Dow 36,000
Brad DeLong produces the following chart:
The Dow-Jones Industrial Average has spent 25 of the past 40 years–62%–between 800 and 1250 or between 8000 and 12500. These ranges comprise roughly 25% of the (logarithmic) total range of the Dow as a function of time.
It really looks like there may something special about the nominal value of a 1 followed by a bunch of zeros.
This is disturbing to me as an economist.
This seems like a proposition we could test by looking at other stock indexes. Nor do I really see why it should be all that shocking. We understand that expectations about market participants’ future behavior are relevant to rational market behavior. And it’s well known that decimal counting systems have an important role in human psychology. The tendency of goods to be priced at $9.99 rather than $10.01 is pretty well-understood.
Meanwhile, countries have been known to conduct their monetary policy in part by lopping off zeros. You’re in a situation where you’ve experienced a lot of inflation over the past twenty, so instead of $1 being equal to 600 francs, it’s now equal to 6,000 francs. To curb the inflation, you need to take some dramatic steps in “real” policy. But creating a “new franc” such that 1 new franc is equal to 1,000 old francs can also plausibly be part of the solution — it changes people’s thinking, and it symbolizes the new determination.
Maybe along these same lines we need to arbitrarily re-mark the nominal values of all our stocks, so that the DJIA will be equal to approximately 36,000 (to pick an arbitrary number) and then we can watch it slowly drift upwards until it gets into the 100,000 range. I’m not familiar enough with the workings of the system to say exactly how you would do that as a technical matter, but surely it can be done — it would just be a question of changing the meaning of labels.
48 Responses to “Dow 36,000”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
October 30th, 2008 at 11:19 am
We can just start pricing our stocks in yen.
October 30th, 2008 at 11:21 am
Speaking of Dow 36,000, has anybody heard from Assman and Asshat lately?
October 30th, 2008 at 11:28 am
Most people with I.Q. scores above room temperature have long since figured out that $9.99 really means ten bucks. Stock market investors, who for whatever their other faults generally aren’t stupid, surely would see right through this re-marking scheme.
October 30th, 2008 at 11:31 am
One of the first economic policies De Gaulle implemented when he came to power with his coup in 1958 was … creating a “new franc”.
October 30th, 2008 at 11:33 am
Interesting that the flatting of the prices coincides with a large increase in the trading volume. Since GNP has not been flat, there must be other reasons why the prices flattened: the wealth moved to other stocks or other types of investments, or profit was paid out in forms other than price appreciation.
October 30th, 2008 at 11:34 am
If you tell people you’re re-norming to achieve some behavioral goal, that kind of kills the efficacy. I think Jon Elster once noted that if you drink a glass of warm milk before going to bed, you can’t will yourself into going to bed just by drinking a glass of warm milk.
October 30th, 2008 at 11:35 am
One of the first economic policies De Gaulle implemented when he came to power with his coup in 1958 was … creating a “new franc”.
Right — I wasn’t picking the example at random.
October 30th, 2008 at 11:38 am
Stock market investors, who for whatever their other faults generally aren’t stupid
All the evidence to the contrary.
October 30th, 2008 at 11:41 am
Benford’s Law? as many of the people on Delong’s blog have pointed out.
October 30th, 2008 at 11:50 am
It’s two data points. Not convincing.
October 30th, 2008 at 11:54 am
Lots of discussion of Benford’s law in the DeLong thread, though this is stronger.
I think it’s clear that round numbers, nominal or not, carry weight. So while technical analysis is basically stock-market entrail-reading, markers like $2-to-the-pound or 100 yen to the dollar aren’t the same as other values.
October 30th, 2008 at 11:55 am
What malraux said. Seems like a typical case.
October 30th, 2008 at 11:57 am
Or maybe it’s just coincidental, and we’ve really reached a tipping point in the economy where we can no longer sustain all this debt, both personal, business, and national. Frankly, I don’t see the American economy getting any bigger at this point. We’ve gone too long on borrowed money, individually, corporative, and national. At some point, the taxman will come to collect. I think we’ve reached that point.
Additionally, I was struck by a point made some time ago about the Japanese NIKKEI. It peaked in the late 80s at like 39,000 points or so. Today it is at 8000-9000. Does anyone foresee the NIKKEI getting back to the 39,000 level anytime soon, or ever in its future? I don’t see it happening.
October 30th, 2008 at 12:01 pm
Benford’s law doesn’t really offer an explanation here, because it doesn’t predict that such a high percentage of the time would be spent in these ranges. It does, however, point out that when you start measuring various real word phenomena with, say, a base-10 number system, there really is something special about numbers that start with 1.
