Market Map model: Tactical Asset Allocation Using Low-Expense Index ETFs – 2015

The Objectives of the Market Map Model

1a. To capture multi-month to multi-year market index price trend appreciation through investment in index products that track the S&P 500 or the Nasdaq 100 indices (SPDR Trust ETF (SPY), Vanguard Total market ETF (VTI), or Powershares QQQ Trust ETF (QQQ)) during “equity” allocation periods, and

1b. To invest in: 1) Long dated bond securities represented by ETF products such as TLT or EDV ( Barclays Long treasury ETF / Vanguard Extended Duration Treasury ETF ) and 2) Short term Treasury market instruments during “cash” allocation periods .

2. To add long term value to the portfolio’s total return by utilizing low expense and low tax ratio investment products (this under auspices of investment principles set forth by John Bogle, founder of Vanguard Group of funds).

3. To make infrequent asset allocation changes occurring on time specific, predefined dates during the course of market cycles.

4. To show consistent and comparable performance accompanied by reduced volatility relative to the S&P 500 index over many decades of diverse stock market environments and cycles (using historical testing).

5. To strategically decrease equity allocation as upside price appreciation occurs and increase equity allocation as price depreciation occurs.

Basis for and Components Applied Towards Asset Allocation Decisions

The model utilizes calculations on 7 main “price + time” based variables * ( variables’ data series not subject to revision ) which are then constructed into a chronological sequence of explicit signaling steps or components.

The underlying basis of the signaling adjusts equity / bond / cash allocation exposure for inevitable periods of below-average returns after empirically defining periods of above-average performance. Conversely the signaling adjusts equity allocation exposure for inevitable periods of above average returns after empirically defining periods of below average performance.

Shown below is the historical record of the model which, during entry dates, allocates 100% capital to the SP500 when “SP500” is indicated and 100% or two 50% capital allocations during certain signature market declines and “bottoming” processes; and 75% capital allocation to Long dated bond equivalents when “BOND” is signaled, or short term treasury equivalents when “cash” is indicated.

Shown below is the historical record of the model which, during entry dates, allocates 100% capital to the SP500 when “SP500” is indicated and 100% or two 50% capital allocations during certain signature market declines and “bottoming” processes; and 75% capital allocation to Long dated bond equivalents when “BOND” is signaled, or short term treasury equivalents when “cash” is indicated.

Note: in an attempt to replicate the present interest rate environment, the interest rate on short term treasury equivalents shown during all historical “cash” periods has been set at 2% per annum.

Because of the historical efficacy of the key components and calculations, the vast improvement on equity drawdown versus buy and hold, and the “long term” investment philosophy on which the model is based, stop loss methods have been deemed unnecessary.

    • SP500 price history supplied by ” S&P500 Dividends Reinvested Price Calculator ” + using VFINX

Signal Table 1924 - 1966 Sp500

Signal Table 1966 - 1999 SP 500

Signal Table 1999 - 2015

Performance Statistics

Statrs table Map 1924 - 2015

Stats table Map 2000 - 2015


20 Year Performance Periods

Model vs. SP performance 1924 - 1943

Model vs. SP500 1944 - 1963

$1 growth Model vs. Buy Hold SP500

Model vs. SP500 1983 - 2003

Map model 2004 - 2014

Signal dates

Predefined “equity” ( via SP500 / QQQ ) allocation dates occur on:

a) the last trading day in December

b) ( beginning of ) select 4th quarters

* A minority of  all signals are not “predefined” but still mechanically derived

Predefined “bond” ( via TLT ) allocation dates occur on the first Friday after the 4th of July ( 3rd quarters ):

Equity Allocation Table

Predefined “cash ” allocation ( treasury bills, short term cash equivalents ) occur on:

a) the 1st “trading” Day of February

b) the 1st Friday after July 4th

c) the last “trading” day of the year

d) ( beginning of ) select 4th quarters


 Cash Allocation Table Map

Benefits of using a systematic and mechanical process


    • Fees, expenses, and commissions are low ( with use of a low cost broker, low expense index etfs, elimination of “excess” fees paid to portfolio managers )
    • The chronological format gives a “map” and “advanced knowledge” of fixed signal dates as a guide in making the appropriate allocation transactions.
    • Human judgement and subjectivity are removed from the investment process
    • The account holder has some control over position management and knowledge of the signaling.
    • “Sequence risk“ is alleviated.
    • Because of the empirically derived nature of the research and robustness of data sample, a level of confidence and precommitment in executing the allocation transactions in gained, thus avoiding the behavioral / cognitive biases and ego driven mistakes that appear during discretionary portfolio management (ie. we let the process do the work).
    • We procure “long term” thinking instead of being distracted by short term “randomness”, financial news flow, and media “noise”.
    • “Losses” are avoided                                                                                                         . . . . . . . . . . . . . . .

