Finance: An Industry Based on Psychology
By Riva-Melissa Tez (firstname.lastname@example.org)
Edited by Peter B. Clark of Kalytix Partners
- Holding Venture Capital in the Wrong Regard
- Testing Old Assumptions in Venture Capital
- LPs Need to Not Outsource by Necessity
- Why Investment is a Psychological Game
- Investment from First Principles- the Need for Systematic Game-Changers..
- … Means Investing in Improving Scientific Research and Artificial Intelligence
- Aids for Playing the Psychological Game of Finance
- Epistemological Tools to Help Find the Right Problems
- Design Thinking Applied to Mission Orientated Investment Decisions
- More Research into Tools to Improve Decision Processes in Finance (Bayesian Mapping)
- Creating a Bayesian Causal Map for Investment Decision Making
- Fallible Minds Equals Fallible Approaches to Humanity’s Future
1) Holding Venture Capital in the Wrong Regard
A year ago I shared a common outlook about investors. We attribute some sort of unusual mystery to them, a direct affinity to power and success. They are our contemporary decision makers, often holding more power than politicians in shaping our current world. Acting as a sort of Charon figure, they steer the boat that enables the lucky CEO contenders to cross the River Styx of the start-up ecosystem. The CEOs don their start-up t-shirts and jump up and down all over Silicon Valley, degrading themselves and their companies by sycophantically doing whatever it takes to win over a potential investor. If you ever want a true human-safari experience, head over to a pitch competition. It’s much more exciting than The X-Factor.
Our current mentality is to regard VC as the all-powerful asset class of investors who define the future of start-ups and the return of capital to their Limited Partners (LPs). LPs hand over capital on poor terms, blinded by narrative fallacies of the fund potential. Acknowledging the wrong problem, there’s been much press of the last decade to say that the VC model is ‘broken’, but the fundamental issue is the low quality of LP’s decision making. Let’s think rationally: typically a General Partner commits only 1 percent of partner dollars to a new fund while LPs commit 99 percent. LPs shoulder the risk of failure.
LPs are getting a bad deal. Start-ups are getting a bad deal. What’s worse is that the future of humanity is getting the worst deal. There isn’t enough research on the fundamentals of the entire industry to find the pinch points or the potential catalysts to re-boost and re-shape finance for the upcoming eras. There hasn’t been a great change in funding and asset classes even as the technological world has dramatically changed. We are still holding onto old financial models that struggle to support innovation. LPs deserve the same kind of level of insight into the funds they contribute to as VCs do with due diligence for potential portfolio companies. It is common for institutional investors to make investments in VC funds without requiring information about general partner compensation, carry structure, ownership, firm-level income, expenses, or profits. What’s even more horrifying is the LP ‘bucket’ style attitude- to say that LPs are making weak decisions implies that they even think enough about the decision process at all.
2) Testing old assumptions in Venture Capital
The news has taught us how easy it is for individuals to dip into the start-up ecosystem and leave with large financial rewards. We read stories of investors who fund early-tech and get 5x, 10x or even 20x valuation only a few years down the line. In reality, these successes are rare. The Kauffman Foundation published a 52-page paper in 2012 that detailed the metrics of the foundation’s investments into VC funds. The Kauffman Foundation, created to encourage entrepreneurship, has more access to top –tier funds than pretty much anyone else. Having invested in over 100 such funds over the last 20 years, the datasets provided in the 2012 paper gave a viewpoint of VC returns from the last two decades from an LP perspective.
The Kaufmann Foundation did a deep analysis of their VC portfolio and claimed that the research ‘challenged our long-held assumptions about investing in venture capital and forced us to acknowledge that most of what we believed to be true is not supported by data from our more than twenty years of investing experience.’
The current market standard 2 percent management fee and 20 percent profit-sharing structure (“2 and 20”) pays VCs more for raising bigger funds and, in many cases, allows them to lock in high levels of fee-based personal income regardless of fund performance. The fact that LPs support such a misalignment of goals shows how irrational LP capital deployment has become.
