In a Radical Uncertainty world, be careful how we use data.

In this third issue I share 3 publications with different approaches, but all have the same line in common, how could we better understand the current scenario of uncertainty.

The first post seeks to understand the limitations of many of the predictive and prescriptive models that are developed. Today, the complex reality, the uncertainty, the risks and the current dynamism make us be extremely careful about how we use advanced data analysis.

In the second post I make a brief summary of an excellent book by John Kay and Lord Mervyn King that gives us a constructive criticism on how are we interpreting, communicating and understanding uncertainty. Over the years there has been an erroneous implication of this concept. And finally, in the third post, something that worries me very much is how Data-based products are being designed. I see that there is a worrying simplification in the data monetization process, so I wonder if we are aware of why we pay what we pay.

I hope you like it, and again your comments are welcome.

The danger of Data, and of the advanced Models of analysis, without having a context in social sciences.

This is a small reminder, after having worked as a Principal Data Scientist for several years, and having made several simple, complex, static, dynamic models, without learning, learning from the new incoming data flows (machine learning models), about the limitations of these models that are used today by many companies, executives and in the design of public policies.

For example, this week we see how those same algorithms, which understood that covid was something normal, now have understood that the inflation data is fixed. In the absence of new news following yesterday’s CPI, algorithms trading on technical indicators likely drove reversal, which resulted in 2nd largest trading range in 2022 so far, at >2.5 times YTD average daily trading range.

I like to differentiate it from two perspectives.

The arrogance of data. The acquisition of ordered and distributed information in tables and data sheets has flooded the tables of senior management. The data indicates. The data sign. The data ensures. This is a transcendental error to the extent that data hides a research design, a question, a paradigm. The data does not operate in a vacuum, and thinking that data is objective, pristine or honest prevents us from understanding the structure that operates behind it. Data are systematic answers, but they are not the truth. In its political dimension, political data is already concerned about the cognitive biases that are generated in the creation of public policies based on data, but without public or social context. Similarly, those organizations that believe they are right because they have good data are unaware of the impact of radical uncertainty on the bottom line.

The social criterion. Access to databases tends to become a commodity, as numerous multinationals offer the same service. difficult to differentiate through the accumulation and capture of data, which is why I defend the recovery of the notion of managerial intelligence. The manager has to broaden his knowledge in social disciplines, such as history or anthropology. The machine already mines data, but it lacks the social knowledge to lead projects, empathize with customers, or build social trust.

The company's strategic decisions cannot be hidden behind a veil of documents, data and probabilities. The false sense of security that expansion plans, spreadsheets or forecasts give harms the organization and turns the manager into a mere executor of programs. There, in addition, artificial intelligence offers better results. That is not, therefore, the way to resolve crises or to transform institutions.

-------------------------------------------------------------------------------

IMF and other projection-driven houses must read “Radical Uncertainty: Decision-Making Beyond the Numbers” written by John Kay and Lord Mervyn King.

In the last 4 months the IMF, to name one multinational organization that has a global impact, has changed its monthly projections for the global economy (and for each country), 4 times. This does not mean that the IMF technicians do not know how to make GDP projection models, they have been doing it since July 1944. The clear problem is uncertainty. Therefore, I read thoroughly about uncertainty (the real one, not risk disguised as uncertainty).

In 7 days, I have "eaten" this book. Although it is not new (well, it was published in 2020) it has a number of concepts and conceptualizations that are very important when we try to understand reality, and above all, to be able to approach a possible future through projections. Here the key word, which we see repeatedly every day by politicians, analysts, journalists, is uncertainty.

The focus of the question is the authors' dissatisfaction with the conventional treatment that economics has given to the concept of uncertainty: assigning subjective suspicions to all possible future events. This has been the way in which conventional economic analysis has succeeded in encapsulating the notion of uncertainty as a matter of risk management.

This approach, initiated in the American Academy in the 1980s and later incorporated into university economics curricula throughout the world, rests on some crucial assumptions. First, it is possible to know all possible events and assign them a probability. In other words, we do not know the future, but we can discern all possible future courses and assign them a subjective probability of occurrence. Second, all those situations that fall outside this map are treated as exogenous shocks. Third, from this representation of the world, individuals, households, and firms maximize expected utility. Fourth, another central element of the argument is that, during the period for which the agents formulate their expectations, there are no changes in the possible future paths identified - there is an assumption of "steady state" of such horizons.

In opposition to this analytical battery, the authors rescue the classic concept of uncertainty, which they qualify as "radical uncertainty". This deep uncertainty, the one that has the usual meaning of the term, includes what we do not know, what we can only imagine in a very weak and imprecise way, options that in turn offer infinite ramifications (even difficult to model and calculate computationally)

One of the sections of the book is dedicated to analysing what would have probably been the reasons for moving away from the classical approach that had the consequence of distorting the most evident interpretation of the concept of uncertainty. The authors maintain that the Chicago School, through the contributions of Friedman and Savage, had a decisive transcendence, and in particular of the latter. However, they note that while Savage warned that his formulation applied to "a small world", where it was feasible to better limit the scope of the proposal, Friedman generalized the theory. In this way, and to summarize, the real uncertainty disappeared from the scene and everything was reduced to risk management and subjective expectations about the future.

