Monday, September 1, 2008

Explaining the brokers – Part III

This is proving to be a hard series to write. The brokers are complex. To really go through the Lehman 10K is a schedule clearing event. (The same is true for Barclays and other institutions I have made the attempt on.) But I hope I have painted the world appropriately.

In Part I I described just how big the brokers have got. Enormous. Some criticisms said that I was overstating the case because I compared US brokers to possible US assets – but the brokers were global. However that criticism is unfair – I did not include the foreign investment banks which are just as big – note UBS (which long had the biggest balance sheet in Europe), Barclays (the second biggest balance sheet in the world), SocGen (which has about half the trading revenue of Goldman Sachs), BNP Paribas and Calyon (both also large), Deutsch Bank and Nomura.

In Part II I explained just why there was such a glut of excess savings in the world – and why those savings are in such a risk-averse form.

My assertion is that the brokers – more than any other institution – have become the intermediaries between the Anglo/American (and Spanish) economies and copious pools of savings.

The first way they did this was through the now thoroughly discredited securitisation market. What securitisation did was take pools of assets (some junky, some good) and create AAA rated securities by segmenting and selling off the most senior cash flows. This left equity tranches (which had high returns and natural owners) and BBB tranches which were waste at the end with no natural owners.

So investment banks worked out a trick – which was to wrap the BBB tranches up into “collateralised debt obligations”. The purpose of these was to make the BBB bits go away. They would tranche the BBBs and produce some more AAA (which was sold to “natural owners”) and some more BBBs and equity tranches.

Never failing to do something to excess Wall Street then tranched the BBB bits of CDOs to produce the notorious (and mostly worthless) CDOs of CDOs or CDO Squareds. Surprisingly they sold some of these as AAA (admittedly with a few billion of guarantees from the monolines).

Some brokers (notably Goldman and Deutsche Bank) were pretty darn good at avoiding holding the “super senior” crap at the end of this. “Super senior was a euphemism to hide the fact that the AAA stuff was really contingent on BBB stuff that was already levered.” Other brokers (notably Merrills and UBS) were particularly bad at ducking this stuff. Similarly poor were National Australia Bank in my home market. [See the forthcoming Part IV.]

None of this explains why the broker balance sheets are so big (other than UBS and Merrills). The brokers don’t need to hold the assets for any length of time to securitise (though sometimes they did). The securitisation business is “moving business” not “storage business” and assets that are held when the music stops are ubiquitously bad – but are not necessarily huge in number. [The losses from holding assets at the end however will total tens of billions – maybe even 100 billion if you count all investment banks.]

It’s the “storage business” that fascinates me. They call it “trading” which implies a “moving business” but trading as done by most investment banks involves considerable storage.

The strangest thing about it is that the people actually doing the trading had no idea the economic function that they were performing.

Put yourself in the shoes of some prop trader. Some of them do “risk arbitrage” on takeovers and the like (see the description of Goldies trading business in Robert Rubin’s “In An Uncertain World” ]

What I think the traders did – and I would love some trader to confirm or deny this via email – is sit there and say “I can buy this security, sell this security, buy this interest rate hedge, buy this credit default swap and I am balanced”. So he did it.

And over time he kept putting the positions on one way – he went long the exotic instrument (say a mortgage or a piece of credit card or CDO paper), purchased a CDS or a monoline guarantee or some other credit protection, and funded this by shorting a vanilla instrument (some AAA paper – especially a Treasury).

And as they did it the balance sheets got bigger and bigger and bigger.

The economic function that the trader was performing was turning exotic instruments issued by the West into vanilla instruments favoured by the East. They were intermediating on the trading floor. And they were the biggest intermediators in town.

Goldies prop trading profits got to 30 billion dollars. If you added up the prop trading profits of the street – and they were always profits until recently – the numbers were well over 70 billion and probably over 100 billion (depending on who and what you counted). There are approximately 110 million households in America (sorry I have not checked that number for a while) and maybe that again in the indebted developed world. The trading profits were say 500 dollars per household.

Well if they were profits to Goldies et-al they were losses to someone – somebody paid $500 per household. And struggle as I may the only service that I can think that is worth $500 per household that was provided by the trading desks of investment banks was intermediating debt. So that is what they did. The “losing side of the trade” saw the 70 billion plus paid to the trading floor of investment banks as just something embedded in their monthly interest payments. The trader saw it as a mathematical game – but it was just service provision.

Now the reason this post is hard is that the data to support this proposition is thinner than I would like.

At a first cut if the trading was really just the bring-forward mark-to-market of the book then the trading revenue as a proportion of assets should fall each year.

When I first studied this (a few years ago) that was true. But it is no longer true. Trading revenue as a proportion of assets was particularly high in 2006. One person I talked to suggested that was just because the broker balance sheets just grew faster than ever – they just did it on the OTC market and didn’t actually need to back anything with the physical. I don’t know about that – and I do not have the data to either prove it or not.

But here is a simple model which details my quandary. In it I assume Goldies balance sheet would shrink by 30% each year just through natural attrition (loans repay etc). Then I look at the addition to the balance sheet and trading revenue as a percentage of additional revenue.



This seems to suggest that trading margins (ie a bring forward of all the profits) seems to be about 6% of incremental assets added. It’s a nasty profit front load if the business shrinks because you need to hold capital based on the whole balance sheet. 2006 looks to be a particularly good year.

But finding enough data to support this thesis generally I am finding hard to do.

If anyone has a better explanation I would love to hear it.

John Hempton

2 comments:

  1. There is basically one position in banking whether it be old fashioned retail and commerical banking or broker/dealer style investment banking - and that's positive carry. The 'customer business' is just a smokescreen (albeit a reasonably profitable smokescreen when volumes are high). What b/d prop traders do (and all traders at a b/d are to some extent prop traders) is buy high yielding assets and, err, hold them. Yes, there's lots of 'arb' trading around to provide more smoke - vol arb, credit arb, and so on. Yes the b/ds do need to carry inventory in order to have something to sell when they get a call. But basically they are an enormous positive carry position. What is of course interesting about this is now their funding cost has gone to the moon what used to be positive carry isn't any more...

    ReplyDelete
  2. Fascinating conceptualisation of the process JH. My own thinking would probably overcomplicate the process somewhat and look for losses at smaller institutions who end up holding the structured bags that the big guys can pass on from their mkt-making/sell side desks, rather than prop. My own very limited knowledge is that the market making desks tend to do most of the arb/intermediary work you point out, and the prop guys tend to take more directional punts. Macro Man would no doubt have something to add to this. What happened to the less frequent posting? Cheers, JL

    ReplyDelete