When you mention diversification, most people will immediately recognise that it is a good idea. Most people have heard of the old saying “Don’t put all of your eggs in one basked” and I fully agree.Â
Of course you should hold more than one stock – I mean if you hold Apple, Facebook, Google, Bank of America and GE then you are diversified right?
Well in a sense yes, because if one of those companies doesn’t do so well then you have a couple of others that might make up for the difference. But this is just one level of diversification and it is has a major flaw.
The fundamental flaw in the typical approach to diversification is that it is only taking advantage of a single level of diversification.
When you hold multiple stocks in your portfolio you reduce company specific risk because you have different companies with different risks in your portfolio. But you are not diversifying against other risks like economic, investor sentiment, currency fluctuations and so on.
Thinking about all of these risks and diversifying against each one is VERY HARD, and chances are you will miss the one risk that you really need to address!
I propose that you think about it from the angle of diversifying how you make your money. When you do this there are 6 levels of diversification to consider.
Positions
Just like the “don’t put all your eggs in one basket” example above, all traders should hold multiple positions. Company specific risk and unexpected announcements can make individual positions swing wildly, so the more positions you have the less this will be a problem for you.
No matter how good the story is, don’t bet it all on one or 2 hot stocks. Even the market darlings can fail – Just look at Worldcom, Enron, Goldman and hundreds of others. You just can’t rely on a small number of positions to generate consistent returns!
Parameters:
Regardless of whether you use indicators, a trading system or fundamental analysis, you will have to choose the parameters that will trigger entries and exits.
A simple example is a moving average crossover system – you enter when the 20 day moving average crosses above the 200 day moving average, and exit when it crosses back below. (I am not advocating that as a system, it is simply an illustrative example). In this example there are two obvious parameters: 20 for the short average and 200 for the long average.
Most people would search for the combination of parameters that work best in the past and use those for their trading. A better way is to choose parameters that are in the middle of a range of well-performing parameters. But an even better way is to diversify your parameters across the range that works.
So in the diagram above, say the range of good performance was between 24 and 30 days for the short term moving average. Here you would diversify by using several values within this range (say 25 & 29) as slightly different versions of the same system.
If you diversify a couple of key components of your trading system so you have a couple of different versions of the system then you will most likely outperform one single version over time.
Trading Systems
There are many Trading Systems that work. Each different system will enter and exit the market at different times. They will also experience profit and drawdown at different times. This makes a huge difference if you are profiting from one while the other is losing.
When you combine two different trading systems into a portfolio you can reduce the volatility of your returns OR achieve higher returns with the same volatility. This is diversification gold!
Time frames
Trends & patterns can exist on Weekly/Daily/hour/minute charts. Most people will have one natural center when it comes to timeframes. If you create a trading system for your preferred timeframe and then diversify by adding a second timeframe then you can find similar benefits as you get by adding diversified systems as shown above.
Markets
If you are a stock trader who only trades in your home market then you are missing another big source of diversification. It is now cheap and easy to trade on all sorts of exchanges around the world so there is no longer an excuse to not consider diversification across other markets.
While most of the world’s stock markets are correlated, they still move somewhat independently day to day. Depending on what your home market is the right market for you to diversify into will differ. You will need to explore and find the right market for you to diversify into.
You may consider going even further – if you a stock trader now you could even consider adding futures or forex or options to your portfolio.
Return Drivers
The final level of diversification is based on return drivers. This means you diversify what type of market behavior actually generates a profit for you. There are many types of market behavior that you can profit from including (but probably not limited to):
- Upward Moves
- Downward Moves
- Statistical Relationships
- Mean Reversion
- Increasing Volatility
- Decreasing Volatility
- Time decay
As you get comfortable trading one of these return drivers you can add another one for diversity.
The diversification combinations based on these 6 levels are potentially endless. The more uncorrelated ways of making money you have the better your long term performance will be. You won’t be able to take advantage of them all, but just remember, there is a lot more to diversification than just adding a Google position to you Apple and Facebook holdings!
Learn The 10 Commandments For Profitable Trading
Discover the 10 most important lessons from 18 years of profitable trading & reading over 150 trading books.
Get your FREE Report & 10 Training Videos NOW!
Related posts:
How To Survive and Profit Trading Stocks in a Bear Market
Bear Market Stock Trading Risk Management Survival
Are you using your stock trading software the wrong way?
You Don’t Need to Be Right to Make Money – You Just Need a Good Trading System
Can I Apply One System to All Markets
The Ultimate Guide to Backtesting
An area of diversification that I can improve? Definitely return drivers. I have too many long only systems that trade trend and momentum. I have some mean reversion but they are a lot smaller in size. I definitely need to add more systems for downward moves and increasing volatility.
