Saturday, October 04, 2008

Two lessons I'm taking from current world financial crisis and US government bailout.

1) US-style capitalism doesn't work without government involvement. As Ayn Rand once put it, real capitalism is an unknown ideal. Who knows how ideal it is? Certainly not the US, which offers us no evidence, one way or the other.

2) Markets "don't know nuffink". Seriously, all the arguments about mark-to-market etc. reveal that these grand distributed information processing machines are pure fraud. What exactly is the worth of XYZ asset? Well, we don't know. We must stop mark-to-market because what the market is telling us is "wrong". That's why we must pump more money in, to convince the market to "know" something else about the value of these things.

Yeah, right.

Markets do not discover the "correct" value of things. We've just let free-market ideologues paint us into an epistemological corner where we assume whatever the market tells us is the correct value. Except when a crisis comes along, like today, and the result is so painful that we try to convince markets of something else.

The crisis is not because the price is too high or too low. The crisis is because markets are too volatile to be relied on as generators of knowledge. The world didn't dramatically change between the pre-crisis and now. The knowledge of the world which is incorporated in markets didn't dramatically change either. What dramatically changed were the expectations of how other agents in the market would behave. In other words, the dramatic re-valuation of assets reflects only endogenous feedback loops inside the market, not an epistemic process responsible for enabling beliefs to reliably track the world.

Now we realize that such a mechanism for generating accurate knowledge about the world barely exists. The information gathering function of the market is swamped by the "betting on other agents' behaviours" part.

Of course, none of this will stop the hypocrites who will pocket their government hand-out and go back to selling us the idea that the market is the best information aggregator that exists.


Oli said...

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

My take is that regulators should re-examine the current idea that "speculation = useful liquidity"

Instead I think the assumption should be that "speculation = market distorting instability" and therefore should (in general) be banned.

I haven't thought about this much but maybe we could remove the dangerous short term speculation from the market by doing something like insisting that people can only buy shares if they intend to hold them for a minimum of (say) 1 year.

As with bonds you could only get your money out before this time by paying some kind of significant penalty cost, thereby removing any incentive to undertake short term speculation.

This kind of 'buy to hold' trading would be perfectly fine for the perfectly 'good' forms of long term investments (like pension funds) that help churn the money usefully around the capitalist system.

I presume too that such an approach would make investors be more cautious about who they invest in and then only invest in companies that they have reasons to be confident in over the longer term.

So, IMHO, it's the speculation that enables the rapid sentiment driven feedback loop that then distorts the market away from finding a ''''correct'''' value.

But, I find it hard to know what other kind of 'correct' price a good can have other than the price that other people are willing to pay for it.

Hence, IMHO, some kind of mark to market accounting is the only sensible way to value a given asset.

However, what seems absurd is the idea that this mark to market accounting has to happen in near real-time, which appears to be the case in the US and that article you link to shows the destructive downward spiral that such real-time mark to market accounting has been causing.

The suggestion in the article of using a 3 year rolling average sounds like a sensible kind of approach. Maybe for faster moving goods you would need a shorter time span, but clearly using anything near real-time causes dangerous feedback loops.

What doesn't make any sense to me is any kind of approach that effectively allows the companies to 'make up' the value of the asset based on what they hope to be able to sell it for in the future.

Making up asset prices would surely be another recipe for disaster.

Oli said...

I've been thinking about this more - and realising that I really want (and need!) to read and learn more about economics as it's so non-trivial and so important.

In particular I kept thinking how I should have said something more like: "the 'correct' price for something is a price that the seller is willing to sell at and the buyer is willing to buy at". Hence the 'correct' price is the price at which a trade occurs within some kind of market.

But of course the more I think about this the more it's clear that at the very least such a definition of 'correct' leads to a range of price points at which trades will occur within a market.

... and then of course it's really a dynamic market with feedback loops.

I wonder if 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?