It would be fascinating to see the integral of this one. It looks like the dow is seesawing around the middle, visually, but we know that in fact it's growing, so it's spending proportionally more space above the middle than belo.

It's a little bit tricky, because it's entirely a matter of coincidence that I was able to use the same units for both graphs on the Y-axis. The Dow a year ago was 14 thousand, while GDP was 14 trillion. It's beautifully convenient. I just consider the Dow to be shown in dollars while GDP is in billions of dollars.

So, I could arbitrarily move either line up or down in comparison to the other without changing anything meaningful. This is one reason I like the graph that's based on growth rather than absolute values; there's nothing arbitrary about the units.

I have to head out now, so I'll do the integral graph in a bit.

If I'd done a graph that showed, not percentage change, but rather simply showed slope (absolute change), then the integral of the graph would be the original graph itself.

My first worry about making an integral of a percentage-change graph is that the results could be misleading. For low values, a negative change of 50% means very little. For high values it means a lot. So we're not necessarily looking at a meaningful aggregate change if we take the integral of percentages.

For example, suppose you're looking at the average percent change over 6 months. The initial value is 1000. At the end of each month the values look like this:

100, 10, .1, .01, .001, 1000

The percentage change for each month looks like this:

-.1, -.1, -.1, -.1, -.1, 1000000

The average percentage change over the six months is:

166,700

Which is meaningless because the actual value both began and ended at 1000.

... so it would make more sense to do an absolute value integral rather than a percentage change, true. Possibly just for the Dow because that's the sawtooth one.

Ah! I think I may see what you're getting at. Here's the integral of nominal GDP and the Dow over time. It's like the volume of money that's accumulated in each one.

Is this what you were looking for? Let me know if it isn't so I can fix it.

I can also obviously make this in log-scale inflation-adjusted dollars, if you like. It's actually pretty quick and easy :)

I'm afraid it's been a long time since I posted this so I can't be sure. I usually I use sites like Google and Yahoo! finance for stock quotes. GDP was almost certainly from bea.gov. (See http://www.bea.gov/national/index.htm#gdp).

For non-stock data to be used in serious numerical analysis I generally limit myself to US government sources like bea.gov, bjs.gov and usda.gov (http://www.ers.usda.gov/Data/). These sites all have spreadsheets available for download.

The Dow isn't very many companies. Depending on what you want to find out, try doing it with the S&P 500 if you just want to look at big "blue chip" companies or Wilshire 5000 if you want to see them all.

Theoretically, that's true, but in reality if you look up the Dow, the S&P 500 and the Wilshire they mirror each other extremely well. I generally prefer the Dow because 1) everyone has heard of it and 2) it's been around in its present form for much longer than the other two.

I do not have a comment but an observation. I could be wrong. But the dip @ 2001 in the neighborhood 8,000 was not present in several other charts I've seen. For the DOW

Is it possible you were looking at charts used by financial advisors? These usually show, not stock prices, but rather what you will actually earn if you invest in the markets... so they factor in dividends.

Supports your GDP graph with the DJIA graph and growth. Thank you for these. I am a teacher and will be using your site/blog as reference. If you have any other graphs, please post.

## Comments

cossoffistiqueTill I get the log scale working right, here's the data displayed in terms of growth:

soffistiquecoslookslike the dow is seesawing around the middle, visually, but we know that in fact it's growing, so it's spending proportionally more space above the middle than belo.soffistiquethousand, while GDP was 14trillion. It's beautifully convenient. I just consider the Dow to be shown in dollars while GDP is in billions of dollars.So, I could arbitrarily move either line up or down in comparison to the other without changing anything meaningful. This is one reason I like the graph that's based on growth rather than absolute values; there's nothing arbitrary about the units.

cossoffistiqueI have to head out now, so I'll do the integral graph in a bit.

If I'd done a graph that showed, not percentage change, but rather simply showed slope (absolute change), then the integral of the graph would be the original graph itself.

My first worry about making an integral of a percentage-change graph is that the results could be misleading. For low values, a negative change of 50% means very little. For high values it means a lot. So we're not necessarily looking at a meaningful aggregate change if we take the integral of percentages.

For example, suppose you're looking at the average percent change over 6 months. The initial value is 1000. At the end of each month the values look like this:

100, 10, .1, .01, .001, 1000

The percentage change for each month looks like this:

-.1, -.1, -.1, -.1, -.1, 1000000

The average percentage change over the six months is:

166,700

Which is meaningless because the actual value both began and ended at 1000.

cossoffistiqueIs this what you were looking for? Let me know if it isn't so I can fix it.

I can also obviously make this in log-scale inflation-adjusted dollars, if you like. It's actually pretty quick and easy :)

K, I really have to head out now. Damn parties...

soffistique(Last one. Really. :)

cossoffistique(Anonymous)http://vicakin-thedow.blogspot.com/2009/03/djia-3-9-09.html

(Anonymous)soffistiqueFor non-stock data to be used in serious numerical analysis I generally limit myself to US government sources like bea.gov, bjs.gov and usda.gov (http://www.ers.usda.gov/Data/). These sites all have spreadsheets available for download.

(Anonymous)soffistique(Anonymous)in the neighborhood 8,000 was not present in several other charts I've seen. For the DOW

soffistiqueHere I've zoomed in to show more detail:

And more:

Do you have any thoughts as to what's going on?

Edited at 2009-03-08 12:08 am (UTC)(Anonymous)(Anonymous)http://2.bp.blogspot.com/_eCZSJq_Wcpw/SbXCG7k7yhI/AAAAAAAAAfw/qRNZIBKutew/s1600-h/djia-gdp-1885-2008.JPG

http://vicakin-thedow.blogspot.com/2009/03/djia-3-9-09.html

soffistiqueIs it possible you were looking at charts used by financial advisors? These usually show, not stock prices, but rather what you will actually earn if you invest in the markets... so they factor in dividends.

soffistiquehttp://2.bp.blogspot.com/_9SoEE9d_aQo/SRZGBzx-GUI/AAAAAAAABV8/AnLsKCsZZ4k/s400/graph08-div.gif

(Anonymous)soffistique