WIXLIB

CLIMB YOUR MOUNTAINQ GROUP MEETING – LAJOLLA, CAOctober 29, 2019

THE EXISTENCE AND PERSISTENCE OF FINANCIAL ANOMALIESHarry Markowitz, PhDHarry Markowitz & Co.San Diego, CAJohn B. Guerard, Jr., PhDMcKinley Capital Management, LLCAnchorage, AKOctober 2019

PORTFOLIO CONSTRUCTION AND MEASUREMENTResearch Includes: Markowitz (1959) Mean-Variance Portfolio Selection Sharpe (1964), Lintner (1965), Mossin (1966) CAPM Treynor and Mazuy (1966) CAPM Performance Measurement Fama (1970, 1976, 1991) Efficient Markets Hypothesis Ross (1976) and Roll and Ross (1980) APT Friend, Dhrymes, and Gultekin (1983) APT Blin, Bender, and Guerard (1997)APT Applied Investment Management and Portfolio Measurement Guerard, Markowitz, and Xu (2015) APT and Axioma Risk Models Tested3 3

EXPECTED RETURNS MODELING TO EXPLOIT FINANCIAL ANOMALIESResearch Includes: Dimson (1988) Size Anomalies Jacobs and Levy (1988) Financial Anomalies Bloch, Guerard, Markowitz, Todd, and Xu (1993) Financial Anomalieswith the Markowitz Risk Model Blin, Bender, and Guerard (1997) Financial Anomalies and the APTModel Guerard (1997) CTEF Introduced with BARRA Risk Model McKinley Quant (2006) MQ Introduced with APT Risk Model Guerard, Markowitz, and Xu (2015) MQ and CTEF Updated and Verifiedfor Axioma Stock Selection4 4

T H R E E TA K E - H O M E S F O R O U R A U D I E N C E1. The Financial Anomalies stock selection models recommendedin 1991 have worked to produce statistically significant ActiveReturns and positive Specific Returns (or Asset Selection),including transactions costs;2. The Japanese and R1 models initially estimated produce higherActive Returns in R3, Non-U.S., and EM stock markets;3. Real-time portfolio statistically significant Active Returns can beachieved, but only by implementing large active stock weights inportfolios.5 4

F ig u re 1 : M C M P o r t f o lioCo n st r u ct io n a n d M o d e lin g P ro c e ss6

F ig u re 1 : M C M P o r t f o lioCo n st r u ct io n a n d M o d e lin g P ro c e ss7

Public Model Expected Returns Stock Selection Model, GLERTRt 1 a0 a1 EPt a2 BP t a3 CPt a4 SPt a5 REPt a6 RBPt a7 RCPt a8 RSPt a9 CTEFt a10 PMt et ,where: EP (1)[earnings per share]/[price per share] earnings-price ratio;BP [book value per share]/[price per share] book-price ratio;CP [cash flow per share]/[price per share] cash flow-price ratio;SP [net sales per share]/[price per share] sales-price ratio;REP [current EP ratio]/[average EP ratio over the past five years];RBP [current BP ratio]/[average BP ratio over the past five years];RCP [current CP ratio]/[average CP ratio over the past five years];RSP [current SP ratio]/[average SP ratio over the past five years];CTEF consensus earnings-per-share I/B/E/S forecast, revisions andbreadth;PM price momentum; ande randomly distributed error term.8

Public Model Expected Returns Stock Selection Model, GLERTable 1: Top/Bottom decile spreads of FSGLER1997 - 2011VariableTop 3 decilespreads (t)AnnualizedTop 3 decilespreadsTop decilespread (t)AnnualizedTop decilespreadEP0.42% (1.66)4.43%0.20% (0.66)1.40%BP0.50 (1.67)5.210.96 (1.45)9.54FEP0.50 (1.95)5.380.54 (2.02)5.11CTEF0.72 (5.85)8.851.16 (7.27)14.54EWC0.70 (3.67)8.371.06 (3.55)12.36GLER1.12 (4.54)13.551.48 (3.67)17.19Source: FactSet and APT. Past performance is not indicative of future returns. Guerard, J.B., Jr., Markowitz, H.M., & Xu, G. (2015).Earnings forecasting in a global stock selection model and efficient portfolio construction and management. International Journal ofForecasting, 31, 550-560.9

