Felix Salmon

Eisenhower charts of the day

Felix Salmon
Jan 21, 2011 22:09 UTC

My fabulous editor, Jim Ledbetter, had a party a couple of days ago for his new book about Dwight Eisenhower. He asked for the most famous passage from Eisenhower’s 1953 “Chance for Peace” speech to be turned into updated charts, so here you go:

schools.png power.png
hospitals.png highway.png
wheat.png homes.png

Sources: The price of a bomber, according to the Air Force, is $1.2 billion in 1998 dollars, which works out to about $1.6 billion today. It costs $18.5 million to build a school. For the power plant, I’m assuming energy usage of 11.4 kW per person (obviously this is up sharply from 1953) and a cost for building a power plant of $1,050 per kW, which works out at about $700 million. Hospitals are coming in at about $260 million apiece. Highway costs are about $10 million per mile.

A fighter costs $150 million; a bushel of wheat is $8. Destroyers run about $1.75 billion apiece; and construction costs on a new single-family home are $222,511.

And here’s the passage in question, which still carries enormous force:

Every gun that is made, every warship launched, every rocket fired signifies, in the final sense, a theft from those who hunger and are not fed, those who are cold and are not clothed.

This world in arms in not spending money alone.

It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children.

The cost of one modern heavy bomber is this: a modern brick school in more than 30 cities.

It is two electric power plants, each serving a town of 60,000 population.

It is two fine, fully equipped hospitals.

It is some 50 miles of concrete highway.

We pay for a single fighter with a half million bushels of wheat.

We pay for a single destroyer with new homes that could have housed more than 8,000 people.

This, I repeat, is the best way of life to be found on the road the world has been taking.

This is not a way of life at all, in any true sense. Under the cloud of threatening war, it is humanity hanging from a cross of iron.


I’m not sure this ends up saying what I think you intended.

You’re arguing that military expenditure is destruction of capital (true) and is becoming increasingly expensive (which may or may not be true), but you are ignoring possible increases in efficiency in the production of consumption goods.

The cost differentials between relative expenditures could be caused by the increase in the bomber cost, or by a decrease in the the cost of producing the compared resource. I’d go back to the drawing board and include a comparison between the [real] cost difference between the 1950s item cost and the today cost, and then compare it to the bomber.

Posted by ARJTurgot2 | Report as abusive

Chart of the day: California taxes

Felix Salmon
Dec 7, 2010 18:02 UTC

ARJTurgot2 left this comment on my chart of US taxes:

You are, of course, going to follow up this chart with a second one that comprehensively reflects the changes in State and Local taxes, especially including sales taxes, that have changed since 1950. And that data is going to include things like registration and usage fees, especially gasoline, telecommunications and sin taxes on things like liquor and cigarettes. I understand that is going to vary widely from state to state, so two, perhaps, should be instructive: say New York and California?

If someone wants to point me to a dataset which gives me that information, I’ll happily chart it. But in the meantime, I pulled table D1 (Californian GDP) and table M13 (Californian state tax collection) from the California statistical abstract. That only gives data from 1967 to 2007, unfortunately, and the GDP series changes slightly in 1997. But in any case, here’s the result:


It seems to me that tax revenues have been floating pretty steadily around roughly 5.5% of GDP since the mid-70s, with a brief blip up to a high of 6.8% during the dot-com bubble. I’m sure that the recession has brought the ratio down of late. But what I’m not seeing is any indication that the decline of federal tax revenues is made up for by a concomitant increase in state tax revenues.


For a long time California was the reddish state that gave us Reagan, and present tax boundaries locked into California statute are a reflection of this.

As the influence of public unions increased, we moved to our present situation where prison guards in California earn six figures and eye-popping retirement packages are the norm. The result is a rather Greek-like combination of generous public spending and low tax receipts with massive bond sales to make up the difference.

At least California has Google, Apple and wonderful and productive agriculture. Under higher taxes the farms at least would have to stay put.

Posted by DanHess | Report as abusive

Volume-based stock chart of the day, flash crash edition

Felix Salmon
Sep 24, 2010 16:17 UTC

Here’s the volume-based stock chart you’ve all been waiting for: the one for May 6, the day of the flash crash. Since the big spike in volume was concentrated at the end of the day, in the final hour of trading, the time-based chart squeezes a huge amount of activity into a relatively small horizontal space. The volume-based chart gives the crash a bit more space.

