Where do we stand with the ’07-’08 crisis, and why is the CIO job so important?

1.Introduction: Storytelling and its insights.

One of the puzzling questions in today’s asset management industry is why the role of Chief Investment Officer (CIO) is still viewed as one of the most important. Indeed, nowadays, artificial intelligence evangelists – believers who trust “only” mathematical algorithms- claim that this job does not have a future: A well calibrated (in terms of past data analysis) software is already allocating assets under management in a more rational and cost-efficient way than humans could ever do as we are more likely to issue misleading interpretations from the analysis of big data sets.

In this first -of two- essays we will shed light on some factors that answer this question. The main theme will be to justify the CIO role by referring to a neglected feature, the CIO’s ability to elaborate and write down allegorical stories. In other words, the neglected aspect of the CIO job resides in their storytelling abilities, the skill at finding the right words to present complex phenomena by creating an allegorical narrative that can easily be shared with clients inside and outside his/her financial institution. This story is key, as it is needed to convince the public about the quality of the allocation scenario offered at a given time and about the timing chosen to modify this allocation. Besides, as we will demonstrate this ability allows to enhance a CIO’s major quality: Creativity.
It is because CIO is a failing creature and error prone, that they will invent scenarios, non-conventional rules and paths: Human creativity is what matters and AI is far to having a gram of that.

2. An example of CIO narrative based on ’07-’08 Crisis or how pointing forward some central facts.

As a starting point, let’s take the role of the CIO who must create a nice and powerful narrative to describe today’s global economic situation.
Now, it has been exactly ten years since the beginning of the worst financial and economic crisis since the Great Depression: What is the best metaphor to describe the current situation?
Without being very original we can build a scenario by referring to the ’07-’08 crisis as a fire declared in a modern residential area, in which different types of houses (i.e. modest and complex ones) have been built closely enough to expect flames propagating from one house to the next.
In this residential complex, the fire simultaneously started at the ground floor of two major houses. The fire department came swiftly, which did not prevent some major pieces of furniture from being fully destroyed in both houses, but saved both houses ‘structures and several items. Few minutes later, a third very complex house was also impacted with a fire problem: We disregard this case because the fire department is still working on it and no clear evaluation of the damages is yet available.
The intervention of the fire department implies two major shortcomings (a third one is presented in my next essay):
A. Given the degree of urgency, the fear of a propagating fire, the amount of water used was too important. Moreover, in the heart of the action, no one evaluated how much water was absorbed by the walls, how much ran away into the basement and whether some finally got absorbed by the ground: These effects would likely weaken the solidity of the buildings, therefore changing their structures.
B. The fire department did not have enough expertise: They saved some items and let others burn without a clear idea of each item’s value and, even more inconveniently, without knowing all the consequences of letting some burn. Moreover, despite the huge effort, we are still unsure whether all the sources of the fire are dead.

3. From a narrative into its analysis: why human creativity matters so much

As stated, our allegory is not very original, but it allows us to highlight two main sets of the crisis’ consequences in simple terms. Let’s now use financial/ economic terminology to obtain an equivalent translation of these two sets:
To fight the crisis, point A, the monetary authority injected too much liquidity (i.e. the famous QE experiment) to avoid a generalized credit crunch: the goal of this liquidity creation was to calm down the financial/banking system. Here, the point is to know when the excess of liquidity and the associated price distortion will eventually be digested. Nevertheless, a major unanswered question remains, even ten years later:  At what speed does the monetary authority need to retreat?
In theory, i.e. rational expectation and general market equilibrium, we all know that any liquidity excesses get corrected -in the long run- thanks to an inflationary process. But, even this long-term wisdom is challenged nowadays because, on the one hand, globalization and technological progress are generating a strong deflationary pressure, and, on the other, the bank industry rules have changed due to the crisis (e.g. new requirements in terms of capital – the Basel agreements-) so that large monetary creation has been prevented.
In this fundamentally new environment, how and when will price increases be registered i.e. at what speed are houses walls are supposed to dry up? Is inflation likely to come soon? And without the inflationary guide how is a central bank supposed to guide itself?
Now, what really matters is not to own a set of definitive answers to all those questions, but instead, to add constantly new puzzling queries, which also implies to assess those new questions so to review and rewrite future scenarios.
The CIO’s minds, not algorithms, are what matters: Humans have a unique ability of re-analyzing and re-interpreting a phenomenon, despite having no new data. It is because we consider our conclusions as being true conclusions that we constantly reevaluate them: An algorithm will provide correct (logic and rational) conclusions given the data, but which cannot be discussed in terms of true or false!

Merely, by referring to his allegorical story and his holistic and critical view of the economy new questions will be created and new future challenges defined to justify both the current asset allocation choices and how to modify them. The key factor is human creativity: Only a human being, a being prone to mistakes, is always ready to rethink and rewrite his narrative of a financial and economic phenomenon.

