Tuesday, October 07, 2008

Oli had a couple of good replies to my last "markets" post. I'm putting my response here 'cos it's more discoverable and readable than in the comments there. But read what he says.

"Surely you're being a bit too broad in your comments here."


Nah. I think it's just you who's getting bogged down in the minutiae. :-)

But good comments, though ...

My thoughts :

1) No one thinks speculation is a good idea. People have always been against it. The problem with speculation is that there's *no* principled way to distinguish it from use of the market in the "right" way.

In both cases people buy now, hoping to sell later for a profit. In both cases people look to the behavior of those around them to learn about the world. (The whole point of a market is that aggregates information publicly, so you can't distinguish "speculators" from "honest investors" by whether they base their judgment on prices.)

"Speculation" can be suspected when prices are increasing rapidly. But there can usually be a plausible story told about why prices *should* be rising rapidly this time. (Eg. the internet changes everything.)

Then speculation is only widely acknowledged *later* when there's a crash and people realize that the prices were "wrong". (By which time, by definition, you've admitted that it's possible for there to be difference between the market price and the "reality")

2) You could *assume* that the "speculators" were the high-frequency (day-traders) and try to filter out the high-frequency movements (ie. by forcing a minimal holding time; or by, say, imposing a tax on purchases which reduces the viability of frequent trades for small gains.)

What you're up against here is a) whether the high-frequency signal really *is* the speculation. And b) whether the market loses certain flexibilities by filtering out the higher frequencies.

As I understand it, the market seems to exhibit "drunkards walk" properties (ie. random at all frequencies). The only real "trend" is a steady, underlying increase in share prices. But I suspect that tells us very little about the world and mainly tracks the increase in the money supply, so it's just a kind of inflation. [Update : See update at end of for me pulling back from this claim.]

As an aside, remember that when people talk about the stock-market "outperforming" the rate of inflation, it's nonsense. Shares are a "good" just like anything else. If their price is increasing, that's inflation. All "outperforming" means is that they're increasing at a faster rate than salaries, goods, services and the rest of the economy. In other words, the government is increasing the money supply and the capitalist class is grabbing an increasingly large proportion of it.

3)

Oli:
the current financial crisis has partly arisen because not enough theoretical work has been done on the manner of these feedback loops and therefore the kind of system dampening that is required to prevent wildly destructive effects.

At this point I realise that I should read more before pondering more, but Phil, is this the kind of thing that you've been exploring with your OPTIMAES project?


During the 20th century economics has focussed on a particular kind of mathematical model : ie. analytic treatment of equilibria. Only recently has the mainstream starting moving away from this to think about other kinds of models (informational ones, agent based ones, cybernetic ones.)

I'm particularly interested in looking at markets from a cybernetics / dynamical systems perspective. There's a rather fascinating underground current of thinking in the 20th century which is agreeably "left wing" and holistic, one that seems to encompass cybernetics, a more liberal organization theory, some psychoanalysis, spirituality, alt.money etc. (Think Stafford Beer or Francisco Varela)

What distinguishes this tradition from mainstream economic theory - which is far more game theoretical, methodologically individualist and appealing to the right - is a belief in feedback, non-linearity and emergence (not just of order but also emergent catastrophic events). I think that economic thinking needs (and is going to get) its next breakthroughs in understanding from this cybernetic / dynamical systems tradition. And the method of discovery will be computer simulation. So, of course, OPTIMAES is all about this :-)

Update : there's an interesting diversion on the comments where I'm discussing the long term trend of share-prices rising. I hypothesized above that this may just be in line with the growth of the money supply. In fact, Darius later convinced me that shares can rise in line with growth of goods and services rather than purely inflation due to the money supply increasing. (I guess I was running with a sort of gold-bug prejudice that I picked up somewhere.) So, ok, I pull back from that earlier suggestion.

11 comments:

Oli said...

Wasn't the original purpose of the stock exchange meant to be an easier way to enable investment in companies?

If a venture capitalist or other kind of investor is asked to invest in a privately held company (rather than a publicly traded company) it would be very odd for them to buy and then sell shares in that company within a single day.

Indeed, I would find it odd for them to invest for a period shorter than about a quarter.

So, although it's great that stock markets allow people to invest much more easily in a wide range of companies for much smaller values, I see no reason why some kind of scheme encouraging a minimum preferred holding time couldn't work.

Or, in other words, your saying that 'they' find it hard to define speculation. Why isn't it simply that there is 'bad' short term speculation and 'good' long term speculation=investment?

Composing said...

Wish I could find a good, easy and reliable reference to the idea that share-prices are actually a random walk.

Here's a start.

http://knol.google.com/k/sam-vaknin/technical-and-fundamental-analysis-of/2ji9lwjy5dg4e/25#

Oli said...

But are you (or 'they') arguing from theory that stock prices will always end up being a random walk, or that in practise when people do analysis of actual stock market prices then they find them to be statistically equivalent to a random walk?

