The iM Minimum Volatility (USMV) – Investor

Approximately five years ago we launched at iMarketSignals the four quarterly displaced iM-Best12(USMV)Qx,  x=1, 2, 3 or 4, investor models each holding 12 positions. The purpose of these models was to check whether our hypothesis (that a portfolio of top ranked USMV stocks in each of the four models will provide higher returns than the underlying ETF)  is supported by the actual performances of the models over a longer period of time.

USMV’s inception was Oct-20-2011, and by end of Jun-2014 when we launched the first quarterly model the fund’s net assets were about $2.7-billion. Now, five years later, the fund’s current net assets had increased 10-fold to $28.6-billion, attesting to this ETF’s popularity.

Over the last five years the performance of iM-Best12(USMV)Q1+Q2+Q3+Q4-Investor resulted in an excess return of 34.8% over USMV and 54.5% over SPY, thus corroborating our initial hypothesis. (Figure-3 in the Appendix-A refers.)

Over time the holdings of the four models have converged to hold only 18 different stocks between them, so future performance of each of the models is expected to be very similar. There is not much to be gained by following four similar models and we will be replacing them with the iM Min Volatility(USMV)-Investor. This model holds 10 equal weighted stocks from the iShares MSCI USA Minimum Volatility ETF (USMV), rebalanced every 13 weeks, and retains a position for a one year minimum period. Every three months the stock universe from which stocks are selected will be updated to correspond to the then current holdings of USMV.

The iM Min Volatility USMV – Investor.

We have backtested this model from Jan-2013 onward with point-in-time data of USMV’s holdings. The simulated performance is shown in Figure-1 relative to the benchmark SPY. Over 6.5 years the model would have produced about double the total return of SPY. Estimated trading cost and slippage were taken into account.

Since inception the simulation shows an annualized return of 22.0% versus 14.3% for SPY and an annual turnover ratio of 60%. There were only 38 realized trades with an average holding period of 481 days.

width="640"/(click to enlarge)

Also, over the time period 7/1/2014 to present, it outperformed iM-Best12(USMV)Q1+Q2+Q3+Q4-Investor, see Figure-3 in Appendix-A.

The model currently only holds large-cap stocks in five sectors, as shown in Figure-2.

The historic performance statistics are in Table-1, and the risk measurements are in Table-2.
/ /

Current Holdings (7/5/2008)

Ticker Weight Buy Date Qty. Shares Avg. Buy Price
incl. fees
incl. fees
Value now Gain to date
HD 9.95% 1/12/15 171 $105.13 $17,977 $36,269 $18,292
ROST 9.22% 5/11/15 336 $48.04 $16,141 $33,600 $17,459
CDNS 11.37% 3/27/17 554 $24.81 $13,746 $41,445 $27,699
MSFT 10.48% 6/11/18 278 $101.16 $28,123 $38,214 $10,091
JNJ 9.40% 10/8/18 241 $139.12 $33,529 $34,256 $727
LLY 9.57% 11/12/18 305 $114.80 $35,013 $34,898 -$115
PEP 9.41% 11/19/18 256 $121.88 $31,201 $34,301 $3,101
CSCO 9.57% 12/3/18 618 $47.51 $29,359 $34,905 $5,545
BAX 10.29% 3/4/19 457 $75.56 $34,529 $37,506 $2,977
AFG 10.58% 6/17/19 366 $103.59 $37,915 $38,565 $650

Note the buy dates. The minimum one year holding period is timed from the dates listed.

Following the model

At our website iMarketSignals one can follow the performance of this model, updated weekly.


One should be aware that all results shown are from a simulation and not from actual trading. They are presented for informational and educational purposes only and shall not be construed as advice to invest in any assets.


Figure-3 shows the actual historic performance of iM-Best12(USMV)Q1+Q2+Q3+Q4-Investor comapred to SPY.  Also plotted is the simulated performance of iM Min Volatility(USMV)-Investor

width="640"/(click to enlarge)

Posted in blogs, featured
15 comments on “The iM Minimum Volatility (USMV) – Investor
  1. randyfloyd says:

    is the model always long? is there any risk management for bear markets?

  2. geovrba says:

    This model is always long. We have plenty of market timer models and recession indicators at iM to provide timely signals to get out of stocks.

  3. stbrock says:

    Looks like a real improvement and a simplification as well. Thanks. The table of correlations you did a few years ago was very helpful. Would it be possible to put together an updated table covering this strategy and some others that weren’t in the last table? VDIGX trader, VIX/ZIV, 7 HiD LoV, 5 ETF Trader leveraged, 12 USMV Trader for example.

    • Anton Vrba says:
      Correlation Matrix A B C D E F G
      Best12(USMV)-Trader A 0.83 0.53 0.59 0.71 0.58 0.84
      Best10(VDIGX)-Trader B 0.83 0.54 0.57 0.74 0.54 0.83
      Best7(HiD-LoV) C 0.53 0.54 0.29 0.55 0.37 0.53
      iM-5ETF Trader D 0.59 0.57 0.29 0.21 0.73 0.62
      iM Tax Efficient Capital Strength Portfolio E 0.71 0.74 0.55 0.21 0.5 0.7
      iM VIX-Timer with ZIV F 0.58 0.54 0.37 0.73 0.5 0.6
      iM Min Volatility USMV-Investor G 0.84 0.83 0.53 0.62 0.7 0.6
  4. stbrock says:

    Thanks, much appreciated

  5. jlombard says:

    Will this model continue to have a rule to sell a position if it declines 15% from its high mark?

    • jlombard says:

      Is there any draw down protection in lieu of the 15% rule? The original USMV 12’s had the 15% rule did they not?

    • geovrba says:

      We have added a rule to sell if return of an existing position is less than -15% (dividends are not included). i.e. sell a position if it gained less than -15% from buy price.

      This improved 2013-2019 simulated annualized return by about 0.5% and resulted in two sales.

  6. randyfloyd says:

    i’d like to propose a model just like this, but for a small cap low-vol fund, such as XSLV. the results could be extraordinary. check out this article from Ploutos on seeking alpha, you will see what I mean:

  7. sqpei says:

    hi Georg,
    How about combining this model together with Composite Timer or super timer? For example, in between Oct 2018 and Jan 2019, when signal from composite timer is to sell SPY and buy IEF, we sell all holding position and then buy IEF? Would like to know this performance and the max D/D since 2013. Thank you!

    • randyfloyd says:

      this would be interesting. i’m guessing you could probably just implement it yourself. the analysis above says that the correlation with SPY is 0.90. I plan on using one of those timers as a hedge for this strategy

  8. Tom Bria says:

    Why the model table is no longer present in the first list ?

Leave a Reply

With reference to Section 202(a)(11)(D) of the Investment Advisers Act: We are Engineers and not Investment Advisers, read more ...
By the mere act of reading this page and navigating this site you acknowledge, agree to, and abide by the Terms of Use / Disclaimer
Share This