Friday afternoon thorts ... as I've mentioned before, absolutely central to finance theory is the concept of the "risk free rate of return". This makes an appearance in Classical economics as "the price of waiting", "the intertemporal rate of substitution", the "rate of time-preference", etc. It's an absolutely indispensable analytical tool for any economics which deals with the concepts of time and uncertainty (or to put it another way; it's hardly used in mainstream economic thought at all).
In so far as mainstream Neoclassical economics gets to grips with time at all, it deals with it by assuming that (since markets are efficient), the rate of time-preference, or the premium one is willing to pay to have "Bread Today" rather than "Bread Monday", is taken to be equal to the rate of return on a risk-free asset. After all, since markets are efficient, the rate of return on a risk-free asset has to be the price of something, and since it can't be the reward to risk-taking, it has to be the reward to pure waiting, right?
Well ... maybe. In actual fact, I have big philosophical problems with the concept of "pure waiting". It seems more or less incoherent to use the normal economic tool of ceteris paribus in this context. If you hold everything else the same except that you allow time to change, then you're saying that nothing changes, but time passes. But is time without change really a sound concept? How could we tell it had passed? But that's not what I'm on about right now.
For the time being, let's accept that "the reward to pure waiting is the rate of return on a risk free asset". What is a risk free asset? Name one. Sorry, MBA students, you just gave the answer "ten year government bonds", you are the weakest link, goodbye. As I've noted before, it's perfectly easy to lose your shirt trading ten year government bonds; they fluctuate quite a lot. They give you a more or less risk free return in nominal terms and over a ten year horizon, but those two provisos include a whole load of items which we don't want to count as part of the pure reward of waiting. Probably best to make a list first of potential investments which can't possibly be risk free:
- Any investment in physical capital (return depends on the rate of profit, which is variable)
- Any investment promising a fixed return in nominal terms (doesn't guarantee the level of consumption you can afford with the returns)
- Any tangible item not directly consumable (what you can buy with it will depend on relative prices, which are variable) -- this includes gold and land.
The most promising candidate left would be index-linked government securities, which promise that, on some future date, you will be able to buy the basket of goods represented by the CPI index, in a quantity which grows at the yield on the index-linked security. But this brings us face to face with the problem indicated above; as a matter of empirical fact, you can get carried out trading TIPS. A ten year bond will let you postpone your consumption for ten years, but your money is locked up in the meantime, and all sorts of things might change over the next ten years. You could get a bit more clever and buy TIPS strips, effectively "locking in" today's entire forward curve of expected inflation, but you're still not really protecting yourself in this way; you're closer than anything else on this list to locking in a guaranteed stream of consumption goods, but that's not the same as postponing your consumption in a risk-free way. It doesn't get you round the fact that in carrying out any such financial transaction, you're exchanging certain consumption for claims on deliveries in the future, during which time anything could happen. Or to put it more bluntly, when I swap an apple today for an apple tomorrow, I'm always at least taking the risk that I might be run over by a bus this afternoon, in which case I'll never get to eat that apple. An asset can't be risk-free in the sense one would need for its return to be the "price of postponing consumption" if it's illiquid.
So what are we left with? Well, basically, that staple of the investment portfolios of millenarians and loonies, canned food. You can guarantee consumption at a future date by storing canned food (by the way, a quick tip to "child-free" whiners; this is the only way in which you can fund your retirement in a way which doesn't make your moaning about having to pay taxes for the education of "other people's children" utterly hypocritical), and you can open the cans and eat it any time you like. But what's the rated of return on canned food? Probably negative if you consider that the quality of what you end up eating is lower than if you'd spent the same cash on fresh food today.
So, I'm turning this one over to the collective wisdom of my readers. Do please feel free to post schemes for securing risk-free exchange of today's consumption for consumption at some future date, which deliver a positive implied rate of return, and I'll report back in a couple of weeks (on the other hand, as a sort of Dadaist joke on the theme of this comment, maybe I won't). The best I can do so far is buying an apple tree which is already bearing fruit, but I can't help thinking this is some sort of cheat .....
1As is apparently customary in what I refuse to refer to as the "blogosphere", I would like to point out that the titular phrase was conceived of entirely by me, in a single act of creativity with no input from any outside source ever. It is also copyright me for ever, as are its component words, which nobody can now use for any purpose at all.
