don’t belong in utility functions.”

The brake that we need to apply immediately to address climate risk is the price of emissions.

If we knew exactly when and how damages would follow from increased emissions we might indeed ease on the brake, in particular we would if we knew for sure that those damages would be small and far enough into the future. But given the uncertainty, society will soon recognize that it has no choice but to react. It will then have to consider seriously what the appropriate incentive to reduce emissions is.

I don’t know what that appropriate price is, nor does anyone else. But I do know one thing, which is that the uncertainty about the future impacts of greenhouse gases, and the potential for catastrophic outcomes, means that a price high enough to quickly reduce emissions is inevitable. That will be a high price relative to current expectations, and it will likely be imposed very quickly because it never makes sense to smooth prices.

It’s long past time to hit the brakes. Don’t be surprised if that scenario plays out very quickly.