Ubs Group Ag · SIX:UBSG
Ubs — UBS Quant Insight Series webinar (January 2021)
15 slides from Ubs's conference presentation, January 2021
Slide 1: Real economy portfolio
Gerry GarveyG Systematic Active Equity January, 2021 Real economy portfolio
Slide 2: Disclaimer
Disclaimer This presentation was produced solely by Gerald Garvey. The opinions and statements expressed herein are those of Gerald Garvey and are not necessarily the opinions of any other entity, including UBS AG and its affiliates. UBS AG and its affiliates accept no responsibility whatsoever for
Slide 3: Overview
Overview • 2020 Disconnect of Economy and Equities: Mayers and Roll wrote about about this long ago – Conclusion: Even MSCI ACWI is not the true market portfolio • What can we do about it? – V1: Don’t expect too close of a connection – V2: try to make portfolios that get closer to the theoretical id
Slide 4: Origin
Origin: • GE on the Goodradigbee, 1994 • Fisher and Marshall (2011). Make extended IO table into a GE model that links shocks to rewards to different types of capital and labor • Hey! Our investable universes only contains claims to a subset of the capital. LIMITED DISTRIBUTION 4
Slide 5: Our key raw input: IO table from BEA or OECD or WIOD
Our key raw input: IO table from BEA or OECD or WIOD • Here is the coarsest US BEA table, 2019 • Existing finance research uses only the square portion with a naïve momentum thesis • Last three rows turn our focus to the division of surplus within the industry • Inputs from other industries are just
Slide 6: The table we use in practice
The table we use in practice US 2016 from WIO.org rolled up to seven sectors • Sectors have at least 20 listed firms in US allcap; also can be done for Japan and China • Map directly into old-school SIC codes, GICS not so much (see appendix slide) • “LAB_XS” denotes educational level of labor: model
Slide 7: The big pivot: Leontief Matrix
The big pivot: Leontief Matrix • IO tables show only direct, first-order input requirement; “Leontief inverse” sums up all direct and indirect input requirements • Here’s the inputs required to produce a unit of output for each sector • Industries with similar labor inputs effectively compete along
Slide 8: What REP does
What REP does • Think of economic dynamics as shocks to industries or some subset thereof. • If we could pinpoint shocks we have thematic portfolios • Here we try to be completely agnostic to shocks. Get unit exposure to them all. That’s the equivalent of “Market Portfolio” in CAPM LIMITED DISTRIBUT
Slide 9: SAE's experience
SAE’s experience • Used in globals since 2014 • Ignores countries, so over-weights those with more service and consumer firms • Cap-weighted within industry, to avoid a big “Fundamental Index” footprint • If you go truly global (include EM), – Overweights are US, Mexico, Japan, Thailand – Underweigh
Slide 10: Some names: Cheap expression with ETFs is GICS Staples + Tech
Some names: Cheap expression with ETF’s is GICS Staples + Tech OVERWEIGHT UNDERWEIGHT • Amazon • Saudi Aramco, Exxon • Microsoft, Alphabet, Facebook • JP Morgan, Wells Fargo • Walmart, Home Depot, Kering, Diageo • Visa, Mastercard • Altria, BAT • Roche, Merck, Pfizer • AT&T, Verizon (REP has a very
Slide 11: Some US results in academese
Some US results in academese (using FF data so anyone can replicate) LIMITED DISTRIBUTION 11
Slide 12: Old school performance regressions
Old school performance regressions Variable Full Sample WIO data available Post Research (starts 1973) (starts 1995) (starts 2014 Alpha 0.000715 0.00101 0.00823 (2.33) (2.76) (2.21) MKT 0.0013 0.032 0.038 (0.187) (1.03) (0.98) SMB 0.11 0.038 -0.0133 (10.31) (2.89) (1.81) HML -0.034 -0.019 -0.028 (2.
Slide 13: Is it just mean-reversion in weights?
Is it just mean-reversion in weights? Here is EW 10 sectors, cap weight within sectors LIMITED DISTRIBUTION 13
Slide 14: Next steps
Next steps • Better ancillaries? Current ones are just ok • REP’s dividends or earning are only a bit more correlated with one-year macro (GDP, consumption). • Active positions quite strongly forecast increases in book (retained earnings plus net issuance plus IPOs) • Multi-asset • How to add PE? Th
Slide 15: Appendix 1: REP vs GICS sectors
Appendix 1: REP vs GICS sectors • R_Buildings • A slice of Financial • Construction slice of Industrials • R_Consumer • Staples, some Discretionary • Also some Industrial and Tech • R_Finance • Similar to MSCI • R_Manufacturing • Industrials, Tech, Materials • Discretionary too, e.g. auto parts • R_
UBS Group — UBS Quant Insight Series webinar (January 2021) on 6ix