Sunday, January 10, 2010

Making society forecast-proof

The trouble with forecasts is that they are almost always wrong. That's in part because the accuracy of forecasts deteriorates rapidly with time. I might predict quite well what I will be doing tomorrow. But predicting accurately what I will be doing exactly one year from today or exactly 10 years from today is exceedingly difficult if not impossible, even if I already have something planned.

How much harder it is then to predict the state of complex systems such as the world's energy delivery systems 10, 20 or 30 years hence. There are many factors that make such predictions hard including:
  • the inaccessibility of audited data such as oil and natural gas reserves for many of the largest producing countries in the world
  • the uncertainties about future discoveries
  • the uncertainties over the rate of depletion for fossil fuels
  • the uncertainties concerning future technological advances in extraction and energy efficiency
  • the growth and viability of alternative fuels
  • and the future course of energy prices and the world economy just to name a few.

And yet, we have premised our entire future on business-as-usual forecasts made by leading energy consulting firms and government agencies. Strangely, some of these forecasts come without even the slightest hint about how unreliable they may be. A forecast from the highly influential energy consulting firm Cambridge Energy Research Associates tells us we have precisely 3.74 trillion barrels of remaining recoverable oil in the world. It concludes that a peak in world oil production is still at least 20 years off and is to be followed by an "undulating plateau that may well last for decades." Neither a range nor an error bar can be found in this forecast.

Because such forecasts have been embedded in public policy and business planning worldwide, we have made our entire global society dependent on getting these forecasts right. If they turn out to be too optimistic, then we could all be in for serious trouble. Since long-term energy forecasts--and really any long-term forecasts--are difficult if not impossible to get right, perhaps we should consider making society forecast-proof insofar as that is possible.

When it comes to energy supply, the surest way to make that happen would be to power society with 1) solar energy and its derivatives including wind, waves, and biomass, 2) energy derived from gravity such as tidal power and 3) hydroelectricity which is made possible by gravity and the water cycle. These sources fall under the heading of renewable energy, of course, and it more a scientific fact than a forecast that the sun will burn for a few more billion years and that the force of gravity will be with us indefinitely.

A global society powered in this manner would be immune to any forecasts about future supplies of fossil fuels, be they pronouncements of plenty or warnings of imminent supply shocks. Yes, the future of the petrochemical industry might be at stake, and given the current nexus between agriculture, oil and natural gas, this is no small issue. But if food production were also made independent of fossil fuels, then we would create an even more resilient society. (Such a move might mean many more people would be involved in growing food in such places as their yards, community gardens, and unused urban, suburban, and rural lands. But again, the upside would be an agriculture freed from the dangers inherent in fossil fuel supply forecasts.)

Naturally, the industries and interests which benefit from the world's addiction to fossil fuels would not be in favor of such a society. And, that is why they entertain us with their frequent forecasts of fossil fuel abundance far into the mists of the distant future.

If we were to attempt to make all major activities in global society less exposed to forecasting errors, many of those who currently make their living producing essentially defective forecasts would lose their jobs. No doubt they would forecast a tremendous cataclysm if this were to occur.

Forecasts may, in fact, be desirable in many parts of our lives. We like to know what the weather will be today and tomorrow and perhaps for the week. We seem to have an appetite for the weather report regardless of how many times it is in error. Companies need to project what their prospects will be before they attempt a new venture, release a new product or enter a new market. In each case the forecast a company makes may be right or wrong, but a wrong forecast won't bring down all of society!

And, that is my central point. I am not suggesting that we somehow eliminate forecasting. Rather I am proposing that we structure our lives and society so as to make forecasts largely irrelevant. A resilient, robust society would be one that is not dependent on finite, nonrenewable energy sources. It should be highly decentralized, diverse, redundant and mostly small in scale in its critical functions such as the procurement of energy and food.

