Scientia Potentia Est, "Knowledge is Power", is what Sir Francis Bacon claimed in the early 17th century. A timeless lesson that survived to our days, now more than ever critical to understand our world and to invest with clarity. Yet, although one may rush to associate “power” with “profit”, institutional investors know this adage hints to a more pragmatic and valuable consequence of possessing knowledge: prudence.

‍

In this sense, if surely prudence emerges from sensible risk management, it is primarily the result of a correct representation of the expected payoffs and risks portfolios are exposed to. We know this can be achieved by analyzing data, yet to take action something more is needed. In this paper, we discuss how this gap is being bridged by Portfolio Diagnostics, a game changer for asset and wealth managers to scale and improve their decision-making by converting portfolio intelligence into actionable investment insights.

‍

Rethink, Research, Refine: Is Data the New Oil? Or the New Fuel?

Finance movies often feature crowded trading floors and genius-like managers yelling at flashing monitors, depicting an image – sometimes the myth – that investing is like fighting in a battlefield.

‍

However, all the emphasis that is put on taking risks and relying on gut-feelings may distract us from acknowledging what is the actual role that asset and investment managers play in the economy: prudent decision-makers, not fortune-tellers.

‍

Regardless of style, objective or risk tolerance, following a rational process and appropriately weighing the outcomes of any decision seems a more realistic image of what investing looks like – and most importantly, the reason why prudence is the key to sustainably succeed in this world.

‍

In this sense, the nature of investing stands at an ideal intersection between a certain past and an uncertain future. Investors evaluate the information available and then choose among investment options whose payoff may not always be easily determined. 

‍

Information can be indeed a double-edged sword: being informed is key, but information overload can easily lead to decision paralysis. That is why the question “what information is needed to guide our decisions?” appears not only of epistemological relevance but also of practical matter. Typically, the answers to this question vary depending upon the scope, the duration and the complexity of what is taken under consideration. However, by taking the “negative route” – as Nassim Taleb would say – we can find a precious piece of insight.

‍

The absence of clarity is one of the major factors that put our decisions at risk. Even if we had access to every piece of information available, it would be indeed useless in most cases – toxic in the others – if not represented in a meaningful, concise and actionable way. 

‍

Indeed, with the increasing amount of data, information has become a commodity. However, just like oil, information must first be first refined – to grow in value and become usable – and then consumed to fulfill a specific need. As oil barrels eventually become energy, financial data must be transformed to provide insights on how market dynamics are going to unfold. 

‍

The asset management industry has long leveraged technology to gain a deeper understanding of portfolio construction and risk management. Indeed, it has opened the opportunity to increase the frequency of portfolio checks to uncover hidden risks in portfolios and untap pockets of performance previously gone unnoticed. 

‍

As we will see, reaping the benefits of these portfolio “pit stops” means converting data intelligence into actionable insights to assist our decision-making. A process – called Portfolio Diagnostics – that appears to be suited for a variety of portfolio management applications to assess and reduce the impacts of unexpected events that cause strategies to drift from expected results.

Portfolios Need Diagnostics – Just Like Hamilton’s Mercedes

The high-speed world of Formula One racing and modern investing have many touchpoints: they both have tons of data available, require a massive amount of planning and scenario analysis, and involve a continuous revision of the initial strategy to achieve the best results. 

‍

Indeed, even if the audience gets mostly drawn to the cars’ aerodynamic design, winning in Formula One is just as much about speed as it is about making the right calls assisted by state-of-the-art software.

‍

Over the years, technology has indeed disrupted traditional racing. During the 1970s, for instance, cars were equipped with on-board computers capable of transmitting real-time data from the vehicle back to the engineering team. It was a real gold rush: more data meant more control, enormous time savings and the chance to run computer simulations beforehand. 

‍

Today, it would be nearly impossible to race without the assistance of thousands of cutting-edge computers that provide continuous feedback on every angle of the race. Billions of data points are analyzed every second so that the racing team can precisely make subtle changes such as, how and when to adjust wings, change tire pressure, tweak suspensions, and so on.

‍

Furthermore, these actionable insights – called diagnostics – have contributed to dramatically reducing the time spent in pit stops, timely revisions of the car balance during a competition. In this way, it has become possible to ensure the greatest performance level throughout the race without sacrificing too much time spent at the boxes. 

‍

We find that similar concepts – like control, revision, and simulation – hold true also when, instead of a car, we consider a portfolio’s asset allocation.

‍

Indeed, over time the mix between asset classes may need to be realigned with the original target allocation, or we may want to know if the portfolio is poised to meet its objectives, before the end-of-period comparison against its benchmark. 

‍

Indeed, if we compare running a Grand Prix to investing, investors are the pilots entitled to make prudent decisions with the best technology available. As a matter of fact, it is right in the application of scientific rigor combined with accurate Portfolio Diagnostics that asset and investment managers are finding ways to harness the power of data to adapt to the complexity of financial markets. 

‍

In this sense, converting information into insights is how investors can gain an edge. Learning how to do this is related to our ability to turn what appears to be seemingly chaotic into manageable information. The end of a long road that initially brought investors to turn uncertainty into risk and luck into probability, and now ambiguity into clarity.

‍

‍

How to Convert Intelligence into Investment Insight

As we recently discussed in a paper entitled “The Right Call”, combining a disciplined, data-driven approach with the computing power of Artificial Intelligence is enabling investors to make sense of the huge amount of financial data available and develop valuable investment insights.

‍

In this sense, a thorough Portfolio Diagnostics explores a wide range of heterogeneous possibilities not only in terms of time horizons but most importantly in terms of risk levels. 