Public Model Expected Returns Stock Selection Model, GLERTable 1 cont.: Top/Bottom decile spreads of FSGLER2003 – 2011VariableTop 3 decilespreads (t)AnnualizedTop 3 decilespreadsTop decilespread (t)AnnualizedTop decilespreadEP0.23% (0.80)2.2670.25% (1.08)2.65BP0.80 (1.78)8.740.47 (1.57)5.21FEP0.38 (1.33)4.120.23 (1.39)2.66CTEF1.07 (6.48)3.370.58 (4.88)7.12EWC0.97 (5.79)12.070.60 (1.83)7.41GLER1.33 (3.98)16.320.96 (4.59)11.78Source: FactSet and APT. Past performance is not indicative of future returns. Guerard, J.B., Jr., Markowitz, H.M., & Xu, G. (2015). Earnings forecasting in a global stock selection model andefficient portfolio construction and management. International Journal of Forecasting, 31, 550-560.10

P u b l i c M o d e l R i s k P re f e re n c e s , 1 9 9 9 - 2 0 1 1Table 2: Efficient Frontier of the Global Stock Selection Model withVarious Portfolio Optimization Techniques1999 – 2011APT Risk ModelEarningsModel orComponentGLERBenchmarkMeanAnnualized Standard Sharpe InfoVarianceLambdaReturnDeviation Ratio 0 0.7813.1150016.3424.850.590 0.8212.0820016.3724.380.610 0.8512.6810015.9024.610.580 0.8112.66510.1119.360.440 0.518.815.590.240Source: FactSet and APT. Past performance is not indicative of future returns. Guerard, J.B., Jr., Markowitz, H.M., & Xu, G. (2015). Earnings forecasting in a global stock selection modeland efficient portfolio construction and management. International Journal of Forecasting, 31, 550-560.11

P u b l i c M o d e l R i s k P re f e re n c e s , 1 9 9 9 - 2 0 1 1Table 2 cont.: Efficient Frontier of the Global Stock Selection Modelwith Various Portfolio Optimization Techniques1999 – 2011APT Risk 800.850.800.4311.0710.8710.049.8412.18Source: FactSet and APT. Past performance is not indicative of future returns. Guerard, J.B., Jr., Markowitz, H.M., & Xu, G. (2015). Earnings forecasting in a global stock selectionmodel and efficient portfolio construction and management. International Journal of Forecasting, 31, 550-560.12

Table 3: Axioma Statistical Risk Model and OptimizerJanuary 2003 – December 2016STAT RiskModelTracking ErrorModel: XUS GLER 4.005.00 6.00 7.00 8.00 9.00 10.00Ann. Port Return13.18 14.13 14.47 15.22 15.80 15.95 15.95Ann. STD20.16 20.56 20.66 20.99 21.20 21.63 21.63Ann. Active Return5.336.296.627.377.948.108.10Ann. Active Risk6.227.147.557.969.168.648.64ShR0.573 0.605 0.619 0.645 0.668 0.662 0.662IR0.856 0.880 0.876 0.925 0.975 0.937 0.937Source: FactSet and Axioma. Past performance is not indicative of future returns.Guerard, J.B., Jr., & Chettiappan, S. (2017). Active quant: Applied investment analysis in emerging markets”, Journal of Investing 26, 138-152.13