(?xml version="1.0" ?)
ReplyDelete- (data)
- (comments)
- (thread id="81869241")
- (comment)
(datetime)Friday 2002-09-20 15:25:21(/datetime)
(name)Zizka(/name)
(email)zizka[at]vanitysite[.]net(/email)
(uri)vanitysite.net(/uri)
(text)I've posted your "Oligarch" piece on my front page, together with links to other stuff. Take a look if you feel like it. I'll read some of the stuff you mentioned.
PS. You could also post your email on your site??
=============
- (comment)
(datetime)Friday 2002-09-20 15:26:38(/datetime)
(name)Rob Schaap(/name)
(email)rschaap[at]iprimus[[[[[[[[[[.]]]]]]]]]]com[[[[[[[[[[.]]]]]]]]]]au(/email)
(uri)http://blogorrhoea.blogspot.com/(/uri)
(text)'k'n' brilliant, big fella!
Again.
That said, stuff like 'efficient markets' and 'pure waiting' are surely taken as stuff you have to learn to get a degree rather than stuff you're allowed to believe at yer actual place of gainful employment,, no? It's more part of a system-legitimating liturgy than Homo Armanirolexicus's real belief system, ain't it?
=============
- (comment)
(datetime)Friday 2002-09-20 15:54:35(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)cheers mate. My linking & email policy is a bit obscure since, for a variety of reasons, I don't want this weblog to become high-profile enough to be "noticed".
it isn't hard to get hold of my email, btw; check out DeLong's comments section. This is my compromise between anonymity and whatever the opposite of anonymity is.
*LOVE* the piece on your site.
=============
- (comment)
(datetime)Friday 2002-09-20 16:08:40(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)>>That said, stuff like 'efficient markets' and 'pure waiting' are surely taken as stuff you have to learn to get a degree rather than stuff you're allowed to believe at yer actual place of gainful employment,, no?<<
Well this is the question that I'm struggling with. I honestly believed up until about six weeks ago that, although there were more problems with economic theory and neoclassical analysis than you could shake a stick at, the foundations of finance in the mathematics of compound interest were sound. Ie, I thought that discounted cash-flow was a legitimate way of analysing the world.
Then I got into a short but bloody flame-war with Paul Davidson on the PKT list on the subject of actuarial tables and ergodicity, and kind of had my foundations rocked. This flirtation with post-Keynesianism is pretty new for me; if you check out the LBO archive, you'll see a few occasions on which I spring to the defence of finance as being quarantined from the problems in the rest of neoclassical macro.
What I'm doing in the unpopular techie bits of this blog is trying to make a few base camps for an assault on the Everest of a proper reading of Keynes' General Theory. I'm warming up at the moment by trying to play myself into Davidson's latest "Financial Markets, Money and the Real World", but it's f**king difficult. The Keen book helped a lot.
So no, you'd be surprised, "they mean it, man". It's even warned against in Brealey & Myers (the standard MBA textbook) that "an MBA student who has mastered DCF analysis is like a baby with a hammer; everything looks like a nail".
=============
- (comment)
(datetime)Friday 2002-09-20 17:24:19(/datetime)
(name)b(/name)
(email)brennan[[[.]]]peterson[at]stanford[[[.]]]edu(/email)
(uri /)
(text)One great investment over history has been lots of children, as long as you instill in them a proper work ethic and a parental care ethic. And an early retirement age coupled with a youth-funded social security like program....
B
=============
- (comment)
(datetime)Friday 2002-09-20 18:07:57(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)yeh, but not exactly either risk-free, or liquid :-)
I just realised that buying a house probably counts, as it allows you to pay for a stream of housing services up front, as it were, but I have problems with the fact that for a finite-lived individual, the return in terms of consumption sacrificed in order to buy one might be negative unless you count the capital appreciation, in which case it wouldn't be risk-free.
=============
- (comment)
(datetime)Friday 2002-09-20 19:22:56(/datetime)
(name)Maynard Handley(/name)
(email)name99[at]redheron[[.]]com(/email)
(uri /)
(text)Daniel I'm not sure quite what you are getting at here. Perhaps it would be worth a second post clarifying exactly your problems.
