Monday, March 31, 2008

Fight Hard when the Stakes are Low

Sheez, Paulson rolls out this plan to consolidate various regulatory agencies, and people hate it. The WSJ said
Groups ranging from small banks to state attorneys general criticized Treasury Secretary Henry Paulson's proposals to consolidate regulatory agencies and revamp oversight of a financial system bruised by crises in the credit and housing markets.

Merging the SEC and CFTC, the OTS and the Fed, getting rid of state-by-state insurance regulation. Clearly the latter one will get rid of lots of little fiefdoms, but there is no reason these things should not be merged, it just makes it easier to do business when you have lots of different assets, and work across state lines. What if the FDA worked state by state? It's good to see these guys use this crisis to make some overdue reshuffling. Corporations have major reorganizations every 5-10 years or so, implying the government can reorganize every 50 years or so.

Friday, March 28, 2008

Taleb on Bloomberg

Bloomberg has a big story on my favorite literary-philosophical-mathematical flâneur, Nassim Taleb. It appears his book, The Black Swan, is a huge best seller, and supposedly he gets $60k per speech now, all for his new theory: 'shit happens'. He's also saving us from the oppression of the Normal distribution, which statisticians believe exactly describes the world (fools!).

Why is Taleb so popular now? Perhaps you have heard of the Sub-prime debacle? Nassim called it. Whatever unexpected that happens you'll find that most of the experts didn't expected it--just as Taleb predicted. Space Shuttle? Berlin Wall? Britney's meltdown? Again, these thing were big events, and most experts didn't expect them, so Taleb did. Well, actually all he said was that unexpected things happen a lot. Is that a correct call? If you believe so, you are a Taleb fan.

The article starts by noting that Taleb was lecturing to a group of Morgan Stanley risk managers, lambasting 'stress tests'. This is strange because one would think that, to a guy that hates the normal distribution, or really any parametric distribution, he would love the stress test. Stress tests can be anything you want: what happens to your portfolio if the S&P goes up 10% and the dollar falls 10%? Improbable, perhaps, but these are nonparametric, their scope is only limited by your imagination. That he would say these are bad, leaves one to wonder, what, exactly, he thinks risk management should focus on. But that is not Taleb's oeuvre, which is merely to criticize any forecasting tool because it is imperfect. The only positive advice he gives, is trivial, things like, go to cocktail parties, because you might hear a good idea in such a nontraditional environment; or that you should invest in wacky investments that have Powerball-type upside. Good luck with that.

Success is mainly marketing, and you have to hand it to Taleb, he's slyly presents himself as a prescient speculator, someone who makes money taking positions, without ever really saying so. For instance, it is remarked in the story he 'made' $35MM on eurodollar options on Oct 19, 1987 (the market crash). As a trader, that just means he didn't have his risk hedged, because remember, we was primarily a market maker then, a guy who makes money off customer flow, and so generally you don't want these guys taking sides, they make money off the bid-ask spread. And so the 200 point move in Eurodollars that day (unprecedented) exposed him. That's simply a mistake, and others who didn't hedge their risk probably had the opposite pnl. But, elsewhere, if I remember, Taleb admits that his Oct 19th fortune was luck. Nevertheless, by putting out he 'made' $35MM, and that traders called him 'Nassim the dream', clearly it suggests he has the Midas touch. Subtle.

And then there's his hedge fund, Empirica Kurtosis, a fund he ran for 5 years, from 2000 through 2004. There was a joke when I was at one fund, where a bad idiosyncratic trade is always called a 'hedge' after the fact. That is, the money you lost on that punt on volatility, or the oil bet you made that goes against you, you say was hedging something else in your portfolio. It's a nice way to explain away a bad idea. So after the fund starting grinding out losses, Nassim started calling his fund a 'hedge', not a fund, later, a 'laboratory'. Now he says about the fund:
`Our aim was not to make money,'' Taleb says. ``I make no claims of being able to beat markets.'

But he makes sure any article that mentions his fund notes he made 60% in 2000. The only record of his total fund was a WSJ article on him in 2007, which notes he lost money in 2001 and 2002, made single digits in 2003 and 2004. That averages out to around 12%, and as the risk free rate was about 4% over that period, and the volatility was probably around 17% on a monthly basis, thats a Sharpe of 0.47. Not so good. And that's with his unaudited returns, so it's probably biased high (people have a tendency to round unaudited results upward significantly).