The other point to make is that it’s nearly impossible for a traditional VC fund to truly think long-term, even if they claim that as a differentiating factor. A 2 and 20 model over 10 years will find it hard to support long-term innovation. We don’t have good structures that create the right short-term incentives to support long-term projects. VC term sheets promise to return money back to LPs within rigid time frames. A way to overcome these limitations would be for a new VC fund to strategically pick time-aligned LPs. Representatives of a Sovereign Wealth Fund will want to demonstrate returns on their decisions within their career at the fund, whilst large multi-generational family offices are often more accepting of long-term structures that could generate later rewards and also the potential for legacy.
Something that compounded my concern was to find that VCs were often ‘selling’ their advisory board seats to the largest LP commitments. Does it make sense to give some power back to LPs, especially if they are neither goal nor time aligned with the philosophy of the fund?
3) LPs Need to Not Outsource by Necessity
Glancing over memorandums from new and old VC funds, nearly all claim the desire to ‘change the world’ whilst promising ‘exceptional returns’. In reality, there are very few funds who actively engage in investing in risky yet crucial companies and ideas. The challenge of recognizing and valuable ‘real’ innovation lies with a subset of maybe 2–3 funds willing to demonstrate with their portfolio companies that they actively act on their manifestos. Those funds often remain oversubscribed and a challenge for many LPs to buy into. For the LPs unavailable to gain access to these funds, in particular family offices, we need to crush the belief that deploying capital to VC is always preferential. Instead these LPs need to consider the possibility of improving their decision-making skills, hiring people with industry knowledge, and doing something in-house.
We asked the CEO of one University Management Company why they didn’t just build their own VC fund in-house instead of using external managers, and whether the numerical value of the management fees spent on the outsourcing of capital deployment is worthwhile in terms of the value of risk-mitigation incurred from not doing it in-house. His reply shocked me. We’re not Venture Capitalists he said. We’d much rather write the cheques for other people to manage the decisions. It would be hubristic to think we can do it ourselves, and I want to avoid being in the newspaper. The best, however, was his belief that ‘we’re too small to hire the right people.’ (As of 2014, the fund has $25bn of assets under management.)
It seems absurd how few Limited Partner groups look into solving their supposed self-acclaimed limitations. They doubt their decision-making process and outsource the capital to VC, yet don’t work on improving their decision-making and framing processes, or try and build their own internal networks and team.
The value of psychology is overlooked. The more we come to understand finance, the more we realize at every level the power of psychology. And it goes even further, if we’re to acknowledge to whatever degree that minds move markets and that markets move minds then, by investing into understanding minds, we learn how to move markets better.
4) Investment: Behaviorism/ Psychology
We’re going to claim that understanding the mind is one of the most undervalued financial strategies. This is not a claim just in terms of market success. By understanding the psychological traits around good decision making, investors could correlate these with investment successes and failures, and develop tools and technology to improve the methods used.
Researching into how minds work would only go on to give us better data points to create more valuable companies. We can go even further with this. Through the A/B testing of these sorts of products and companies, we could create an iterative process whereby by knowledge would grow deeper. There could be no such thing as non-profit psychological research, if implemented correctly. Maybe a psychology-motivated investment fund already exists. The potential power of such an entity, if carried out correctly, could be huge.
5) Investment from First Principles- the Need for Systematic Game Changers….
Some investors are mission orientated. Lets say there is a family office or high-net-worth individual who recognizes there are huge problems in healthcare. Let’s compare the healthcare industry to a broken wall. Now some insightful investors will not only see the broken wall, but will also want to play a part in rebuilding it. Instead of looking at the broken wall (and, by that I mean studying the shape of all the individual bricks,) that investor would regard the entire architecture. Most investors throw money at some supposed messianic entrepreneur, and that entrepreneur approaches the broken wall with spackle putty (polyfiller), instead of a new blueprint and cement. Predictably, since putty is not designed for fixing walls, inevitably the latter falls and we’re left with either the same problem or, even worse, a greater problem due to the putty. Systematic change is required.