The intervention, point B, has dramatically reshaped financial markets, without anyone really knowing whether this will assure or prevent some future problems or whether we are now following a path that will ensure better allocation of our savings. For instance, the crisis had serious consequences in terms of public debt (i.e. some needed furniture got substituted to assure a functional ground floor):
How will this fact influence the future allocation of private savings?
Besides, while some investment vehicles have disappeared, others are now dictating the financial tempo in the asset management space, e.g. passive investment vehicles. However, what about a critical analysis of these new instruments? Are they as transparent, cheap and liquid as pretended? What about the ETF’s judgement being biased by roughly 8 years of bull markets in the USA? What will be the long-term consequences of a generalization of these vehicles? Are we out of trouble? Are we sure that no bubbles are currently running in the financial markets?

Once again, to consider and evaluate all these questions we need a more holistic and critical approach, which only the human brain can offer.

4. Conclusion and few steps into our future analysis: Strategic consideration and investment timing.

As we saw above, the main objective of the CIO is to explain why a given asset allocation is chosen at a given time. The narrative is there to prepare and justify further choices. Namely, only time provides the answers to all open questions, so what matters is to implement choices over time that maximise the expected gains once these answers have materialized. As we have already point out these choices are likely to be based on a very holistic and critical approach about the current situation.
Still, and this will represent the essence of my second essay those moves are also dictated by strategic considerations: A CIO is not isolated.
He knows that other CIOs share a close financial and economic outlook, i.e. the data used and economic and financial notions are common knowledge among them: therefore, the timing for a reallocation is more likely to be set by strategic considerations, a variant of Keynesian beauty contest type of game. But then, once again, a robust and well-structured allegorical story may reinforce the chances to be among the winners because the analysis is likely to widespread.

More details will be explained in the second essay. It is the presence of this strategic environment which will represent a further main obstacle in substituting CIOs by machines.

2 thoughts on “Where do we stand with the ’07-’08 crisis, and why is the CIO job so important?

  1. With the AI tools we possess now, most CIO positions are not at risk. They are, however, in need of some change. New means are at the disposition of the finance world, which should learn to control them. AI is only a tool and therefore has to be manned to work properly. The operator must use and understand the right algorithms, finely tune them for the desired purposes, know what features to include, train the model on the right training dataset, avoid overfitting and so much more. The person behind the computer also has to have a deep investment understanding. This operator must know what he or she is doing, and this is where I believe where plenty of CIO jobs are headed.

    “The CIO’s ability to elaborate and write down allegorical stories” should not, in my view, hold any weight. Telling stories is unreliable, lead by bias and intuition. Everybody can fabricate a story as to why an investment has been made after we choose to make it. Decisions should be data driven. Moreover, not all algorithms are black boxes that shroud the motivations underlying investment picking.

    Note that some types of investing; such as value investing of the type done by Warren Buffett, cannot be done by computers yet. The low amount of hard data and the need for human interactions in processes such as due diligence make this kind of investing beyond the reach of computers.

    “Human creativity is what matters and AI is far to having a gram of that.” This is no longer true. The most recent machine learning algorithms make computers more than mere calculators. Probably the most famous example is AlphaGo’s 37th move against Lee Sedol in their second game (see https://www.wired.com/2016/03/sadness-beauty-watching-googles-ai-play-go/), when the algorithm played a move that was unthinkable at the time. This new intelligence can truly be considered computer-driven creativity. Other examples can be found in the artistic applications of TensorFlow. The applications of this types of AI are now slowly making it into finance through platforms such as Numerai or Quantopian and within hedge funds.

    In the second essay, you explain that the private investigator does a better job than the engineer because he or she collects more data (mostly soft, qualitative data, such as opinions, suggestions, interpretations…). There’s no reason as to why this data could not be incorporated and understood by the machine (such as through market sentiment from news or web scraping with natural language processing or similar tools). While these tools are not perfect as of yet, they are improving at an ever-increasing rate. Therefore, the machine does not have to be the engineer in the case of the house-on-fire story.

    Like

    1. Hi Edgar,
      Thank you for your very interesting comment. To be honest I really believe that these two essays should be viewed as work in progress materials and not definitive answers, i.e. some aspects and arguments in there continue to be badly defined and presented, and I am constantly challenging myself to find more satisfactory formulations.
      For instance, I am not saying that decision should not be data driven, my concern it is more should not be only data driven: Sometimes, row data are tricky and despite appearing as being clearly defined, they can lead to misjudgment. Often in economics, data series are living entities, which definitions need constant review.
      Therefore, it is important to create a theoretical discourse, which can justify why a given data matters more than another one at a given time. I will present all these considerations in a further essay that I am planning to publish soon in this blog.
      To conclude, in this further essay, I am discussing the notion of creativity as well: I agree with you, AI has made unbelievable progress in this area, but I still believe we tend to use this term too broadly, i.e. you are referring rightly to a narrow form of it by writing a computer-driven creativity. But, is this really what we meant when we think to a human creativity? This is my point. Once again, I will discuss this in the next essay.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s