Like almost everything he writes, he is inconsistent, which makes taking him seriously pointless, because he can say that he said anything: his fund was a hedge, it made a ton of money; he was a lucky trader, he was a skilled speculator. I'm clearly a minority in my assessment of his insight, but then again, I didn't like Confessions of an Economic Hit Man or Nickel and Dimed either. Basically, my favorite books tend to be #347 in their category.

The story mentions his former assistants are starting funds based on capturing the Black Swan, a fantastic plan. If you write a best selling book about investing, clearly some of those readers will be rich hedge fund investors. Now pitch them with the following story that is totally consistent with your revolutionary insights: I get 2 and 20 fees. Your returns will be near zero, until we catch a financial Black Swan, whose return cannot be quantified, but think Google or Harry Potter. Of course, 'absence of evidence' is not 'evidence of absence', so if nothing happens after 10 years, that proves nothing. Heck, even a lifetime of zero alpha proves nothing. Meanwhile, on $1B, that's $20MM per year. Brilliant!

Thursday, March 27, 2008

Merriweather does it again


The basic advantage of a hedge fund investing in corporate bonds is to take advantage of the stale pricing in these illiquid securities. You can easily game the Sharpe ratio because they hide the true volatility of these assets (just look at the NAV of convertible bond funds like JPC vs their market values. If you could mark at the NAV, that is, the market value of the bonds, they are great trades. But if you trade the market value of the funds, which tries to put more realistic, and thus more volatile market pricing, on these portfolios, they are losers.

Thus it seems that Merriweather's bond fund merely applied leverage on corporates, and lo and behold, that blew up. Given their relative strength in the last recession, it was probably weighted towards financials, which did relatively well in the 2001 debacle, but is doing horribly now. I would stay away from any corporate bond hedge fund because theres simply isn't enough return in these things to justify their fees. Either they have an angle on the debt-equity arb or play the distressed game, where I do think one can make sufficient hedge fund returns. But if they are merely playing traded corporates, there's no there there. The average spread over time in B rated bonds, after transaction costs, is not measurably above libor. You would have to foresee every credit crunch to make money, a systematic trade that is implausible, to make money

Tuesday, March 25, 2008

Tyler Cowen for Regulation, Details Unimportant

Banking is a very tricky business because modern finance has specialized to the degree that it is no longer as simple as when the bank held the savings of the community, and lent it to businessmen. A financial institution is a small piece of a long list of specialists from saver to investor and back again, involved in one or more of the following: origination, underwriting, repackaging, custodian, and warehousing of both sides of the transaction. Its basically like the molecular biology of a cell: there are many important, complicated processes going on, and details matter a great deal.

Further, because financial contracts have virtually zero marginal costs, sometimes an inflated 'face value' arises because it is easier to keep adding offsetting ones, rather than extinguishing some and starting over, the way a futures open interest works. The reason why derivatives are not transparent is that they come in some many different flavors they escape any reasonably finite taxonomy--they are 'over the counter' and thus have idiosyncratic terms.

So when Tyler Cowen makes his bold opinion on the current banking crisis, by saying we should apply capital requirements to derivatives, such a proposal leaves innumerable questions that will make regulation almost surely miss its mark. But first, Cowen seems to think regulation is a reasonable quid pro quo for guaranteeing these guys. But the bail out of Bear Sterns was not an obligation of the Fed. Unlike the S&Ls this is like bailing out Long Term Capital management. They were afraid of the contagion effects. If the Fed decides to impose regulations on derivatives for everyone, not just those institutions currently regulated by the Fed, what a huge power grab by an already overextended institution. Remember when Thatcher remarked to Gorbachev that it was nice that she didn't have to decide how much, and at what price, steel was going to be. Competitive markets decentralize what are really very difficult questions that no one person should have to answer (hmm, maybe they should read more Hayek at George Mason). But lets say we are going to do it, because Federal regulation of the energy, education, and drugs has been so successful.

How much capital for derivatives? Good question. Should it be weighted by risk? If so, how does one measure risk? Considering that risk is a function of the collateral, which comes in many different flavors (traded debt, pools of mortgages, pools of bank lines), and then are structured very differently, with differing levels of subordination, differing rules for the waterfalls of cashflows depending on various metrics of collateral quality. It's a mess. If you think market participants are confused, make it a political issue, so not only do you have to explain this to dunderheads like Hillary, Nancy, and Dennis, but then you have to expect them to add something like the Community Reinvestment Act, so disabled lesbians of color can start lobbying for more equity (i.e., loans given out the way Huey Long gave out construction contracts).