Table 3 cont.: Axioma Statistical Risk Model and OptimizerJanuary 2003 – December 2016STAT RiskModelTracking ErrorModel: GL GLER4.005.00 6.00 7.00 8.00 9.00 10.00Ann. Port Return12.42 14.17 14.78 15.88 16.30 16.80 17.24Ann. STD17.82 18.92 19.60 19.98 20.12 20.54 20.64Ann. Active Return4.296.046.657.758.178.679.11Ann. Active Risk6.047.888.008.578.699.149.41ShR0.601 0.659 0.667 0.710 0.726 0.775 0.753IR0.710 0.852 0.832 0.905 0.940 0.949 0.968Source: FactSet and Axioma. Past performance is not indicative of future returns.Guerard, J.B., Jr., & Chettiappan, S. (2017). Active quant: Applied investment analysis in emerging markets”, Journal of Investing 26, 138-152.14

Table 3 cont.: Axioma Statistical Risk Model and OptimizerJanuary 2003 – December 2016STAT RiskModelTracking ErrorModel: EM GLER 4.005.00 6.00 7.00 8.00 9.00 10.00Ann. Port Return18.79 19.79 20.15 20.76 21.16 21.82 22.67Ann. STD26.09 26.22 26.34 26.45 26.74 26.95 27.25Ann. Active Return8.489.479.84 10.46 10.45 11.61 12.36Ann. Active Risk8.999.169.36ShR0.655 0.689 0.708 0.721 0.751 0.746 0.769IR0.944 1.033 1.062 1.095 1.085 1.128 1.1809.55 10.09 10.22 10.40Source: FactSet and Axioma. Past performance is not indicative of future returns.Guerard, J.B., Jr., & Chettiappan, S. (2017). Active quant: Applied investment analysis in emerging markets”, Journal of Investing 26, 138-152.15

F igu re 2 : Re t u r n v s. R e a liz e d Tr a ck in g E r ro rSource: FactSet and Axioma. Past performance is not indicative of future returns.Guerard, J.B., Jr., & Chettiappan, S. (2017). Active quant: Applied investmentanalysis in emerging markets”, Journal of Investing 26, 138-152.16

F igu re 3 : Re t u r n v s. Tr a c kin g Tr a ck ing E r ro rSource: FactSet and Axioma. Past performance is not indicative of future returns.Guerard, J.B., Jr., & Chettiappan, S. (2017). Active quant: Applied investmentanalysis in emerging markets”, Journal of Investing 26, 138-152.17

REGRESSION MODEL DEFINITIONSREG8 OIF99Regression(EP,BP,CP,SP,REP,RBP,RCP,RSP)REG9 EG10 M71)18

Table 4: Portfolio DashboardMean-Variance OptimizationAXIOMA Fundamental Risk Model1/2002 - -StatRiskTotal TransactionEffectEffectTotalEffectJAPAN REG8 8TE 0.56 0.54 5.64 4.00 -0.18 0.44 5.46 -2.243.26JAPAN REG8 6TE 0.54 0.51 4.73 3.74 -0.23 0.20 4.50 -1.493.04JAPAN REG8 4TE 0.50 0.46 2.75 2.88 0.170.63 2.92 -0.792.17JAPAN REG9 8TE 0.42 0.28 2.14 1.65 0.741.34 2.87 -2.000.951.59 0.55 1.45 2.09 -0.821.32JAPAN REG9 4TE 0.44 0.31 1.5419

Table 4 cont.: Portfolio DashboardMean-Variance OptimizationAXIOMA Fundamental Risk Model1/2002 - 11/2018InfoRatioRiskRiskStockRiskStock SpecificRiskFactorsSpecific Effect Factors 000 REG8 6TE0.69 0.30 2.52 2.07 -0.55 0.16 1.97-0.881.10R1000 REG8 8TE0.58 0.13 2.36 1.57 -1.40 -0.18 0.96-1.10-0.15R1000 REG8 4TE0.71 0.27 0.81 1.00 0.44 0.84 1.25-0.430.82R1000 REG9 8TE0.58 0.15 1.25 0.99 0.07 0.59 1.32-1.150.17R1000 CTEF 4TE0.75 0.44 0.63 0.89 1.53 2.34 2.16-0.691.4720

