?

Log in

No account? Create an account

Previous Entry | Next Entry

Keynesian Experiment: The Explanation

OK, time for an explanation, especially as someone got all four correct answers.

The experiment you have been participating in is what is known as a 'Keynesian Beauty Contest'. Keynes first described it in his 'General Theory of Employment, Interest and Money'. The point of the beauty contest is that it explains movements in share prices.



I first asked you to say who you thought was the most beautiful woman from two lists of ten. I also pointed out that there was a prize for picking the two "most popular", but I didn't explain what I meant by that.

Now instead of thinking of beautiful women, think about the most valuable shares on the stock market. At this point, we have to consider what makes a share valuable. Fundamentally, what makes a share valuable to an investor is the share of the company's profits (in the form of dividends) that the investor expects to receive because he holds that share. There is a secondary factor, which is the capital gain that the investor hopes to realise by selling the share in the future for more than he bought it for. However, hopefully you can appreciate that the future selling price also depends upon the dividend income that the next shareholder is going to receive, and so on. So the capital gain aspect of the share's value is always secondary to the primary aspect, which is the dividend stream.

At least, that is the basic theory. Unfortunately, many stock exchange movements are not explained by a simple function of future dividend income. Keynes noticed this.

Let's get back to our beauty contest. Many of you probably voted in the first part of the experiment for the women you actually thought were the most beautiful in the two lists. Personal preferences no doubt played a part in this just as they would in picking a company to invest in. Picking the women "with the least slap" might be equivalent to picking the companies with the most respected CEOs or the most ethical behaviour or the best write-up in brokers' briefings. Some people (certainly at least one going on the comments I saw) didn't necessarily go for the women _they_ thought were most beautiful but (perhaps with their eyes on the prize) went for the ones that would rank as "most popular" - whatever that meant.

Looking at the voting in part one, we had a broad range of votes in the first half of the poll, with only poor Marisa Miller failing to pick up a vote and Kristen Stewart getting slightly more than Keeley Hazell and Frankie Sandford. The second half had a clear leader in the shape of Charlize Theron, who picked up seven out of the twenty votes.

Now when we move to part two of the experiment, the questions change subtly. I revealed that the first list was actually from FHM's 'World's Sexiest Women' poll and the second list from a now defunct women's fashion magazine, Eve. When I asked you to name the women who you thought would have been top in FHM and Eve respectively, I'm giving you some information about the market. You have some idea of the sort of women that FHM readers find sexy and fashion magazine editors consider beautiful, and that extra information shapes your pick. That's the equivalent of selecting shares based on what you think investment analysts like (for example if you know that brokers tend to like companies with substantial cash reserves or a strong presence in China).

Finally I asked you to vote for the woman that the most readers of this journal thought was most beautiful, and I suggested that you could look at the results of the first poll. If you look at the results of this final question, you see two very clear winners - Kristen Stewart and Charlize Theron, both by a landslide. These are _possibly_ the correct answers because no doubt in the very first question many people voted for the woman they thought most beautiful. But, we also know that some people voted in the first poll on the basis of what _other_ people thought. We also know that not everyone who voted in the first poll voted in the third.

The number of votes for Kristen Stewart and Charlize Theron in the third poll depends not on the inherent beauty of the two ladies but on voters' perception of what other voters' think. Moving back to the stock market analogy, the price of shares in Kristen Stewart plc and Charlize Theron plc has shot up - not because they have become more beautiful / likely to pay more dividends, but because investors think that other investors think they are. You can take this further an additional step - investors think that other investors think that other investors think they will pay more dividends in future.

If we were to ask ten more people the third question and also showed them the results of the first question, what do you think they would answer? Most would probably vote for Kristin Stewart and Charlize Theron, so the two women's lead would continue to increase, without their underlying beauty increasing. In the stock market, the share price for Kristin Stewart plc and Charlize Theron plc would keep on rising, without any underlying increase in the two companies' profitability.

What we have created is a speculative bubble in Kristen Stewart and Charlize Theron!


