Imagine you are in charge of a futuristic Formula racing team. The rules are as follows:
- Up to 10 cars permitted in any one race.
- The course will only be revealed AFTER the start of the race. Courses vary from the old type Formula 1 tracks, to extreme mountain off-road tracks to tracks with unusual features such as trenches etc. Height may vary markedly on the track, sometime precipitously.
- No car changes are allowed after the start.
- Points are awarded to all finishers of the race, the winner taking 10, the second 9 points and so on.
- Cars which are too severely damaged are out of the Championship, no extra cars are allowed once the year has started.
What is the strategy involved to win the Constructor’s Championship? Easy; use the analogy of portfolio management but do the OPPOSITE of most of what Modern Portfolio Theory (MPT) recommends.
MPT would tell you to always pick the fastest cars with the best track record on most tracks (highest return/lowest risk). No need for more customized cars as most tracks are like the old style Formula 1 tracks. Diversify to about 10 cars and that will ensure you always have a potential winner no matter what the type of track. The cars will seem uncorrelated until you get to a really rough track and they will all correlate by NOT finishing!
Modern Racing Theory (MRT) says almost the opposite.
- Firstly, you cannot predict what sort of track you are going to get so don’t even try. There will be a lot of people telling you after a long run of “easy” Formula 1 tracks to choose cars which have the most speed (return) at the expense of the cars which are better suited for more extreme tracks. Do not follow the advice of these people… the year is long, they will eventually pick the wrong track with the wrong type of cars and crash them all, unable to finish the race and ultimately the Constructor’s Championship.
- Pick a balanced portfolio of cars whose characteristics differ markedly from each other (low correlation) which offer an ability to thrive in any type of race track. The number of wins won’t be as big, but neither will the number of crashes and poor places.
- Inform the drivers to be aggressive if the track suits them and conservative if it doesn’t. That way all the cars should live to race another day on a track which suits them (risk control).
- Inform the drivers that should the track become dangerously volatile height-wise, they are to become extremely conservative.
- Don’t ever knowingly risk a total write-off of the car. The unexpected crash will happen, just make sure it doesn’t happen to too many of the cars.
- Drivers to be in communication. Should the lead car come across an unexpected trench or such-like, it will be conveyed immediately to allow the rest to circumnavigate it.
Most of the teams will be overly aggressive, hiring forecasters at 2% to predict what the next course will be like and what kind of mix of cars would be optimal. They will ignore the possibility of rare occurrences like extremely curve-ridden, up and down tracks in bad terrain. Most will be overly optimistic about their drivers’ capability. And most won’t end up with more than 3 cars left in the Championship by the end of the year.
Portfolio management is like this Championship. You have a portfolio of assets which you need to shepherd to retirement age. You are faced with an uncertain future. You have a finite amount of money that cannot be replaced easily. You have variables which can trade-off against one another, i.e. risk, return, correlation and convexity. How you manage them will determine in what state the portfolio reaches the “end”.
Having funds in a portfolio which are negatively correlated to “traditional” portfolios are like having Hummer off-road vehicles in the racing analogy. They won’t win race on good tracks but you already have cars for those scenarios. They will however shine when the tracks turns ugly and “traditional” cars cannot cope.