Better Second Opinion
Cover Story, by Mark Robertson, Managing Partner October 1st, 2010
Trust, but verify. We've made subtle shifts over the years in characteristics like EPS Stability and Financial Strength ... we make another subtle change this month.
Trust, but verify. Long-time subscribers will remember and recognize that we’ve made subtle shifts over the years in characteristics like EPS Stability and Financial Strength.You also know that this evolution has led to a form of “conditioning” with respect to things like P/E ratios — forming a consensus around sources like Morningstar and S&P in addition to the opinions and forecasts from Value Line. And we’ve tracked our median forecast (MIPAR) vs. VL’s median price appreciation (VLMAP). Effective immediately, we’ll make another subtle change, leaning on something else … and it’s better.
Value Line calculates a high and low forecast for the 3-5 year annualized total return for the stocks in the standard edition. We’ve long wondered how the low forecast compared to our projected annual returns. So we checked …
It kept happening again and again.
I’d complete the calculations on an Equity Analysis Guide (EAGLE) … take a quick glance at the Value Line company report … and notice a great deal of similarity between the PAR for a given company and the low forecast conditions detailed in the Value Line 3-5 year projections.
And we wondered … and curiosity swelled. Just how often does our PAR come very close to the VL low forecast? How close? How often?
We’ve counseled using the average VL forecast for years. The thinking was that the average of the high and low forecast conditions (according to VL) can provide a guiding beacon or checkpoint as we audit and consider the results of our stock studies.
A New Dawn for the Crystal Ball & Barometer
Based on our findings, that all changes. We’d discovered some time ago that the amount of variation between actual returns and the forecast known as the Value Line Median Appreciation Projection (VLMAP) routinely was on the order of 3-5 percentage points. To account for this, we made an adjustment to historical VLMAP. We also noted that the conditioning (including the influence of Morningstar and S&P et al.) had made a gentle reduction in the average PAR. This was a result of lower forecasts making a modest reduction in the growth rates and projected annual P/E ratios when the overall consensus became a routine part of our approach. (The adjusted MIPAR values in the shaded area in the accompanying figure were a result of this research. We’ll be taking a closer look at this as we build out the historical PAR data for all companies in the VL universe.)
MIPAR vs. the VL Low Forecast. MIPAR has “routinely mirrored” the low annualized total return forecast according to Value Line. The shaded area reflects a period of adjusted PAR calculations based on the influence of ACE starting in early 2008. This correlation is stronger than VLMAP and possibly more representative.
Neither VLMAP or VLMAP* included the effect of dividends. The VL total return forecast does.
Our Market Barometer
Going forward, we’re going to adopt the Value Line low total return forecast as VL’s projected annual return and we’ll be auditing MIPAR vs. the VL low forecast continuously. We will continuously calculate this average forecast. This change has been made on this month’s Sweet 16 screening results.
Whenever we refer to VL’s PAR, we’ll be talking about the low forecast whether we’re talking about a single company or the collective “index.”
When it comes to gauging overall market conditions, very little changes. We believe that the multi-decade range (going back to 1960) for VL’s PAR would prove to be on the order of 2-20%. A MIPAR or VL PAR reading in the 10% realm, give or take a percent or two, would reflect a fairly valued market. At the depths of the bear market in Nov 2008 or March 2009, both indicators hovered in the 20% range — signaling one of the most powerful ‘back up the truck’ moments in the investing lifetimes of most community participants. The extended period where our barometer(s) hovered at 5-7% during 2005-2008 is something that we want to take a closer look at.
It’s About the Returns (Stupid)
We have long wondered what we could have seen or might have used to guide our investing efforts back in 2008. MIPAR had been stuck in a range at relatively low historical levels and many of you will recall that we urged accumulation of the deepest blue chips back in that time frame.
We now know that a blue chip concentration served us well. During a period when the market swooned over 40 percent, our collections of high-quality stocks sagged a mere 10-20 percent … enabling a faster recovery. Tin Cup is something of a living example.