Instead of investing in people offering solutions, we should be investing in systematic game-changers — especially things that make it easier/cheaper/faster to test out hypotheses. A lot of start-ups and charities claim to be systematic game-changers, but if the company is not enabling others then it’s not a systematic player. Sometimes looking for these companies takes us out of the industry that we’re originally trying to fix. Massive advances in artificial intelligence would create a catalyst for nearly all industries. Investors need to reverse engineer the thing they’re trying to fix (if at all) until they find the true pinch-points and catalysts. The same goes for philanthropy. We throw money at superficial nonsense. Should we be giving dollars to a charity helping gangs in New York or into neurotechnology research? It seems that most charitable donations work as putty on broken walls. No-one looks for the systematic game-changers, because our psychological set-up restrains us from going meta enough. Ironically, we would have fewer worries about gangs in New York if we also able to understand how humans thought. We should be looking for the things that truly create change. Once we start testing our assumptions on what the answer is, and seperating that answer into core, fundamentals components, we become scientists and need scientific research. The benefits incurred from the reincentivizing of science is substantially larger than the benefits of helping the small subset of people who are directly targeted by charities. We could go as far to say that actually most charities (and start-ups) are actually contributing to problems. By taking on the superficial, we distract from the underlying causes. We tackle the superficial because we aren’t mentally auditing our biases out enough.
6) …. Means Investing in Improving Scientific Research and Artificial Intelligence
One of the greatest yet under-looked problems faced by humanity in our current time-slice is that we allow ourselves and our loved ones to age and die. The cells that make up our children, the DNA of our parents, lie on unknown paths, within unknown frameworks, unregulated. Yet if we reverse engineer the issues around ‘aging’, we learn how the bottlenecks comes from our lack of understanding of the make-up of the biological systems that define us. Here’s an analogy. If we had tried to create a megafund in 1860 to attempt to reach the moon, we would have ended up investing in balloons and kites and not rockets. We would have to first understand thermodynamics and manufacturing, which would derisk the entire endeavour of aviation. We can draw similiar parallels to current investments in healthcare. What analogous areas should we be investing in to derisk the entire endeavour of extending lives? Approaching these sort of issues from first principles often takes us down the path of realizing that small improvements into the structure of scientific discovery or in the field of artificial intelligence may create a bigger impact than more directly-related ideas.
Looking out over the next 30 years, we’re expecting a world of deeper interconnectedness, as well as the development of artificial intelligence and (hopefully) huge advancements in healthcare. The more we work with others to look at the technological horizon, building overviews whilst continously testing for assumptions, and reversing the process to the bottleneck, we nearly always get to basic problems in the current structure of the advancement science. The technological, commercial, biological and societal value created from realms like artificial intelligence and neuroscience stand to create some of the biggest paradigm shifts in human history. If we can create a 20% acceleration on the growth curves of these industries, the effects are dramatic. A greater investment allocation needs to be given to companies or ideas that cause scientific discoveries to be leaner, quicker or cheaper. Advancements in machine-learning can have huge knock-on effects to other industries currently relying on human intellectual capital.
(The Serious Bit)
1) Aids for Playing the Psychological Game of Finance
Part 2 of this article will demonstrate ideas for techniques that would make explicit salient features of decision makers’ mental models and inference processes in a bid to contribute towards a public toolbox for investment decisions. The financial industry continues to avoid tracking the psychological traits around decision-making, ignoring the potential value of correlating an investor’s successes/failures and the development of tools to self-improve.
If we want to build the future then we need mission-orientated, self-improving investors, looking into innovative financial models that fund pinch-points and catalysts that create systematic changes. This is by no means an easy request, but if we want to truly improve the world, then the call to action is necessary.
Talking and working with Limited Partner groups- in particular Family Offices- the more we realized that they genuinely want to create a valuable legacy that contributes to the lives of future generations. Venture capital promises to be the asset class to fulfill those dreams. Family offices need to equip themselves with the psychological tools to work out if/when they should trust that the VC has the ability to even find the right problems.
2) Psychological Tools to Help Find the Right Problems
Imagine we find a systematic problem. How do we go on to solve it? There’s a multitude of psychological and ideation tools that investors don’t use. For those investors who are specifically mission-orientated in tackling global problems, processes such as Design Thinking are particularly useful for addressing so-called “wicked” problems. ‘Wicked’ means that in some-manner these problems are ill-defined or tricky. Solving and funding global programs is no doubt an ill-defined area that’s tricky. For ill-defined problems, both the problem and the solution are unknown at the outset of the problem-solving process (as opposed to “tame” or “well-defined” problems, where the problem is self-evident and the solution is possible with some technical knowledge.) Even when the general direction of the problem may be clear, considerable time and effort is spent on clarifying the requirements. Therefore, in Design Thinking, a large part of the problem-solving activity is comprised of redefining and shaping the problem.