Volume vs Time - SPY - 20100506.jpg

On the other hand, it’s worth nothing that most of the day’s trading still took place before the crash happened.

On thing that strikes me about this chart is not the crash itself but rather the run-up to it: the initial drop from about 1,160 on the S&P down to about 1,120. On the time-based chart, the decline starts slowly and then rapidly speeds up; on the volume-based chart, however, it’s much steadier. And in fact we saw roughly as much volume in the normally-quiet hours between about noon and 2:40 as we did during the craziness of the crash itself and its aftermath. I’m not going to hazard a guess as to what this means, but I do think it shows that May 6 was a pretty unusual day in the markets even before the flash crash happened.

Many thanks to Omer Uzun at Proteus Financial for putting this together: it’s only one tiny piece in the puzzle, but surely every little bit helps.


The top axis and bottom axis use different timescales, if you look closely.
The top axis divides the day into ~ 40 min increments, and shows the amount of trading done there.
The bottom axis divides the day into 10% of TRADE VOLUME…so the time areas are variable. The % adds up naturally, but what it really shows is what % of trades are getting done in what time periods, very precisely. The top bar doesn’t convey volume, just activity.

As for reason…the only thing people can point at now is a sell off on options relating to the S&P 500, I believe. The order was so massive it borked the chain, and the HFT’s who comprise 70% of trading volume exited the market, crashing liquidity and driving prices down to unreal levels as counterparty demands evaporated.

There’s probably more to it, but we’ll see.


Posted by REDruin | Report as abusive

The new type of stock chart

Felix Salmon
Sep 22, 2010 20:13 UTC

A couple of weeks ago, I wondered whether it was possible to see what a stock graph would look like if it split up the x-axis according to volume rather than according to the time of day. After all, when trading is concentrated at the beginning and end of the day, those are the areas worth concentrating on, right?

Wonderfully, Omer Uzun of Proteus Financial rose to the challenge. And here’s the result:

Volume vs Time - SPY.JPG

The chart splits the 390-minute trading day between 9:30am and 4:00pm into ten chunks. On the normal stock-market chart, seen in blue, each chunk is 40 minutes long. But on the volume-based chart, in red, the first 10% of trading is already over after 17 minutes, while at lunchtime it takes over an hour to see 10% of the daily volume change hands.

I’ve only seen one of these charts so far, but at first glance it does seem as though the volume-based chart seems less volatile — smoother, somehow — than the time-based chart. I wonder what the equivalent chart for May 6 would look like.


I invented volume bar charting 7 years ago. Having researched them and published a few articles on them I can honestly say they are far superior to time based bars because they do not contain any inherent variable aspect like typical charts contains. The markets are traded in volume, NOT TIME. The problem with getting the industry to look at this chart type is that a majority of “market professionals” are closed minded. Thanks for opening the door a bit.

Wm Schamp

Posted by ProfLogic | Report as abusive

Adventures in probability, market forecasting edition

Felix Salmon
Sep 20, 2010 19:27 UTC

Carl Richards has a cute graph:


The basic idea here is right. And in fact Richards understates, in his graph, just how bad things are when it comes to market forecasts: his graph curves the wrong way.

Let’s say there’s a 10% chance of any given forecast being right, and let’s say (for the sake of argument) that all forecasts are independent of each other. Then what’s the chance of at least one forecast being right? Here’s the actual graph:


By the time you get to 20 forecasters, there’s an 88% chance that at least one of them will be right. At 40 forecasters, there’s a 99% chance.

In the real world, forecasters aren’t all independent of each other — but at the same time, there’s a hell of a lot more than 40 of them. (And given the squishiness involved in what counts as “being right”, the chances of being able to say that you were right are probably closer to 50% than 10%.)

If you add together the fund managers and the economists and the TV pundits and everybody else telling you where the economy and the markets are going, you’ll get a number somewhere in the tens of thousands. The question isn’t whether one of them will turn out to be right, it’s just how many of them will turn out to have been right.

In fact, given the thousands of people in the market, it’s a statistical certainty that many of them won’t just be right once, but will be right time and time again. Such people are generally lauded as being fabulously smart and prescient, and lots of money gets thrown at them. As a general rule, it’s a good idea to make sure that money isn’t yours.


