Here is an example of constant value investing, as applied to Amazon.com (AMZN) yearly prices. The prices were obtained from Barchart.com, which is on my list of links.
Constant value is the basic principle behind my successful stock trading system, which I describe in my book, Stock Trading Riches, which is available on Amazon.com.
To show an example of how constant value investing self-corrects a case of bad timing, we will start with January 2000, at the peak of the internet bubble.
Here, then, are the yearly January 31 closing prices:
| 2000 | 64.56 |
| 2001 | 17.31 |
| 2002 | 14.19 |
| 2003 | 21.85 |
| 2004 | 50.40 |
| 2005 | 43.22 |
| 2006 | 44.82 |
Here is the table:
| price | shares | cash pool | total value | amount invested |
| 64.56 | 30 | 63.2 | 2000 | 2000 |
| 17.31 | 115 | 0 | 1990.65 | 3408.15 |
| 14.19 | 140 | 0 | 1986.6 | 3762.9 |
| 21.85 | 91 | 1070.65 | 3059 | 3762.9 |
| 50.40 | 39 | 3691.45 | 5657.05 | 3762.9 |
| 43.22 | 46 | 3388.91 | 5377.03 | 3762.9 |
| 44.82 | 44 | 3478.55 | 5450.63 | 3762.9 |
At the end of January 2006, we had invested $3762.90 and had a total value (stock plus cash) of $5450.63. This is +44.85%
Buy and hold, on the other hand, would be down 30.5% (44.82/64.56)
Your comparisons aren't accurate. In order to compare the results, the cash flow INTO the investment should be the same for both methods. The B&H investor just doesn't sell.
ReplyDeleteThat means the B&H investor would have bought the same 140 shares your method did, since you put money into the investment three times. At the end, those 140 shares are worth $6,275 instead of your method's $5,451.
Praveen, I also have a keen interest in this topic. I would like to correspond via email. My address is vincent.apa@gmail.com. Please send me an email if you are interested and we can discuss.
ReplyDeleteHi Randy,
ReplyDeleteEither the B&H investor could have invested the same cash flow into AMZN ($3762.90), or he could have bought the same number of shares (140), but not both.
If he invested $3762.90 in 2000 at $64.56/share, he would have 58 shares and $18.42 left over. In 2006, at $44.82/share, he would have a total of $2617.98 (down 30%).
If he had bought 140 shares in 2000, he would have invested $9038.40 and, in 2006, it would be worth $6274.80.
The cash inflows for your model are:
ReplyDelete$2000.00 in 2000
$1408.15 in 2001
$ 354.75 in 2002
Those same cash inflows need to be applied to the B&H model.
$2000 in 2000 buys 30 shares, leaving $63.02
$1408.15+$63.02 in 2001 buys 85 shares
$354.75 in 2002 buys 25 shares
Total -- 140 shares
Two problems with AIM is that it can take money out of an upward-trending stock and put more money into a downward-trending stock. Any gambler can tell you that the Martingale betting method (doubling up after a loss) can be a disaster in the long run.
You have to make sure you apply the model to the right kind of stock -- one that is volatile, but with no significant trend long-term. For example, over at the iHub discussion board, one screening criteria I've seen looks at the ratio of high price of the year to low price of the year for up to 10 years of history, as well as the overall stock trend.
And, true, if the B&H investor had invested the full $3762.90 in 2000, he would have underperformed your model. However, if he invested it in 2001, 2002, or 2003, he significantly outperforms your model. And, depending on the returns he got from a CD or whatever in the meantime, using the proceeds to buy in 2004 or 2005 give results pretty close to your model.
In this case, you would be right. If, as a buy/hold investor, you bought at the same time, and held thru the rise in the stock price, you would be better off than in my case, where you sold off as it rises (unless, of course, the stock sinks again).
ReplyDeleteBut, if you had invested the original $2000 in 2000, you would have needed some type of analysis and courage to get you to buy in at those other points.
This is a system that, psychologically and phiosophically, works for me. I have faith and confidence in it.
As a technique, constant value investing would take money out of a rising stock and put it into a under-performing stock. That could be dangerous.
But, in my system, I think that I account for the risks. My system is slightly different than AIM. I never increase the control value for each stock. I am investing over time - not a lump sum - so I keep adding new positions. I rebalance each position yearly, not monthly, and then only if it has changed by at least 10%, so I try to capture longer trends.
I am looking to grow the portfolio as a whole, not each individual position. My model is that each stock will fluctuate: some will rise, some will sink, some will oscillate. I make no assumptions.
But, overall, I will pump out cash from the gyrations. When cash gets excessive - say over 30% or so, I use it to add still more positions.
Hello,
ReplyDeleteI understand the workings of your system and of the AIM system but I do not understand why you rebalance based on a time schedule rather than on a performance schedule. In other words, why not take appropriate action when a stock rises or falls to a specific milestone (e.g., +10%, -10%) rather than take action on some arbitrary date? Stocks do not perform based on the calendar. Over the course of a year, a stock could rise or fall 25% or more granting you the opportunity to buy or to take profit. But your system would bypass these opportunities in favor of waiting for January 31.
Hi! Thanks for leaving a comment.
