Market Map Model With “Buy In Nov, Sell In May” (“Sell In May”) Allocates To Equities

As the “Sell in May” variable requires allocation to equities in the beginning of November ( “Buy in November” ), the model now allocates to small cap value, emerging small cap, and large cap value universe ( applicable ETFs may be: Vanguard Small Cap Value ETF (NYSEARCA:VBR), WisdomTree Emerging Markets SmallCap Dividend Fund (NYSEARCA:DGS), and Powershares FTSE RAFI U.S. 1000 Portfolio (NYSEARCA:PRF)). Using the revised “Buy in November, Sell in May” strategy stockmarketmap.wordpress.com/2016/10/25/model-with-sell-in-may-revised/, assets would be allocated from cash to stock universes listed on Tues Nov 1 .

The strategy previously signaled cash allocation on April 28th posting.

Model With Sell In May Revised

As the upcoming actionable signaling within our quantitative tactical strategy suite involves the “Buy in November, Sell in May” component ( buy equity class on November 1, sell on May 1 ), further investigation into “Buy in November, Sell in May” combined with Market Map model variables, has uncovered some benefits that can be useful in the realm of diversification via the use of additional stock universes.

Our research expanded the strategy towards the use of an equal weighted “blend” of the small cap value, emerging small cap, and large cap value universes ( versus a singular use of small cap value previously) during the November 1 through April 30th investment period, and the utilities sector ( or bonds, cash / money market equivalent ) during the May 1 – Oct 31 investment period. Over the historical sample period from 11/01/1953 to present, risk adjusted excess returns ( alpha ) were produced over the buy & hold of the “blend”.

“Improvements” on the strategy entail:

1) the use of a variable in determining “high” risk in forward year and instruction towards bond / cash allocation

2) bond allocation from May 1 through Oct 31 using a moving average cross rule applied to bond prices *.

These improvements streamline the number of transactions and lock the transaction dates to just May 1 / 1st trading day in May and November 1 / 1st trading day in Nov. The use of (up to) three stock universes ( stock universe blend choices left up to the user ) helps diversify across a broader investment spectrum. Additionally, historically, there have been periods when emerging equity markets have become “undervalued” and domestic equity markets were fair to “overvalued” ( as is presently the case ). Exploiting this valuation discrepancy can provide possible dampening of the returns volatility accompanied with the “mean reverting” of valuations while simultaneously providing diversification over many markets and companies.

CAPE valuation measures, Emerging vs. U.S.

 

Allocation Heuristic =

1) buy any of / equal weighted blend of Small Cap Value, Emerging Small Cap, Large Cap Value on November 1st of each year.

2) ( In mid January of next year ) calculate return on S&P500 from mid November to mid January ( risk profile ).

3) on May 1st of next year, sell Small Value, Emerging Small cap, Large Value

4) calculate the 10 period moving average, monthly basis on the Vanguard Long term Treasury fund (MUTF:VUSTX)

5) if risk profile calculation from #2 was > 0, then allocate to utilities sector from May 1 to November 1

6) if risk profile calculation from #2 < 0 ( forward high risk ) and if price of VUSTX > 10 SMA, then allocate to intermediate/long term bonds * from May 1 to November 1. Otherwise ( if price of VUSTX < 10 SMA ) allocate to cash.

* Our preference is 75% capital allocation to bonds

As has been the case in the past, updates will be provided

The table below demonstrates returns using 2 and 3 stock universe blend / utilities sector / bond / cash strategy and buy and hold of 3 stock universes.