Now at some point, someone is bound to say "Charlize Theron ain't all that; Angelina Jolie is much better looking" or "Kristen Stewart is all right I suppose, but she's no Megan Fox". And then someone will agree, and so will someone else and it will become clear that Kristen Stewart in particular is overvalued. And that's when the speculative bubble bursts and prices plummet back towards (or even below if there is an over-correction) a level more appropriate to the intrinsic value of the investment.

"I only voted for Kristen Stewart because everyone else was; really I think Abigail Clancy is the best looking."

"I bought dotcom stocks because the prices kept on rising even though few of the companies ever looked like making any money".


Keynes was mostly concerned with equities (i.e. shares), which when you consider he was writing not long after the Wall Street Crash, isn't surprising. However, you can get speculative bubbles in other investment assets. The event generally considered to be the first recorded speculative bubble was the Dutch 'Tulip Mania' of the early seventeenth century, when the asset was forward contracts to buy tulip bulbs (that is a contract to buy a set number of tulip bulbs for a set price at the end of the season). Where the asset has a less obvious intrinsic value, speculative bubbles may well be less likely.

Before I reveal who it was who got the answers correct, I will briefly mention something called the 'Efficient Markets Hypothesis', As you might expect (at least you will if you've been paying attention to my earlier economics posts), economists disagree on the subject of speculative bubbles and how common they are.

The Efficient Markets Hypothesis (very much a 'freshwater' theory) says that because financial markets are 'informationally efficient', it is not possible to consistently 'beat the market', at least not when you take risk into account. There are three forms of the hypothesis, 'weak', semi-strong' and 'strong'.

'Weak' EMH says that asset prices always reflect all past publically available information and will eventually adjust to new information. 'Semi-strong' EMH says that asset prices always reflect all past publically available information and will instantly adjust to new information. 'Strong' EMH says that asset prices not only reflect all publically available information, but also all non-public 'insider' information. There have been lots of empirical studies aimed at proving or disproving the different forms of the Efficient Markets Hypothesis. Saltwater economists would generally harken back to Keynes and point out that EMH does not allow for behavioural effects like the beauty contest. Nevertheless, there is _some_ evidence to suggest the weak and semi-strong forms _can_ apply. There is also evidence that over-reliance on EMH and a failure to recognise behavioural effects played some part in the recent credit crunch. Both sets of economists would agree that inadequate or uneven information played an important part. I'll come back to that in a future post.




OK, time to name the winner. Congratulations chris_maslen for correctly identifying Cheryl Cole as FHM's 'Sexiest Woman in the World' and Kristin Davis as Eve's 'Most Beautiful Woman in the World'.

Your prize Chris? Well, it's one worth thousands of pounds, maybe even more.

It's a piece of advice:









Learn to spot a speculative bubble and don't invest in it.

Tags:

Comments

king_of_wrong
Jan. 18th, 2011 11:47 am (UTC)
The weak and the strong forms I have no problem with, just the semi-strong one... it annoys my inner physicist as it seems to imply time travel.

I realise it probably makes the maths much simpler, and I can certainly believe that the propagation delay is only a fraction of a second, but...
philmophlegm
Jan. 20th, 2011 02:47 pm (UTC)
I suppose the logic of the semi-strong (and I'm not saying that I believe the semi-strong EMH) is that many markets are supposedly so efficient nowadays that information does propagate phenomenally quickly. As soon as someone offers to pay over the odds for a share, everyone selling that share _instantly_ thinks "Hang on, if he's willing to pay more, it must be worth more" so the price _instantly_ rises.
king_of_wrong
Jan. 20th, 2011 07:13 pm (UTC)
Makes sense.

Of course, I'm coming at it from the other angle, and seeing the milliseconds of lag between someone seeing the over-the-odds price and getting their orders/cancellations to the exchange, or the (extremely temporary) spread opened between prices in Frankfurt and London due to a large order taking a bite out of the liquidity before the market makers can re-fill. In that case, their mental prices may change instantly, but the actual bids/offers on the exchange are still out there...
philmophlegm
Jan. 20th, 2011 07:23 pm (UTC)
I think you're right. I also think that any implementation of EMH beyond weak will inevitably fall over at some point because of behavioural effects.