Entropy and sustainable investment beliefs: A simple metric for - - PowerPoint PPT Presentation

Entropy and sustainable investment beliefs: A simple metric for analysing belief organisation

Dane Rook University of Oxford D.Phil. Researcher, Geography & the Environment dane.p.rook@ouce.ox.ac.uk

[How] can investors make long‐term decisions that are more sustainable, more responsible, and maximise risk‐adjusted returns?

Expectations drive decisions…

• Financial decisions (sustainable or otherwise) stem

from expectations

• Expectations blend the mathematical and psychological

(homo economicus meets ‘animal spirits’); balance shifts with:

– Individual preference – Context – Timeframe – Degree of uncertainty/know‐ability

• Should desire that expectations be:

– Accurate – Precise*

…and beliefs [partly] drive expectations…

• Investment expectations emanate (at least partly)

from how investors believe markets function and price paths are generated

• Investment beliefs are :

“…conjectures and working assumptions about the investment world (including the economy, the workings of the financial system, and social, environmental and other risks) that underlie and inform investment decision‐making” (Woods and Urwin, 2010: p. 7)

• Any investment decision is therefore an (explicit
• r implicit) expression of (possible combination
• f) investment belief(s)

How to judge investment belief(s)

• Beliefs + Context  Expectations  Decisions
• How to judge the quality of beliefs?

Clues from Foxes and Hedgehogs (Tetlock, 2005):

– Accuracy (Correspondence) – Precision (Coherence)*

• The plural: (eco)systems and networks of beliefs:

– Dynamism – Interaction*

How to judge investment belief(s)

• Predictive validity

– Robustness across contexts – Change in composition

• Coherence measures

– Subjective/qualitative

• Gedankenexperiment
• Scenario analysis

– Objective/quantitative*

• Beyond correlation?

Sustainable investment beliefs sampler

A: Environmental, social, and corporate governance (ESG) factors will impact short‐term investment risks and returns (i.e. less than one year) C: The benefits of incorporating sustainable investment principles into the investment process are unlikely to outweigh the costs of doing so B: ESG factors will impact long‐term investment risks and returns (i.e. more than three years) E: Investors are over‐sensitive to short‐term factors and not sensitive to long‐ term factors K: Companies can gain significant competitive advantage through their strategic response to climate change L: Investors can successfully incorporate carbon risk into portfolio management P: Countries (and governments) can gain significant competitive advantage in their economies through their strategic response to resource scarcity Q: Companies can gain significant competitive advantage through their strategic response to climate change

Belief interactions

Organising principles of beliefs

• Converse’s (1964) constraint:

“…the success we would have in predicting, given initial knowledge that an individual holds a specific attitude, that he holds certain further ideas and attitudes.”

• Informational content (order versus entropy)
• The internal (logical) and the external (social)

Organising principles of beliefs

• Martin’s (2000) constraint:

“…the inverse of the degree of arbitrary movement in the space of all possible beliefs”

• The negative of entropy (amount of order)
• Overall constraint =

Tightness (logical constraint) + Consensus (social constraint)

• Social constraints imply that belief holders:

“…frame their understanding…in the same way” (Baldassarri and Goldberg, 2010: p. 5)

Belief interactions redux

High Consensus, High Tightness Low Consensus, High Tightness High Consensus, Low Tightness Low Consensus, Low Tightness

The pith of it all…

• Tightness:

– How one’s beliefs fit together

• Consensus:

– Homogeneity of conceptualisation of ESG topics in financial markets? – Market distribution of beliefs

• Belief ‘densities’
• Implications for investors and policymakers!

– If only we could measure it…

Measuring belief networks

• Informational entropy (Shannon’s entropy):

– And the sign said [small samples] need not apply! [with apologies to the 5 man electrical band]

• Thermodynamic entropy from statistical physics suits

smaller samples and can provide some preliminary intuitions:

LN(W / (M^N)) where: W = N! * (N1! * N2! * … * NM!)^‐1 M is possible belief positions N is the number of overall respondents/believers Ni is the count of believers assuming a belief position

Entropy in action: A preliminary test‐drive

• Unique dataset of asset manager beliefs
• Sample size of 22 respondents, with 18 beliefs

examined across 3 categories:

– Sustainable Investment – Climate Science – Resource Scarcity

• Test of the entropy (thermodynamic version)

measure on various belief combinations

Entropy in action: Some early findings

• Belief networks become more ‘disorganised’

(random) with higher dimensionality

• Some beliefs appear to dominate others and

exhibit an enduring influence as other beliefs are incorporated/removed

– ‘Core’ beliefs?

• Possibility of ‘gestalt’ or ‘emergent’ properties
• f belief configurations

Extensions: Why this matters + where to go next

• Help practitioners understand their own

beliefs on financial markets and ESG factors through more objective and ‘periscopic’ methods of analysis

– Particularly CORE beliefs

• Strategies for change and opportunity
• Large‐sample analysis and

tightness/consensus decompositions…

• WE NEED MORE (GOOD) BELIEF DATA!!!