I have 2 questions in case you have time to respond:
1. In my experience I have found it much harder to develop short trading strategies, especially when compared to long only. I guess because the markets tend to have an upward bias. I was wondering would you adopt a short system even if it produced an overall small negative return but decreased the overall drawdown and volatility of your portfolio.
2. When running multiple systems, how often do you re-allocate/rebalance the capital into the different systems. Also if a system has mostly gone to cash, which could be for an extended period, do you allocate this into other systems or just put in a money market fund awaiting for the system to turn back on?
Thanks.
It just dawned on me that I can check out trading all four market stages as follows:
Stage 1: Accumulation using a mean-reversion strategy
Stage 2: Markup using a Bullish:
– trend following strategy,
– momentum strategy,
– breakout strategy, and
– pullback strategy
Stage 3: Distribution using a mean-reversion strategy
Stage 4: Markdown using a Bearish:
– trend-following strategy
– momentum strategy,
– breakout strategy, and
– pullback strategy
I now have a plan to back-test each of these strategies individually and also as a portfolio of strategies.
Thank you so much for writing this article; it has given me a whole new set of insights!!
Hi Richard,
This is great, sounds like you now have a lot of clarity on how to think about your diversification and test your portfolio – well done.
Adrian
To be perfectly honest I did not know that diversification was anything but by asset class. This was a very interesting enlightenment for me. I will certainly re-read the entire essay in the morning while I am fresh, as there is a lot there to digest. The most interesting one is diversification by return drivers. Thanks for the cogent insights. Don
Awesome – I am so glad you have this realization, this is a huge insight which will help your returns a lot going forward!
For me I think diversification of Parameters is one of the strongest diversifications you can apply in a strategy. I liken it to high frequency trading – if you really understand the conditions your working within, then it should be possible to extract profits even from the most minor of conditions if your parameters are good. I think if you are highly in-tune with the parameters your strategy framework requires, you will understand the psychology of why they work in the first place and thus why the system needs to apply them within the bounds that profit most. I think this can be both a strength and weakness of automation because if you overly rely on optimization functions without understanding why the optimizated parameters work better in the first place. Having done environmental research, the first thing you learn is how to explore your data. The heart of any good research is the question asked – i.e. your idea, and then the factors you observe/manipulate (i.e. the parameters studied). Do this well and the rest should fall into place… you can always diversify positions, markets and timeframes, but if the factors the strategy relies on are weakly defined or understood then your system won’t be robust to answer the question well and problems will compound across all other areas of diversification. I imagine this should ultimately prevent a system from failing despite decades of positive backtesting or at least set boundaries for when it should be used. The idea will always have boundaries of operation inherently because of the parameters imposed – it’s your job to understand what those boundaries are and when the system is being extrapolative for the current market conditions.
My question would be:
Despite the benefits of reducing risk, when does can stock diversification become undesirable for a portfolio of investment strategies? Provided the portfolio has consistent profits and drawdown is kept to tolerance, why dilute the overall edge by reducing equity allocation unless you can prove the diversification also generates higher profitability?
Great insights John. I like your thinking on parameters, well done. In answer to your question, diversification across multiple strategies like this reduces the risk of luck in your portfolio a lot. If we have one strategy then we might have a lucky year and happen to get some good trades out of it but the strategy performance will fluctuate around the expected returns. Risk like this can also be to the downside… But the more we diversify our strategies the more we remove the role of luck and the more we stabilize our portfolio. It is not about dilution the edge of one system, it is about reducing the exposure to each system so we can fit more systems into the portfolio and thereby achieve our goals with a higher degree of certainty. Let me know if that clarifies or if you have additional questions – happy to answer 🙂
Adrian
I enjoyed reading the article and found the examples provided insightful. I have a few questions:
1. Is it fair to say that the diversification approach is grounded in the principle that not all assets/markets will move in the same direction at the same time, providing a level of protection against significant losses?
2. How can one address or mitigate some of the pitfalls associated with diversification, such as:
a. Over-diversification
b. Correlation risks: The possibility that assets initially thought to be uncorrelated may become more correlated during market stress, compromising the intended risk reduction benefits of diversification
c. Increased complexity, which may necessitate ongoing monitoring of each asset’s performance
d. Market-wide events: The potential for increased correlations among various assets during extreme market events or financial crises, leading to widespread losses in diversified portfolios
e. Underperformance of Diversified Investments: Instances where certain assets or sectors outperform while others underperform, causing a diversified portfolio to lag behind more concentrated portfolios.
