Perspective on Risk 1/24/22
Blockchain and DeFi; How Complex Systems Fail; A Twist on Factors Driving Equity Premia; Is VC in a Bubble? How to Increase Risk-Taking; Improving Judgments of Existential Risk; & lots more
OK, big post; I’m back in the game. Lot’s of interesting stuff.
Blockchain and DeFi
Moodys’ Analytics and Gauntlet have written a primer: Block by Block: Assessing Risk in Decentralized Finance
DeFi is a catch-all term referring to the range of financial services that exist on public blockchains that mirror the kinds of services that exist in the traditional financial system: borrowing, lending, asset creation, and more. It uses smart contracts to eliminate the need for a trusted intermediary to facilitate transactions, allowing anyone to transact on these protocols, which in this jargon, refer to the ecosystem centered around decentralized applications like smart contracts or a grouping of them that mirror traditional financial functions. Some current prominent examples include automated market making, lending protocols, and option exchanges.
The paper discusses several interesting risks associated with the current DeFi platform structure: adverse selection, principal/agent issues, liquidity and collateral risks.
DeFi seems yet another push towards collateralized lending (DeFi lending transactions are currently heavily collateralized) and a further challenge to the existing Merton-based view of lending.
This is a worthwhile read for those who want to understand DeFi and how it changes traditional intermediation risks.
How complex systems fail
I know a number of my friends are fans of Normal Accidents, Ron Perrow’s book on operational disasters. That work posits that system risk is related to the degree of interconnectedness, and the tightness of process coupling.
Richard Cook has produced a short three page monograph How Complex Systems Fail that is worth a skim or read. He argues that complex systems are composed of equipment and people, and that they operate safely in a permanently “degraded mode.”
Overt catastrophic failure occurs when small, apparently innocuous failures join to create opportunity for a systemic accident. Each of these small failures is necessary to cause catastrophe but only the combination is sufficient to permit failure. Put another way, there are many more failure opportunities than overt system accidents. Most initial failure trajectories are blocked by designed system safety components. Trajectories that reach the operational level are mostly blocked, usually by practitioners,
The complexity of these systems makes it impossible for them to run without multiple flaws being present. Because these are individually insufficient to cause failure they are regarded as minor factors during operations. Eradication of all latent failures is limited primarily by economic cost but also because it is difficult before the fact to see how such failures might contribute to an accident.
A corollary to the preceding point is that complex systems run as broken systems. The system continues to function because it contains so many redundancies and because people can make it function, despite the presence of many flaws.
An Interesting Twist on Factors Driving Equity Premia
Hat tip to @theRealKiyosaki for highlighting What Do the Portfolios of Individual Investors Reveal About the Cross-Section of Equity Returns?
This paper constructs a parsimonious set of equity factors from the cross-section of individual investor portfolio holdings. We show theoretically that portfolios of stocks sorted by the age or wealth of their individual investors should produce powerful pricing factors. Using the complete stockholdings of Norwegian retail investors, we verify empirically that a threefactor model consisting of a mature-minus-young factor, a high wealth-minus-low wealth factor, and the market factor price the cross-section of stock returns. Our three factors span the size, value, investment, profitability, and momentum factors and perform strongly in out-of-sample tests. We also uncover a rich set of links between investor characteristics and portfolio tilts toward the age and wealth factors.
In a twist on Merton, they identify factors that pertain to the investors, controlling for the traditional Merton factors that apply to companies. They find significance in investor age and wealth levels. Older and wealthier investors have better Sharpe portfolio, and this appears to derive from less turnover and a lower beta and volatility portfolio.
My factor geeks, have fun.
Is VC in a bubble?
One risk factor that you learn about as a bank examiner is to look where there is massive growth in volume. That can often lead to weaker underwriting and operational issues.
CBInsights has published State Of Venture GLOBAL | 2021. In it, they state:
Global venture funding reached a record $621B in 2021, more than double the 2020 mark of $294B. Q4’21 closed the year with an all-time high of $176B in funding – the 6th straight quarter of growth.
The global unicorn count hit 959 in 2021 — up 69% from 2020 — with the year seeing a staggering 517 unicorn births. The increase was driven primarily by rapidly rising valuations at late-stage deals. Today, there are 44 startups with decacorn ($10B+) valuations.
$1 out of every $5 goes to fintech. Fintech startups raised $132B in funding in 2021 – accounting for 21% of all venture dollars. Fintech funding is up 169% compared to 2020’s $49B. In addition, 1 in every 4 unicorns is in fintech — the most by far of any industry.
Lots of good detail for those interested.
How to Increase Risk-Taking
In Emotions and Risk Attitudes, Armando Meier finds that happiness and anger lead to an increase in willingness to take risks, while fear leads to a reduction in willingness to take risks.
