Mechanical investment strategies

I could not find a thread on mechanical investments, so I start this one.

Here is a simple definition: Mechanical investing - Investment strategies - Moneyterms: investment, finance and business explained.

The book https://www.google.ch/books/edition/What_Works_on_Wall_Street gives some ideas and testing results. However, they vary from edition to edition and may be kind of biased.

Most people here invest in ETF and that is OK. There are ETF that use mechanical investment strategies, some indices do too (I think the NASDAQ100?). But I always felt that I can do better by adjusting some of the rules. Most indices contain companies that I would not touch with gloves and just leaving those out should (and did for me) give better performance results.

I tried out a lot over the past decades and Internet tools help a big deal today.

Since 11 years I do a mechanical dividend strategy with a target of low volatility and high cashflow. The XIRR performance (included 15% dividend tax) over this period was exactly 10%.

Since 5 years I do a growth-and-momentum strategy with a target of high performance with high risk. The first year was a loss of 2% but even with that loss the XIRR performance is 27.54%, meaning it made me tons of money.

The most important thing on mechanical strategies as to any strategy is to stick with 'em. You should define almost every possible detail, every outcome and it should always lead to exact instructions of the action you have to take. As detailed as possible.

I’m happy to answer all questions related to mechanical investment strategies. I probably won’t give away every detail of my growth-and-momentum strategy, did cost me a lot.

BTW: if you find an ETF that has all the mechanics you need… go for it. Doing it yourself may be more profitable, but everybody, really everybody including me myself, makes mistakes. And those are expensive…

Hi. Both are mechanical stock only strategies. XIRR is the eXtended Internal Rate of Return per 12 months over the whole period in USD as calculated by the XIRR function in google calc. I use the S&P500, Dow, Nasdaq-100 and Russell2000 for the same period as Benchmark.

Some things to consider:

• Indices are without dividends and of course the dividends strategy is about cashflow and therefor dividends.
• At first I was surprised by the bad performance of the Russell2000. Since 2012 the XIRR was only 8.51%, last 5 years 5.42%.
• XIRR is calculated like your bank account interest in reverse, considering all the amounts you put in or take out with the respective date.
• There are heaps of strategies that would have beaten mine, not only bitcoin. After the fact.
• 5 and 11 years are not very much, but it seems I am on the right way.

I try to present the performance numbers for each year and the XIRR over the whole period (sorry, my first table here):

Performance is a bitch. Many people would have stopped the growth-momentum strategy after one year, losing 2% while the Nasdaq made 47%. I did dry-tests for many years and knew this could happen…

Probably the holy grail of toxic waste.

Maybe some of the rules for my divi portfolio help: I want to be able to hold stocks as long as possible, but not longer. Some of those stocks are with me for 10 years now and did multiply it’s value.

I check quarter and year data from SEC’s Edgar database for some numbers. If quarter or year data are not OK for me I don’t invest any more there, set to “hold”. If quarter and year data are not OK I sell, but only if they are in the worse half of momentum. I don’t want to sell a stock just because it got a little expensive, those tend to get a lot more expensive sometimes…

Those rules are simple:

• FCF payout ratio under 100%
• EV/FCF <34
• OCF/Debt >0.1
• Dividend over 2%. This only leads to a sell if dividend + treasury stocks bought is less than 2%

FCF= Free CashFlow, OCF= Operating CashFlow, EV=Enterprise Value (market cap + debt minus cash).

The quarterly data is extrapolated to 12 months for those checks. So I need to do 4 checks per year on every investment. I want the company to pay a nice dividend (or at least buy their own stock), be reasonable priced and have a reasonable amount of debt. I focus on cashflow, not on earnings, because cashflow is what I am seeking for myself, this is my only source of income. The dividend should be covered by the free cashflow.

This is the actual dividend portfolio:
finviz link

Of those stocks currently TRI and AVGO would be on sell, but are too good performers, so they stay until that changes. DOW, NUE, DD, KLG, CLX, CAT, EMR and ATMU are on hold. The rest is on buy and the dividends are re-invested round robin in all positions that are worth less than 4% of the portfolio.