3. Is it feasible to design a portfolio system where the trading strategy itself can be a dynamic parameter that the system adjusts? Can the system dynamically select a trading strategy (from a list) based on market conditions/behavior for a specific asset? For instance, if an underlying asset in a market is trending (Up or Down) or moving sideways, could the system automatically adopt a Long or Short trading strategy or a mean reversion strategy, among others?
Great questions – thank you!
1. Partially this is right, but diversification is also about different directions and different return drivers. If you are long and short then if all markets become correlated in a certain direction you have some diversification. You can have different return streams – like mean reversion and trend following. markets may be correlated but you can profit from the dips and the trend at the same time. So it is about more than just the different markets moving independently, it can even work within the one market.
2. a. Overdiversification is an issue if you have too many systems that are too highly correlated, or your position size is too small and commission drag becomes a problem, or if you can’t manage it well because of the number of trades and you make mistakes. We can avoid these in fairly obvious ways – Use uncorrelated systems, monitor position size relative to minimum commission and track our execution for mistakes.
2. b. Yes, this is absolutely possible – that is why Long / Short diversification is so important. Plus think about your portfolio in terms of Risk on / Risk Off assets and make sure we are not too highly concentraded in each.
2. c. Increased complexity of managing this is really an issue for automation to solve in my view. This dramatically reduces the burden of managing it.
2. d. Yes, this is possible too – same approach as 2b. Also the more broadly diversified the better. If you are long stocks in ASX, US, HK, TSX… in an extreme event these could be highly correlated. Make sure you have other assets too.
2. e. This is more tricky, great question. Rules can be created to turn systems off when they are not performing, but in general if you think about the whole portfolio there will be systems that are under and over performning at any one time, all the time. We are after high confidence we will get good performance and this means some times we will underperform one concentrated system or portfolio, but that comes and goes – there will be other times where we will dramatically out perform concentrated portfolios as they crash and burn.
3. This is certainly possible though non-trivial. Simple is often best, and that approach would leave you vulnerable to sudden changes that hurt the strategy you are in, but this is a nice area for further research.
Great questions! I hope my thoughts helped
Adrian
Adrian, I want to express my gratitude for taking the time to address all my questions. Your prompt response is greatly appreciated, and the answers provided are immensely helpful to me.
Very happy to be of service! Thank you for being part of the EST community!
I can see the benefit of a long and short type of strategy but what would you do with a sideways market that doesn’t show a clear direction and could be volatile? Do we just stay out of the market and only when we see clear indicators we move?
Hi Alf,
Great question! Long strategies can still make money in a sideways market because not all stocks are sideways when the index is going sideways. Mean reversion can also work… my favourite method is actually to trade multiple markets (ASX, US, TSX, HK, Crypto) so that generally they won’t always be doing the same thing at the same time.
Adrian
Hi I am still in the beginning stages of my trading journey.
The main area that I find particularly challenging is the diversification of parameters in a trading system. The suggestion to choose parameters within the middle of a range of well-performing values is interesting, but it leaves me wondering about the practical implementation.
Questions
1: How would one determine the range of well-performing parameters, especially in dynamic market conditions? Additionally, could you share any strategies or tools that can help traders effectively diversify parameters without overcomplicating their systems? Would one simply back-test different parameters?
2: When it comes to diversifying return drivers, the concept of profiting from different market behaviors makes perfect sense to me but I wonder if this might complicate a system too much and lead to negative returns, are there specific methodologies that I can help me manage these drivers without too much complexity?
Thank you
Hi Kimon,
Awesome questions – thank you. The first one is really a backtesting issue rather than a diversification issue. I have a long article on backtesting that explains a lot about this.
In the second question the issue is that each system is not trying to profit from different market behaviours, you have different systems for different market behaviours. Each system is simple and focused on one market behaviour ideally and they are combined into a diversified portfolio.
All 6 levels of diversification are important for trading profitably and consistently.
For me, I could improve my returns by increasing the number of markets I invest in.
My question is on optimizing the parameters of individual system. I come across lots on systems on the net that suggest the parameters to use for day traders. How to I optimize the parameters for my longer-term investing? It takes too long doing this manually. I believe this can be automated using amibroker (but coding this may be too onerous) or Real Test. Any preferences?
Thanks for your great comment Alberto. I would do this in either Amibroker or Realtest – both are pretty straight forward. Good systems are simple and simple systems are relatively easy to code. Day trading strategies may or may not translate to longer timeframes, but you can absolutely try them. Alternatively start with end of day systems (daily or weekly systems) like what you find in The Trader Success System and use or adapt those for your purposes… this is probably a quicker and more profitable approach than trying to get daytrade strategies to work on longer timeframes… but you can certainly try both.
