Equity Risk Factors for the Long and Short Run: Pricing and Performance at Different Frequencies

Transcription

1 Equiy Risk Facors for he Long and Shor Run: Pricing and Performance a Differen Frequencies Terri van der Zwan, Erik Hennink, Parick Tuijp 11h Financial Risks Inernaional Forum March 27h, 2018

2 Abou Orec Finance Orec Finance is a global provider of soluions for measuring and managing boh long- and shor-erm risk and reurn for insiuional and privae invesors Founded in 1981 Over 200 employees Offices in NL, UK, CH, CA, HK Independen: 100% owned by employees Mission is o improve financial decision-making and monioring using: Mahemaical models Informaion echnology Marke knowledge 1 / 13

3 Inroducion Discouned fuure dividends Macroeconomic facor model CRR model (Chen, Roll and Ross, 1986) Advanage: clear inerpreaion Firm-specific characerisics Fama-French facor model 6FF model (Fama and French, 2016) Advanage: good performance This research: Insigh in relevan risk facors Comparing wo ypes of mulifacor models a differen frequencies 2 / 13

4 Conribuion Merging specral analysis and mulifacor models This research: The cross-secion of US equiy porfolio reurns a differen frequencies Monhly daa Aggregaed daa Annual Three-year Filered daa using frequency decomposiion and specral analysis Shor-erm: exracing flucuaions > 12, 18 monhs Mid-erm: exracing flucuaions > 32, 36 monhs Long-erm: exracing flucuaions > 5, 8, 16 years 3 / 13

5 Daa Shor daa se ranges from July 1964 unil December 2016 Long daa se ranges from January 1927 unil December size-value sored porfolios of French s websie 30 indusry porfolios 40 mixed porfolios (10 size, 10 book-o-marke, 10 indusry and 10 momenum equally weighed) 4 / 13

6 Frequency decomposiion Frequency decomposiion using Chrisiano-Fizgerald Random Walk filer Figure: Filered log index indusrial producion Original Low frequencies High frequencies / 13

7 Frequency decomposiion (a) Low frequencies (b) High frequencies / 13

8 Cross-secion CRR model r i, = α i + β IP,i x IP β T S,i x T S + β UI,i x UI + β DY,i x DY + β EI,i x EI + ε i, + β RP,i x RP + 6FF model r i, = α i + β m,i r m + β s,i r SMB β c,i r CMA + β h,i r HML + β w,i r W ML + ε i, + β r,i r RMW + 7 / 13

9 Cross-secion of equiy reurns Idenify risk facors of equiy reurns using Fama-MacBeh procedure r i, = λ 0, + j λ j, βj,i,1 + ɛ i,, for every ime, where β j,i,1 is he esimaed exposure for window 1. Tes for average λ j. 8 / 13

10 Cross-secion monhly frequency For monhly frequency Table: Monhly cross-secional resuls (expanding window exposures) λ 0 λ IP λ UI λ EI λ RP λ T S λ DY R 2 λ 0 λ m λ SMB λ HML λ RMW λ CMA λ W ML R 2

11 Cross-secion mid-erm frequency Fama-French ouperforms macro, Table: λ j Pass band λ 0 λ IP λ UI λ EI λ RP λ T S λ DY R 2 (0, π/ 6] Three-year (0, π/ 18] Pass band λ 0 λ m λ SMB λ HML λ RMW λ CMA λ MOM R 2 (0, π/ 6] Three-year (0, π/ 18]

12 Cross-secion long-erm frequency Varying λ j over horizons, value facor mos imporan Table: (expanding window exposures) Frequency λ 0 λ IP λ UI λ EI λ RP λ T S λ DY R 2 Frequency λ 0 λ m λ SMB λ HML λ RMW λ CMA λ MOM R 2

13 Summary resuls Figure: Average R-squared for differen frequencies. R 2 increasing over frequencies Fama-French models higher R 2 han macroeconomic models R CRR 6FF 4CRR 4FF Flucuaions (monhs) 12 / 13

14 Key findings Fama-French facor models are more suiable o explain he cross-secion of equiy reurns Differen resuls for aggregaed and filered daa: differen signs in, for example, marke risk premia The average effec/price of risk varies over ime and horizon in magniude Indusrial producion, he credi spread and he erm spread end o play a bigger role on lower frequencies The value facor seems o be he mos imporan facor and he size facor he leas on all frequencies The average cross-secional R 2 increases over frequencies for boh macroeconomic and Fama-French facor models Differen facors are relevan o explain equiy reurns a differen frequencies 13 / 13

15 Conac Terri van der Zwan Researcher Roerdam Orec Finance bv Boompjes XB Roerdam The Neherlands Tel. +31 (0) Amserdam Orec Finance bv Nariaweg BP Amserdam The Neherlands Tel. +31 (0) London Orec Finance Ld Suie 9.10, Ciy Tower 40 Basinghall Sree London,EC2V 5DE Unied Kingdom Tel. +44 (0) Pfäffikon Orec Finance AG Possrasse Pfäffikon SZ Swizerland Tel. +41 (0) Torono Orec Finance Canada Inc 250 Universiy Avenue Torono, ON M5H 3E5 Canada Tel. +1 (0)

16 Disclaimer Please noe ha his repor has been prepared wih care using he bes available daa. This repor may conain informaion provided by hird paries or derived from hird pary daa and/or daa ha may have been caegorized or oherwise repored based upon clien direcion. This repor is no inended as an invesmen advice. Orec Finance assumes no responsibiliy for he accuracy, imeliness or compleeness of any such informaion. Orec Finance acceps no liabiliy for he consequences of invesmen decisions made in relaion on informaion in his repor. All our services and aciviies are governed by our general erms and condiions which may be consuled on and shall be forwarded free of charge upon reques. 2