You may think this is no different than regular lending, but you would be wrong. For example, lets say you have two swaps, but they both offset each other almost exactly for interest rate risk, but as they have different counterparties, they have differing credit risk. How about swaps from the same counterparty, but differing interest rate exposures, partially netted. How much should capital be netted? And if the US banks have capital requirements greater than economically necessary, how many seconds before all swaps would move offshore?

If something is too complicated for the market, it goes double for government. And as finance is really about contracts, ideas, it doesn't have to put up with heavy handed regulation, especially when New York is not the origin of the capital (we aren't capital suppliers). London, Hong Kong, or Tokyo would be glad to pick up the slack. As it stands now, mortgage regulation has led to loan documents where you have to 'sign here' about 15 times, and you still have to pay several ridiculous processing fees that look like your phone bill and add about $500 too much to these things, and in the end, no one reads all those 15 things they initial and sign (which are helpfully put in 8-point font because we at least want to save trees, and no one even pretends to read them). A couple more pages to initial and ignore, perhaps? They can't even get rid of mandatory title insurance which is way overpriced, and other antiquated parts of the process.

Tyler has the seemingly reasonable but totally naive recommendation of someone from 30,000 feet, which given the breadth of things he comments on, is unsurprising.

Monday, March 24, 2008

How much can models help?


I have data on about 600 asset swaps per year, and track their year ahead change in yield. I then bucketed the data into 20 sections by asset swap level, then sorted within those buckets by the Merton model (ie, a model that treats equity as an option on the value of the firm, the basis of KMV's EDFs). I did this going back to 1999.

You can see than the Merton model correctly predicts the higher risk credit rise in yield by about 150 basis points per year in bear markets (bear market: prices go down, swap spreads go up). But surprisingly, in bull markets, there is less power in the Merton model. This is surprising because default prediction gets worse in a crisis than in good times, that is, standard power metrics are worse trying to explain the 2002 defaults, than the 1998 defaults. But this exercise suggests that in predicting prices, there is more power in a model in the bear markets, than the bull or non-bear markets. I remember the big boom in credit in 2003, and it was insane. Everything rallied, and the crappier names rallied most. You do not want to be short 'crap' when the market is rallying credit, a situation I imagine will happen soon in nonfinancials.

Predicting default is not the same as predicting prices, but of course they are highly correlated objectives. The trick is to find a simple, but strong pattern that others have not.

Sunday, March 23, 2008

Depression? What about defaults?


Check out the graph. As spreads have widened to historical highs, the default rate for speculative grade securities in the US has not been lower since 1981! Who'd a thunk it? [default data from Moody's] I don't things are nearly as bad as people think.

Paul Krugman writes that we are potentially at the beginning of another Great Depression, and therefore, we need to regulate more. I'm not clear what particular regulation he considers so useful. The main effects of the 30's regulation on securities markets is basically generating greater costs of entry, by mandating lots of silly bureaucratic formalism to the extent you promise not to lie, cheat and steal. All the indigent and illiterate criminals were now excluded, but that's not much. And separating investment banking from commercial banking, did nothing (never adhered to in Europe).

I think people need some context. Unemployment is only 4.8%, below the average during the Clinton years. There have been no bank runs, just the failure of a few hedge funds and their ilk. This is the wonderful thing about deregulation, in that financial excess, and it opposite, have smaller real effects.

Thursday, March 20, 2008

Pay Risk Management More?

During this tumult, I often see someone say, "we should pay risk management as much as the dealmakers". When I was in risk management and went to conferences, there were three common themes:

1. Risk management is not a model, but good judgment (don't quiz me!)
2. Good Risk management starts at the top (I should be on the Management Committee)
3. You need to pay for good risk management (give me a raise)

But the problem with paying risk management as much as the business lines is that risk management is about preventing improbable events from happening, and so it is very difficult to know if they are doing a good job. Someone might say, this is a great risk management system, but the problem is that for 8 years out of 10, there is no difference between a great and poor risk management system. Then, when the cycle does crash, and your crackerjack team of risk managers prevents the Commercial Real Estate disaster of 1990, you will find your team overconfident, and constraining Commercial Real Estate when it flourishes (ie, post 1990). The next area of interest, Telecom, had a different risk profile, and so your team will probably be focused on poor Commercial Real Estate--i.e., past risk management success on the 'big themes' is not as relevant as it may seem.

Any job that is difficult to monitor with hard data, does not encourage excellence, or highly motivated, smart people, to go there, for the simple reason that, without objective metrics to assess performance, a huckster who focuses on irrelevancies looks too much like someone who focused like a laser on the essentials. Excellence is relatively unrewarded in risk management, so the excellent tend not to go there. Not that there are not many excellent people in risk management, merely that, on average, the business lines are better.