In the Design Thinking process, investors are encouraged to avoid making early judgments about the quality of ideas, which helps to keep a broad outlook. As a result, this minimizes the fear of failure and maximizes input and participation in the ideation (brainstorming) and prototype phases. “Outside the box thinking” (“wild ideas”) is encouraged in the earlier stages, since this style of thinking is believed to lead to creative solutions that would not have emerged otherwise. As a methodology or style of thinking, it combines a) empathy for the context of a problem, b) creativity in the generation of insights and solutions, and c) rationality and feedback to analyze and fit solutions to the context. All this helps derive a potential solution that meets user needs and at the same time generates revenue, that is, drives business success.
3) Design Thinking Applied to Mission-Orientated Investment Decisions.
· Understand the problem: Get an initial understanding of the problem e.g Want to improve Healthcare/Need to understand constraints for biotech industry
·Observe users/field: Observe the industry at hand, talk to those involved, observe physical spaces and places. Take a hands-on approach and document the problems found
· Interpret the results: Interpret the empirical findings, look for pinch points/ problems that if solved could work as industrial catalysts
· Generate ideas (Ideate): Engage in cross-discipline brainstorming sessions to generate as many ideas as possible (in other words, expand the solution space) e.g What ideas/tools existing or otherwise, would create a paradigm shift?
· Identify Areas for Possible Collaboration: Prevent overlap, shorten roadmap
· Prototype, experiment: Look for companies working in these areas (may not be original industry space)
· Test, implement, improve: Test these tools, implement them with industry, and refine the design (narrow down the solution space again; solution-driven phase). Through this, we can verify the market size.
We need to work on broadening investment horizons to find those outlier companies and research groups that could work as paradigm shifts.
4) More Research into Tools to Improve Decision Processes in Finance
Improving investors’ decision processes is key to reducing failure rates for venture capital backed companies and to improving portfolio return and to support investor decision-making. Investors continue to rely on a vast span of faulted assumptions and biases, for instance that VCs have a special understanding that LPs do not have or are able to learn. Approaches like Design-Thinking help us to broaden our horizons in locating the fundamental problems, but we are also yet to see funds that bring on tools to improve the decision process. An example of this would be to construct investment- specific Bayesian causal maps. Seeking this form of probabilistic model would make explicit salient features of decision makers’ mental models and inference processes.
Bayesian causal maps have the potential to support decision-making through bias reduction, as well as reduction of unsystematic error, assumption surfacing, what-if analyses, and by facilitating systematic learning from experience, both individual and collaborative. Bayes net-modeling techniques make it possible to integrate expert knowledge as well as data from past projects. It’s possible to collect knowledge to construct the structure of a Bayes net and use either subjective assessments or statistical data to estimate the numerical parameters of the model. The development of such models and their subsequent incorporation in the decision making sequence of investors should increase the likelihood of picking winners. This should lead to a more effective and efficient allocation of capital to available projects. Decision models can enhance systematic learning. The underperformance of funds and the high failure rates investor-backed business demonstrates the need for improvement.
First, assumptions and decision rules are made explicit in the model building process. This is a form of cognitive feedback, providing investors with insights into their cognitive processes. Thus, decision rules move from being tacit, taken-for-granted assumptions to visually and quantitative explicated. These models then can be actively and consciously scrutinized by decision makers or expert coachers.
There is evidence that investors have imperfect insight into their own decisions, and that further gains in experience are associated with actual reductions in reliability. Experienced decision makers begin to rely on automatic information processing that leads to cognitive error. Instead of evaluating all of the pieces of information surrounding the proposed venture, experienced investors may focus on those characteristics that match past successes or failures. In other words, they may show increased susceptibility to an availability bias. To the extent this is so, the key to improving experienced investors’ decision processes may lie in reducing this reliance on automatic processing- somehow “forcing them outside their comfort zones”. A number of actions can be taken to rejuvenated this process, including requiring them (at least occasionally) to make decisions outside their regular domain, asking them to justify their decision in front of other investors, and then also using bootstrap decision aids and encouraging them to engage in counter-factual thinking.