ReplyDeleteRebalancing by a specific % movement is an option. Other options are to rebalance in a shorter time period (i.e. monthly or quarterly), or use some form of technical analysis (i.e. the crossing of a movign average or an oscillator).
I encourage people to develop their own system - something they believe in and feel comfortable with.
You may even like this optional growth rule.
In my case, I diversify over a lot of stocks/ETF's, and want to keep my system simple and un-optimized. So, I keep one control value for all the stocks. I then rebalance once a year, and if the portfolio cash level gets above 30%, I buy more stock positions.
Otherwise, I buy more positions as I save money and add it to the account.
If we look as "trader<---->investor" being a spectrum, I want to be more towards the investor side, and capture bigger trends. From my testing, rebalancing yearly is a compromise that does a good job of capturing long term trends, and well as capturing volatility.
But, if someone wants to capture more volatility, he or she can certainly go on trigger points.
Hello Praveen,
ReplyDeleteI'm the "Anonymous" who just posted the message concerning rebalancing based on stock-price performance versus waiting for a specific date.
My name is Randall and I appreciate your fast response and your honest, insightful website.
Recently, I designed and tested a system for investing my retirement accounts . I select the stocks with VectorVest using strict fundamental criteria and add a stock to my watchlist only if its 10-year chart shows the ability to push higher with each cycle. I currently have 84 stocks on my watchlist and I update at the end of each month, adding and eliminating stocks in accordance with my selection criteria.
Overall, these stocks do not really respond to market cycles so I tossed out market timing as a trading trigger. Instead, I sort the stocks by ascending 6-month price performance -- this puts the biggest 6-month losers at the top of the list. I then initiate the portfolio by buying the top 12 stocks, hold all positions for a 35% profit, and replace each sold stock with the top candidate on the sorted watchlist. I coumpound by reinvesting at the average value of the remaining portfolio (i.e., total market value of the 11 remaining stocks divided by 11).
Each month, with the updated watchlist, I run rigourous backtests on this system. I vary the start dates and try slight variations in selection criteria. The results vary but are always excellent. The 10-year tests show that the system will complete 100 to 120 trades with a typical holding period of four months to two years. Total return averages around 1,500%. There are, however, always laggard stocks that hang around for three to six years. This is a big problem in a 12-stock portfolio.
I'm confident I can solve the laggard problem by implementing your constant value method. I can reduce my initial investment to create a cash reserve, then rebalance on progressive -10% triggers -- this will hasten the turnover of the slower-moving stocks. Implementing constant value will also reduce the significance of
the starting date of the initial portfolio.
My example shows the benefit of synthesizing constant value as a component of a larger system.
Thanks again for your help, Praveen.
Hi Randall,
ReplyDeleteI'm glad that constant value investing helped you with your system.
Your system sounds really interesting! You are holding a relatively small number of stocks(12), but you are protecting yourself through thorough fundamental analysis.
You are then buying high quality at low prices, and then selling out at a higher price.
I'm not surprised that your testing shows great returns.
Praveen
Hello Praveen,
ReplyDeleteYes, the system is interesting, probably because it evolved through my own thought processes and through lots of simulator testing. I did not borrow the concept and customize for my use.
Using 12 positions worked well in backtesting -- fewer positions resulted in too much variability in total return while more positions resulted in much higher drawdown. Implementing constant value should decrease the drawdown and smooth the equity curve and perhaps increase return.
The system does enforce a lot of disciplines -- as you put it "high quality at low prices." I have come to think of analysis as the foundation of higher returns but, until you mentioned it, I did not consider it to be a form of risk protection -- you have a good point.
Currently, there are 78 watchlist stocks -- I gave the wrong number in my previous post. The current list is diversified across 23 business sectors. The market caps are balanced with a bias toward midcap: 20 large cap, 36 midcap, and 22 small cap.
Right now, the "buy" candidates are concentrated in the beaten-down oil sector -- there are a lot of very solid companies that are 30 to 40 percent below their 52-week highs. I'm also finding candidates in the retail and healthcare sectors.
Take Care,
Randall
Hi Randall,
ReplyDeleteDid you choose the mid-cap bias and the 23 business sectors because of your analysis, and then pick your stocks?
Or, did you pick the 78 stocks individually (based on your criteria) and the midcap bias and sector distribution evolved on their own?
Its apparent that you really enjoy trading and analysis. You have really combined creative insight, fundamental knowledge, and hours of testing to develop your system.
Praveen
Hello again Praveen,
ReplyDeleteThanks for your kind words. Yes, I do enjoy research, analysis, and testing because the effort has built the trading systems I use. I discovered the hard way that taking random shots at the market without a complete plan was, for me, nothing more than gambling. A trader's system is his/her interface to the market so it must be comprehensive.
I choose the watchlist stocks based on their individual merits; no forethought is given to market cap.
Currently, the smallest cap on the list belongs to Drew Industries (DW) at $524 million and the largest cap is United Health Group at about $70.6 billion. The averge cap of current list is $8.3 billion.
Randall
But, now-a-days most of the Investors are doing gambling/speculation. People invest just to make easy money. They are not bothered about these calculations, theories, statistics and analysis. Nice Articles, thanks.
ReplyDelete