Under these conditions forecasting would be far less critical, and errors in forecasting would be far less likely to lead to a civilization-wide catastrophe. This is true for two reasons. First, the energy sources for such a society could not suffer depletion. Second, due to the highly decentralized nature of such a society, parts of it could fail without bringing down the entirety of global civilization. Naturally, there should be arrangements for mutual aid when catastrophes occur. But that aid would be much more assured in a system where most of the parts of the system survive and thrive even in the face of a localized tragedy.

Making society forecast-proof doesn't assure us that no problems will arise. It only helps us to keep those problems smaller, more localized and therefore much more manageable when they do arise. This may one day be seen as a virtue in a world experiencing ongoing global financial turmoil borne of the excessive size of financial institutions and the excessive concentration of risk in them, risk that was poorly understood and therefore poorly forecast.

A simpler, more robust world is possible, and we would be wise to choose it before that simplification is forced upon us by circumstances that we may find exceedingly unpleasant.
_____________________________________________________________

The inspiration for this piece was a presentation featuring Nassim Nicholas Taleb, author of The Black Swan, who now seems to be on a mission to make the world safe from forecasters. He discusses his views alongside Daniel Kahneman, a Princeton psychologist who has studied economic behavior and found it to be far different from what neoclassical economics predicts. Kahneman received the Nobel Prize in economics for his work.

The discussion between these two men focuses primarily on the fragility of the current global financial system and how to make it more robust. It is a stimulating hour of listening.

Listeners might want to consider one obvious implication of Taleb's conclusions. The value of one's investment and/or retirement portfolio is entirely dependent on frequent accurate forecasts from today up to the date of planned liquidation. Should nonprofessionals be risking their savings in this way?

                                                                                               --KC

8 comments:

colinc said...

First, thank you very much for the link to the Edge.org presentation! It wasn't quite as edifying as I was expecting, having read "The Black Swan," but I could listen-to/read Taleb, and now Kahneman, any time!

I concur with your (and Taleb's) observation that "forecasts... are almost always wrong." However, I am a bit dismayed that, despite your implications, you don't actually state the "reasons" for the preponderance of "bad" forecasts.

First, poor predictions stem from the prognosticators' overly limited knowledge, focus or both. In other words, most people making predictions do not possess enough knowledge about the topic and related influences/mechanisms or constrain their focus on too little of that breadth. As an example, look at how the observed glacial and ice-cap melting around the planet are far outpacing even the "worst-case" projections of the IPCC from just a few years ago.

A close companion to the above is a "reliance" on poorly obtained or stated data. That is, the methods used to obtain the relevant data on which to base a forecast were erroneous or, as may be the case regarding peak oil for example, quite likely fraudulent.

While I agree that trying to structure a "forecast-proof" society would be a noble endeavor as well as your recommendation regarding energy, I don't see this happening anytime soon given the selfish, greedy "nature" of the preponderance of the population. Most certainly it won't/can't happen "soon enough" to avert the coming cataclysm.

mattbg said...

Another problem with forecasting: the modified behaviour of people who take the forecast seriously may make the forecast invalid in the end.

And then the people who believed it feel like fools and go back to their old behavior, by which time there's a new forecast based on the old behavior :)

Obviously, it's not really that simple, but I think it plays a part. It plays a part in the formation of bubbles, too -- the consideration of your own behaviour in isolation without thinking about its impact on the bigger picture.

Y2K is a good example. The forecasts of what would go horribly wrong precipitated sufficient action to prevent it from going wrong.

mattbg said...

Also, I'd like to keep weather forecasts because they are quite accurate in the short term :)

But we can get rid of climate change forecasts, I agree.

colinc said...

mattbg, you are absolutely correct in noting "the modified behavior of people" as a factor in "incorrect" forecasting. However, here in northern OH, even the "current" weather forecasts (=< 24 hr) have been wrong more than 80% of the time for nearly a decade now. In other words, weather forecasts have been as, if not more, useless than any other.

Regardless, I'm curious why you want to "get rid of climate change forecasts"? That was most certainly not _my_ contention and, correct me if I'm wrong, I don't think Mr. Cobb would concur, either. My [attempted] point was not merely that the IPCC predictions were wrong, but they were wrong in such a way that paints a much bleaker, nearer future for us all. Even as _that_ "omits" the potential "black swans" of a 5-10% collapse of the Greenland glaciers and the WAIS, which would raise sea-levels 2-3 meters "overnight." The implications of that are much greater and more severe than I've seen "reported" anywhere.