A final point is that risk managers tend to be generalists, looking at a variety of businesses. For example, any risk manager with a modest amount of responsibility will be in charge of several very different products, like consumer and commercial loans. These loans have very different characteristics, as consumer loans tend to have high average charge off rates that very 2 or 3 fold over the cycle, where many commercial loan areas have near zero losses in good years, and 5 or 20% losses in bad years. The collateral is different, with different legal recourse when going after an individual’s assets versus a business where you often hold their property or inventory as collateral, leading to different recovery rates and timing. But this difference is true for really any level of detail. A loan to a Health Care provider would be very different than a loan to a hotel, even though they are both commercial’ loans. While there are common themes, such as standard financial ratios to examine, the key checklist of of bona fides vary by business line, and the business line managers tend to understand these nuances better than the generalist risk managers who independently evaluate them. Risk management is necessarily more big picture, and ignores many important line-specific information that materially affect a line’s prospects. Nonetheless, it’s value remains because of its independence.


So you have ‘risk management’, whose value as an independent oversight is necessary to keep people honest, but they are generally not as smart or knowledgable as the business line they are monitoring. Their quality is difficult to evaluate directly, as the infrequency of the events they are avoiding makes it impossible to tie a bonus to, say, the avoidance of a large loss. Thus ‘risk management’ professionals are more like audit than the business line, more rationalizers, and information reporting. Just as everyone says nice thing about police officers, but these people generally are not paid particularly well and they are generally not insanely bright, so it is with risk management.

Monday, March 17, 2008

Rapping Economists



What's funny is the status hierarchies for economists and rappers are real, in that people really do care alot about their ingroup ranking in a specialized field, and view other status hierarchies as kind of a joke. That rappers are not making jokes about economists highlights the fact that the rapper hierarchy is far bigger, and its easier for economists to empathize with rappers than vice versa.

Friday, March 14, 2008

Bear Stearns and Financial Company Default

The problem estimating financial company default probabilities is that as long as an asset is performing--ie, not past due--they all look the same. A bank could have a bunch of mortgages where their borrowers had FICO score above 600, and 20% equity in their houses (ie, good loans), or they could have mortages on borrowers with no money down, FICO scores below 500, and negative amortization loans. You have no way of knowing based on the data they present in standard financials.

But a firm like Bear should have a complete top-down risk report that aggregates their risks in a meaningful way, and so things like FICO scores, or subordination of the debt, should be available. If they can't pull this together, either they are hiding something, don't have this information, or think they have valuable trade secrets within their balance sheet. I truly doubt their firm's value is significantly generated by something as top-down as their risk exposures, but eventually they might have to to save their existence, and if they can't produce a concise summary, they should fold because this is a dramatic dereliction of duty.

I'm amazed how some big firms are really quite stupid, like when American Express lost over $1B in 2001 on junk debt. The president at that time, Kenneth Chenault, expressed shock at the losses, though the returns on the debt of 13% necessarily implied a large default rate risk.

I have no sympathy for Bear, because they should have the ability to convince people they do not have risk, and if they can't present information to that end, they are inept.

update (3/17): I think this is a case of a liquidity event forcing bankruptcy. Bear would have made it if they could mark assets at book rather than market value, but the market value made them insolvent, and if they sold for $2, they knew they were insolvent if one looked at mark to market. Thus, a great value add for JP Morgan. It's no different than if you all the sudden made Ford or GM recognize the size of their off-balance sheet liabilities (unfunded pension and health care), they would need a white knight.

Thursday, March 13, 2008

The Importance of Geometric Averaging


The graph above shows the total return to my High and Low Beta indices. Every 6 months, starting in 1998, I took the top 1500 market cap companies, and put the companies with the highest 100 betas in the 'high beta' portfolio, and those with the lowest 100 betas in the 'low beta' portfolio.

Over the long run, the high beta portfolio is a stinker. The summary stats over 10 years (up to last weekend) are listed below.

Note that the arithmetic average return is about the same for both portfolios, but the geometric much lower for the more highly volatile high beta portfolio. This is because the difference between the geometric and arithmetic averages is the variance of returns divided by 2. The trick is to keep everything in the same time units, so that if we are looking at annualized returns, we need the annualized variance. Note that the adjustment for the high beta portfolio is about 10%, which in turn matches what is actually calculated when I merely used the following methods:

Arithmetic Annual Average: Average(C2:C2498)* 252
Geometric Annual Average: {=PRODUCT(C2:C2498+1)^(1/10)-1}

For a long/short strategy, an assumption as to when one rebalances is key.

