Sunday, September 25, 2011

Is High-Frequency Trading (HFT) Harming Progress?

In part, yes. My thoughts below.

There are many definitions of progress, but I am going to skip over the more politically convenient ones. I view progress as truly novel ways to do old things (i.e. cell phones, internet, alternative energy), getting rid of widespread diseases (i.e. polio), order of magnitude cost reductions (i.e. chips for calculations), and the like. One can include things we can't imagine yet (past examples include plastics, telephony, flight). Twist-off beer caps, heated car seats or seaweed yoga pants do not count as progress.

For thousands of years, individuals have been able to make various important discoveries and inventions by and large on their own, or with relatively modest capital investments. These days are gone, for the most part: no one truly expects to mix up a cure for AIDS in a basement or to sequence the human genome on the kitchen table over yesterday's NY Post. The next large steps in progress will- in all likelihood- require that the person is a part of a large organization such as a university or a corporation. The next large steps would also likely require deep specialization in a certain field: after all, most- if not all- of the low-hanging fruit is gone, and the new generation has to step on the shoulders of the giants of yesterday.

Progress is in some aspects also a product of luck: you need lots of people doing lots of things, and some will be successful. These people nowadays are usually hard science PhD-level individuals- and, while I am mindful of individual choice and preferences, it seems that society (and progress) would be advancing faster if the high level of specialization was put to work in the proper conditions.

The problem is that less and less of this is happening. Big Pharma has been shrinking for years. The old tech giants seem more interested in behaving like patent trolls rather than funding R&D (for example, what used to be a highly innovative company, Hewlett-Packard, has cut its R&D from 6% of sales to almost 2% of sales in about a decade). R&D is viewed as a cost and is high on the list for outsourcing, creating even less incentives for people to take risks by working on what might be career and scientific dead-ends for a disloyal employer. Who is picking up the talent?

One industry is high-frequency trading. The low-to-negative societal value of co-location, quote/cancellation tricks in nano-seconds, order sniffers and the like is established beyond doubt, in my view. But these companies are dangling $300k++ packages to physics and engineering PhD's straight out of school to come up with even more innovative ways to milk fractions of pennies from your orders. This effectively removes talent from what might be the "unimaginable" developments and puts them to addictively lucrative work that one might argue contributes to regress, not progress, by undermining the faith in and the integrity of the capital markets (the flash crash last May being a recent popular example). So, it does seem to me that HFT is impeding human progress.

Sunday, September 4, 2011

Andrew Redleaf Lecture at Yale Fall 2008

This is a lecture by Andrew Redleaf, co-founder of Deephaven and Whitebox Advisors. It was given in Fall 2008 to Prof. Robert Shiller's class at Yale (Shiller of the Case-Shiller index and Irrational Exuberance fame). Redleaf is not a natural public speaker ("graduated from Yale University in three years with a BA and MA in Mathematics and was recognized as the top mathematics student of his graduating year in 1978"), and there are no slides, so here are my notes. I have bolded what I found interesting.
Efficient market theory: two camps, yes/no.
A theory has to be predictive, and EMH is not: dual class stock, CEFs discount/premium to NAV, stubs, less volatile outperformance, certain outperformers, bubbles all refute EMH
Once the hardcore academic believers die, it will not even be an open question whether markets are efficient

But doing better than the markets is non-trivial (not impossible but hard)

For individuals, should stay within area of information advantage but they will not be diversified; also individuals have more limited access to products and information; less assets

For institutions, there is the constraints: liquidity requirements by outside investors (i.e. 30-day notice); institutional biases stemming  from individual biases; problems with decision making in the areas that have a mix of randomness and skill; understandable areas are fully random are easy; poker is similar to markets because of the randomness/skills mix; often someone who plays will admit to not being that much into it, why would they do tell you? Pet theory: setting up in advance to lose and easier to deal with internally . Other individual biases are causality just because one thing happened after another, recency, anchoring

