Felix Salmon

Chart of the day: Goldman VaR

By Felix Salmon
July 15, 2009

Remember September, when Goldman Sachs and Morgan Stanley became bank holding companies? The WSJ reported at the time that the banks were “taking steps to reduce their leverage”:

It had become increasingly clear to Fed officials in recent days that the investment-banking model couldn’t function in these markets…

Goldman — and to a lesser extent, Morgan Stanley — has maneuvered through the credit crisis better than other investment banks. But its business model, which relies on short-term funding, is under attack. Some stockholders worry that its strategy of making big investments with borrowed money will go wrong someday, which would make it more difficult for the firm to get favorable borrowing terms…

The most fundamental problem is how to generate profit growth in a world that no longer tolerates high leverage.

Now my colleague John Kemp has published a wonderful little chartbook of Goldman’s value-at-risk, which includes this. The annotation is mine:


I guess Goldman Sachs worked out how to generate profit growth in a world that no longer tolerates high leverage. It just increased the amount of capital it puts at risk every day.

11 comments so far | RSS Comments RSS

Just because VaR has gone up does not mean that the capital being put at risk has gone up. It could just be greater volatility on the same capital employed.

Posted by Dan | Report as abusive

I think the GS approach is exactly correct. The right mindset is to think in counter cyclical terms.

The crash has already happened. The environment now is probabilistic wise significantly less risky on the downside compare to the years prior to the crash.

GS is putting more value at risk but the probability of downside risk is significantly lower. The product of this two measurement is likely to show GS is really not taking more risk.

VaR is simplistic and easy to measure. I would be interested in insights on measuring the probabilistic downside risk I’ve described.

On a side note, commercial banks should really crank up lending. The borrowers that are still standing are way better quality than a few years ago. This on top of the unprecedented interest spread means banks are well compensated for the risk they take by lending.

Posted by silly things | Report as abusive

Silly – you need capital in order to lend. Commercial banks facing a tidal wave of CRE defaults are struggling to preserve what they have.

Posted by ab | Report as abusive

Ab – banks are sitting on a huge pile of capital. Bank capital ratio as measure by Tier 1 and by common equity are near all time high. Furthermore, the country’s savings rate went from around 0% to above 7%. Just imagine how much money that is flowing into the bank.

You are right that there are still problem asset on the books. The media is all over this and the populists have been wiped up into a frenzy. The public has overshot on the downside significantly.

Banks aren’t lending more for 2 reasons:
1. The economic outlook is still murky. Default always rises during every previous recession.
2. Banks also follow the herd. This is also why so many banks got into trouble in the first place.

Anyway, if you deep dive into large banks, you’ll find that they are just sitting on a pile on cash.

Posted by silly things | Report as abusive

Agree they have the cash. My point is that they’re holding that cash mainly to protect themselves from further losses on their loan books (your point #1).

I’d disagree that some populist conspiracy is overstating likely CRE defaults. From what I’ve seen, there’s a lot of debt coming due in the next couple of years, and much of that was funded by the regional banks.

Posted by ab | Report as abusive

VaR assumes a normal distribution of asset returns; relying on the VaR measure as indicative of anything is a bad idea.

Posted by Dave | Report as abusive

The time series historical data reflects increasing volatities over the graph’s time line. Var would naturally increase even with no change in risk profile.
Impossible to tell much var analysis.
For example program trading would have extremely low var, but contributed significantly to the bottom line.
Liqudity risk is not captured by the var and drawing the optimal use of its capital conclusion based on var is just silly.

Posted by mcnet | Report as abusive

Ab, your view of the over all banking system is now in the minority. Amusingly, my view was in the minority a few months ago.

At any rate, you should research the percentage of the banks that are in trouble. The question than becomes are there sufficient number of healthy banks to take over the businesses of the weak banks.

One data source is the FDIC web site. Another is the result of the stress test for the big banks. Yeah, I know there were a lot of noise in the media and in blogs. However, the evidence is now quite clear.

Posted by silly things | Report as abusive

What on earth is this supposed to be? VaR is not leverage.

GS’s leverage ratio has gone from 27 to 14 over the course of a year. That is a remarkable change, and it doesn’t have anything to do with VaR.


It just annoys me how you are discussing things as if you know exactly what it means and then come to the completely wrong conclusion, just following the screaming Zero Hedge boys who do the same. As some others have pointed out, VAR is not a leverage ratio, but it uses the last x-days of trading data to calculate the possible 1 day loss on your portfolio. Anyone who had the slightest notion of the markets the last two years would have noticed the huge increase in volatility. That is the explanation of an increase in VAR, far more than anything else. VAR in 2006 would have consisted of market data of very calm markets, only moving a few basispoints a day. You can double your risk in such markets without much impact on VAR. Now you can reduce risk a lot and still have large increases in VAR.
Learn your basics, mr Salmon

Posted by M | Report as abusive

As a currency trader named DeRosa said back in the ’90s, VAR is a lighthouse for the soon to be shipwrecked. It gives me the jim-jams to believe that major-league trading operations are still depending on it to estimate market risk.


Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/