For many futures trades, you can assume daily rebalancing, because transaction costs are so low, and people try to keep their first and second order exposures square up (e.g., flat delta and vega), but for equity strategies, a long holding period is appropriate, and this has drastic affects on the returns, especially for highly volatile strategies (note the effect on the 13% vol strategy was pretty minor).

Muni Problem in 2009

I just got my annual property assessment used for my taxes, and for the first time ever, it was down about 7% from the prior year. Property taxes make up most of the city's revenue, and are a linear function of the assessed property value. This is for taxes payable in 2009, which are paid in March and September of that year. My city has gotten used to an increasing tax base of 7% per year for 20 years, plus growth in the number of houses, so actually cutting the budget by 7% will cause a lot of political problems. I imagine a fraction of cities facing this issue won't make it without getting more money to finance the shortfall. It won't hit this year, but it seems like a train wreck coming. Poor cities, I'm sure, won't get more money, especially now that we know the muni insurers (Ambac, MBIA) can not handle a broad set of defaults.

Tuesday, March 11, 2008

Great Spitzer Quote

I don't care about Spitzer's sex life, but I spit my coffee out reading this take on the high end escort service Spitzer used:

For its members, Emperors' Club isn't a whorehouse. It's a whorehome"

Monday, March 10, 2008

Michael Lewis Exposed

I remember way back when I read Liar's Poker, and thought, good read, but wrong. That is, to the extent Michael Lewis stuck to personalities, anecdotes about the rich and famous, and allusions to his major of art history, he was great. When he made riffs on economics, such as, that it's all based on hype, or everyone is a big faker, it was an overstatement to a considerable degree, and degree is ultimately important. It's not all a game of BSDs playing macho games, there are some real issues about pricing and risk management, which Lewis was oblivious to in his brief career there.

His latest article in Portfolio.com notes that the subprime mess is because...Black-Scholes is wrong! I worked with some guys who thought Black-Scholes was wrong, and would write papers to journals, and didn't get any traction. The key is, it is not 'wrong'. It has been proved several independent ways. It is--sacre blue!--based on incorrect assumptions, and since day one, people have been addressing what happens when this assumption is changed, etc. Models are always wrong, sometimes useful. Lewis doesn't really understand finance, which is why he would be shocked to know that people were layering B-S with outside the box adjustments before he went to college. But even if Black-Scholes is predicated on the idea than you can sell a stock at its price, and this is wrong (er, so the price is not a price?), what the heck does that have to do with subprime? The subprime mess was about as related to Black-Scholes as the Dividend discount model, or the Miller-Modigliani theorem, or Fisher's breakdown of the interest rate into a real rate and inflation expectations. They all involve numbers, finance even, but not really the same. If they were relatives, they would be 'third cousins once removed'.

Sunday, March 09, 2008

Hedge Fund Genius Exposed

Andy Lo wrote a neat article on how stat-arb has done in the recent downturn. What I found most interesting was that the strategy he examined as the archetype of 'stat-arb' was mean reversion using daily horizons: past day's returns to determine winners and losers, hold for a day. I knew pairs was a common strategy, and generalized into a long/loser and short/winner strategy, but I had no idea people ran this at the daily horizon I always assumed 5-15 day. That's a lot of transactions, clearly a strategy that would benefit from great order execution systems.

But what is really neat is it highlights that trades that work are often quite simple. It's always funny to see that a trade of going-long-losers-and-short-winners, is often managed by a PhD in math. I guess it's no different than a 3rd string violinist, playing easy scores, but trained to do much more difficult work.

Saturday, March 08, 2008

Cash Flow Still Working



Cash flow is the most powerful single factor in explaining cross-sectional equity returns (see references here). Like mean reversion 10 years ago, the question is, has there been too much entry? Is it done? In an interview with some anonymous trader, there's the suggestion that cash flow has been a key to the underperformance of funds like Goldman and AQR. Think again. Some factors are getting stepped on, but thusfare, Cashflow is not one of them.

Thursday, March 06, 2008

DefProb vs. Merton



Just visually, you can see the comparison between the scatterplot of DefProb and spreads as compared to the scatterplot between the Merton model and spreads. It isn't easy to see, but DefProb is more powerful. But on one level it gives you the sense of power in these models. The key is a realistic benchmark, such as naive alternatives like Merton and Agency ratings. You can definitely do better. Not amazingly so, but you can do better.

DefProb Business Plan Revealed

The wisdom of the gnomes!