(click to enlarge)

CAGR Table
Sell in May Buy & Hold Sell in May
3 Stock Universe * 3 Universe 1 & 2 Stock Universe
Utilities / Bond / Cash Utilities / Bond / Cash
1954 – 2016 19.8% 15.6% 20.8%
1954 – 1985 20.6% 14.3% 22.1%
1986 – 2016 18.3% 16.3% 18.9%
2000 – 2016 17.3% 11.8% 19.1%
T-stat 2.08 2.12
Largest Monthly -18% -54% -20%
Decline 2008 2008 2008
* 1 Small Cap Value 2 Emerging Small Cap 3 Large Cap Value Utilities Bond ( VUSTX 10 yr Yield )

Signaling History:

Blend Utilities Annual returns
Bond Nov – Oct
Cash
Nov 1 – Apr 30 May 1 – Oct 31
1954 14.1% 4.6% 19.4%
1955 31.7% -4.2% 26.2%
1956 15.4%  2.0%    Cash 17.7%
1957 5.2% -20.6% -16.5%
1958 14.2% 15.6% 32.0%
1959 9.8% -8.6% 0.4%
1960 5.5% 8.4% 14.3%
1961 22.0% 34.0% 63.4%
1962 9.3% -1.7%   Bond 7.4%
1963 21.4% 0.0% 21.4%
1964 11.7% 20.0% 34.0%
1965 11.1% -1.6% 8.5%
1966 14.9% -1.8% 12.8%
1967 33.0% -23.0% 2.4%
1968 14.1% 14.6% 30.7%
1969 0.7% 3.5%    Cash 4.2%
1970 -16.5% 3.2% -13.8%
1971 32.6% -10.0% 19.4%
1972 23.0% 16.0% 42.7%
1973 -2.0% -13.2% -15.0%
1974 -9.8% 4.0%    Cash -6.2%
1975 31.8% 25.0% 64.8%
1976 26.4% 24.8% 57.8%
1977 19.4% 3.0% 23.0%
1978 18.6% 3.5%     Cash 22.7%
1979 13.9% 1.0% 15.1%
1980 5.2% 5.4% 10.9%
1981 14.1% 7.5%    Cash 22.7%
1982 2.0% 9.4%    Bond 11.6%
1983 33.7% 13.3% 51.5%
1984 10.0% 16.4% 28.0%
1985 13.7% 10.4% 25.5%
1986 32.5% 14.7% 51.9%
1987 24.4% -2.6% 21.2%
1988 23.8% 11.8% 38.4%
1989 23.5% 15.0% 42.0%
1990 2.6% 4.9%    Bond 7.6%
1991 32.7% 8.4% 43.8%
1992 17.6% 9.3% 28.6%
1993 18.3% 8.3% 28.1%
1994 6.0% -1.8% 4.1%
1995 -0.9% 14.8% 13.8%
1996 16.1% 3.3% 20.0%
1997 8.0% 19.0% 28.5%
1998 10.1% 9.4% 20.4%
1999 26.2% 6.4% 34.3%
2000 9.8% -6.5% 2.7%
2001 4.6% 9.9%     Bond 14.9%
2002 28.3% 8.1%      Bond 38.7%
2003 9.2% 15.3% 25.9%
2004 13.2% 13.6% 28.6%
2005 7.8% 6.2% 14.4%
2006 24.8% 15.4% 44.0%
2007 17.5% 1.2% 18.9%
2008 -10.2% -0.9%    Bond -11.0%
2009 2.4% 11.0% 13.7%
2010 26.8% 7.9% 36.9%
2011 17.6% -2.7% 14.5%
2012 10.9% 3.0% 14.3%
2013 15.1% 1.3% 16.6%
2014 6.1% 5.1% 11.5%
2015 2.4% -1.3%    Bond 1.0%
2016 -1.6% 4.9%      Bond 3.2%

Charts below show longer term investment return periods. Note that although returns on the three different portfolio examples have declined in the past 15 + years, declining returns of the S&P500 index ( the most referenced investment vehicle used for the growing “passive investing” movement ), have been accompanied with increasing and bourgeoning popularity.

