Designing better futures

VaR explained & the BS model (not Black Scholes)

Value at Risk is pretty simple.  Someone is making a guess about what they expect to happen 95% or 99% of the time.  i.e. 99% of the time you will lose less than $1000 a day. etc.

What many people forget to focus on is the fact that what happens the other 1-5% of the time is what counts.

What many people don't know is that there are 3 types of VaR.  The Risk manager presenting the report usually just mentions the number and that is all, but scratch below the surface and you can learn what your risk manager is really thinking.  Next time ask them what type of VaR they are using and keep this handy explanatory list nearby to decipher what they are saying.

1. "Well, basically we use Historical Var compiled from the data."  (I am going to let history bullshit you.)

2. "We use our mighty computer resources to calculate a Monte Carlo simulation for VaR."  (I take history, blow it into a million little pieces and then each piece will bullshit you ever so slightly.)

3. "We have a sophisticated financial model creating a robust Parametric VaR."  (I am going to personally bullshit you with a tweaked Garch model or overly sensitive Levy Distribution)

Not knowing much you will nod and agree that, "the number looks about right."  and sign off on the risk limits, now you are bullshitting yourself and the whole team.

I am not Against VaR or any other standard risk tool or view.  I am against any tool which becomes institutionalized.  Once a risk tool or approach becomes a wink and a nod, sign off or a CYA calculation it loses value.  Risk management is meant for curbing human behavior, otherwise individual or collective greed wells up and overwhelms an organizing system of faith, which is what finance and money are.

Examples of institutionalized risk management beliefs that have or will become rubber stamps:

• Ratings Agencies (NRSRO ministry of truth)
• Securitisation magic (of course we all know that now)
• Basel II Capital Accords
• Probably the next financial innovation which will yet again revolutionize access to credit
• Any company wide policy which creates a ...Yeah but it passed the X test. Or well they signed off on it...type of mentality.  See Dilbertian risk management.

The most dangerous risk manager or system is the one that becomes the de facto agreed standard we can all agree will stand the test of time.  Once a unified risk belief is shared in aggregate all participants will start gaming or trading it the same way.  When all participants share a similar view, the system is susceptible to harmonic dynamics.  See the video below.

US Yield Curve Mambo 3 month to the 30 year Dance

I am looking at taking a role as a prop trader/PM with RCG (Rosenthal Collins Group) in Chicago.  They do a bit of Yield curve trading among other things, so I thought maybe I could liven the data up a bit.

The typical yield curve looks boring and doesn't show relationships over time or the whole yield matrix.
Being interested in getting a feel for the data means understanding all the US swaps spreads. So I put them into excel and created a nice scrollable  matrix.
 
But hey even that isn't fully helpful and discovering anomalies in data (which is fun for me) means getting a feel for the relationships.  So I ended up with something like this video.

Trading the curve still means ferreting out the anomalies, but I have already found some things in currency swaps. I am hopeful the US  Mambo will yield some alpha (insert bad pun groan).  One caveat, the data matrix didn't have data for the full duration, for example 7 yr. issuance is fairly new etc., and the curves aren't interpolating the null data.  This is more for fun than information value.  The flattening of the curve and then the spread gaps are interesting to watch from 05-09.

If you want the yield curve mambo get it here from on the Godsdog album from De Phazz

If you are an educator you may use a simple video editing program to illustrate the flattening of 2005 and the risk expansion during the crisis.  Here is the Raw AVI file for editing and educational purposes. Enjoy.Download Testfilm3 (50MB).

Philanthropy evolves: important

Philanthropy changed with Andrew Carnegie and Rockefeller.  Today there are new forms of making a difference.  Katherine Fulton in this video highlights some of these shifts in a good TED talk.  She evens mentions Wikipedia as philanthropy etc. and how little acts are contributing to big things in interesting ways.

Her Monitor institute is here.