You seem upset that there is no absolutely secure way to replace gold today with gold tomorrow that can ensure that the gold you receive tomorrow is "really" what you deserve, rather than possiblt being screwed by events that happen between now and tomorrow---inflation, new gold mines, confiscatory taxes on gold. So the future is unpredictable---what else is new?
One imagines with non-treasury financial instruments a split into the "genuine interest rate" plus some extra interest as compensation for the uncertainties of the future---default etc. One can likewise imagine a treasury factored that way.
Or if you prefer, why should the "price of waiting" be what it is---well it has a "risk free" waiting component, plus a "compensation for uncertainty" component---but since you can't ever create a perfectly risk free environment, all you will ever see are the two components linked, as in the price of treasuries.
In more technical terms (if you know the language of differential geometry), it seems to me you are asking for a connection in some space, a mathematical construct that tells you how a scalar (value today) changes when you move it to a different point (value tomorrow). You make the underlying assumption that such an EXACTLY-DEFINED connection exists. But the mathematical theory doesn't have to be constructed that way. The connection could be stochastic. Now iit is obvious that there is not a single well-defined value for gold tomorrow---there is a range of values as specified by some pdf attached to the connection, and the range gets wider as time goes by. To collapse that range into a single value, you can take the mean, or perhaps a mean weighted by the personal tastes of all participants in the market (who may weigh a downside more strongly than an upside).
So I can't understand quite what your concern really is? Are you just pointing out that "risk-free" as used by the finance guys is exactly so? Are you making some philosophical point that bad things can happen in the future. Are you complaining that the mathematical machinery people use (at this simple-minded level) to discuss the future is inadequate (but bette machinery exists).
Maynard
=============
- (comment)
(datetime)Friday 2002-09-20 19:27:32(/datetime)
(name)Jeremy Leader(/name)
(email)jleader[at]alumni[[[[.]]]]caltech[[[[.]]]]edu(/email)
(uri /)
(text)Heh, I mis-read the footnote at first, and thought you were claiming copyright on "blogosphere" and were going to withdraw that term from circulation.
=============
- (comment)
(datetime)Friday 2002-09-20 21:28:35(/datetime)
(name)a different chris(/name)
(email /)
(uri /)
(text)Maynard, read the post again. IMHO he isn't "concerned" somehow about risk-free assets, like we're supposed to be his investment advisers, he's trying to build a model to take us somewhere and he's missing a piece.
D, I think you might have to constrain things to economic forces. For example, leave out early death (which screwed up the canned goods) or uninsured acts of God. For instance, now the house can't blow down in a tornado but it can still be affected by some unexpected social trend. Like everybody decides it's best to move back into the trees and start civilization over again with hopes of a better outcome.
=============
- (comment)
(datetime)Friday 2002-09-20 22:53:04(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)Maynard: different chris (whose weblog I would surely read if he had one ...) has it abut 80% right.
Your points are all correct but slightly orthogonal. I'm guessing you're a hard scientist or mathematician of some sort rather than an economist ... the point is, as you say, there is no reason why a sensibly constructe mathematical theory should require a single necessary connection between today and tomorrow. The point is (as I allude to in the comments, this post is rather self-indulgent and not really for an audience), several of the most important results of finance theory depend absolutely on the correct theory being the special case that you describe, and I'm poking some gentle fun at them by pointing out how unrealistic an assumption this is. So it's the third; I'm complaining about the adequacy of the mathematical toolkit.
As to whether a better kit exists ... I don't know, and I suspect that the answer lies either in Paul Davidson's book or in a closer reading of Keynes. I will make more posts on this subejct, and I may try to riff off your own insightful remarks about the connection between past and present as being a geometric relation in state-space.
I'd note that your assertion that, say, the future gold price can be determined today as a draw from some pdf is actually a quite restrictive assumption to make about that state space. Specifically, you're assuming that the problem of predicting the future is ergodic. Paul Davidson certainly denies this, and I'm currently in the process of taking his view seriously.
Bur anyway, my point, if I had one, is that you're right; this post is clearly incomplete and unsatisfactory and the reason is that I haven't quite grasped the underlying theory. I'd be very grateful if you could continue to make these comments on future posts in this series.