‍

On the one hand, risk can be understood by looking at concentrations in asset classes, sectors, geographies, and so on. On the other hand, it can be seen as the mix of different risk premia, such as the equity risk premium, or the Value, Size, or Momentum factors. 

‍

Understanding how different risk exposures influence each other leads to a better comprehension of the worst – and best – scenarios for our portfolios, enabling us to act before a real risk arises or an investment opportunity disappears.

‍

Moreover, this process contributes to shed a light on what is the real risk-return profile embedded in the positioning of the portfolio. By understanding how a portfolio will react under different market scenarios it is possible, for instance, to investigate the connections with its benchmark in terms of performance, exposure, and weighting. 

‍

Considering all these factors is indeed fundamental to best define what investors can expect from their investments, ensuring they deliver what they promise.

‍

Overall, effective Portfolio Diagnostics is turning traditional investment forecasting into a probabilistic and holistic understanding of potential shifts in market dynamics and portfolio behaviors. In this sense, this sophisticated approach – schematically depicted in Exhibit 2 – allows investors to enter an ideal “control room” of their portfolios to evaluate their positioning and discover new diversification opportunities, defining a clear path to actively meet their objectives.

‍

As we are going to see, AI-driven scenario analysis represents the cornerstone of effective Portfolio Diagnostics, as it reconciles all the information we need to gauge the expected behavior of our portfolios. 

‍

Such a comprehensive and forward-looking process ensures a better understanding of their inherent risks and enables us to verify if what we expect is coherent with our objectives. 

‍

Eventually, combining data analytics, scenario analysis and a careful scientific approach leads to the construction of stable portfolios capable of efficiently withstanding the possible unfolding of financial markets.

‍

‍

If You Want Peace, Run an AI-Driven Scenario Analysis

Investors often wonder how their portfolios would behave in different circumstances. It can be intuitive to guess how an all-equity portfolio would perform during a stock market crash. Yet, more complex portfolios require running a scenario analysis to know how things can play out eventually.

‍

This approach appears to be more prudent than trying to exactly predict the future. Indeed, just as the old saying “if you want peace, prepare for war” suggests, a thorough understanding of what we can expect from investments in different market scenarios is essential to achieve robust positioning and avoid unexpected surprises. 

‍

Over the years, this need has led to progressively refine how we can gauge the behavior of a portfolio across various market conditions. 

‍

For example, traditional static what-if analyses (e.g. how a sudden interest rate hike would affect the portfolio) and stress tests (e.g. evaluating the impact of a global recession) have found large application in portfolio management for hedging and assessing the hidden sources of risk. 

‍

However, two major factors have limited the degree to which these approaches provide insights for asset and wealth managers on how to take action.

‍

First of all, a correct evaluation of how a portfolio would have reacted to a past scenario (e.g. during the 2008 crisis) required all its securities to have a track record long enough to be analyzed. 

‍

However, until recently, this was not always possible because securities are issued at different points in time. As a consequence, the results were often either partial or not completely reliable.

‍

Secondly, excessive reliance on backward-looking scenarios ignored the possibility that future market developments could differ significantly from the ones previously observed. In other words, implying that the future would exactly resemble what happened in the past left investors dangerously exposed to never seen scenarios.

‍

Today, the unprecedented computing power of Artificial Intelligence is helping investors overcome these limitations. In particular, AI is proving useful in sifting through billions of data points to identify a common set of risk exposures among securities. As a consequence, it has become possible to generate proxies that replicate the risk profile of unavailable securities and obtain more precise risk assessment. 

‍

Moreover, AI is also allowing investors to consider forward-looking scenarios (e.g. an upcoming highly volatile bear market) and thus have a more precise picture of what we can realistically expect from our portfolio positioning under several circumstances. 

‍

The chart in Exhibit 3 shows an example of how this can be achieved. The chart in Exhibit 3 shows the over/under performance of a multi-asset portfolio under forward-looking market scenarios (strong negative, negative, lateral, positive and strong positive).

‍

Each scenario assumes three levels of volatility (i.e. high, medium and low) and considers an expected duration that ranges from 20 to 125 days. By looking at the graph we can immediately see how this portfolio is positioned with respect to its benchmark and how we can expect it to perform in different scenarios. 

‍

Indeed, the portfolio shows a quite robust positioning, with a tendency to overperform its benchmark in most of the scenarios considered, especially during lateral or moderately volatile rising markets. On the contrary, in the case of a medium to highly volatile market downturn, the portfolio is expected to lag behind its benchmark.

‍

‍

If Knowledge is Power, Prudence is Control.

In conclusion, the reason why investors benefit from turning data intelligence into investment insight appears extremely relevant – if not even foundational – to win the technological pivot the asset management industry has recently experienced. In other words, to confront uncertainty as active players, not as idle spectators.

‍

Now more than ever, clarity and stability are the answers to succeed and thrive in a financial landscape that has become increasingly data-intensive and complex. Yet, clarity can only be achieved as the outcome of a process that puts a meaningful – and actionable – representation of risks and payoffs at the center of decision-making. 

‍

Combining Artificial Intelligence with a scientific approach to risk identification, mitigation, and ongoing monitoring is enabling investors to build portfolios structured for success, unlocking their full potential. In this context, Portfolio Diagnostics can assess how portfolios are poised to meet their objectives and evaluate their robustness over several scenarios. 

‍

As a consequence, it has become a crucial tool for asset and wealth managers to understand what risk drivers affect portfolios and how to improve their positioning.