Table 4 cont.: Portfolio DashboardMean-Variance OptimizationAXIOMA Fundamental Risk Model – Non-U.S. 5000 Universe1/2002 - ransaction TotalEffectEffectPortfoliosSharpeRatioXUS CTEF 6TE0.85 1.45 9.35 6.89 5.58 6.80 14.93 -4.16XUS CTEF 8TE0.91 1.44 11.51 6.60 5.24 5.27 16.74 -4.98XUS CTEF 4TE0.70 1.26 6.40 5.77 4.66 7.71 11.05 -3.07XUS REG10 6TE0.65 1.21 4.16 3.53 7.23 6.53 11.39 -4.27XUS REG10 4TE0.49 1.07 2.66 3.09 4.67 5.85 7.33-3.0610.7711.767.987.124.2621

Table 4 cont.: Portfolio DashboardMean-Variance OptimizationAXIOMA Fundamental Risk Model1/2002 - RiskPortfoliosSharpeRatioEM REG8 4TE0.51 0.31 3.99 2.38 1.79 2.11 5.78-2.053.73EM REG9 4TE0.55 0.49 3.45 2.20 3.28 3.60 6.73-2.114.63EM CTEF 8TE0.59 0.61 3.54 2.18 5.76 3.64 9.30-3.585.72EM CTEF 4TE0.58 0.72 3.20 2.16 3.84 4.42 7.04-2.104.94EM CTEF 6TE0.60 0.65 3.67 2.13 4.67 4.02 8.34-2.755.59InfoRatioFactors EffectEffect T-StatTotalEffectTransactionEffectTotalEffect22

Table 5: CTEF and REG9 Portfolio Analysis using Axioma STAT andFUND Risk Models Mean-Variance OptimizationAXIOMA Fundamental Risk Model12/ 2002 - 11/ T-StatEM CTEF STAT 6TE1.92XUS CTEF STAT 2-3.043.583.893.664.243.878.13-4.773.37ACW CTEF STAT 6TE4.473.672.953.087.42-4.063.37XUS CTEF FUND 6TE3.633.503.983.837.61-4.583.02EM CTEF FUND 6TE1.901.933.824.045.72-2.862.86ACW CTEF FUND 6TE 4.133.322.312.816.43-3.802.63Portfolios23

Table 5 cont.: CTEF and REG9 Portfolio Analysis using Axioma STATand FUND Risk Models Mean-Variance OptimizationAXIOMA Fundamental Risk entumGrowthEarningsYieldPortfoliosDividendYield12/ 2002 - 11/ 2018EM CTEF STAT 6TE0.43 0.70 -0.01 2.18 0.14 0.67 0.78 -0.91XUS CTEF STAT 6TE0.09 0.70 -0.15 2.66 0.18 0.90 0.67 -1.78ACW CTEF STAT 6TE0.00 0.70 -0.09 2.75 0.05 1.27 0.91 -2.14XUS CTEF FUND 6TE0.08 0.73 -0.19 2.60 0.27 0.89 0.54 -1.88EM CTEF FUND 6TE0.35 0.67 -0.02 2.08 0.05 0.60 0.75 -1.01ACW CTEF FUND 6TE 0.04 0.66 -0.15 2.61 0.07 1.09 0.80 -2.0124

Table 5 cont.: CTEF and REG9 Portfolio Analysis using Axioma STATand FUND Risk Models Mean-Variance OptimizationAXIOMA Fundamental Risk Model12/ 2002 - 11/ T-StatEM REG9 STAT 6TE0.52EM REG9 FUND 6-3.181.780.600.943.553.394.15-2.861.29ACW REG9 STAT 6TE0.791.052.932.513.72-3.520.20XUS REG9 STAT 6TE0.050.313.022.943.06-3.98-0.92ACW REG9 FUND 6TE -0.31-0.102.352.332.04-3.19-1.15XUS REG9 FUND 6TE0.262.472.552.35-3.85-1.50Portfolios-0.1125