(A version of this post was published as an essay on Medium.)
(This rule set was updated on August 14, 2019 by replacement of the Fisher Transform with the DMI as the trading signal. See the new version here.)
Last week, as I watched with sadness as the return on 13-week Treasury bills decline yet again, I scratched my head and thought, yet again, about where to put funds for a return that would meet my two goals: 1) Beat inflation, and 2) make a reasonable profit. The Federal Open Market Committee has turned to stimulating the economy, lowering interest rates to make that happen, and lowering my return on my T-bill cash reserves.
We who lived through the Great Recession of 2008 are quite familiar with the problem. Interest rates fell rapidly, closing in on zero percent, what the Federal Reserve folk call the ZLB, the Zero Lower Bound. I think of it as something like an absolute zero for central bankers.
We had retirement savings to manage. We couldn’t allow it to be chewed and mangled by inflation, a clear danger in the minds of we who lived through the horror show of the 1970s. But with interest rates bouncing along near zero, we had safe haven to turn to.
It was that realization that convinced me to start trading heavily managed options trading. It proved to be a learning curve, and I took losses. And as I got better at it, options trading proved to be a good way to make higher risk trades.
Of course, it’s usually not very wise to put all funds at high risk. I divide my funds into three groups: High risk (options, which are leveraged), mid-risk (stocks, which aren’t leveraged), and low risk (Treasury bills, whole-life insurance policies with a cash value that can be borrowed against or cashed in). If I owned property, I would add a fourth category, real property, which is low risk, leveraged and somewhat illiquid, with high carrying costs.
One consequence of the lower interest rates environment is that the low-risk Treasury bills become inadequate to my needs. Some of that money needs to go to elsewhere in order to earn a return.
So last week I checked out the Big New Thing of the last couple of years in financial management, the robo advisor. They’re a species of fin-tech that is pitching heavily to adults of the Millennial generation who are flooding into the workforce.
The robo advisors I check out use modern portfolio theory to construct a high diverse mix of stock exchange-traded funds and bond ETFs, both U.S. and global, with the percentages set according to a customer’s age and risk aversion. Some of the funds set it up so that the customer can fully invest, rather than having money left over because of low granularity imposed by a share price. Some also will automatically adjust the holdings to maintain a balance among the various types of holdings in the account. The cost is low. Most that I saw charge 0.25% of holdings annually. Some charge 0.5%, and I saw some that are free. All to the good.
The downside, it turns out, is that none of the robo advisors I checked out won’t allow funding from a brokerage account, only from a checking account. Think of it. An investment platform that won’t allow funding from a brokerage account. It seems senseless to me, an unforced error, perhaps an effect of highly restrictive work rules imposed by the robo advisors union. And since my personal finances rely heavily on brokerages, I sadly turned from robo advisors, convinced that if I wanted a method of managing funds for greater return but no more than middlin’ risk, I would have to invent it myself.
Robinhood came to the rescue. Robinhood Markets Inc. is a brokerage founded in 2013 with the goal of easy trading online without trading fees. I underlined the no trading fees part because that change overturns the long-standing preference most private traders have to holding stocks for the long term. In traditional brokerages, each trade into a position produces fees, and each trade out produces fees. Stocks generally aren’t leveraged, and trade in and out enough, the trader will find his exits have been eroded away.
Robinhood, by eliminating trading fees, liberates the trader from the buy-and-hold restriction, making it possible to follow trading signals closely, even if it turns out to be a buy signal on Wednesday and a sell signal on Thursday. It puts an end to the two ancient laments, “Well, if I had bought Amazon back in the day, I’d be a millionaire today,” and “I know the markets down 10%, but if I just hold on long enough, it’s sure to come back.”
Adherence to trading signals in Robinhood’s first advantage.
By allowing quick trading without a financial penalty, Robinhood also allows for diversification over a period of time rather than static diversification. Exchange traded funds provide diversity in a holding, but under a buy-and-hold scenario the diversification is limited to the holdings of that fund. If I hold fund A for a week (during a buy signal), and then switch it for fund B for three weeks (during a buy signal), and so forth with funds C, D, E and F, then I’ve achieved a higher degree of diversification than a single fund will allow.
Diversification over time is Robinhood’s second advantage.
And so I came up with a plan, and trading rules, for managing stocks by closely adhering to trading signals on exchange-traded funds, providing diversification both statically and over time.
Here, the, are my trading rules for a self-managed alternative to the robo advisors.
Stock Trading Rules: Mid-Risk
The goal of my Robinhood rules is to create a highly diversified managed weekly according to changes in the Fisher Transform technical analysis tool that signals trend changes. A trend change signal is generated with the Fisher Transform crosses its signal line, which is a moving average. (See the appendix, below, for a description of the Fisher Transform.) Any unambiguous trend analysis, such as a moving average or the MACD, could be used in place of the Fisher Transform.
My holdings consist of a portfolio of five exchange-traded funds picked from a pool of funds. For signalling, I use the Fisher Transform applied to a daily chart. If the Fisher Transform is above the signal line, then it is a buy signal. If it is at or below the signal line, it is a sell signal.
I shall begin the method with a pool of nearly 90 exchange traded funds, including U.S. general index funds, sector funds, international global and country-specific funds, and a few futures-oriented funds in metals and agriculture.
Each trading day, I do the following tasks:
Update the pool with new Fisher Transform trend readings (which are binary: Above the signal line or at or below the signal line).
- Compare with the final trend signal of the day.
- Exit any holdings whose signals have changed from buy to sell.
- Bring the holdings count up to five positions by selecting according to these criteria:
- The Fisher Transform is showing a buy signal.
- The most recent date that the signal for each symbol switched from sell to buy is preferred over earlier signal dates. If the number of symbols on the most recent signal date is insufficient fill out the portfolio, use the next most recent date. For any selection date where there’s a choice of symbols to use, make each selection using a random number.
- Each fund in the portfolio represents a unique sector compared to the others.
More briefly, the selection criteria for my five positions:
1) Buy signal. 2) Newest trend. 3) Unique sector.
Appendix: The Fisher Transform
The Fisher Transform, created by John F. Ehlers, converts prices into a Gaussian normal distribution, highlighting when the prices are at an extreme based on their recent range. The goal of the conversion is to spot potential turning points.
An article in Investopedia (“Fisher Transform Indicator“, April 19, 2019) describes the steps in calculating the Fisher Transform:
- Choose a look-back period, such as nine periods. This is how many periods the Fisher Transform is applied to.
- Convert the prices of these periods to values to between -1 and +1 and input for X, completing the calculations within the formula’s brackets.
- Multiply by the natural log.
- Multiply the result by 0.5.
- Repeat the calculation as each near period ends, converting the most recent price to a value between -1 and +1 based on the most recent nine-period prices.
- Calculated values are added/subtracted from the prior calculated value.
I use a derivative metric, the FTtrend, that I coded using the ThinkOrSwim programming language, ThinkScript. It returns a 1 for a buy signal and a -1 for a sell signal. Here is the code:
plot status = if(reference FisherTransform().FT > reference FisherTransform().FTOneBarBack,1,-1);
And that’s it. I anticipate that updating the fund each trading day will take 5-10 minutes, max. It seems like time well spent.
I shall list the initial holdings using this method on Monday in my Live post, and update the list whenever the trend signal changes on a position, causing an exit from the old and replacement by the new.
By Tim Bovee, Portland, Oregon, July 28, 2019
Read More »