IMHO I'd be surprised if it's the former, but the latter would suggest to me that the short-termist speculation (drive by roughly random human sentiment) is completely masking any underlying signal from the value of the company.

This to me would be proof that stock markets are not working in a healthy way at all.

You also suggested that the rise in value of a share of a company is equivalent to a form of inflation, but this again suggests to me that there's something deeply wrong with the way that our stock markets are working.

Imagine I buy for £10,000 a 25% stake in a company that makes enough profit each year to distribute £4,000 in dividends to its shareholders. I therefore will get £1,000 a year in return for my £10,000 investment - which is a 10% return, hence (currntly) better than putting this money just in a bank.

Imagine that 2 years later the company has now grown (or become more profitable in some other way) and is now able to distribute £8,000 dividends from profit per year. Hence I am now getting £2,000 per year which is now a 20% yield on my initial invested capital.

If I were to now sell my shares I think it would be reasonable to hope to find a buyer willing to pay £20,000 for these shares so that the new buyer will get back a 10% return on THEIR initial investment.

So, overall I've had a better return on my money than leaving it in a bank and I've also managed £10,000 capital gain because the share price went up.

Although some of the increase in the value of the profits and share price might be accounted for by inflation, certainly not all of it.

I can't see why the stock market doesn't follow the same logic when talking about share prices over years. Companies that grow above the rate of inflation should also see their share price grow above the rate of inflation - and this should be a clear signal in their share price.

Hence there really should be clear signals (at least over the medium term) in share prices.

Composing said...

1) Re: drunkard's walk. I believe that some people have claimed to have found the randomness from analysing the data, rather than simply deducing from first principles.

But I also believe it's a contested claim. I'm not sure what the state of the art research says. Trying to find something.

[quote]but the latter would suggest to me that the short-termist speculation (drive by roughly random human sentiment) is completely masking any underlying signal from the value of the company.[/quote]

That's definitely the corollary. Personally, I find it very plausible. That's really the claim I'm making here (and some time ago on ThoughtStorms

[quote]This to me would be proof that stock markets are not working in a healthy way at all.[/quote]

Yes.

2) I think my point about share-prices and inflation is this. If your company goes up, but that gets balanced by a corresponding fall in other profits and other shares so that the average share-price or aggregate stays the same, then there may not be inflation.

But if the averages and aggregates are all going up ... then clearly that extra money is coming from somewhere. Either the government is "printing it" or it's being siphoned off from somewhere else in the economy.

Actually, I don't know if this is the case (ie. that the average is going up), or if it is, which of these is the cause (I suspect a bit of both), but that is really the only way it can be.

It seems disingenuous to try to compare stocks with inflation, when stock prices have to be figured into the calculation of inflation.

Oli said...

Now I'm confused (again :( ).

I thought inflation was (roughly speaking) the average rate at which prices rise for things that ought to remain at the same price.

This would IMHO exclude things like share prices which are expected to really change price because of fundamental changes to the companies.

But you're suggesting that inflation is (or should be) the rate at which the average price for all things rises.

.. but then I don't see why all growth isn't rendered irrelevant by inflation.

I'm confused and intrigued by how baffling even these basic issues about money are.

As I've said already, I think it's high time that I pick up some basic economic text books (to sit on a shelf with all those other books I'd love to read one day :( )

Composing said...

Agreed. It's fascinating, and very hard to get your head round. And there are actually quite different schools of thought (eg. Keynsians, Monetarists / Austrians etc.) who disagree about quite fundamental issues.

I tend to think of inflation as (roughly) the money supply outstripping the goods and services available. And that you *can* increase money as long as the goods and services are increasing.

But then I'm brought-up short by the commodity-money people eg. people who think we should use (finite) gold-backed money. These people ask why you need to grow the money supply at all.

A sandwich may cost 2 Reais today. If we double our productivity and make twice as many sandwiches tomorrow, why shouldn't we just drop the price to 1 Real each? Why increase the money supply in order to maintain the sandwich to Real parity?

It's an interesting question. I can imagine certain answers. If productivity goes up, but the quantity of money remains fixed, then prices will keep dropping.

But does that mean no-one will ever *invest*, because it's better to hoard your gold than put it into factories and make widgets?

Once you start trying to think through this stuff, the weirder it gets. :-)

dariush sokolov said...

"I tend to think of inflation as (roughly) the money supply outstripping the goods and services available. And that you *can* increase money as long as the goods and services are increasing."