Let’s monetize that.
Send out say 20,480 newsletters.
Half predict the market will rise, half say it will fall.
Whichever way it goes, you will be correct to 10,240.
Repeat the process with that 10,240. To half you say “rise,” and to the other half you say “fall.”
Soon enough, there is a core of people for whom you have been right 10 times.
Announce a hedge fund.
Check laws on extradition before absconding.

Posted by RobertArvanitis | Report as abusive

Adventures in information design, WSJ edition

Felix Salmon
Sep 13, 2010 22:28 UTC

Last week, Justin Lahart presented an interesting thesis in the WSJ:

For American business, it has become a two-track economy.

While global players like industrial conglomerate 3M Co. and burger giant McDonald’s Corp. are getting ever-bigger boosts from their operations in fast-growing economies like China and Brazil, companies dependent on the U.S. market are hemmed in by recession-scarred consumers who are hesitant to spend.

The accompanying chart was one of the most incomprehensible things I’ve ever seen on newsprint:


I doubt that one reader in 20 actually understood, on looking at this chart, the information it was ostensibly trying to get across.

There are three axes here, and two colors, and all manner of confusion. It took me a while to work it out, but I got there in the end: the height of the bars represents a healthy year-on-year increase in projected revenues. They increase in volume the further that number is from zero, so that the smallest bars aren’t the companies with falling revenues, but rather the companies with flat revenues. That’s silly. And because of the 3D effect, the biggest bars have much more volume than the smallest ones, as though the product of the two variables is somehow important. (It isn’t.)

The amount of color in a bar represents the proportion of those revenues that come from outside the US. The main axis we see when we look at the chart — the one marked with a bold black line — doesn’t actually represent anything at all. Oh, and the chart claims to be “an analysis of the 30 companies in the Dow”, but it only features ten stocks.

The weird thing is that Lahart’s thesis would come out loud and clear from a simple scatter chart. Put the change in revenue on one axis, put the percentage of revenue coming from overseas on the other axis, and look for two clusters: one at the top right, with the companies active in fast-growing countries, and one at the bottom left, with the companies hemmed in by recession-scarred consumers.

It took me a while, but I eventually got my hands on the data that Lahart was using. So I decided to plot that scatter chart, to see what it looked like:


I’m not a professional designer, and this could certainly be a lot prettier, especially if I could get my fonts to work and if I remembered to get rid of those silly decimal places on the y-axis. Still, the message of this chart is much clearer, I think, than that of the one in the WSJ. It does show a clear correlation between revenue growth and foreign sales, but I can’t see much of a case that American businesses have diverged onto one of two tracks.

All of my numbers are exactly the same as the ones that the WSJ used — even the 0% of foreign sales for Verizon, which resulted in a correction. (Verizon claims it gets some revenue abroad, although it’s less than 10%.) So I’m puzzled why the WSJ didn’t include Caterpillar as one of the “five companies with largest share of overseas revenue”: it’s up there in fourth place, with 67%, higher than both Hewlett-Packard and McDonald’s.

So, what really happened here? Did Lahart know that a clear chart would somewhat undercut his headline talking about a “two-track economy”, and therefore contrive something messier instead? Or was there just an overenthusiastic new kid at the WSJ who hasn’t read his Tufte? Either way, it would be great to see a stronger emphasis on clear information design at the WSJ. Financial publications have an especial responsibility to do this kind of thing well, given the highly-quantitative nature of what they’re writing about. The NYT has fantastic information designers; let’s hope the WSJ tries hard to compete on that front.

Update: Andrew Burton points out there’s actually a Wall Street Journal Guide to Information Graphics, by Dona Wong, available on Amazon in hardback for $19.77. I wonder what she would have to say about the WSJ’s chart.


Of course, this would be MUCH better if they simply showed the growth rate of overseas revenues and the growth rate of US revenues.

Posted by MitchW | Report as abusive

A new kind of stock-price chart

Felix Salmon
Sep 10, 2010 13:54 UTC

I love Kristina Peterson’s profile of Briargate, an algorithmic prop-trading firm (it’s an anagram of “arbitrage”) which has more or less given up on trading during the middle of the day.

