Weekly Update

The Challenge of Building Investment Models


Tagged: , , ,

Published March 3, 2017

timingcube_cartoon030317

Over the 15+ years we have been publishing our investment models much has changed. Exchange-traded funds (ETFs) were in their infancy when we started back in 2001 with only a handful of major choices, all focused on major indexes like the S&P 500. The notion of shorting markets had been around a long time. But primarily among professionals and avid traders. Ever-rising computation power and internet speeds made crunching data quicker, easier, and cheaper putting into every hobbyist traders hands the power to build investment models.

What looks easy isn’t. This week’s blog post sets out how changing market conditions can render a model obsolete. There are no alarm bells going off to tell you the market has changed its character and left your prized model behind.

Take a widely-followed market indicator – the ratio of the number of advancing stocks to the number of declining stocks. It seems common sense that a rising stock market would be characterized by more stocks going up than down. The past ten years have been very kind to this indicator. Chart 1 below shows two lines – the ratio of advancing and declining issues zigzagging upward with a simple moving average smoothing out the movement of the zigzags. Every time the indicator crosses notably above the smooth moving average line we issue a BUY signal (shown with the blue vertical lines), executed by purchasing a broad market ETF like the SPY or QQQ. Every time the indicator crosses below the smoother line, we have a SELL signal and sell our position in SPY or QQQ, expecting further drops ahead.

Chart 1 – Buying at the blue line and selling at the red produces great results!

Good Timing

On Chart 2 we have the price of the stock market, the S&P 500. You can see that there is a very high correlation between the direction of the above indicator and the stock market. Buying on the blue lines and selling on the red lines is a highly winning strategy! We would have sold our stock market position in the summer of 2007 missing entirely the 2008 market plunge. We would have re-entered the market in April 2009 and rode stocks higher until the brief market selloff in August-September 2011. Back into the market in late 2011 and on board for another long uptrend until about May 2015 when we would have sold our position returning to buying stocks again around March 2016. We would still be invested as of February 2017. We would have participated in almost the entire market move higher over the past decade while avoiding the crushing loss of the 2008 bear market. How awesome is that!?

Taking our winning formula discovery further, we look to apply it to the prior market cycle where stocks ran up powerfully until the 2000 market crash. Then, a new market rally in 2003. How did our hotshot indicator do then?
This time our indicator sadly falls apart – putting us into the market from the beginning of 2001 through mid-2002, a period when the market is steadily dropping. The indicator told us to sell our stock position in mid-1998 and remain on the sidelines to the end of 2000. This would have caused us to miss out on the final 100-200 points of the S&P 500’s gains of the late 1990s. We would have lost money over a 5 year period before finally getting it right with the market turn in 2003.

Chart 2 – Buying at the blue line and selling at the red produces a disaster!

Bad Timing

And so it is that some very successful market timers and experts get things so very wrong. They rely on a market indicator or model that worked for them for some period of time. During that time they gained great confidence in the model and came to believe it foolproof. Some very notable investors have sat out the past years of a bull market because of their mistaken models and flawed assumptions. For a classic and high-profile example, read this piece:

http://awealthofcommonsense.com/2014/02/curious-case-john-hussman-understanding-biases-process/

We let price rather that any particular viewpoint drive our models. We use multiple inputs in order to avoid the problem above where an indicator stops working. Price is the best indicator. When stock prices are rising, get invested. When they are falling, sell and step aside or go short. Simple as that. Building a model that accomplishes that simple premise is brutally hard. We think we’ve achieved consistency, reliability and performance, however, and hope you have profited and will continue to profit from our models.

Note: While TimingCube offers “single-point” models that are either in, out, or short the market, our FPResearch Multi-Asset portfolio combines eight models applied to various market sectors and asset classes into one portfolio. Using eight separate models will further lessen the impact of a model going rogue and losing its way. The result is almost 20 years of always-positive returns. Find out more at www.fpresearch.com.


Market Update

Stocks continued their winning streak as investors looked to close February with solid gains. Monday’s session saw strength in defense and biotech names with defense companies the beneficiaries of President Trump’s proposed boost in spending on that sector. The market broadly closed with only a slight +0.1% gain however. Investors took money off the table Tuesday ahead of President Trump’s speech with small-cap stocks off almost -1.5%. Large-cap indexes performed better with losses limited to -0.3% as those indexes closed the month with better than +3% gains. In earnings news, the destruction of “brick-and-mortar” retailers continued with Target (TGT) shares plunging on weak earnings and outlook. Caution ahead of President Trump’s speech gave way to a furious short-covering rally Wednesday with stocks racing to +1.4% gains. Home improvement retailer Lowe’s (LOW) bucked the above-noted trend in retail by posting strong results while investors continued ramping up expectations of a March interest rate hike, with the 10-year yield back near 2.5%. Some of the Wednesday gain reversed in Thursday’s session with Caterpillar (CAT) weighing down large-cap indexes on news of a search of its facilities by federal agents. The IPO of social media company SNAP was very well-received with a +40% pop in its first day price. Broad markets gave back -0.6% on the day with oil inventories continuing to add to their record levels which is keeping pressure on that sector. Fed Chair Yellen added to March rate hike expectations in Friday remarks noting that the economy and pricing has firmed.

Stocks held flat on the day to close their sixth straight positive week. The Nasdaq 100 (QQQ) index has been positive eight of nine weeks without a loss of any note all year. For the week, the S&P 500 added +0.73% while the Nasdaq 100 lifted +0.60%. The small-cap Russell 2000 (IWM) continues to lag with a +0.07% move on the week.

Warm wishes and until next week.