The explication and scrutiny of beliefs is a key first step in belief updating and revision, whilst Bayesian causal maps may also allow for easy what-ifs and scenario analysis providing investors with a deeper understanding of the nature and implications of their decision rules
Second, models can be used to systematically evaluate outcome feedback. Once the outcome of a venture can be observed, the initial assumptions about the states of the cue variables can ban be compared to actual outcomes and “wrong” decisions can be traced to specific failures in either (1) the cue variable state assessment (e.g. management quality was overestimated) (2) the overall decision rules (e.g. the model did not consider a key variable or interdependency, or did not weight its impact correctly), or (3) the implementation (marketing strategy was correctly assessed as good, but the implementation of the marketing campaign was mishandled). As failures in variable assessment or inference processes become clear, decisions rules can be revised.
Third, formalised decision models enable collaborative learning. Models can be compared between novices and experts, or between partners of the same firm. Due to their explicit nature they can serve as a point of reference and departure for both individualised coaching as well as group learning in an organizational context.
5) Creating a Bayesian Causal Map for Investment Decision Making
- Elicitation of a raw causal map from successful investor (interview)
- Preprocessing of the raw causal map – conditional independence (creating a perfect map, both D- maps (concepts) and I-maps (independence map). Studying underlying reasoning – deductive/abductive. Direct and indirect relationships- eliminating circular relationships.
- Assignment of states to variables directly from interviewee, using verbal and experiential anchors iteratively to make up for the fact that people do not think in probabilities.
- Assignment of probabilities to states
- Refinement and validation
Evidence from prior studies suggests that in over 90% of cases, decision makers are outperformed by their own bootstrapping models due to the elimination of unsystematic errors. Although decisions aids are just that, and are not designed to replace the decision maker, a Bayesian network makes transparent the drivers behind the overall assessment and the software implementation allows easy what-if analysis by changing variable states and observing the automatically updated investment probability.
Further, the simple act of explicating and formalizing decision models surfaces hidden assumptions, which can now be scrutinized. We are animals full of assumptions. High levels of perceived learning can come simply from the elicitation process itself and testing of assumptions.
What’s more, unconfident novices can become experts over time by learning from well-analysed experience. This sort of analysis can also improve the collaboration of partners within a firm by enabling them to understand their differences in terms of decision models, biases and focus. LPs could learn to make capital deployment decisions better than most VC fund middlemen.
Obviously Bayesian causal maps are just one of a whole multitude of psychological tools to help improve investor decision-making. There is no lack of psychological tools. The problem is that, until investors realize how deep-routed the game they are playing is a psychological test of themselves- their own assumptions, and their world-model- then these psychological tools will continue to go under-utilized. Those investors who want to find the fundamental bottlenecks of the world, and then act upon them, need to understand their own fundamental bottleneck- which unfortunately in 100% of cases, is the fallibility of the human mind.
6) Fallible Minds Equals Fallible Approaches to Humanity’s Future
It doesn’t seem apparent that we can have true technological innovation, without financial innovation first. Our research into the financial areas is for the purpose of demonstrating that there are inherent flaws in the LP/VC model, and to show that there are a whole series of tools that LPs could use by highlighting two simple ideas- Design Thinking and Bayesian Causal Maps. As an industry, we need to destroy the idea that VCs have a special toolbox of powers. The toolbox is public knowledge. LPs need to work with people who want to strengthen their power and their capital pool and not feign a secret spectrum of special skills and charge for access. In this manner, LPs can become better investors of their own money, whilst keeping the pool that they lose to management fees to pay for funders with misaligned goals. We’re calling for rational integrity in finance.
The problem is just as investors and donors don’t go meta enough in finding and funding the fundamental bottlenecks, so too do investors not get meta enough when looking at financial models. It’s not simply a choice between Venture Capital and Philanthropy. It’s possible to combine ROI and Impact. Unfortunately, ‘Impact Investing’ is a new industry marred by the lack of definition for ‘impact’.
LPs can choose to demand more alignment in terms of goals and risk with the assets they already invest in, or they can work with people who encourage them to surface their assumptions.
We can fix the biggest problems of the world. But we contend that the most-overriding problem is that we don’t see what the biggest problems are or the ways to solve them.