October 30th, 2008 at 12:03 pm
What’s “special” is that if you elect a Bush to the Presidency, things turn to shit on a major scale.
October 30th, 2008 at 12:07 pm
Nah.
October 30th, 2008 at 12:07 pm
I’ve looked at all the other indices I can find: it only works for the Dow–which is the salient index…
October 30th, 2008 at 12:10 pm
Not necessarily. They would need to function with significant independence from the DJIA.
You could claim support for positive evidence if they flattened at round numbers at different times than the DJIA, but you couldn’t necessarily claim the converse.
While it might seem that multiple indices flattening at the same time would indicate a genuine economic root cause, it might be that the Dow flattend for psychological reasons, and other indices were then influenced by the flattening of the Dow.
You would need to observe a sample of indices in which flattening occurred as often just before flattening in the Dow as just after.
Then you can have subjective arguments about what exactly constitutes flattening.
Somehow, I don’t think I’ll have to face the dilemma of whether to sell stocks and buy gold when the DJIA hits 100,000.
October 30th, 2008 at 12:14 pm
just so any economists, or PR or advertising people are reading this – the question isn’t “$9.99 vs $10.01″, its “$9.99 vs $10.00″ flat, no $.01. why on earth would anyone add a .01? why subtract it either? back in the stone ages, some people may have fallen for the “$9.99 is less than $10!” gambit, but it doesn’t fool anyone anymore. and if I’m in a store and one product is priced $9.99 whereas a comparable one is $10 – I buy the $10 one, because at some level, you’re dealing with more honest people/policy/pricing. just saying. subtracting the penny does not fool anyone.
October 30th, 2008 at 12:20 pm
Puh-lease. First, the chart highlights two periods as a kind of norm, but in one the index fluctuates between 569 and what looks like around 1100, or in other words 100% or its low, while in the other it fluctuates by roughly an equal percentage, from around 7500 to 15,000. Yet it uses these periods of relative stability to attribute magical significance to 1000 and 10,000.
Second, the period of growth in between is longer than the rest put together, suggesting at the very least that more than the psychology of the endpoints is at work. Third, it makes no reference to the years at hand, as to when such sudden, er, irrational exuberance may have occurred and why it ended. At the very least, it’s not grappling with what actually happened.
And of course as others suggested it’s preposterous anyhow. You don’t look for secret codes to the human psyche without some evidence of what people do or say or write. Maybe a little more time thinking and less time playing arbitrarily with log and other scales might suggest some actual respect for statistical evidence.
October 30th, 2008 at 12:49 pm
You don’t need to reprice the stocks. You just need to change the index divisor to get the index to change prices. It wouldn’t do anything, since indicies are composites. I really doubt that stocks move or stop moving based on where the index is. Do you really think this affect IBM’s or AT&T’s stock movement?
October 30th, 2008 at 1:16 pm
Oh please.
Let’s assume that the phenomenon is real — which would totally shock me, but I’ll go with it.
All we need to do is publish the findings in a persuasive way. Problem solved. Why? Because then all sorts of quant-type hedge funds and other investment vehicles will read the findings and immediately start shorting straddles around the relevant numbers — in other words, betting that the market will hang around that area (http://en.wikipedia.org/wiki/Straddle). Derivatives like straddles exert a pressure on the underlying security, and this will quickly eliminate the effect.
Markets are really good at equilibrating. All you need to do to eliminate a particular irrationality once you’ve discovered it is to announce the irrationality.
October 30th, 2008 at 1:23 pm
By the way, this has nothing to do with Benford’s Law. Benford’s Law would apply if DeLong had noticed that the index value usually starts with a “1″ in the leftmost digit — not that it starts with a 1 AND ends with a bunch of zeroes.
October 30th, 2008 at 1:44 pm
If the example is not random, the conversion rate seems: the new franc was worth 100 old francs, not 1,000.
October 30th, 2008 at 1:58 pm
I agree with JohnH — this is pretty meaningless. The DJIA is also one of the worst stock indices to try to do comparisons over time. It represents a small and essentially arbitrary selection of stocks relative to the S&P500 and others.
October 30th, 2008 at 3:04 pm
This seems like a proposition we could test by looking at other stock indexes. Nor do I really see why it should be all that shocking. We understand that expectations about market participants’ future behavior are relevant to rational market behavior. And it’s well known that decimal counting systems have an important role in human psychology. The tendency of goods to be priced at $9.99 rather than $10.01 is pretty well-understood.