What then, is the conclusion reached by Kay and Mervyn King? First, the need to recognise that we do not know the future and that the artifact of potentially subjective ones is of very limited application, but that it loses potency if it is taken beyond those limits. Second, the foregoing does not invalidate making scenarios, and we use information, data and models, but any conclusion must be accompanied by a "narrative". The purpose is to contextualize the analysis and, above all, to examine the limits within which certain conclusions are reached, especially if it is about making policy recommendations, a strategic business plan, or investment portfolio management. Third, and perhaps more importantly, the authors suggest the convenience of revaluing the old concept of uncertainty, removing it from the treatment of subjective suspicions and rational expectations. The latter is also a message from the authors that tends to bring awareness of the importance of the practical sense of economics. It is worth a separate consideration of the question.

* For those who are interested in diving into these concepts and works on Uncertainty, I recommend reading "Psychology is Fundamental: The Limitations of Growth Optimal Approaches to Decision Making under Uncertainty". A new research program in ‘Ergodicity Economics’ has reinvigorated interest in growth-optimal approaches to decision making under uncertainty, here

-------------------------------------------------------------------------------------- 

Are we aware of why we pay for what we pay?

I don't know about you, but in my family we highly value going to the movies together. It doesn't matter if it's Black Panther: Wakanda Forever, Guardians of the Galaxy Vol. 3, or Loki, season 2. And if the opportunity arises, going to dinner or a snack after the movie is excellent (it all depends on the time of the movie, or the appetite of the oldest child).

Now, a topic that always comes to mind every time I go to the movies or out to dinner, is the price of popcorn, and the price of soda. Why do I have to pay n times more for soda or popcorn at the movies than at the supermarket? Why do I have to pay x times more for soda at a restaurant? I'm sure you too have asked yourself these questions several times. It is clear that the first answer that comes to mind is “because once you have crossed the door of the premises, they can charge you whatever they want”.

But, from behavioral economics, we know that this answer is not entirely correct. I don't rule out that it isn't rational to think that way, but believe me, I go to the movies and have dinner often, so he analysed the issue financially several times (I know, we all have problems!). Price-sensitive moviegoers will surely find substitute products outside the theater. They will have an ice cream, they will buy a drink in the supermarket, and they will not consume anything in the cinema, only the film. For their part, people who are not price sensitive will consciously be willing to pay a higher price for popcorn and soda.

Now, after reading several papers by Dan Ariely (who is a Professor of Psychology and Behavioral Economics at Duke University) or Richard Thaler (Nobel in Economics in 2017) or even George Loewenstein (Professor of Economics and Psychology at Carnegie Mellon University), we know that many times we act irrationally in the face of certain events (at least in the short term). And we also know that coincidences do not exist, that the world is full of causalities (we do not need to read many economics papers for this, our grandparents told us so).

When the dots are "joined" it can be inferred that what is behind "this situation" is the clean strategy of theaters/cinemas or restaurants to discriminate against people based on price. It is clear at this point that prices are a signal, and within the economics of signals there is much to learn and test. Companies clearly know which attributes are valued by people, those where the person is really sensitive to price. Particularly in the cinema, it is the quality of the image, the quality of the sound, the quality of the seats, the quality of the environment, and the quality (premiere or not) of the film (obviously, the quality of our companions is critical). That's what really determines the price of popcorn and drinks.

If the determining attributes are correct, the prices of the complementary attributes dramatically reduce their elasticity. In other words, very few definitely buy popcorn not because of the price of popcorn, but because of the quality of the movie experience. Finally, we as clients do the work for companies by segmenting ourselves, and clearly defining something that the cinema did not know, how much we were willing to pay for popcorn and soda, based on how we value the determining attributes of the cinema and the restaurant.

Perhaps the next time you go to the movies, or have dinner in a restaurant, you can understand why the price of soda or pop is x, y, n times respectively more expensive than outside those places, and know that you are not only paying more for them (in absolute terms) but that he is giving clean Data to the cinema and the restaurant about your willingness to pay. Everything that follows is an advanced analysis of data in order to improve profitability decisions.

Therefore, the Product Director of solutions or products based on Data must not only understand technology, data, but also, and above all, must understand that psychology, economics and sociology play an important role when to be able to define the product strategy, to know if the product will be successful, and if the product will be able to generate, appropriate and distribute value (it is critical that the Theory of Value is part of the requirements of a PD). It's not just Data and Technology in the life of a Data-Product Director.

Note: all opinions are personal and do not represent the organisations I belong to.

#business #design #data #event #algorithms #complexity #statistics #markets #economy #data #storytelling #pricing #uncertainty