We don’t regret the suggestions we made at the time. Since then, we’ve carefully explored NAIC co-founder George Nicholson’s concepts of asset allocation including embracing relatively large “cash equivalent” positions when MIPAR approaches historical low conditions. That said, the multi-year period at historical lows for projected returns strained the patience of many. We’ve wondered if there could be fundamental indicators that may have nudged quality (and cash) higher in our portfolios in a timely manner.
Might there be a fundamental indicator to serve alongside the technical analysis-based “death cross” and relative strength index peaks? Answer: “I think so.”
Our emphasis on projected returns at MANIFEST is because we’re “here” for the returns. It follows somewhat naturally that the answer probably lies in the returns … or the forecasted returns. It’s also logical that any threat to those returns or ‘disturbances in the force’ play a critical role. In the market, this plays out as uncertainty. Disruptions in expectations generally have a damaging effect on stock market averages.
Measuring Uncertainty? Might it be possible to gauge investor angst? The bars reflect the amount of variation in forecasts and the highlighted range could be considered “normal.” The blue shaded area is a stock market index, specifically the VL arithmetic index — so it includes a mix of small, medium and large companies. The graphic suggests that forecast volatility exceeded the upper limit in early 2008, well before the financial crisis and only a few months after the start of the market swoon but before the cliff diving was fully underway.
In the accompanying figure, we take a closer look at the variability of return forecasts. In a nutshell, we’re gauging the range of high and low forecasts and how much the group of forecasts deviates from a long-term average. The highlighted area represents the “normal range” (6-7.5%) leading up to 2008. As shown, forecast volatility first exceeded the upper limits in early 2008. Volatility continued to increase for months before the demise of Bear Stearns or Lehman and/or the advent of the great recession.
The takeaway is that forecast turbulence (in the form of exceeding typical upper limits) appeared several months before the largest market declines. The variability in forecasts finally receded and was restored to normal levels from December 2009 until recently. As of 9/30/2010, this metric is back at the “upper limit.” Needless to say, we’ll continue to monitor this closely and work to add more historical data as part of our ongoing research.
Our Core vs. Non-Core Debate
What is a core stock? How would you define it?
This is a fairly common question. I’d reinforce it with a follow up question: “Why should it or does it matter?”
Core vs. Non-Core. This is an example of a poll conducted in the MANIFEST Forum to gauge audience thinking.
The answer to the second question first is that we design and enforce selling disciplines around the core/non-core characteristics of a given company in our EAGLEs as we set the selling PAR criteria. Yes, our “zoning” is return-based. Imagine that.
Basically, the more “core” a particular holding is, the lower the PAR for the selling condition in our analysis. As the figure suggests, Johnson & Johnson (JNJ) and Abbott Labs (ABT) rank among the “most core” among the holdings of the Challenge Club and/or Round Table. We’d basically hold them until their PAR is less than money market rates or 5-year yields (currently approximately 1%.)
At the other end of the spectrum, we suggest that it makes sense to be quicker on the trigger — selling a non-core holding — at a much higher PAR.
A Big Role in Core vs. Non-Core. This graphic displays Challenge Club and Round Table stocks graphing core “perception” versus forecast volatility. It suggests that forecast uncertainty is a potentially significant influence on our perceptions of core vs. non-core holdings.
During a recent Round Table session, my colleagues stated that they believe all eight of our first selections were core holdings. They balked when I suggested that our audience would disagree. As shown in the figure, core vs. non-core is in the eye of the holder. The difference obviously is related to quality ratings but there appears to be another dimension. The amount of forecast volatility appears to play a significant role for most of us in our perceptions of core vs. non-core.
Bottom line? We believe that the VL low forecast for companies and the overall average as gauge for our MIPAR barometer is a better approach. We’ll incorporate the VL PAR in the Sweet 16. We’ll continue our work with tracking forecast volatility and its potential as part of of our market barometer.