 7 main non subjective, price based variables:

    1. Mean reversion measurement using consecutive occurrences of annual overperformance or underperformance of SP500 versus a fixed baseline.
    2. Time series calculation conducted on performance of SP500 over specific months, categorized into risk profiles ( Favorable, Neutral, and High) applied towards upcoming year, indicated on 1st trading day of Februarys
    3. Relationship of long bond price and short length moving average of bond prices to long length moving average during select 3rd quarters
    4. Statistically significant equity performance in 4th quarters based on Presidential cycle
    5. SP500 price vs. long length moving average combined with the use of price oscillator
    6. *  “Sell in May / Halloween” anomaly  ( used for map model applied to small cap universe )       7. Generational strings of consecutive years and derived heuristics (as calculated by #1)


Disclosure:  Model is in cash position since 1/20/2015

Model design change: cash allocation dates

Through an “ease of use” improvement suggested by users and colleagues, we are changing the “cash” allocation publishing dates that have occurred previously on the “3rd week of January(s)” to the 1st trading day of February(s).  This improvement does not affect the structure of the model and past performance was minutely affected. Also, this transaction date will concur with the transaction date of another sector investment strategy ( thus consolidating strategy transactions onto one date ) that we have implemented for our clients. The change will be reflected in the subsequent post.

Model With “Sell In May” Component Allocated To Small Cap Value

As the “Sell in May” anomaly requires allocation to equities in the beginning of November, the model now allocates to small cap value universe ( we use Vanguard Small Cap Value ETF (VBR). It previously signaled cash allocation on May 3rd posting

Primer on Map model with “Sell in May”

Market Map Model Indicates Cash Position For 4th Quarter 2015

As the Market Map model is composed of different variables or “components” that guide tactical allocation decisions, the model’s algorithm has not confirmed an equity allocation for the 4th quarter of 2015 via variable # 4 (“Statistically significant equity performance in 4th quarters based on Presidential cycle” here

Model = Cash position ( since 01/20/2015 )

next calculation and update will be Dec 31 2015 …

Market Map Stays In Cash

As another predefined date for actionable asset allocation as defined by the Market Map model has been reached, at this time the model is indicating holding the allocation in cash. As there was a possiblity that the model could allocate to bonds during this July, the negative trend in bonds cancels the opportunity. In a single late signalling case, ( 9/3/1982 in table below ) out of 14 occurences back to 1924, did the model indicate an allocation to bonds outside of the early weeks of the July time frame. If the model should indicate this late behavior again, we will update the blog ..

The next actionable allocation date will be 12/31/2015 …

Map Model
All signals BOND
Buy Date Sell date BOND return
07/11/46 10/30/46 -6.1%
7/5/1960 10/03/60 2.6%
09/03/82 10/02/82 5.5%
07/10/92 10/02/92 3.2%
07/11/97 10/03/97 6.6%
7/2/1998 10/02/98 9.7%
07/02/99 10/02/99 1.1%
07/06/00 9/28/00 3.4%
07/06/01 9/29/01 7.5%
07/05/02 10/04/02 15.9%
07/05/08 10/03/08 4.5%
07/09/10 10/01/10 6.6%
07/08/11 09/30/11 21.3%
07/11/14 09/29/14 3.50%
average 6.1%

Model With “Sell In May” Component Goes To Cash

As May 1st occurred last week, the Market map model combined with the “Sell in May” component using the small cap value universe, has signaled a cash allocation. The model previously signaled an allocation to small cap value on 09/27/14 ( here: Previous instances of buy and hold returns during cash allocation periods are reflected in the chart below.

Small Cap Value returns during  Sell in May Cash

Further explanation on the Market Map model using the “Sell in May” component can be found here:

Probability of allocation to TLT (iShares 20+ Year Treasury Bond ETF) on 7/1O/15 = 5O%

Market Map: Staying the Course for the Long Term 2015

As we head into Spring 2015 with the model in a cash position, we look back at the model’s performance in 2014 and note it as a year of “underperformance” vs. the S&P 500 (as the model produced a return of 8% and the S&P500 benchmark produced 13.5% ). From time to time, it is possible for an occurrence, such as this, to happen and for us to react with skepticism, uncertainty, and doubt, especially when it involves our investments. We are “hard wired” and  persuaded by societal cues into thinking in terms of “short term” results and we tend to define the equity market in terms of “winning” and “losing” ( we are much more sensitive to “losing”).

We are fortunate to have 9 decades of  historical model performance with which to analyze and review past occurrences of when the model “underperformed” the S&P 500.


ScreenHunter_540 Apr. 07 17.52


Model underperf years

As we can see in table 1 and the chart above, these years of underperformance have happened occasionally and are evenly distributed over the sample with “consecutive” years of underperformance occurring in sporadic fashion.




ScreenHunter_539 Apr. 07 17.43

Table 2 shows the ratio of outperforming vs. underperforming occurrences and cum % performance.

The stock market’s price is the discounted present value of it’s future earnings and the market  can “overshoot” this present value for periods of time. During this part of the investment cycle, this “overpricing” can be caused by: 1) portfolio analysts’ upwardly adjusted revisions in earnings estimates and  2) investors’ optimism and “fear of missing out”  as the market rises or has, in retrospect, produced many consecutive positive years returns. We must expect that these periods will exist and can expect to leave some “gains on the table” with the knowledge that we are accumulating and compounding assets over the “long term”. We can gain perspective by examining the “long term” performance statistics, and adjust our reactivity and concern accordingly . 

We can also gain perspective by viewing, in a graphic sense, how much equity exposure towards historical market up moves or down moves has been realistically “captured” by the model in order to achieve portfolio alpha (in this presentation) :




Asset allocation since 01/20/15 = cash
Probability of allocation into TLT (iShares 20+ Year Treasury Bond ETF) on 7/13/15 = 5O%