Re: Lyle's Laws

Hal Finney (hal@rain.org)
Mon, 6 Jan 1997 15:39:36 -0800


From: "Peter C. McCluskey" <pcm@rahul.net>
> hal@rain.org (Hal Finney) writes:
> >You then form an infinite sum, adding each estimated future dividend
> >multiplied by the discount rate appropriate for that distance in the
> >future. This gives you the total discounted future return for the stock,
> >and that, according to this model, would be the stock's value for you.
> >Average this over all market participants and you get the stock price.
> >
> >Now consider how this model works as we approach an era where industrial
> >production grows explosively, due to nanotech, or AI, or self-reproducing
> >cheap robots, or some other breakthrough. This can cause returns from the
> >stock to begin increasing at a rapid rate, even faster than the discount
> >factor reduces them. The result could be that the infinite sum above
> >diverges! The stock will in fact, today, have an infinite value to you.
>
> This assumes the discount rate is unaffected by the explosive economic
> growth. But what sane person would use a discount rate that is significantly
> lower than the economic growth rate?

I was using a different model of the discount rate, where it is a
psychological measure of the pain inherent in postponing present
gratification for future reward. A dollar today is more fun than a
dollar next year, and the degree to which this is true for you will
determine your own personal discount rate. People whose discount rates
are higher than average (who hate waiting) will tend to be borrowers,
and those with lower than average rates will tend to be lenders.

It is true that this model does not obviously explain why discount rates
tend to equal long term growth rates, so it might be more plausible to
assume that they track each other. But it is not clear to me how the
existence of rapid growth will affect people's willingness to forego
pleasures until tomorrow.

> It makes no sense to pay more for a stock than it would cost to
> create a similar new company. While there some assets that can't
> be quickly duplicated (Bill Gates' brain, Nabisco's reputation, etc),
> I don't see how they can explain why people would pay orders of
> magnitude more for stock shares than it would cost to start a
> similar company.

Good point. By my reasoning this would suggest that people would expend
resources to start new companies as well, until some limiting resource
raised the cost of doing so.

> >This would not be a conventional speculative bubble, where people invest
> >just to get temporary gains from the ongoing price rises, but rather a
>
> The majority of people who buy near the peak of a bubble think they
> are long term investors buying at the start of a new era (usually with
> rationalizations like those in Lyle's parody).

That may be true, but what I meant was that those people were wrong all
those other times, and this time they would be right. So it really is
fundamentally different from a speculative bubble in actuality, even if
it may appear superficially similar.

> >I would expect the financial explosion to occur a few years to perhaps a
> >few decades before the productivity explosion, depending on how farsighted
>
> What would cause the value of labor to decline relative to capital
> years before the cost of duplicating human-level intelligence is
> reduced?

I don't understand what you are getting at here. I argued that the
value of productive investments is more than commonly recognized, and
that eventually the value would therefore increase. This would imply
that the value of labor will decline relative to capital, not because
labor is less valuable, but because capital would be more so. To take
an exaggerated example, spending your money on a robot which will build
a million more for free would be better than spending it on a hairdresser.

> What would cause the value of cash and cash-like assets to decline?
> One of the reasons that these assets have value is the ability they
> provide to quickly move one's assets into new (unanticapted) investments.
> Since my intuition is that the period you are talking about will
> produce enormous uncertainties about which companies will produce
> or benefit from breakthroughs in nanotech or AI, I expect I will
> value cash fairly highly.

Cash is not a productive asset, so to the extend you hold it you
are foregoing the income that you could have made by investing it.
If investments are growing much faster than today, then holding cash
will be more costly.

However you may be right that being able to move deftly into various
investments will be important for some people. Others will prefer to
adopt a diversified position. It's hard to say now how evenly the
benefits from the new technologies will be spread. However surely
in the present day the argument implies that the best strategy is to
invest productively so that you benefit from the future "catastrophe"
(as in catastrophe theory, a sudden change to a new set of stable points).

Thanks for your comments. I am not really adhering to this theory, but I
am interested in its plausibility.

Hal