I did start with 4% per position and every time a position reaches 6% I sell down to 5%, I call this the market dividend. Market dividends and the reinvestment of dividends do a kind of buy low and sell high and that works very nice on cyclicals.

There are some diversification rules too, only 20% per sector, no new positions in sector with >20%. Initial position is 4% of portfolio value, which gives an initial size of 25 positions.

There is also a “crash recovery program” which I may explain later. Worked nice in the last few bear markets, but adds risk.

No macro environment (like interest rate etc.) except for the crash-recovery mechanism. There I use the actual S&P500 difference to it’s last high. The definition changes to “Bear Market” when it is at 80% or less, which may trigger some additional rules which I will describe later.

And yes, you could program a robot to do it. But as I don’t work anymore (since 11 years when I started with this project) I have time and it is only a few minutes every quarter for every position. February tends to be the busiest month.

I could automate it easily, as I use a google sheet where I put in the data from Edgar and then it tells me if the criteria are still OK or not. But finance stocks are tricky, as their trading good is money you have to check the cashflow and decide what to include into the FCF and what not. So I prefer actually doing this by hand. Takeovers are a problem too, there the numbers have to be corrected.

As promised here the “crash recovery strategy”. It is basically a market timing and I really don’t like market timing. The point is to add risk when everybody else is taking out risk: in a bear market.

Now I am always invested 100%. The only way I can add is with a margin loan. At Interactive Brokers this is very easy and convenient and quiet cheap compared to other brokers.

One can just leave out the crash-recovery part and be fine.

I did use it the last 3 bear markets and it worked 2 times perfectly (good timing) and one time I had to suffer a lot until it worked out again (bad timing, entered too early). But the payout then was even higher. As the old saying goes: “The stock market is pain and gain, first the pain then the gain”. The crash recovery plan may be active for many years and you may suffer a lot during this time.

Here are the rules: Trigger is the S&P500 closing under 80% of its last high. I measure the lowest point in percent. That defines the leverage in the following way: for each 1% I take 1% of margin credit. If the lowest point is 70% I go to 130% margin, 30% on credit. The maximum is 150% when the S&P500 loses half of it’s value or more.

I do this when I “feel” (of course there are rules for that too) the market turns up again. There are tons of measurements one could use for that. First I check the 50 days average price. Once at least 250 of the 500 companies in the S&P500, the S&P500 itself and the Nasdaq100 and the Dow are all over the 50 days average the first requirement is fulfilled. I don’t check the Russell2000 for that, maybe that would have been a good idea the last time.

The second requirement is based on expected volatility. I check the VIX Future for contango. Once at least 3 months of the VIX Future are in contango, the second requirement is fulfilled. I buy on margin credit.

Then a three phase plan starts. The first phase is the time until either a new high in the S&P500 is reached or the credit is completely paid off by the dividends. Until then dividends are not invested and the “market dividend” concept is paused.

If the S&P500 reaches a new high before the debt is paid back, the barrier for the market dividend is set to 133% instead of 150% (5.33% of portfolio value instead of 6%) to pay back the credit faster.

While the credit is open I have a stress tolerance test active: whenever the credit reaches 300% I start selling until I am under 300% again. This is the emergency exit, didn’t happen until now, but you never know.

One may need "portfolio margin to execute this plan and have enough reserve margin. Portfolio margin at Interactive Broker gives you up to 800% margin, which is insane. As I start selling at 300% this is more than enough.

Once the credit is paid back the crash recovery phase is over.

OK guys, now you know all the rules of my dividend investment strategy, which is my pension plan and my only source of income.

I increase the debt.

As the dividend yield is higher than the margin cost that is good business. This can enlarge the time I need to pay back the credit of course.

I do this always, not only when in crash recovery mode. I don’t sell just because I need money, that is a bad reason for a sell. The credit is paid back with dividends, market dividends or when my system tells me to sell something.

Actually I am not in crash recovery mode and have a little margin debt because I needed money in December. It will be paid back with this months dividends.