Unsexy but vital: Bureaucrats without Borders

The title of this blog Designing Better Futures reflects an idea.  The idea is that Design (the thoughtful act of production) might help make the future better for many.  I am interested in design with a high ROI in sustainable economic development for poor countries.  My beliefs are more in line with William Easterly than Jeffrey Sachs.  To put it simply Easterly wants it aid to accountably to work, while Sachs wants to throw money at it until it works and feel good about it. 

If you really want to make a difference, please think before acting. Feeling good about yourself and doing good for others aren't necessarily the same thing.  The best results rarely are the ones that make you simply feel good. Real development impact is unsexy hard work, like building or designing a latrine to reduce infant mortality.  

Read Easterly's fantastic book, The White Man's Burden and then compare it with the feel good money pit approach from Sachs entitled The End of Poverty.  Another thoughtful book on world problems is The Copenhagen Consensus.

In looking for high payoff sustainable development ideas, I have been very interested in micro-nutrients, iodized salt, Vitamin enhanced golden rice etc. 

Basically I look for cheap projects with huge positive externalities,mega-bang for the poor persons buck.  These are things that can cheaply add years to life capabilities or raise the collective IQ of a village sustainably without creating aid dependencies etc.   The trick is in the distribution and self sustaining of the enhancement in the poor's lives over time.

Another high value add, that I am deeply committed to is Property Rights and the wonky delegation of those rights via efficient registration processes.  Recognizing Dead Capital* as defined by Hernando de Soto and properly bringing it to life could create Trillions of $'s for the world's poorest and enfranchise them in their own political systems.  An enfranchised middle class is a powerful force. 

To that end, the unsexy stuff needs technical help.  Forget the easy emotional appeal of feeding babies in times of crisis, put the unsexy infrastructure (soft(bureaucratic) & hard(project based)) in place to mitigate mass starvation in the first place. 

Medecins Sans Frontier (doctors without borders) is very high profile do good stuff, but what could even be more important would be plumbers, engineers, lawyers and tradespeople without borders.  The resources (motivated people) are probably there, the organized charities aren't. 

Imagine plumbers or civil engineers or Quality Assurance people helping out in countries on the "soft" infrastructure, either training or jsut learning about it.  My own goal would be land registrar consultants without borders.  I view it is a $9 trillion unsexy problem.

A good example of an unsexy but important charity is Reporters sans frontiers.  Good reporting acts as a feeback loop and curb the abuses that may lead to socially inflicted humanitarian disasters.

The video below hinting at the need for Bureacrats without borders, warning it dry and wonky.  A council on best practice for bureaucracy inter governmental with benchmarks such as the World Bank's excellent doing business report would be a very good thing as it measures the "output" of a govt. service and allows for govt and citizens to hopefully seek best practices from peers.

Imagine the same for the planning processes for infrastructure, planning and deployment of basic govt. services.  Bureaucrats without borders sounds horrible, but could be wonderful, it may even get rid of some corruption which acts in some instances as the only effective way to progress in some states.

*Unbelievably there is no wikipedia entry for such an important concept as "Dead Capital".

Full disclosure Bernard Kouchner, founder of Medecin Sans Frontiers spoke at my MBA commencement in Paris and my undergrad at the U of Iowa was anthropology where I independently focused on economic sustainability in West Africa.

Minsky Moments, Wilmott's Magician and a lack of imagination.

Money is not a physical reality or absolute,  it is a collective social belief tool that works very well.  My anthropology is showing.  Money in large collections reflects trust and faith pooled by a group.

The art of managing risk is currently messed up as it is considered a physical science when in reality it is significantly a social science.  The tools, language and metaphors used to describe risk management are mostly 2nd rate physics and introductory probability.

Paul Wilmott has an excellent example of people thinking in boxes and it involves a simple question about a deck of cards and a magician.  The Original post is here.  The all too scary follow up post is here. 

Willmott's magician is as important if not more important than Taleb's Swan in my opinion.

