A story that is told about banks is this: Bankers are paid to maximize short-term results and screw the risks. All they care about is this-period earnings. They dance until the music stops. (Then they sue.) In this story, banker pay drives risk and volatility and other terrible things.
A counter-narrative that sometimes floats around is this: there's no particular evidence that bankers are paid in line with short-term risks. The jury is sort of out on this one because the data isn't really there one way or the other; see this post from the Epicurean Dealmaker for why you shouldn't take seriously claims that bank executive comp (who cares about executives?) doesn't closely track stock prices (who cares about stock prices?). He tentatively endorses a form of the traditional view:
My industry’s pay practices and culture were built over decades when the vast majority of business investment banks conducted was agency business. Business like M&A, where you earn a fee for helping a client buy or sell a company, or security underwriting, where you earn a fee for placing client securities with outside investors, or securities market making, where you earn a spread for standing between buy- and sell-side investors as a middleman and temporary warehouser. None of these businesses entailed any material amount of persistent or hidden financial risk to investment banks ...
The problem arose when investment banks (and their bastard cousins and often ultimate owners, commercial or universal banks) began conducting business as principals, either explicitly and in full knowledge, or—most dangerously—in total ignorance. Mouthwateringly profitable leveraged lending, structured products, complex derivatives, and proprietary investing of all kinds meant that investment banks no longer conducted business as short-term conduits of temporary risk, but began accumulating long-term financial risks on or off their balance sheet, often without their own knowledge. But when this happens, the old view that Joe in Structured Products should get a massive bonus in February because he brought in $100 million of fee revenue to the firm this year cannot cope with the fact that Joe’s fabulous trades expose the firm to $1 billion in potential losses over the next five years.
Lest you reply "but you get paid in stock so ooh long-term incentives ooh," ED has an answer to that that is I think irrefutable. And he ends up calling for an empirical study of line-trader-and-banker pay that determines how much risk-taking is and is not incentivized by comp.
Maybe we'll do that when we get an intern. Until then, a data point you might look at is this paper that Oliver Faltin-Traeger of Blackrock and Christopher Mayer of Columbia Business School presented this weekend at a fancy soirée in Chicago. It is not at all about banker comp. It is about the fact that CDOs were shit, pure shit, basically. I know, you knew that already, but really:
We find that CDO assets tend to be lower rated securities from the lowest quality asset classes and vintages, and with a higher spread at issuance. CDO assets performed much worse than comparable securities that were not included in a CDO. When we control for the initial rating, CDO assets have a downgrade severity that is at least twice as bad as comparable ABS not included in a CDO. Synthetic CDOs assets perform worse than cash CDO assets, but assets included in both cash and synthetic CDOs perform worst of all (with a downgrade severity about two and one-half times worse than the average downgrade severity). Even when we include controls for a wide variety of observable characteristics, including initial yield, CDO assets still underperform comparable ABS by between 50 and 100 percent. These results suggest that CDO originators successfully sold securities and insurance against the worst performing ABS assets, but also that buyers of CDOs would have had a hard time analyzing these securities based on observable characteristics alone.
That's number of "fine notches"* that various things were downgraded between July 2007 and July 2009. For example, the average AA rated asset-backed security as of 2007 was rated A- as of 2009 (down 4.08 fine notches) if it never found its way into a CDO. If it was in a cash CDO it was around BBB-/BB+ (down 7.62 fine notches); if it was referenced in a synthetic CDO it was closer to BB. Things in both kinds of CDOs did even worse, which makes sense, as those things were basically the all-star team of securities that everyone wanted to bet against.
Now, in one light, this looks kinda bad. Certainly you can spin this as "banks systematically screwed their customers," and Bloomberg did that, and the authors kind of did too. Also it's kind of true, no? You can argue the moralities here where the customers could actually go look at the collateral and decide for themselves if it was crap or non-crap, but before you get all defensive consider:
Evaluating CDO performance is complicated by several additional factors. Liabilities from one CDO can be repackaged within a second CDO (referred to as a “CDO-squared”). The performance of one CDO tranche may therefore depend upon not only the cash flows from the ABS in that CDO but also the ABS underlying any CDO tranches that the CDO of interest owns as collateral. Each CDO in the dataset used in this paper invests in an average of 119 securities and about 5% of those are themselves CDO liabilities. Going only one “level” down in a set of 120 securities, an investor must evaluate the 114 ABS directly underlying the CDO and 6 additional sets of 114 securities underlying the 6 CDO tranches in the pool for a total of nearly 800 ABS.
Which is, like, a lot of work to do before investing your $10mm in a CDO.
