Systemic Risks, Ahoy There!
Be an investor or trader long enough, and you eventually start to collect moments that go along the lines of “what were you doing when the market did X?” Maybe I’m just getting old, but it seems that there are more and more of these moments that I can recall. Yesterday’s market action is definitely going to be one of them, assuming that something more dramatic isn’t on the agenda for next week.
In case you’re reading this blog sometime down the line, what happened yesterday was that the Dow dropped approximately 1000 points (about 10%), much of it over a period of about 5 minutes. Then it rapidly recovered about 700 points in about the same length of time, ending down about 3% for the day. In normal times, this would be considered a bad day, but most investors and traders were probably just breathing a sigh of relief that the 10% drop did not stick.
During this time, the S&P 500 also collapsed about 10% and then regained the bulk of its loss back. The S&P 500 represents the market capitalization of about 75% of publicly traded companies, and was valued at about $10 Trillion at yesterday’s (6 May 2010) close according to S&P’s website. So, over the course of 15 minutes or so, the economy erased ONE TRILLION DOLLARS of market capitalization, and then restored about 700 Billion of it back.
What a Trip! Are we Levered yet?
It will take some time to sort through exactly what happened, but the scuttlebutt around the net and news sites is that a Citibank trader accidentally typed B (billion) instead of M (million) on what was supposed to be a $20M sell order that was then executed at $20B. Now, selling 1000x the quantity that one is supposed to sell is definitely a major mistake, and presumably that trader no longer has to worry about showing up for work anymore, but even $20B is only a tiny fraction of the total market movement yesterday. To clarify: $20B / $1T = 2%.
Let’s get this straight. A $20B sell order kicked off a chain reaction that temporarily eliminated $1 Trillion ($1000 B) of US market cap. That’s conceptually similar to 50-1 leverage in the system, although the number is probably much, much larger when you consider that selling a stock shouldn’t make the stock’s market cap go down by the entire amount of the trade - in fact, it should drop by a tiny fraction, if indeed at all.
Presumably what happened is that a portfolio sell trade that is normally managed to minimize market impact was simply so large that it pushed the price of all the component securities down. Then, high frequency traders using intra-day or other short term momentum / trend-following strategies jumped on board and served to magnify the selling process, putting further downward pressure on prices and starting a mini-panic.
But why the bounce back?
What’s interesting is what happened next. Over the next 10 minutes, the market returned most of what it had taken away. It gave $700 Billion back. How did this come about?
Some could argue that this shows how markets are efficient, and there is a reasonable rationale for this. Prices clearly shouldn’t correct 10% in 5 minutes based on no new information, and therefore we could say with a good deal of confidence that - at yesterday’s depths - prices were clearly out-of-line.
At some point computerized value investors stepped in and bought shares that the computerized momentum traders were selling. Perhaps some fundamental value investors also stepped in, but since this all happened over the span of 15 minutes, it is much more likely that automated trading has more to do with the price action.
Eventually, enough value algorithms stepped in to halt the downsliding, and then, as the prices started to rise, the automated trend followers started pushing UP the price, returning most of the loss.
In addition, the Citibank trader at the start of this process may have noticed that they’d sold $1.98 billion too much and may have tried to buy the order back. I’ve seen no mention of this, but presumably this would have pushed prices back up and at least contained Citi’s total exposure to the problem.
If one wants to believe that markets are efficient, yesterday suggests that since there was relatively little other important news yesterday to affect prices, we might just conclude that markets are about 70% efficient. This rough estimate comes from the idea that since there should have been little price change, a $10T market cap loss (a mispricing due to a trading error) was 70% restored by clever computers on the lookout for mispricings.
That still means that 30% of the market change was not restored. Presumably some of that 30% market drop is justified by increased fear and risk aversion, but it is very difficult to argue that the bulk of it is.
Blurred lines between Systemic and Asset-Specific Risk
Assuming this analysis captures most of what happened yesterday, a few points are critical, and it it is also significant that this event happened as Congress is debating financial reform, regulatory issues, and the “too big to fail” problem.
First, a trading error - even a big trading error like $20B - is normally considered a firm-specific risk. Yes, $20B is a lot of money - more than the GDP of some countries - and yet it seems odd that a $1T swing in the stock market should depend on a single keystroke of a single trader. Remember that a $20B dollar sell order shouldn’t even reduce the market cap of its index by even $20B, not even close, and yet the result of this error was an “erroneous” drop of $1T in the S&P.
But when a trade error from a single employee at a single company spreads through the economy and an entire index, it is almost by definition a systematic risk.
We normally think of systemic risks as things like economic growth, interest rate changes, regulatory or tax changes, natural disasters, terrorism, etc. These are mostly “big things” that seem like Acts of God, or perhaps Acts of People who Think They Are God; things that clearly affect large numbers of companies simultaneously. Company-specific risks are things that have to do with company management decisions, product quality, and mistakes that companies make that affect their bottom line. Now we have to include mistyped trade orders from a single individual as a source of systematic risk (not to mention oil platform explosions and their environmental consequences in the petroleum field).
Surely this will feed into the “too big to fail” / “too big to live” debate going on on Capitol Hill. But how do you determine when a simple trading error will create a feedback loop like that.
The key conclusion is something that many of us probably have suspected all along, and which first became obvious during the Lehman failure in 2008: our economy and financial system is so tightly interlinked with competing (nonlinear) feedback loops, that it is exceedingly difficult to tell when a $20B dollar mistake will erupt into a $1000B mistake in a matter of minutes.
Many investors may have comforted themselves with the thought that Lehman was a 1-in-100-years flood, and simply want to move on and forget it. Or others point out that “Lehman had it coming” because of bad leverage policies and the ability to hide leverage off the balance sheet. But yesterday is not really about fraud and cover-ups (though maybe that will be discovered). It’s more about how relatively tiny mistakes (relative to the final consequences) by employees can blow up into a true systemic crisis.
In other words, after yesterday, it is hard to escape the conclusion that there is a lot more systemic risk lurking under the surface than we even feared the on the day before yesterday. A bit like Europe before WWI, where one starts to realize that a single assassination in the Balkans can blow up into a world conflagration.
The logical consequence is that any serious investor needs to be re-evaluating how much risk is truly out there, and, while one might be able to make short-term, tactical trades to benefit from temporary mispricings, it is hard to escape the conclusion that the world is much more risky than we had hoped it would be, and that a responsible investor needs to adjust their portfolios accordingly. This means taking risk off the table and exposing oneself to the level one had originally intended. I can’t see how this can’t but push index levels down in the short term, although one could ask whether a 3-5% discount is a sufficient discount to accommodate our newly revealed systemic risk levels. It seems unlikely just now, but time will tell.
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home