But the Dow’s an index, a function of the value of 30 stocks. I could understand your argument if we were talking about a given stock, that it might tend to float around 100 or 10 or some other round number. But people don’t really trade the Dow. It’s just random.
October 30th, 2008 at 3:32 pm
A good example of this would be Brazil renaming its currency the “real” in order to boost confidence in its value.
October 30th, 2008 at 4:22 pm
Pender,
The EMH doesn’t hold quite as strongly as you think. There is a good deal of price clustering, and it has been shown empiricaly by a series of papers published over the period of a decade.
See a write up of the research here[shameless plug of my website].
Doesn’t the idea of quants spending their time constructing complicated Straddle portfolios to capitalize on this stuff seem a bit silly?
Maybe one day, but as of now they have bigger fish to fry.
October 30th, 2008 at 7:39 pm
Not at all. Straddles are not complicated derivatives. They are constructed by writing a put and a call with the same strike price and expiration date. I bet you could do it yourself on e*trade.
The charts on your website are interesting, and I agree that the market is not perfectly efficient with respect to the first decimal place — but there is a huge difference between clustering around integers (for which it would be difficult to construct derivatives without losing all the benefit in transaction costs, and which doesn’t really matter anyway since the distance from the “efficient” price is necessarily very small) and clustering around 10,000 (for which it would be absolutely trivial to construct a derivative and for which the rewards would be large).
Another way of phrasing the difference is as that between the price getting stuck to the nearest 1.0 and the price getting stuck to the nearest 10,000.0. The inefficiencies would not be even remotely comparable, so solving the efficiencies, on top of the difference in transaction costs, would yield vastly different rewards.
October 30th, 2008 at 7:58 pm
Pender,
Thank you, your point was actually the one that I was trying to make. I was referring to the difficulty of constructing straddle portfolios to take advantage of the integer clusters.
But the effect is a bit larger then you think. Since the study analyzed individual stock prices, it’s really rather likely that the sum over all stocks makes the distortions in the index somewhat big.
But, obviously any such structure would be more complicated than “hover over 10,000″.
October 30th, 2008 at 8:55 pm
The solution to this problem is obvious. We should all start counting in binary, so that a 1 followed by a bunch of zeroes happens more often.
It will proportionally multiply our happiness. And, while we won’t have learned how to talk to the animals, we’ll have learned how to talk to our computers. I think that would also be nice.
October 30th, 2008 at 10:25 pm
Given that the DAX for as long as I can think has been somewhere between 4000 and 6000, and (insert other world index hovering in arbitrary integer range here), I think either the phenomenon is not real, or there is something peculiarly base-10 about the American mind.
On the other hand, the fact that the Dutch AEX seems to be around 1/10 of the DAX suggests that, um, Hitler was right, or something equally stupid.
October 30th, 2008 at 10:51 pm
I’m skeptical, but in fact you understated the phenomenon. The range of 800 to 1250 is less than 20% of the logarithmic range. log(1250/800)=.1938…
October 30th, 2008 at 11:24 pm
Quite the reverse, I’d think — since each stock would have its own individual random component, and since each random component would be less than one and otherwise independent of the others, you’d expect them to more or less cancel out in a broad-based index thanks to the law of large numbers, and you’d similarly expect them to cancel out over time given the frenetic nature of such small movements in the market.
Regardless, my first point — and the one I’m the most confident in — is that if this “stick at 10,000″ thing were true, and you published a persuasive argument to that effect, then enough people would sell straddles to cause it no longer to be true. The derivatives are extremely simple to construct, would be extremely profitable as long as the irrationality exists, and would be extremely effective at killing off the irrationality.
October 31st, 2008 at 9:33 am
If you meant “people have eyeballs” then yes, you are correct.
If you meant “we know why this is effective” then no, you really don’t. There was a recent study into this and it’s not the answer you assume.
March 1st, 2009 at 6:36 pm
cialis
Very interesting site. Hope it will always be alive!
March 11th, 2009 at 4:41 am
It is the coolest site,keep so!
March 12th, 2009 at 11:28 pm
I want to say – thank you for this!
March 22nd, 2009 at 6:16 am
tramadol
Great site. Good info
March 22nd, 2009 at 10:29 am
buy viagra online
I want to say – thank you for this!
April 3rd, 2009 at 4:14 am
I want to say – thank you for this!
cheap brand pfizer viagra
April 9th, 2009 at 6:08 am
Great site, Good info viagra