Weekly Update

Building an investment model

Tagged: ,

Published October 21, 2016


 This week we pull the curtain back a bit on TimingCube‘s publisher, Fraser Partners, LLC, to discuss some of their non-TimingCube activities. When we build a Model, where do we start? What are we trying to achieve?

Define the Box … in terms of risk
“Most advisers are looking for an answer to a given investment question. We boil that question down to a “box” into which the strategy’s results must fit. We define the box by the typical risk-return characteristics, first and foremost. People look at risk from varying angles, of course. It can be simple volatility. However, advisers focus that by usually responding to their client’s tolerance for draw downs. The draw-down is the maximum loss from the peak. Think of it as how far one falls from the top before stopping. The sooner we stop that fall, the better! If we can keep that fall under 10%, or even closer to 5%, we are pretty pleased with that result. For comparison, the stock market gives up 25-50% during a bear market phase. Advisers find that declines greater than 10% really sap their client’s confidence, even though a 10% decline in the stock market is a fairly normal part of doing business. The market delivers a 10% decline once a year, at least. We figure that if we can prevent their clients from experiencing those 10% declines, they will stick with the investment strategy and capture the resulting long-term benefits. It’s the draw-downs, the perceived losses, that kick investors out of the market more than anything else. We want to minimize the opportunity for that emotional reaction.

… in terms of return
The return objective is fairly straightforward. It’s easy to say that more return is better. But not all returns are created equal. If you get a 40% return one year, for example, that does buy you some time in the strategy typically. But investors that are excited by the big out-performance are likely going to be the same investors that will immediately bail out of a strategy if it under-performs the market the following year. They are looking for the hottest, fastest stuff, and often don’t have the patience to hang around for anything they perceive as mediocre – even though it may be a winner when considering risk. We think this emotional reaction to high returns is one reason options and currency trading have become more popular – this eternal pursuit for out-sized performance. Now, we obviously love to deliver that performance! After all, our new Turbo Model is built for that type of action. But you will note that Turbo‘s results are also pretty “lumpy” with spectacular performance some periods while only solid performance other times. When building a strategy for an adviser to attract clients, however, it’s better to have steady returns than to have returns that spike. Those steady returns allow clients to build up confidence in the strategy, which, in turn, allows them to weather a bad month here or there.

… in terms of diversification
The 3rd side of the box is diversification. Investors have a belief that diversification is a positive attribute. However, many don’t really know how to define diversification much of the time. They see it only as a collection of assets. It’s intuitive that the internet ETF is more diversified than any single company within the ETF. But it’s not so intuitive for most people that investing ONLY in the S&P 500 as a single position in the SPY, for example, is being diversified. It certainly doesn’t FEEL diversified, even though that one investment is representing 500 companies. Thus, we want to build strategies for advisers that not only are diversified, but that they can easily sell and explain as diversified. That means holding a collection of assets rather than the singular “concentration of force” concept that drives the published TimingCube strategies, whereby we might place all our money into the Nasdaq 100 (QQQ). This part of the puzzle is a big difference between building a portfolio strategy compared to building a single Model, as we present on our Web sites.

… in terms of trading
The last side of the box is trade frequency. This preference really depends on the person. Most brokers and advisers prefer to spend the majority of their time interacting with clients, both existing and prospective. Thus, they would rather have a strategy that does not trade all that often – some want only one trade every two weeks, or even once per month. The trade-offs typically become managing draw-downs without having a lot of trades. You can build a reasonably effective market timing Model that avoids a good chunk of a bear market. But that system will typically exhibit some heavy drops before it exits, and will similarly lag upon its reentry. Conversely, you can keep draw-downs very low by selling positions very quickly on the slightest hint of a change in trend to the downside. But you will be trading all the time, and will be frustrated when the market shifts gears and leaves you behind – you could be spinning your wheels an awful lot. In the end, building a system that trades with just enough regularity to keep the portfolio out of danger and responsive to major trend moves is the balance we are looking for.

In the end, building a good investment system is some parts art, some parts science. You are trying to put together a system that won’t get derailed by investor or adviser emotions, but still achieves their objectives. As we all know, the stock market is an inherently volatile world. Taming that volatility while still capturing the bulk of the returns offered is what we are all trying to achieve. The two TimingCube Models, Classic and Turbo, come at that resolution from different angles. They each do a reasonable job getting to the goal line objective of good risk-adjusted returns. Combining those methods in a broader ETF portfolio can enhance one or more sides of that box we outline above. It’s like a never-ending puzzle, in part because the market is always shifting, investor moods are always a-changing, and the foundation we are working from is dynamic. When the markets are doing well, people want more return and tend to think less about risk. When the market crashes, they want less risk in exchange for even a small token return. Building a system that meets both of those mindsets is what we are working all the time to deliver.

Do you provide the list of past signals for the Turbo Model?

Yes, we do! The complete list of all signals generated by our Turbo Model since March 10, 1999 can be downloaded from our Website. To do so, first log in, then go to the “Signal Returns” section of the Turbo Model “Results” page. There, you can download a file that contains all signal and trade dates, along with the corresponding up-to-date returns for the Long and Short and Long Only strategies applied to the Nasdaq 100 (QQQ), Russell 2000 (IWM) and S&P 500(SPY) vs Buy and Hold. The file is in CSV format and is meant to be opened with a spreadsheet program such as Excel but it can also be read by any text editor such as WordPad.

Market Update

Stocks continued their cautious roll toward the U.S. presidential election this week with a flurry of earnings and economic data to consider along with a note from the European Central Bank (ECB). A weak reading Monday from the Empire State Manufacturing Survey pushed yields lower though stocks did not benefit, falling -0.3% while awaiting a full slate of earnings announcements before Tuesday’s market open. One-time market darling Netflix (NFLX) shot higher on its earnings report to boost tech and related consumer shares Tuesday while gains from United Healthcare (UNH) and Goldman Sachs (GS) combined to further elevate the positive market tone. Stocks added +0.6% on the day.
But that would be the best showing of the week. Wednesday brought more strong earnings from financial companies while energy drilling heavyweight Halliburton (HAL) popped +4% on its earnings beat. However, a poor outlook from Intel (INTC) and sour responses to earnings in the healthcare sector kept the market from making much headway with a +0.2% gain all that the broad market indexes had to show for what should have been a decidedly positive day. The European Central Bank’s meeting provided no hint of further easing in European monetary policy. That output pressured the Euro, boosting the U.S. dollar, which all helped keep stocks in check Thursday. Friday brought a mixed message on earnings with Microsoft (MSFT) beating and hitting an all-time high while General Electric (GE) and Schlumberger (SLB) slumped. A rumored bid for Time Warner (TWX) by AT&T (T) generated interest but little follow-on activity. Stocks opened lower, then gradually marched back to flat throughout the day.

Warm wishes and until next week.