 

 

Concluding Points …

In order to produce optimal risk managed asset accumulation ( alpha ) in an empirical fashion, an investor can construct a systematically based investment policy using:

– a diversification of stock universes that have produced highest excess returns historically over a long sample**

– risk mitigation variables that have shown positive outcomes, statistically

– asset classes that have negatively correlated price movement or “safe” assets during periods of high risk ( long dated government bonds )

Supplemental article data here:

Rolling 15 year total returns comparisons and additional data here:

docs.google.com/spreadsheets/d/1tvKoFaFCQhbO5gQSld7i8TMPw_ZTXB1-AA_9velX_tM/edit#gid=1235486791

Some of the funds representative of the stock universes and asset classes for use in the strategy: VBR, VIOV, DFSVX, DGS, DEMSX, PRF, FSUTX, FUTY, TLT, VUSTX

High risk years = 1956, 1962, 1969, 1970, 1974, 1978, 1981, 1982, 1990, 2001, 2002, 2008, 2015, 2016

* data for moving average calculation – 10 year bond yield data 1954 – 1983 and VUSTX data 1983 – 2016

** Fama and French docs.google.com/document/d/1-_nQTfZG1aoIduZuvqCgJws3_ja_zHCuz4h6zLadUoo/edit

Data sources and calculations = IFA Index Calculator, DFA funds, VUSTX, FSUTX, longrundata, portfoliovisualizer.com

As diversification of stock universes is important, diversification of strategies is additionally important. As the “Sell in May” with Map model strategy has shown excellent results, it may not be prudent to invest all of one’s total capital towards it.

Market Map Allocates To Long Bond

Summary

– The Market Map model indicates allocation to long bonds

– Since 1924, there were 16 instances of long bond allocation

– This allocation applies to the 3rd quarter of 2016

As another predefined date for asset allocation action as defined * by the Market Map model has been reached, at this time the model is indicating an allocation from cash (position allocated on 01/29/2016) to a 75% allocation in the iShares 20+ Year Treasury Bond ETF (NYSEARCA:TLT). The model identified 16 previous instances of allocation to long bonds in the investment timeline since 1924.

7/1/1929 9/30/1929 1.70%
7/1/1930 10/6/1930 -0.20%
7/11/1946 9/30/1946 -6.1%
7/5/1960 10/3/1960 2.6%
9/3/1982 10/2/1982 5.5%
7/10/1992 10/2/1992 3.2%
7/11/1997 10/3/1997 6.6%
7/2/1998 10/2/1998 9.7%
7/2/1999 10/2/1999 1.1%
7/6/2000 9/28/2000 3.4%
7/6/2001 9/29/2001 7.5%
7/5/2002 10/4/2002 15.9%
7/3/2008 10/3/2008 4.5%
7/9/2010 10/1/2010 6.6%
7/8/2011 9/30/2011 21.3%
7/11/2014 9/29/2014 3.50%
7/8/2016 9/30/2016 ?

As we can see from the chart below, there is statistically significant negative correlation between long bonds and S&P500 during Map model bond allocation signaling.

The long bond allocations within the full record of model allocation changes can be seen here.

* Objective # 3. To make infrequent asset allocation changes occurring on fixed, predefined dates during the course of market cycles. See objectives.

Market Map Model With “Sell In May” Goes To Cash

As the Market Map model identified 2016 as having a “High” risk profile, using the popular “Sell in May” strategy combined with the Market Map model ( variable #6 https://stockmarketmap.wordpress.com/2015/11/14/market-map-model-tactical-asset-allocation-using-low-expense-index-etfs-2015/), assets would now be allocated from equities to cash., assets would now be allocated from small cap value equities to cash on Monday 5/2/2016.

The next possible action may be a bond allocation held for the 3rd quarter commencing in the second week of July 2016.

Market Map Allocates To Cash

Since the equity allocation instruction of post 12/31/2015 ( https://stockmarketmap.wordpress.com/2015/12/31/market-map-model-allocates-to-equity-etf/ ), as of 01/29/2016,  model variable #2 has identified 2016 as a “high” risk profile year.  Thus, the model allocates from equities ( QQQ / SPY)  to cash equivalents on 02/01/2016.

The present -3.5% S&P500 loss that the model has produced for the month of January falls in line with previous and similar signaling history involving “high risk” year Januarys ( 12/31/1969 – 02/02/1970 -3.2%, 12/31/1973 – 02/01/1974 -2.5%, 12/30/1977 02/01/1978 -5.3%, 12/31/1981 – 02/01/1982 -1.9% ).

The model has identified 16 previous instances of “high”risk profile since 1924 as shown in chart below.

ScreenHunter_1131 Mar. 23 08.07