Looked at from the other side - every market has two sides! - there's some good news here. It's along the lines of "sure, banks offloaded their shittiest assets to clients, but on the bright side, at least they offloaded their shittiest assets." That's something, right? The story you could tell here is that banks were managing risk by moving dangerous inventory off their books.** Notably, they were doing so at the expense of short-term yield: the average spread on non-CDO'ed ABS was 58bps, versus 90bps for ABS in cash CDOs and 120bps for ABS referenced by synthetic CDOs. (This is at least in part due to the fact that ABS that found its way into CDOs was, as you might expect, lower rated. The authors don't seem to distinguish a spread-per-ratings-category variable, which is annoying. So I don't know how much you can read into this - the incentives to put lower-rated paper into CDOs, for capital, ratings-structure, etc. reasons, is obvious; for all I know the banks were CDO-ing the lowest-yielding paper within any ratings grade.)
Still, though, the fact that CDO'ed paper performed so much worse than non-CDO'ed paper, in some cases years after the CDO construction, does suggest that the banks were doing thoughtful medium-term risk management for their own balance sheets through their CDO pipeline. And you could argue that this has good systemic effects. You could argue otherwise too, since that CDO paper ended up somewhere, and that somewhere was often somewhere terrible. But if there are going to be worst bonds, somebody has to own the worst bonds - and from the bank's perspective, getting rid of the worst bonds is a good idea.
A reasonable thing to think about what bankers and traders do is basically "they try to make money for their firm without blowing it up." Pay does matter - if you get no performance-based comp, you will have less desire to make money for your firm; if your pay is based entirely on short-term profits, maybe you'll cut some corners on long-term risks - but basically people expect to have a career and like the people they work with and feel some loyalty and generally don't want to end up being famous for failure. So while you can fiddle with pay structures to better align bankers' interests with their banks' long-term interests, you could also just trust that those interests are already more aligned than you think. The story of CDO performance is maybe some evidence for that view.
But that particular story sucks for clients! Because while the bankers, whether driven by (a) their compensation structures or (b) their general fondness for the home team or (c) some combination or (d) other, managed to offload the worst bonds via CDOs, they offloaded them to their clients, who are now sad, particularly after reading this paper. And it gets back to the distinction that the Epicurean Dealmaker draws between old-line agency businesses and new-style principal businesses. In agency businesses, not only was your comp closely aligned with your bank's interests (more revenue good!), but it was also pretty reasonably aligned with your clients' interests: sure you could get more fees by generating pointless trades or by pushing bad M&A deals, but in the medium term that was not a way to get repeat business. (I mean, in theory.) And clients thought that way: they were paying you for service, so they expected you to be providing service. To them.
In principal businesses, though, you end up really rooting for the home team: yes, offloading shitty assets to clients via CDOs will lose you repeat business when those CDOs blow up, but offloading enough of those assets to miss out on a financial meltdown is worth more than quite a lot of repeat business.*** And clients - well, lots of clients think of banks as counterparties and expect to be screwed. But the fact that CDO buyers bought without doing the analysis of the 800 underlying bonds, or whatever, suggests that those buyers were thinking of the banks on the old-timey agency model, and expecting the banks to be looking out for their best interests rather than just selling them the things that they didn't want to own any more and for good reason. Which seems to have been a tragic misunderstanding.
Goldman, Citigroup CDOs Were Tip of Iceberg [Bloomberg]
* A "fine notch" is what it ought to be: a downgrade from AA+ to AA is one fine notch, AA+ to AA- is two, etc. etc. When I am king this will be the only way that people talk about ratings notches; calling AA to A a "notch" as, for example, many regulations do is silly.
** You could complicate that story by trying to determine how much of the ABS that went into cash and synthetic CDOs came from risk that the banks were already holding versus how much they went out and sourced to build their CDOs. Similarly, you could try to figure out what happened to non-CDO'ed paper: how much was held by banks directly and how much was sold outside of the CDO structure. For instance, one narrative that would explain the "everything in CDOs sucked" in a more innocent way is: (1) banks created ABS, (2) they sold every bond they could sell to people who wanted to buy them, (3) what was left over they repackaged into CDOs and sold to people who wanted to buy in that form, (4) what was left over they repackaged into CDO-squareds, (5) etc. My sense is that this narrative is basically true. Thus the story is not "banks screwed clients by building CDOs" but rather "the people who bought the stuff on the remainder rack got remainders." Anyway.
*** See, e.g., Goldman Sachs, except: the paper also notes that Goldman's shitty CDOs were less shitty than everyone else's. So, go team.