Improving Judgments of Existential Risk: Better Forecasts, Questions, Explanations, Policies
I’m a Phil Tetlock fanboy, ever since Danny Kahneman touted his research. Phil is perhaps the leading academic researching how to make better decisions. His latest paper, Improving Judgments of Existential Risk: Better Forecasts, Questions, Explanations, Policies, tries to extend forecasting to longer horizon, “existential” issues. Can methods raise predictive accuracy above chance or, harder still, extrapolation algorithms?
Forecasting tournaments are misaligned with the goal of producing actionable forecasts of existential risk, an extreme-stakes domain with slow accuracy feedback and elusive proxies for long-run outcomes. We show how to improve alignment by measuring facets of human judgment that play central roles in policy debates but have long been dismissed as unmeasurable. The key is supplementing traditional objective accuracy metrics with intersubjective metrics that test forecasters' skill at predicting other forecasters' judgments on topics that resist objective scoring, such as long-range scenarios, probativeness of questions, insightfulness of explanations, and impactfulness of risk-mitigation options. We focus on the value of Reciprocal Scoring, an intersubjective method grounded in micro-economic research that challenges top forecasters to predict each other's judgments. Even if cumulative information gains prove modest and are confined to a 1-to-5 year planning horizon, the expected value of lives saved would be massive.
It lays out the challenges associated with such forecasts, and proposes statistical and other techniques to improve long run forecast accuracy.
This paper will be of most interest to those interacting in policy-development circles.
Extreme Risk for Insurers
This Willis-Towers-Watson piece covers 2019-20, but I don’t think much has changed. From this piece, one can see why Global Temperature Change is so high on policy-makers agenda: Highly Likely (within 20 years), quite certain, crushing and pan-generational. None of the others rise to this level.
I might argue that, in the tail, quite a few of these are correlated, with Global Temperature Rise as the causal factor. For instance, warming leads to a food crisis, which leads to emigration, which then affects political extremism.
Buying Canadian farmland for the (yet-to-be-born) grandkids.
When investors fear ambiguity,
… they want higher premia on instruments that pay less in bad times. This is the conclusion, I think, from a BIS paper on state-contingent debt: The premia on state-contingent sovereign debt instruments. This, of course, makes sense. We know that when PD systematically rises due to overall economic performance, for many individual borrowers, PD is highly correlated to LGD. The main findings are:
First, the risk premia in state-contingent instruments are high and persistent.
Second, the premia exhibit a pro-cyclical pattern.
Third, the liquidity premium is higher and more volatile than that for plain-vanilla government bonds issued by the same sovereign
I think there is further analogies and insights here, but I’m not sure I’m grasping them at the moment.
Robots vs Humans
It has become clear that in many endeavors machines out-perform humans. They can detect subtle patterns, and they, if constructed properly, can avoid the cognitive constraints, biases and behavioral traps we humans fall into.
In this light, I came across this interesting paper: Human Versus Machine: A Comparison of Robo-Analyst and Traditional Research Analyst Investment Recommendations.
Abstract: … Our results indicate that Robo-Analyst recommendations differ from those produced by traditional “human” research analysts across several important dimensions.
First, Robo-Analysts produce a more balanced distribution of buy, hold, and sell recommendations than do human analysts and are less likely to recommend “glamour” stocks and firms with prospective investment banking business.
Second, automation allows Robo-Analysts to revise their recommendations more frequently than human analysts and incorporate information from complex periodic filings.
Third, while Robo-Analysts’ recommendations exhibit weak short-window return reactions, they have long-term investment value. Specifically, portfolios formed based on the buy recommendations of Robo-Analysts significantly outperform those of human analysts.
Portfolios formed from Robo-Analyst buy recommendations earn abnormal returns that are statistically and economically significant and are between 4 to 5 percentage points higher than human analysts when annualized. … For sell recommendations, however, we find no evidence to indicate that Robo-Analysts’ recommendations are differentially profitable.
Consistent with Robo-Analysts having less cognitive bias and fewer economic incentives for optimistic reports, we find that RoboAnalysts issue a more balanced set of recommendations (less optimistic), revise more frequently, and rely more on large, complex volumes of disclosure in forming their recommendations. They are also quicker to downgrade buys and are less likely to recommend glamour stocks, which tend to underperform in the long-run.
The Perfect Curse Word
We evaluated prediction of tabooness of single words and novel taboo compound words from a combination of phonological, lexical, and semantic variables (e.g., semantic category, word length). For single words, physiological arousal and emotional valence strongly predicted tabooness with additional moderating contributions from form (phonology) and meaning (semantic category). In Experiment 2, raters judged plausibility for combinations of common nouns with taboo words to form novel taboo compounds (e.g., shitgibbon). A mixture of formal (e.g., ratio of stop consonants, length) and semantic variables (e.g., ± receptacle, ± profession) predicted the quality of novel taboo compounding. Together, these studies provide complementary evidence for interactions between word form and meaning and an algorithmic prediction of tabooness in American English.
They find the key is combining a standard swear & an upsetting non-swear ending.
Podcasts
The Journal: The Stock Trading Scandals at the Federal Reserve (17:47)
This is a good summary on what went on with Caplan, Rosengren and, in particular, Clarida.