A chart that I came up with earlier this year. Blue line is the Dow close at the end of each calendar year. Curves are simple geometric plots, increasing at a rate such that it will intersect with the actual Dow as of the year indicated in the legend.
If you want to figure out where the Dow “should” be, you pick the rate of growth that makes the most sense to you, extrapolate to the present day, and you’re done—provided you have a talent for oversimplification.
I’m certainly no expert in finance, but I’m trying to apply a few lessons I learned in anthropology back in grad school. Namely:
- People don’t change much, even when their cultures do.
- We’re no different.
So it seems to me that if you want to believe that the years 1930-1980 have any particular relevance to us, then you come to one of two conclusions about any major delta in the rate of increase of the Dow:
- We are different.
- It’s a temporary aberration.
There are all sorts of reasons why we can call ourselves different; our grandparents didn’t spend their evenings Twittering about the lovely upgrades they made to their Hoovervilles. But most of the time, when we think we’re different, we’re kidding ourselves.
I’m perfectly comfortable with the idea that things changed after 1980; computers were arriving on people’s desks, and whether we knew it or not, the Cold War was about to end. As I understand it, that means all sorts of things about international capital flows and other financial voodoo, all of which is designed to make money too complicated for mere mortals. Hence my oversimplification: yes, the 1980s and early 1990s were different enough to account for that curve.
As for the second bend in 1994, well, sure. The Internet was becoming a major force right around then, and again, whether we knew it or not, economic forces were starting to realign.
The third bend ending in 1999, though, is clearly batshit. Even our massive housing bubble in 2006 couldn’t peak us anywhere near that growth curve.
I can see all sorts of logical fallacies in this kind of thinking, any one of which would mean that I wasted around 18 hours teaching myself how to use iWork Numbers well enough to do this chart. I’m especially unhappy with the timing; I don’t know where to intersect those curves in order to capture a hockey stick change, rather than merely intersect at its endpoints.
That said, here are the numbers at the end of each curve for 12/31/2009, in case they turn out to be illuminating or prophetic someday: 1980: Dow 1,848.40; 1990: Dow 5,124.81; 1994: Dow 6,842.23.
Hey JP,
Your reference to Chiroptera guano in 1999 may be an apt one. If bats were tracking their ability to ascend the heights of trees in the years leading up to gliding or flight, they might record, “moment 1980: 1,848cm high; moment 1990: 5,124cm high; moment 1999: holy shit, we’re frickin’ gliding on frickin’ air currents here! [plop]” No way in hell could pre-flying bats have any way of predicting what life would be like once their progeny took flight. Sitting bass ackward on the tree trunk don’t do diddly to explain it.
One interpretation of your chart is that we actually approached (or are approaching) the mythical asymptotic moment (aka “The Singularity”) in financial markets; where old rules no longer predict near term future state. Certainly that’s what I’m seeing in the mortgage markets this week. Absolutely insane by any other world view, I would seem…
Interesting times indeed.
Dann
First of all, I like this analysis a lot. The curves show the different phases our economy has been through over the past twenty years, and the space between the curves speaks to the relative shift in each phase.
The thing is, the hyperbolic volatility of the last nine years is masking similar volatility in the past. Try drawing the same graph from just 1930-1936, or 1970-1976, for instance. If these were your end points and you drew your geometric curves, you’d see the “most recent” few years as abberations too.
But, as you say, this is just playing with graphing programs. The hard part is drawing conclusions. My take is that the markets perpetually go through cycles – both up and down. Seismic shifts in the economic paradigm, like the ones you mentioned, vary those cycles both in terms of length and amplitude (in the case of the DJIA: 1930-1980: 100-1,000; 1980-1990: 1,000-2,500; 1990-1994: 2,500-4,000; 1994-1999: 4,000-12,000). Five years from now, if the Dow is at 18,000 or more, then this was a hiccup. If it’s at 5,000, then this was the start of a Japan-style “lost decade.”
@Dann —
And the poor bastard who hitched a ride on the space shuttle completely screws with the bat community.
@Brian —
Absolutely agreed that I’d need to post a logarithmic chart to accurately capture volatility that is flattened out on this scale. But the fact remains that 1930-1980 is much longer than 1999-2009.
So what I take from this, purely as social statistics, is that it’s a good retroactive barometer for how crazy things are right now. The 1980 curve is the “true” curve until the early 1990s, when it’s clearly no longer predicting anything. The problem with the 1990 and the 1994 curves is that they never map to actual changes, so while I think they might provide a more reliable prediction at an arbitrary point, it’s completely arbitrary where you might want to pick up their predictive effects. (That said, if the Dow closes between 5125 and 6842 this year, yes, I will call myself a genius.)
I played around with creating a second curve fit from 1980 to 1994 as yet another prediction, but that ran into two problems: 1) if you want to continue adding curves to the data, eventually you’re not modeling anymore. (Steven Wright line: “I bought a map of the United States. Actual size.”) 2) To me, the reassuring aspect of this data is precisely that it doesn’t pretend that the years 1930-1980 never happened. Add more curves, and lose 70% of your data.