If you are at a casino and you are not sure what number is next at the roulette table, you might put €20 on two options instead of €40 to increase the chance that you will win, at the price of having a lower payoff if you do win.
Protection against the unknown
In investing diversification allows you to protect yourself against things you are not sure about at the price of having lower returns. For example, let's say there are 100 start-up companies you could invest in. Statistically one of them will return 10,000%, the rest 0% (because they fail). You could:
- Bet €100 on one of them, which gives you a 1 in 100 chance to make €100 * 10,000% = €10,000
- Bet €1 on each, giving you a 100 in 100 chance to make €100 * 10,000% / 100 = €100
In the above example the high risk strategy is basically a lottery ticket, and the low risk strategy returns the money you put in in the first place.
Neither are ideal. What you want is to pick the companies in the list of 100 that are more likely than the others to be that 1 winner. In investing that means you want to make decisions that decrease risk more than decreasing returns.
In the words of Warren Buffett:
Diversification is protection against ignorance.
The more you know, the less you need to diversify. If you happen to know which 50% of companies in the startup example don't have the expertise to build the product they want, you can ignore those and increase your chances.
Reducing single-stock risks: index funds
Let's say you want to invest in the stock market. You could pick a stock you like and invest all your money in it. That would however open you up to a lot of risk. Perhaps you think you know something the rest of the world doesn't (for example you did deep research on they market and the rest of the investors did not). You still are not protected against unpredictable events like:
- The CEO getting hit by a car
- The whole sector going down because of a disruptive new technology
- Macro-economic factors like a recession the company had no control over
Instead of picking one stock, you could buy a composite of of the entire market. An example of such an index is the S&P 500, which tracks the biggest US companies. Buying one piece of the S&P 500 means you own a tiny bit of all sorts of companies (a bit of Apple, a bit of Unilever, a bit of companies you have never heard of etc).
If you want to diversify (ie spread risk) even more, you would buy an index of the entire world like the MSCI ACWI which covers both the developed world and emerging markets.
Reducing market-timing risk
The stock market moves up and down all the time, which is normal and to be expected. But what if you save €2000 in a year and decide to invest it in december, and for some reason the stock market is up/down 20% that december (for example like it did the christmas of 2018). This would mean you open yourself up to great variability, both to the up and downside.
To manage this risk, we can diversify across time by cost averaging, often referred to as 'dollar cost averaging' because much investment writing is done in the US (so ignore the 'dollar' part). In practice this means that you enable a monthly transfer from your paycheck into your investments.
I highly recommend doing this. The only potential complication is when you have a pile of cash to invest. In that case I recommend reading this vanguard paper from which I quote:
To calculate the average magnitude of LSI [ lump sum investment ] outperformance, we calculated the average ending values for a 60%/40% portfolio following rolling 10-year investment periods. In the United States, 12-month DCA [ dollar cost averaging ] led to an average ending portfolio value of $2,395,824, while LSI led to an average ending value of $2,450,264, or 2.3% more. The results were similar in the United Kingdom and Australia: U.K. investors would have ended with 2.2% more and Australian investors with 1.3% more, on average.
In short: cost averaging is good if you have a monthly paycheck, but needs more consideration if you have a lump sum to invest.
Reducing correlation risk
If you fear risk factors against one type of asset group (e.g. the stock market) you can try to find other investments that are uncorrelated, meaning they go up and down regardless of what the other asset does. For example (these are not factually correct):
- If stocks always go up when the gold price goes down, their correlation is high and inverted
- If house prices always go up if gold goes up, their correlation is high in the same direction
- If house prices and interest rates move up and down no matter what the other does, they are uncorrelated
Some things historically are uncorrelated, and some things tend to be correlated and uncorrelated depending on the time frame.
An example of a correlation-based strategy is Ray Dalio's all weather portfolio (source):
“ You can immunize yourself from the cycle by holding a balanced portfolio of assets,” Dalio told CNBC Make It in September. And Dalio has a simplified “all-weather” asset allocation formula that just about any investor can use.
The makeup is as follows:
This all-weather portfolio, which Dalio created for the Tony Robbins book, “Money: Master the Game, ” involves a mix of 30 percent stocks, 40 percent long-term U.S. bonds, 15 percent intermediate U.S. bonds, 7.5 percent gold and 7.5 percent other commodities. The portfolio needs to be re-balanced annually, Dalio notes.
The idea is that the assets and ratios Dalio chose are designed to balance each-other out in different market conditions.
“When you look at most portfolios, they have a very strong bias to do well in good times and bad in bad times,” Dalio says in Robbins’ book. The typical portfolio has somewhere around 50 percent bonds and 50 percent stocks, according to the book, and will just rise and fall with the market. By contrast, Dalio’s formula produces a diversified portfolio and balances risk rather than wealth.
What risk to cover
Ask yourself what risks apply to you.
If you know nothing about investing, I would follow Buffett's advice and "diversify away your ignorance" by buying index funds.
More specific questions for your case will vary, but it always pays to ask "what is my risk" and "how can I diversify it away". For example:
- Are you unsure a government/company can pay back a bond? Get a bond index
- Are you unsure a house in a region will go up in value? Consider a location-diversified REIT (real estate index basically)
There is no one-size-fits-all approach because risk tolerances differ. Personally my investing experience started in cryptocurrencies in 2013, so up and down swings of 50%, 200% or more are not new to me. But if you would feel anxiety over a 20% decrease of your wealth, perhaps you should invest only in things that will fluctuate in lower amounts.