What they try to do:
3 principles
(1) The investment world divides b/n coupon-clippers (cash flow-oriented investors) vs. re-sellers (owning to re-sell to someone at a higher price). There are successful people in both areas, and they make their money off the less successful people in the group. They are about the coupons: focus on cash flow. VCs are re-sellers: find an idea, sex it up, and get a good valuation from the public.
(2) Risk is primarily about what is the worst thing that can happen. Typical Wall Str risk management is about VAR: how wide is the range of probable outcomes. The VAR approach is fundamentally wrong but it done because statisticians are good at it, and it is hard to imagine the worst that can happen.
(3) Most at odds with the academics/general perception: you don't get paid for taking risk, you get paid for eliminating it. In the history of the planet, very few people have been paid for taking risk. A tightrope walker in the circus is an example: it is paid very well but you have to look at lifetime income, it is not that great if they fall. They are paid well because they have been working on training not to fall (=eliminating risk). Pharma drug development is uncertain risky business. Are they paid for the risk? No, they are paid for applying intelligence and eliminating certain compounds and combinations in advance. Insurance companies: if they are doing their job right, they should be avoiding risk. A casino does not have the risk that they will lose money because the law of large numbers takes care of it (the risk is in traffic/visits).

So we look at what is the coupon, what is the worst outcome, and how can we eliminate it? Sometimes it is diversification (independent positions; or sometimes offsetting positions, such as long a high yield bond and short the common in proportion is a simplistic example of clipping a coupon and reducing risk.

The next part is a discussion of events happening around the time of the lecture (November 2008) and Q&A.

Q: Are markets less efficient if prices are going up vs down?
A: No systematic view, but in recent years, more examples of overpricing but not clear if this is due to money creation.

Q: ??
A: An individual's default position should be cash; should invest only in compelling cases. Equities get cheap periodically by most standards. Recommends buying stocks now (Nov 2008). With the housing bubble taking off a few years ago, apartment REITs were interesting (below replacement cost) with compelling comparative yields (5-9%). Municipal bonds now yield more than equivalent treasurys, that's compelling. Has not happened since the fears of them losing their tax exempt status.

A: The long term argument for stocks in the US has strong survivor bias and has benefited from the US switching from agrarian to industrial. It is not clear at all whether the next 100 years will be like this.

A: Swensen of Yale Endowment has strong preference for equity; Redleaf likes fixed income securities a bit more because they are more analyzable inherently. Stocks have to be really cheap because they don't pay out earnings and in the future, they might do something stupid. Buying stocks at mid-single digit multiple for a good reason, you will get paid twice: multiple expansion and earnings growth. The market is telling you though that the earning are going down if the P/E is mid-single digit. RIMM is risky: very high multiple, it might not matter if they double earnings, and they'll never pay any dividends.

A: Involved investors are good (activists or direct owners). Executive comp is not the result of arms-length negotiations unlike hedge fund comp which is market-negotiated.

Final Q:??
A: Lengthy discussion of special situations where there is an event of default for delayed financials filing, and how they are active in it. They co-owned Sun Country Airlines with Petters Group (now a known massive ponzi scheme)

Thursday, September 1, 2011

Why ZIRP is Bad for the Middle Class

The Fed announced recently that they will continue with their zero interest rate policy (ZIRP) at least until 2013. May be they want to see a wholesale flight out of paper dollars like we saw last fall. Regardless of the market implications, I think that ZIRP is bad for the middle class in many ways, including some that are less quantifiable. As Charlie Munger says, not everything that counts can be counted.

A fundamental tenet of "middle class-hood" is future orientation (vs. present orientation). People with present orientation like to do things without regard to future consequences (driving drunk, having unprotected sex, skipping class, buy now/pay later, etc.) A core middle class value is future orientation: decision-making that takes into account future consequences. This has meant staying in school, not fathering children left and right, and, the focus here, saving: saving for a downpayment on a house, saving for a car, saving for college, saving for retirement. ZIRP discourages saving by stealing the interest earned on savings and encouraging present consumption. This leads to both overconsumption (relative to incomes) and to lower retirement savings (and, in part, to defined benefit plan underfunding).

ZIRP is also bad for middle class jobs. In many businesses there is an optimal choice between capital and labor (sometimes not direct, make something manually internally vs. buy machine-made from outside). A factory owner would love to fire all workers if robots were cheap enough: I know I would prefer to deal with machines than people who stay home when their kid is sick or walk in the office with an offer letter, asking for a raise. ZIRP (and Obama's accelerated depreciation stimulus) heavily favor machines vs. labor. I am all for progress/mechanization but I think that the relationship has gotten a bit distorted.

