The Sand Pile Effect is a thought experiment that I think helps explain a lot of what we are seeing in 2020 politically, economically, and socially.
Imagine a giant grid. One each square of the grid, is stacked a tiny pile of sand.
We can keep track of how many grains of sand there are on each dot by writing a number on the appropriate square.
The rules for how this system works are very simple:
1. A vertical pile of sand grains can only get to three grains high without falling over.
2. Whenever four or more grains of sand are at the same dot, four grains topple off, one in each compass direction.
So if you start with this:
The pile topples and gives you this:
If multiple tiles are overloaded, then you can have a cascade effect. Let’s say you start with 4 grains of sand in one tile and three grains of sand in adjacent tile.
The first pile collapses, spilling over into the adjacent squares.
This pushes the square that previously held only three squares to collapse, cascading into the adjacent squares.
What’s important about this model is that very small changes over time don’t necessarily lead to very small effects immediately following them. Often they lead to nothing noticeable changing and then a seemingly small incident results in an extremely volatile and violent event.
Let us take this model as a very, very, very simplified model of how many types of complex systems work.
An airplane where the first dozen rivets fall out of a wing might never be missed. But, the thirteenth rivet popping out may cause the wing to fall off. The wing does not get slowly less effective over time. Everything seems fine until, suddenly, everything seems terrible.
So too, in the Sandpile example, the board starts empty and, over the course of weeks, months, years, or decades, the sand starts to pile up on different squares of the board with very little noticeable change.
Eventually, the board reaches a configuration that looks something like this.
To the casual observer, the piles have gotten bigger but there hasn’t been much volatility or change. However, a single grain of sand on any pile at this point will lead to a massive cascade across the whole system.
The pattern we see in markets is that volatility tends to cluster around a lower level, transition, and then maintain an elevated level for a period of years. This is what we would expect if markets behaved like the sandpile.
In the lead up to the 2008 financial crisis driven by mortgage backed securities, the changes in underwriting standards and slowly increasing default rates began showing up many years before the actual crisis. The grains of sand kept mounting and to the casual observer of markets, everything seemed fine.
When asked how he went bankrupt, Hemingway replied simply “Gradually, then suddenly.” This is how pretty much everyone goes bankrupt. Like a plane that has lost a dozen rivets on its wing, everything seems fine until suddenly it seems very very unfine.
I believe this is because most investors are, unbeknownst to them, short volatility – that is their portfolios are positioned to benefit from periods of low volatility (e.g. 2003-2006, 2012-2019) but be harmed in periods of high volatility.
In the same way someone can short sell a stock to bet against it, many people are short volatility in a way that is not immediately obvious, but worth unpacking.
Consider a typical individual that has a retirement portfolio full of stocks, a mortgage on their house, and a job or business. All of these assets are effectively betting against volatility – they all suffered in 2001, 2008, and many crises before those.
Home values are tightly linked with the local labor market. What determines rents is the availability and median wage of jobs (Exhibit A: the Bay Area. Exhibit B: Detroit.). The local labor market is linked to the overall business cycle, just the same as stocks. So when stocks go down, home prices tend to do the same.
Your job or business is generally linked to the business cycle and benefits from periods of low volatility. When markets go down, you may be the person getting laid off or with reduced income. As a result, your job or business is effectively additional leverage which is betting that the good times keep on keeping on.
Someone who buys a home where they put 20% down on a mortgage and a career in a cyclical industry holding a portfolio of stocks and bonds could be seen as making a leveraged bet on volatility remaining low.
While it’s anyone’s choice to do that, it doesn’t make a whole lot of sense in my opinion. The logic is akin to playing a game of Russian Roulette where surviving wins you a million dollars.
While it’s possible (even likely!) that you get lucky on the first round, eventually something is going to happen that is going to harm everything in your financial life at the same time.
You might roll the dice and take $1,000,000 to play Russian Roulette one time (though I wouldn’t advise it). But there’s no amount of money that would make you play it 6 or more times.
Since 1957, bull markets tend to last about five years though this most recent cycle has lasted considerably longer. 1
I believe the key from a financial perspective is to prepare for a situation where the markets drop and your income gets hit at the same time. In one lifetime, it will likely happen.
For most investors, that means having exposure to some strategy or asset that is long volatility, that can benefit from volatile events like the 2008-2011 financial crisis. Combining assets that benefit from volatility, long volatility assets, with assets that are harmed by volatility, short volatility assets, an investor should be able to do well in all environments – high volatility or low volatility.
If we look at a simple toy model portfolio that is 80% S&P exposure and 20% Long Volatility exposure (In this case spot VIX, basically an index measuring volatility)2, we see something that performs like we would want a properly diversified portfolio to perform.
When volatility picked up during the Dotcom crash and 2008 global financial crisis and the S&P saw significant declines, the VIX index rose and helped to zag while the S&P zigged.
The results of the combination are pretty amazing. $100 invested in the S&P alone in 1990 grew to $813.61 by September 1, 2019. Over the same period, $100 invested in the combined S&P and VIX portfolio resulted in $14,700.41, an outperformance of 1,861%.
Beyond that, combining short volatility (stocks) with long volatility (VIX) had smaller drawdowns. You can see the large sell offs in the S&P in 2001 and 2008 are much milder in the combined portfolio because the VIX rose during those crisis periods.
The trouble with this particular example is that the Spot VIX is simply an index, not a tradable or investable product and so this exact portfolio couldn’t exist. 3 However, it’s useful as an “intuition pump” for understanding the concept of diversifying between short volatility and long volatility assets.
I have found myself frequently quoting Vladimir Lenin’s remark that
“There are decades where nothing happens; and there are weeks where decades happen.”
Short volatility is for the decades where nothing happens.
Long volatility is for the weeks where decades happen.
Combined, they are ready for either.4
This combination of long volatility and short volatility is very much the philosophy behind Mutiny Fund. If your interest is piqued by this concept, you might enjoy our research on the 100 Year Portfolio or the first 5 episodes of our podcast where we explore these topics.
The Volatility Index (VIX) was conceived after the Black Monday crash in 1987. It uses option prices to gauge expectations for the speed and severity of market moves. If options are expensive, it suggests that market participants are worried and willing to pay more for insurance. If they are cheap, it suggests participants are unconcerned.
You can think about it like a real time reading of prices for insurance on the US stock market, as represented by options on the S&P 500. If it were car insurance, imagine it like insurance re-priced in real time, with premiums going up when you get in the car, go faster, when it’s raining, and going down when you pull into the garage at night.
The options prices that feed into the VIX calculations are only designed to be looking about 30 days out, so it’s an estimate of short-term volatility. It’s historical average is around 20, but it tends to spike during stock market declines so it can be an effective form of long volatility. On October 24, 2008, during the heart of the financial crisis the VIX reached 89.53.
According to the CBOE, From January 1, 2000 to September 28, 2012, VIX moved in the opposite direction of the S&P 500 about 80% of the time. The VIX then, and the options it is derived from, are much more directly long volatility. The Vix and options are more likely to benefit from volatility because volatility is defined by options prices than something like bonds which relies on a historical correlation and investor behavior. This makes it a good candidate for behaving more like the hypothetical zag.
- The VIX is based on options prices as mentioned in the footnote above. Options have a property associated with them, known as time decay, or theta. Let’s say you buy an option on a stock that comes due in six months from now. If nothing else changes, the value of that option decreases a little bit every day because there is less time left until it expires and thus less “optionality.” Any long volatility strategy will have to deal with this and that will necessarily reduce (or eliminate) the outperformance in the example here. So, again, just a toy model
- Or if we want to get real nerdy, they are more ergodic.