=============
- (comment)
(datetime)Friday 2002-09-20 23:11:35(/datetime)
(name)Jason McCullough(/name)
(email)jason[at]hronk[.]com(/email)
(uri)http://zebco.blogspot.com/(/uri)
(text)Shouldn't a "least risky rate of return" suffice just as well as a risk-free rate of return?
=============
- (comment)
(datetime)Saturday 2002-09-21 03:00:05(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)not really, for mathematical reasons I can't go into here because I'm drunk.
Basically, a least risky rate of return would work for practical purposes, but it wouldn't be a rate of time preference, and so it wouldn't serve all the analytical purposes it's needed for.
In actual fact, you can do a lot of finance without a risk-free asset if you believe in a "zero-beta" asset; one whose return is uncorrelated with all other risky assets. But that's as much of a theoretical construct as the risk free asset.
=============
- (comment)
(datetime)Monday 2002-09-23 17:00:57(/datetime)
(name)Froomkin(/name)
(email)froomkin[at]law[.]tm(/email)
(uri)LAW.TM(/uri)
(text)For US consumbers, a passbook savings account is almost riskless for the first $100,000 since it is insured by the government. "Almost" since when the bank goes under there's delay in getting your $$ back from the feds, and you lose any interest owing but unpaid. Close enough?
=============
- (comment)
(datetime)Tuesday 2002-09-24 22:23:43(/datetime)
(name)Jeremy Leader(/name)
(email)jleader[at]alumni[[[[[.]]]]]caltech[[[[[.]]]]]edu(/email)
(uri /)
(text)Froomkin, what if the bank failure is tied to other bank failures, and the government fund can't cover all the losses? Or what if your account is fraudulently emptied by some third party in such a way that the bank isn't held liable? For xample, if the fraud is so clever that only you (and the criminal) know for certain that you didn't take the cash? Low risk, yes, but arguably non-zero.
=============
- (comment)
(datetime)Tuesday 2002-09-24 23:08:26(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)Heh, that's a clever one which I didn't think of. Are there index-linked passboook accounts, though? I wouldn't swear that there weren't, but I'm not aware of any. So far you're in the money position though
=============
- (comment)
(datetime)Wednesday 2002-09-25 03:27:50(/datetime)
(name)froomkin(/name)
(email)froomkin[at]law[.]miami[.]edu(/email)
(uri)http://www.law.tm(/uri)
(text)RE: Jeremy Leader's comment, if there's fraud you have a legal right to the money back, usually from the bank, if it's not your fault. In any case, there is no investment so riskless it protects against your own idiocy. And, yes, the cost of recovery is expensive (that's us lawyers). I think the systemic risk is almost negligible in the US. If that's not the case, then it certainly follows there's no investment denominated in local currency that meets your criteria, since anything big enough to take down the banks takes the rest of the economy with it most likely.
On the index linking point, I know of no such investment other than US government index-linked bonds. Hmm. Why aren't those (1) better than passbooks and (2) better for model-makers since anyone can hold them. Oh yeah, the time problem. That's what we want to solve.
But you know, I would have thought that it would be possible to trade a derivative product, some sort of indexed-T-bill future that would allow payments at any given point. So that's it! you buy bond, top up with a portfolio of derivatives, spend far more money than the interest on the bond was worth, and Voila! you have just proved that the optimal riskless investment loses you money. (It follows that the most risk-averse people take risks, because having no risk is a sure money loser. Ok, got that. This is why I'm not an economist. Twelve impossible things before every meal.)
But wait! We're not done! This money-loser is a bad investment, we must diversify...into cash and just puts and calls (no bond). Now here I have no idea what the cost structure is, I imagine it's not good, but maybe good enough for playing econogames? (And of course, the people who want to price the puts and calls need to know the value of the riskless investment to price them right, which is recursive which, oh never mind.)
=============
- (comment)
(datetime)Wednesday 2002-09-25 17:48:41(/datetime)
(name)dsquared(/name)
(email /)
(uri /)
(text)I can't find it on their website, but I'm pretty sure that the UK's National Savings people run a passbook account which is index-linked. National Savings is explicitly guaranteed by the UK government and pays interest gross of UK tax, so in principle, a non-UK taxpayer could, risklessly and maintaining liquidity, trade future consumption for present at a positive rate of return.
I therefore declare Michael Froomkin the winner. This will be announced in a forthcoming article; if I get round to it, there may even be a prize.
=============