That's right. Basic monetary economics:

There are disputes between keynesians /monetarists etc. about inflation in the short run, but in theory economists agree that in the long run (when we're all dead) the quantity theory identity holds. In the simplest form:

MV = PQ

where M is money supply
V is the velocity of money (how quickly money is actually changing hands, notes being passed around etc.)
P is the price level
Q is the actual quantity of real stuff - shoes, sandwiches, etc. - being traded

if you suppose velocity is constant and, for now, that the quantity of stuff (the real economy) doesn't change - there is no growth - then

dP/P = dM/M

i.e. % change in price level is just the % change in the money supply

inflation is just a monetary phenomenon

but if you have a growing economy then can see that to keep prices constant you need to grow money supply at just the same rate as real growth

otherwise, if money growth doesn't keep up, indeed you will have deflation

deflation is generally thought to be a bad thing, for corresponding reasons to why inflation is thought to be a bad thing

at the very least, it's inconvenient - you have to keep changing the prices of everything in the shops, e.g., which is costly. and i guess confusing economic decisionmaking over time - it's more complicated to make long term financial decisions discounting inflation/deflation, even if the prices changes are constant ... and then if there's uncertainty about the changes ... Also lots of psychological reasons i guess.

now if you fix the currency to gold isn't that just hostaging price change to a completely arbitray factor? it'll just be sheer luck if the gold market moves in step with economic growth. e.g. now to make intertemporal economic decisions i need to predict likely gold discoveries in the future ...

dariush sokolov said...

no one thinks speculation is good?

effiency of markets, we're told, depends on arbitraging away inconsistencies between different markets. e.g., simple cuirrency markets, imagine independent bicurrency trading markets:

GBP = 2 USD
USD = 2 BRL
GBP = 5 BRL

now obviously if i can move between these different markets i can go to the GBP/BRL market buy 5 reais there for a quid, then sell them for 2.5 dollars, and change those dollars back for £1.25, making a quick 25% profit

this is called arbitrage. buying cheap and selling dear, like mercantilist traders.

it has no direct part in investing in real stuff, but it arbitrage trading is (so it's said) essential to creating efficient markets. this is the market's mechanism for removing these distortions - once traders spot the price difference such as that one in real/ sterling marlets, speculative traders rush in until they have shifted pricing to "efficient" levels.

Composing said...

Thanks Darius. Good comments.

1) I'm pretty sure the gold-bugs aren't expecting the discoveries of gold to track the growth in the economy. They like gold *because* it doesn't change (much). And they *do* want deflation in this sense. At least, that's how I read the people who I've talked with.

2) I'd say that arbitrage *is* recognisably different from speculation. My definition of speculation being something like : buying stuff to sell tomorrow *only* because you think there's a trend towards price increases. As opposed to buying stuff to sell tomorrow because you think that the thing will appreciate value for its own reasons.

The distinction is a bit fluffy exactly because there's no *behaviorally observable* difference. That's my orignal contention about markets qua knowledge producing institutions : they can't distinguish speculation from non-speculation.

In fact, *we* can only distinguish by bringing in some kind of folk-psychological explanation with "beliefs" of the agents.

John was speculating when he bought the house *because* he thought house-prices were just going to go up and up, and it would be a great investment.

Jane was NOT speculating when she bought the house because she thought that the new sports-centre that was being built would make houses in this area more attractive to young families and it would be a great investment.

John is bringing no information and no ratiocination to the market, he's just gambling on what the mass of other economic agents will do. Jane *is* potentially bringing new information and ratiocination to the market : knowledge of the coming sports-centre and the inference that it will attract more people to live nearby.

Very, very informally I think people realize this. Speculation was a dirty word for Victorian industrialists (read Dickens). I do think people have a picture of speculators as mere gamblers who are different from "useful" investors (or even arbitrageurs).

The problem is that the *market* doesn't know the difference. It's a fundamental blind-spot of the mechanism.

Composing said...

Update to my last comment. Unless, of course, you follow Oli's line that high-frequency transactions more likely to be speculation than investments, so you filter them out.

I agree with Oli, this is possibly as good as it gets, but it's still a fairly blunt instrument.

Oli said...

I suspect that one could make a principled distinction between speculative and arbitrage trading.

Speculation is when you buy in one market and sell back into the same market.

Arbitrage is when you buy in one market and then sell into a different market that is trading in the same good but at a different price point.

I can see why arbitrage is probably a desirable thing as it allows traders to spot when one market is oversupplied and another is undersupplied with a given good.

Given the clear difference between the two types of trade, can we allow one without allowing the other via a 'back door'?

A first guess is that we could apply a penalty cost to short term speculative trading when you sell back into the same market from which you bought.

But, I think this would just allow people to speculate between two tightly coupled markets. Arbitrage trades would keep the price point in the two markets very similar thus allowing speculators to buy one day and sell the next on the expectation that BOTH markets will have risen.

Hence, this simple approach would still allow high frequency speculation.

So, how about this: If you sell back into the same market there is always a penalty cost. However, if you sell into another market then the penalty cost is discounted in some relation to how different the price points are between the two markets.

So, if the price for the good is identical between the two markets then the penalty cost for high frequency speculative trades is the same for both types of trade.

However, the more the prices differ between the markets, the lower the penalty cost, until at some point there is zero penalty and hence a large incentive for traders to undertake arbitrage trades.

Thus we'd get effecient arbitrage between markets with no back door for short term specuation.

What do you think?