The disconcerting nature of the Willmott follow up article is that it shows people in roles of risk management woefully lacking even the basic skills that it requires, namely imagination and mental flexibility. 

The risk managers who can't wrap their heads around Willmott's Magicians test seem to have some form of cognitive aphasia which has removed all common sense from them.  In a word, they are so enamored with the "laws" of probability that they forget they live and manage a social resource.  Money is a social phenomenon, gamed individually or collectively constantly by people, not diffusion equations with the appropriate drifts.

If you are interested in risk or understand the reason that human behaviour means that the next minsky moment will arrive as soon as we collectively forget the last one, please read the Wilmott Blog posts and think about the skills in a risk manager. 

If your risk management process is lead by PHD's who don't pass the magician test, please put them in a room or island where they can do little harm or better yet, get them into a math dept. where they can do some good.

I am still advocating the book Since Yesterday first written in 1940 as a great way to understand the thinking, events and cultural process known as the great depression. Spoiler Alert:there was a mini echo depression in 1937 when the "stimulus" of the era disappeared.  Since Yesterday is a very rich read and maps disturbingly well to today.

About that social experiment.  This cartoon sums up the interaction between risk management and people's desire quite nicely from Jesse's cafe Americain.

Here is a link to the original source of the comic Abstruse Goose. Tip o' the cap to Noah Yetter.

Minsky moments won't leave us, we are human after all.  Shakespeare figures, bubbles and greed won't go out of style, they are classics and endimic to the human cultural condition, please find risk managers who understand that better than physics.

Top Marginal Tax rates since 1913 & Great Depression

If the shoots are green what color will the next tax increase be?  This chart data comes from the Daily Dose of Excel Blog.  The US Income Tax was started to pay for the Civil war.  More history here.

Trivia for the day taken from the excellent book "Since Yesterday 1929-1939" by Fredrick L. Allen: 

The top Tax payer in the US in 1936 was William Randolph Hearst.  The #2 tax payer was Mae West with $430k in earnings.  The book is a great read and talks about the culture, fashion, economics and politics of a process known as the Great Depression.  It reads like a newspaper from today which is kind of scary.  Economically we are in 1932-33 as near as I can tell. 

Anyone wanting to understand the unfolding economic and cultural process that was "the depression" may want to read Since Yesterday, it is one of the best books I have read in a quite some time.

Which Beta? I vote Null. Beta is overworked and needs a rest.

Here is a new article about beta and the use of Fundamental indexes at Index Universe.

My own bias is to ignore Beta as it is only moderately interesting.  Most people use it to measure results relative to peers or as some silly index. Absolute terminal real wealth appreciation is all that matters. 

Beta users tend to be either academics or fund managers calculating who can underperform an index the least.  Beta is the vulgar naked cult of fund mediocrity proudly boasting of its sub-par achievements behind the fig leaf of academic group think that has gone too far.

Beta gets used too much.  It is the mirror image of VAR in terms of the way it is abused as a tool.  ie. it reflects a threshold, but really doesn't say much more.  Asset allocation across asset classes is more important in regards to maximizing terminal wealth in real terms.  Mean-Variance as a risk budgeting metric is all but useless. 

Consider those lovely CMBS and CDO^2 products that must have had low beta correlations while showing nice low variance.  Common sense as a risk metric would have shown the gamma risk inherent in these things better than any statistical ftool.  CDO^2 was financial obfuscation taken to new and interesting places. 

The cult of correlation has replaced common sense and Beta leads the charge.  Correlation and covariance are mirages that vanish quite rapidly providing no solace or liquidity when they are most needed in a portfolio. 

The practitioners will of course say the model wasn't robust enough, or the data wasn't there. etc.  and up go the cries of "29th Standard deviation events etc.". Quick, blame the Black swans and get back to the modelling.

Rarely does new and interesting equate to better in finance.  Financial innovation almost always equates to a few people dreaming up a lexicon and new semantic set to make money for a while followed by many suffering over a longer period of time when those collective beliefs in correlation or co-variance vanish.