High-frequency strategies like Briargate’s work best when there’s maximum liquidity, and that’s definitely during the first and last hour of the trading day. So instead of babysitting their computers at noon, Briargate’s principals go for long walks, or visit their children’s schools, or go out for pizza — and don’t even notice when something like the flash crash happens. (They were at the movies at the time.)

All of which makes me wonder whether we shouldn’t be presenting intraday stock charts a little bit differently. Right now, they invariably construct the x-axis so that every given unit of time (one minute, one hour, whatever) takes up the same amount of horizontal space. Underneath that you sometimes see a volume graph which shows you the important parts of the chart to look at.

Does anybody publish charts where the x-axis has a constant volume chart along the bottom, spreading out high-volume trading periods and skipping over low-volume periods relatively quickly? Is there a way of publishing data so that every tick, or every 1,000 shares traded, takes up an identical amount of space on the x-axis? The axis could still be labeled by hour or minute, it’s just that those labels would no longer be equidistant.

I’m pretty sure that such a chart would provide an interesting and fresh perspective on how stocks move. But of course it would be hard to generate in Excel, so maybe that’s why I’ve never seen one.


Hi Felix:

The answer is the “tick” chart. Traders have used this format for years. The data can be plotted in increments of x transactions or x contracts/volume. When the market is “dead”, fewer bars are plotted.

Teresa Lo

Posted by Teresa_Lo | Report as abusive

Money supply chart of the day

Felix Salmon
Jul 9, 2010 19:55 UTC

If Matt Yglesias can wonk out with meditations on the velocity of money, then I can wonk out with a chart:


The red line, here, is the total US money supply, and as you can see it’s started leveling off recently. (Source data here.) In fact, in many months it has actually declined — a rare occurrence, historically speaking. The blue bars are the month-on-month change in M2; it declined as much as 0.65% in January 2010, and in the first five months of this year — all that we have data for so far — it has fallen in three and risen in only two. The money supply in April 2010, at $8.5 trillion, was lower than it was in November 2009: it’s almost unheard-of for the money supply to shrink over so many months.

More generally, I’d take issue with Matt’s assertion that the Fed’s response to the crisis has “involved a sharp increase in the M2 money supply”. Yes, M2 rose in the wake of the crisis. But the sharp rise in M2 dates back much further than that — in fact, you can trace it all the way back to the mid-1990s. The red line doesn’t start rising more sharply when the crisis hits, nor do the blue lines get noticeably larger. There’s one big jump in M2 between August 2008 and January 2009, right at the height of the Lehman collapse, during which it rises from $7.79 trillion to $8.32 trillion, a rise of just under 7%. But we’ve seen that kind of thing before: between November 2000 and May 2001, M2 grew by more than 5%, and then between May 2001 and October 2001, it went on to grow another 4% on top of that.

But I do agree with Matt that we should start publishing M3 data again. If America’s economic statistics are “arguably the most robust in the world”, as Emily Kaiser says, then we should be able to know what’s happening to broad money, without using narrower money as a proxy. These things are very wonky, and only one part of a much bigger puzzle. But they’re still important.


I’m not an economist but:

1. Matt is probably thinking of the monetary base, which the Fed did double in response to the crisis. But this was offset by a tanking M1 multipler.

2. I think MZM is often used as a stand-in for our missing M3.

Posted by nedofbaker | Report as abusive

European bond chart of the day

Felix Salmon
May 10, 2010 02:15 UTC

Thanks to Johannes Bruder of Hamburg University for sending me this intriguing chart:


I suspect that the pattern would continue were you to include non-European countries as well; what’s interesting to me is the way in which there’s much more variation among eurozone countries than there is at the bottom of the scale, between Germany and countries like the UK and Switzerland which set their own interest rates.

As for trades, I wonder whether the chart might be pointing to a long-Ireland, short-Italy relative-value trade here. That trade has a positive carry, and if the two countries even so much as converge, you end up making a nice profit. Ireland has already shown that it’s politically grown-up enough to be able to implement tough fiscal austerity. I don’t think anybody really believes that of Italy.


I suspect Chinese premier Wen Jiabao announcement that China will continue it’s investment in Europe will calm European bond prices & nerves!

Posted by FootyR | Report as abusive