ZIRP leads to malinvestments: we had a housing bubble a few years ago. We not only ended up with too much (and highly energy-inefficient) housing but also we ended up with a large % of the workforce in the bubble-fueled real estate economy, from the slick jumbo mortgage broker to the gruff plaster guy. Some of these were formerly middle class jobs, especially in smaller communities, and now the sheer decline in business volume has hurt them. The decline in housing prices post-bubble has also cost the middle class dearly in equity.

ZIRP speeds up job exports. Large international companies, like MCD or WMT, can borrow at very low rates here in the US and invest these money at much higher returns in emerging markets. These create not only MCD or WMT front-line jobs: with every expansion come the accountants, the marketing personnel, the buyers, the logistics, the IT, the management, and, perhaps most importantly, the R&D. GE just moved their X-ray R&D work in China. This probably would not have happened so quickly (if at all) if they could not borrow so cheaply to build or acquire abroad.

ZIRP is an implicit policy of dollar devaluation. People argue about the mechanics but the message is that the dollar will be weaker. This means inflation as the US is an oil importer. The middle class is already already struggling the points above (unemployment, reduced wealth, no interest on savings), this hurts even more, be it at the pump, the heating oil delivery, the airline fuel surcharge or the consumer-level price increases due to higher energy costs. The winners here (exporters such farmers and miners) are a small percentage of the workforce. The middle class is not a winner from $120 oil.

ZIRP discourages real capital formation. This is a off-shoot of the savings and employment point: capital formation for small businesses comes from savings- we are talking about rentals, restaurants, car washes, chips delivery routes, hair salons, food trucks, small shops, etc. Most everyday people are not into social media start-ups. A potential entrepreneur (and his/her family) actually loses money after taxes and inflation while trying to save up for the business. This hurts both entrepreneurship and, consequentially, hiring (oft cited stat is that most jobs come from small businesses). There is an argument that loans are cheaper, too, but the contra here is that higher equity businesses are more robust.

ZIRP needs to end. To give you an idea of where the rates "should" be, for a 1-year CD to earn a meager 1% in real after-tax return, the 1-year CD rate should be at about 5.75% (simplified calculation: 1% real return + 3.6% LTM inflation =4.6% after-tax, at 20% total tax rate-Fed/State/Local- this implies 5.75% pre-tax). The average one-year CD is at 0.84% today per  Yahoo Finance. Who benefits from this? Three administrations have worked to double the size of the government in 10 years via cheap deficit spending, while the banks are getting nicely recapitalized.

A Better Way to Hire?

The labor market has been a challenge for years. A part of the problem is that the market one of the least efficient markets out there. Many transaction opportunities are not advertised, limiting the flow to a small network. And even the more widely advertised positions often have a gatekeeper: HR or recruiter whose short attention span (and, often, lack of ability to discern quality) means that only the "best-branded" resumes get a priority. Like in the supermarket, going for the brand makes the buyer feel more secure and reduces time spent on making decisions. This is a part of the larger credentialism problem that we have as a society.

So is there a better, more meritocratic way to hire? There is a good recent example of that. A NYC-based hedge fund manager, Aram Fuchs, was looking for a junior and a senior analysts to work for him. Instead of playing the "resume game," as he calls it, he managed an open forum, the Capitalist Collective, where anyone, regardless of background, could submit stock ideas, discuss others' ideas or engage in general discussions. Mr. Fuchs recently announced four "finalists", and the process will culminate in a day-long investing conference on September 9th, 2011, featuring speakers, stock speed-dating, and, of course, the grand finale. Along the way, Mr. Fuchs hosted a couple of socials and had calls with some of the participants.

While obviously longer and more involved than the traditional route, the process draws from a much larger audience and the hiring person can observe and interact with the applicants over weeks, thereby developing substantially more valid views on suitability and other important factors. The forum format also allows a lot more contact points with people who are currently working elsewhere. There are unlimited opportunities to show and defend ideas rather than the regular 1-2 stock pitch approach. Finally, the participants know that the opportunities advertised are "real", unlike the phony "openings" that collect stock ideas without any intention to interview or hire.

Let's hope that this approach becomes more widespread with time.