These levels are helping me to understand diversification. The underlaying systematic approach is very helpful to me.
Very useful article providing a foundation for diversification.
My reflection is that one should be mindful of the objectives of one’s trading too. In Australia, for instance, I think (not totally sure though, yet!) that investing for an SMSF does not allow short selling systems. However, this article has provided pointers for other ways an SMSF investing strategy may be diversified.
Brilliant thoughts John – thank you. There are certainly many ways to diversify within an SMSF – it does not have to be the boring buy and hold approach that many think!
Adrian
If you had asked me prior to reading this article, I would have said that you diversify your portfolio by investing in different positions. Period! The other 5 parameters would not have been included in my thought process on diversification. Definitely eye opening to areas where I can improve diversification of my portfolio. I can improve most by the return drivers I use. I currently trade upward moves but that leaves a lot of potential on the table.
As an investor in the US, I haven’t found an easy way to trade on downward moves. I can’t just go into my trading account and short a stock. Do you have recommendations of brokers were that can be done for US traders?
Perhaps consider using put options as an alternative to shorting stocks directly.
This is certainly an option (hahaha – pun intended). The challenge with using put options instead of shorting is that while you have an absolute cap on your risk (this is good), it is hard to backtest the system with options because you need the options prices… often when you want to short a stock , the volatility can be high, which drives option premium up… so you need to be careful to ensure your short selling system remains profitable when implemented with options instead of shorting stocks.
It is doable – it just takes more work to ensure the edge of the system remains when you use options instead.
Great idea – thanks for sharing it Tim!
Adrian
I’ve just learned that I can improve my diversification, by diversifying a parameter. I’d never thought about doing that! Diversifying by systems, markets, timeframes and positions were all ones I knew. Return Drivers was also a new one for me, and a little challenging to get my head around, but I can see the diversification they would bring. However, I like that just by changing a couple of key parameters of a system, but still staying in the range that works, to create a different version of the system is a subtle diversification that can be done fairly easily, and is likely to improve my results over time.
In regards to the Return Drivers, can you please explain what you mean by Statistical Relationships, and Time Decay?
Hi Jennifer,
Fantastic! I am so glad you found the article helpful.
Parameter set diversification is a good one isn’t it 🙂
Time decay is a return driver you can profit from by selling option premium (high risk strategy). Option prices decay over time and most options expire worthless, so selling premium can work… but you have to be very careful as many option sellers blow up in spectacular fashion. This is not one of my preferred strategies – it is here for completeness.
Statistical relationships is where some instruments are generally highly correlated, however sometimes they move a little differently but tend to snap back to their historical relationship over time – you can profit from this by going long one of the instruments and short the other. It can be quite complex and margins tend to be small so it is not a great strategy for retail traders.
Hope that helps!
Adrian
All this time, diversity has meant (to me):
1. Property (with further diversification by states)
2. Super (including international shares)
3. Shares (with further diversification into ETFs covering different industries or major stocks)
4. Cash (held in offset accounts)
After reading this article, I felt a moment of shame, but upon reflection, I don’t believe it’s warranted. This article discusses diversification within shares, specifically in the context of trading. It seems the world of trading is fundamentally different from that of investing. I need to shift my way of thinking a lot.
Hi San,
Shame is certainly not warranted – you have done very well. Within the ‘shares’ bucket there are many things you can do other than just hold many stocks – that is really the part that this article is discussing.
I hope it has helped you stretch your mind to consider some alternatives which can give you better diversification.
Adrian
so much to learn, and I’ve been trading a few years now. This has opened my eyes to what diversification can really be.
If you were trading futures would seasonality be considered a driver?
Hi Ian,
Thanks for your comment – glad you are getting great value from my articles. Yes seasonality can be a driver, and not just in futures. There are seasonal patterns in the stock market that can be quite profitable too. The Ultimo Strategy is a great example of this.
Hope your trading is going well!
Adrian
My current focus is a trend following system but I know if I diversify by introducing a system which captures downward moves then my profitability would improve. My question is would it be better to introduce the new system to my current market ie ASX or would it be more profitable to diversify further by introducing my new system to the US market? Basically, if trading both long and short systems is the diversification better to have them in the same market or two different markets? Also, is the timeframe a factor here as well?
Again as has been mentioned a few times my idea of diversification was to have a few different industries / companies in the portfolio. This has opened my mind to proper diversification.
Hi Vlado – this is fantastic! Now that you have seen what is possible you will be far better off in the future. Well done!
Adrian