Investing for the Long Slump

by Eliezer Yudkowsky1 min read22nd Jan 200954 comments


Personal Blog

I have no crystal ball with which to predict the Future, a confession that comes as a surprise to some journalists who interview me.  Still less do I think I have the ability to out-predict markets.  On every occasion when I've considered betting against a prediction market - most recently, betting against Barack Obama as President - I've been glad that I didn't.  I admit that I was concerned in advance about the recent complexity crash, but then I've been concerned about it since 1994, which isn't very good market timing.

I say all this so that no one panics when I ask:

Suppose that the whole global economy goes the way of Japan (which, by the Nikkei 225, has now lost two decades).

Suppose the global economy is still in the Long Slump in 2039.

Most market participants seem to think this scenario is extremely implausible.  Is there a simple way to bet on it at a very low price?

If most traders act as if this scenario has a probability of 1%, is there a simple bet, executable using an ordinary brokerage account, that pays off 100 to 1?

Why do I ask?  Well... in general, it seems to me that other people are not pessimistic enough; they prefer not to stare overlong or overhard into the dark; and they attach too little probability to things operating in a mode outside their past experience.

But in this particular case, the question is motivated by my thinking, "Conditioning on the proposition that the Earth as we know it is still here in 2040, what might have happened during the preceding thirty years?"

There are many possible answers to this question, but one answer might take the form of significantly diminished investment in research and development, which in turn might result from a Long Slump.

So - given the way in which the question arises - I know nothing about this hypothetical Long Slump, except that it diminished investment in R&D in general, and computing hardware and computer science in particular.

The Long Slump might happen for roughly Japanese reasons.  It might happen because the global financial sector stays screwed up forever.  It might happen due to a gentle version of Peak Oil (a total crash would require a rather different "investment strategy").  It might happen due to deglobalization.  Given the way in which the question arises, the only thing I want to assume is global stagnation for thirty years, saying nothing burdensome about the particular causes.

What would be the most efficient way to bet on that, requiring the least initial investment for the highest and earliest payoff under the broadest Slump conditions?

Personal Blog