Anonymous said...

Good point, Kurt!
I always think that the way many people are dealing with forecasts is like car drivers who drive very close to the car ahead.
If you'd comment to this sort of drivers "If the car ahead slams on the brakes you technically won't react in time to avoid a crash" they would claim "I'm a good driver and my car has excellent brakes, so don't worry". "Are you sure that you can do it?". "Well, no sure, but don't bother me with these ugly stories".
Here a "resilient" way of driving would be to keep distance from the risks. But for many it seems to be "cool" to take more risks - or they just ignore them.

mattbg said...

colinc, on climate change, I did get your point but I don't know who to believe anymore. Even if it is a problem -- and I don't see evidence of that, either -- I don't think funneling money to some government body in Europe is going to do much to solve the problem.

So, yes, some ice caps are melting. And they might be melting faster than the forecasts. But, so what? Is putting them back together going to solve the problem?

I wish someone would just come out and do at least two things:

1. Address Monckton's questions once and for all, rather than ignoring him.

2. Tell us exactly what we have to do to stem the problem, and what it will achieve.

#2 is the big one for me. I don't see any solution other than to funnel money endlessly into a government body, almost as some of us do to third world charities to assuage our guilt.

Far better, I think, to take things as they come and adapt to the changes that occur. As opposed to pouring money down a hole with no idea of whether or not it will do anything.

Anyway, I used to be a believer in this kind of thing but I no longer am. The refusal to take questions seriously is what turned me off. It is a business and a career path, not a solution to a problem. It is a tool used by people that want to strike back against things they imagine are set against them.

As for Taleb, his talks are interesting but his books are a non-stop waffle.

Anonymous said...

Interesting read.. but I see that the link to the CERA report has been removed. http://cera.ecnext.com/coms2/summary_0236-821_ITM

It is true that given a forecast a society could react to make that forecast become incorrect at some future point. Given the negative outcome of a rapid decline in oil reserves (as well as possible declines in gas and coal) in the near future, then the function of the forecast should be to accurately predict oil production at current rates and taking into account the usual 2% increase yoy.

However, oil and gas use isn't just limited to transportation, they have real uses as chemical feedstock and in road and tyre production.

If there is a real peak soon (and I think there is), then China's current explosion in automobile buying and road expansion will come to naught. In fact, given that Iraq and Saudi Arabia and the Canadian tar sands remain the only swing producers for oil and the fact that India and China will continue to see rapid car population rises, we are looking at a permanent increase in oil prices relative to the past.

Since the transition to electric motor vehicles is hampered by production capacity (battery and motor related because of rare earth metal production), there remains no good alternative to mass-produced oil- and gas-powered personal transportation. We won't see electric vehicle growth as the mechanism by which oil consumption is reduced.

mattbg said...

Anonymous, there are other barriers to electronic vehicles.

They may not even let you buy the battery for the Nissan Leaf, for example. This is a battery-only car. Since it is a lithium-ion battery, you can expect the range to gradually go down as the car ages. People with iPods know how much less the battery holds after a few years of use.

Since they will only let you lease the battery and not buy it, you have a fixed cost.

How to get the most out of a fixed cost? Well, you have to use it a lot to get good value for money. But it'll only go 100 miles on a charge (at best) and then you can either charge at home for a few hours or from a fast-charge station that will do it in 20 minutes. Waiting 20 minutes to "fill up"? If you can find a station.

Maybe a new paradigm in physics would provide a solution, but with the current technology I don't see how these electronic vehicles will help. They might be good for urban dwellers, but people who live in good urban environments don't need a car, and that should probably be the goal.

I think these companies are just taking the approach that it's worth trying and a good investment, and maybe they expect each component to incrementally improve to the point that it's viable. i.e. if someone develops a new battery technology, they can drop it into the Leaf and be ready to go.