All the big banks have reported earnings so let’s take a little stock of things, shall we? A thing I did midyear was look at the ratio of trading net revenues to VaR for GS and MS, to prove that MS was winning the trading or something. Let’s expand that a bit. The result is, I think, an interesting chart:
The interpretation you could put on this is “this is how much money GS/MS/JPM/BAC made on an average trading day for every dollar of capital that they said they had at risk on that day.” Technically, this is is a graph, quarterly from 1Q2006, of [Trading Net Revenue / (Average Daily VaR * Trading Days in Quarter)], for four banks. Go read caveats here.*
This is obviously a toy chart, and too aggregated/undersampled to tell you all you might want to know, but still kind of fun. I don’t spend my days cursing VaR as an abomination upon the land, but it is fun to sometimes make fun of it as a quantitative measurement, and if you enjoy that you’ll enjoy reminiscing about the fourth quarter of 2007, where, on average, every day at Morgan Stanley was a one-in-twenty down move and every day at Bank of America was a one-in-a-hundred disaster. What are the odds?
But the numbers are mostly positive, which makes sense, since you should make money in a risky business if you plan to stay in it (hint to those roughed up by MS FID comp today), and you can interpret their trend if you so choose. It seems to be mostly down from the ’06 glory days. That’s probably a bad thing for bank shareholders and stuff, but also – maybe more so – for regulators. Like, here is a business where you should make basically infinite money per dollar of VaR: buying and selling stocks on commission. Here is a business where you should make more like ($1) to $1 per dollar of VaR: principal risk-taking for your own account. As the Volcker Rule noise heats up, you might take a moment to ponder the fact that the big banks’ trading revenues seem to be more and more at-risk – i.e. more and more tied to their using their own capital to support trading, and less and less connected to pure no-risk commission business. Like a lot of people, I’ve got no problem with that. Others do.
In any case, none of that applies to JPMorgan, who keep making more dollars per dollar of VaR. They can do no wrong. Huzzah.
* Data from Bloomberg and filings/announcements where Bloomberg was missing data. I have ignored the GS and MS calendar shift from a November to December FYE, just doing whatever Bloomberg does. Trading revenue is some plausible interpretation of filings meant to get at “flow” trading revenue in the investment bank and probably has comparability issues given different segmenting etc. VaR is reported aggregate VaR or “trading VaR.” Note that BAC reports a 99% VaR; the rest use 95% confidence.
JPM's the new promised land.
VaR, a long, long way to runnnnn…
Where are the other 4 / 5ths of the article Matt?
We don't risk OUR money. We risk YOUR money.
Matt, you and your little toy soldier are in for a real real treat later, buddy. A real treat. Pants are optional.
Where's UBS? The chart aint complete.
The problem is, you have to take into account the one year in a hundred where a fool of a trader who knows the access codes to the reconciliation system bets all the capital in the institution on a forex trade (or a commodities trade, or an ETF trade) that goes bad and takes the whole place down with it. All of these "look how trading is doing" charts should have an asterisk attached, "except for the one time that a rogue trader or a bad company policy takes the whole institution down".
Interesting fact: you can create the same graph for UBS by pushing the Y axis down 5 ticks
VaR is calculated differently by every firm…
His articles are WAAAAYYYY too short for the Y axis to go down that far.
is it bad that i skimmed through matt's post looking for a way to poke fun at him?
Matt didn't want to change the Y-axis scale to scientific notation.
Matt…I'm really upset you didn't add UBS to the chart!!!
Headlines are now irrelevant, only picking articles based on author.
This was awesome, srsly.
Pls now graph the $/VaR against the $bonus/employee and/or number of Zero/employee for even more expert laughs at each firm. Should be intriguing.
No it's not. We run numbers for everyone in the chart.
-JD
What happened in 2008? Lightening hit the transformer?
It must be tough to handle weightlessness!
How Is The Whole “Don’t Waste All Your Time at Goldman When You Can Write For Dealbreaker” Thing Working Out For You Matt?
How is that remedial capitalization course at Harvard Extension School working out for you?
How is the whole "not funny" thing working out for you?
1 year in a 100? That'd be great!
/Barings, SocGen, Amaranth, and, of course, UBS.
Don't forget Bo's two tries.
I'm no fool.
kweku wollinsky
Yours is one of the most honest blog on viral conversion I saw on my new Droid. I signed up to the rss feed to stayed updated when the beta launch happens…