If inflation shows up in a meaningful fashion, then Beta will be seen to be wanting as a be all end all metric.   In the current equity market, Beta has fallen out of fashion as consumers discover they want absolute and not real returns of course, they still haven't been thinking in Absolute Real returns.  Inflation could be the next interesting adjustment to fund metrics.

Ironically the one place financial managers seem to be comfortable letting go of correlation and co-variance is in the relationship between risk adjusted performance and pay.  Pay in finance is still assymetric(sic) and more correlated with peer expectation than the building of real terminal wealth over the long term.

I vote Null on beta.  Unfortunately most pension fund managers etc. subscribe to this group think and many important institutions will suffer accordingly.  The over-allocation to US equities and US debt could become quite an anchor for many, but at least the giant underfunded pensions and institutions will have the salve of correlated misery to sooth and unite their angst.

The only good to come out of the cult of Beta and mean variance risk metrics is that by herding the sheep together it kept most of them from frequently doing stupid things individually.  Instead fund managers get to do stupid things collectively.

Widget for Converting Sharpe to Daily VaR (value at risk)

Here is a an Excel widget I built for converting a Sharpe Ratio to a Daily VaR.  This is helpful for a trader or asset allocator who thinks in terms of Sharpe's which are annual units of risk to reward and having to provide a daily Value at Risk measure.

To use this widget, after downloading the excel file:

1. Move the scroll bar to reflect your Annual expected/historical return - Risk Free Rate
2. Move the Scroll bar to reflect your annual standard deviation. In Excel= Standard Deviation of Daily returns * (Number of Trading days/year ^.5)
3. Move the VaR Daily limit to reflect the VAR threshold you are comfortable with
4. Move the VAR threshold to reflect the confidence or (rate of occurance) you would be comfortable with violating that threshold.

You will see the #of days annually that the threshold will be violated and the annual returns expected.

All the usual VaR caveats apply.  The assumptions being a normal distribution in the generating process which never holds true and the fact that VaR tells you how often you will fall off a $ cliff, but not how far you are going to fall. Caveat emptor.  I am not a big fan of VaR as people tend to get hung up on it. For true buy and hold investing it just plain doesn't make any sense to use at all. 

For a quant trading approaches VaR is an OK starting point but for a buy and hold or value based portfolio, I put Beta and VaR in the same bucket. Both are silly and dangerous.  VaR of course does little if you have a major gamma risk or high convexity low event type of risk embedded in a portfolio or approach.  It can be argued that a pure Stat-arb has a gamma approach similar to selling premium, but that is another topic for another day.

Book Recommendation of the Day. Excel Hacks

Accountability and Attribution Bias in mutual fund returns.

I finished The Fundamental Index by Robert Arnott, Jason C. Hsu & John M. West a few months ago.  It is a useful read  even if one doesn't  agree with the thesis about fundamental indexing. There are interesting points to be made about many aspects of equity behaviors and the profiles that make up the concept of equity returns and risk.

I offer the following chart as interesting example of the outcome of mutual fund returns for the average investor. 

Most investors ranging from hedge fund pros to the average person are subject to buying high and selling low relative to their own investments whether its a Hedge Fund of Funds, or the average person buying a single mutual fund. 

Attribution bias is the human bias of attributing success to oneself and failure to some external source.  Accountability and responsibility are out of fashion these days, but I thought I would submit the following for consideration. The chart below is found in The Fundamental Index.

Apparent Trojan horse in the Bloomberg directory could be an AVG false positive.

I use Bloomberg  on my laptop to round out my algorithmic trading research. 

My copy of AVG Antivirus just popped a window saying I have the Tiny F. Trojan Horse inside the Bloomberg Directory.

Toby (seen in the comments below) indicates he has contacted Bloomberg and the error is with AVG.  It seems AVG issues a False positive.  So this is most likely not a virus in the Bloomberg system.

Hat tip: Toby