The sharp rise in inflation to a 40-year high in 2022 was a turning point, upsetting the optimal scenario of moderate growth and low inflation. Plan managers are now walking a tightrope trying to find a solution that reduces both inflation and recession risks.
With their long planning horizons and multi-decade commitments, pension plans face far too many open and unknown risks. The search for predictable sources of value creation has multiplied. The spotlight is on industries that are being reshaped by life-changing megatrends that are driving disruptive innovations, transforming business models, and reshaping public and corporate policies. In this regard, thematic funds represent an innovative approach that breaks free from the traditional silos defined by countries, industries, drivers or regions.
Pension plans have been primarily focused on operational priorities, and rightly so, but are increasingly faced with changes in reporting requirements on investment governance issues. As a result, ESG investing is increasingly becoming the new normal for pension funds, placing climate change as the top concern for institutional investors, followed by areas such as healthcare, aging populations, and China's economic growth as a superpower.
Funds are no longer just talking about responsible investing; they are now implementing measures to do so. And with the use of machine learning-based predictive models, ESG data can be quickly integrated into an existing portfolio to better reflect the ESG profile of the fund or plan.
The Potential of AI for Pension Plans
While some sectors of the Canadian pension industry, for example, rely on more traditional methods, several vendors are exploring and implementing innovative advances in pension technology. Artificial intelligence is at the heart of many of these applications.
However, there are many definitions of artificial intelligence. According to John Beckett, author of the book "The New Fund Order", data as a point of information does not constitute AI, in order for a technology to be called AI, it must meet one of the following three criteria:
An algorithm or program built in layers, which are commonly called pillars that make decisions;
An algorithm that can move left or right based on input, and
A so-called the naïve Bayes classification based on Bayesian inference. The system can forecast the likelihood of a specific result when given a set of parameters.
According to him if you do not have one of these types of machine learning, all you have is data.
Using AI data adds an element of objectivity. Their material and timely data helps institutional investors to continuously monitor fund managers. According to some observers the culprit for disappointing pension fund performance is not just poor returns, but the inability of active managers to outperform.
In the defined benefit (DB) space, AI and related analytics tools should help improve outcomes and achieve efficiencies. AI can help multiemployer plans with 10 or more employers by conducting proportional reviews of participating employers and summarizing the results on a searchable platform. It is also capable of helping smaller plans by providing proportionate liability valuations in a more cost-effective manner.
On the risk management side, AI could provide pension plan administrators and their risk managers with real-time analysis of the impact of recessions, inflation or employer insolvency. On the investment side, identifying patterns and the likelihood of certain events will help administrators and investment managers make investment decisions. Plan administration could also be greatly facilitated by AI by reducing costs and improving accuracy.
However....the devil is in the details. For AI to help retirement plan administrators and service providers deliver the best possible retirement product to members, plan and member data will need to be accurate and consistent. The adoption of AI-based tools in the retirement industry faces two challenges: practical concerns and emotional apprehension.
Quality data is essential, and while AI tools can be powerful, their results must be interpreted, and their limitations recognized. AI tools can help automate tedious tasks, but humans are still needed to run them. The industry is at a tipping point, with the increasing use of artificial intelligence.
The Future of Artificial Intelligence in ESG Pension Investment
The biggest frustration for portfolio managers is the disagreement over ESG ratings. Mastering the quality and nature of the aggregated data to get the full picture of the portfolio. This is a significant challenge for institutional investors. But there are artificial intelligence (AI) tools available today that can collect and analyze an impressive amount of data on ESG risks and opportunities. These tools increase the quality of the information, make it easier to analyze, and offer exciting new opportunities.
Situated at the intersection of technology, innovation, and sustainability, AI can have a significant impact on ESG investing, taking into account environmental, social, and governance risks and opportunities in investing. In addition to this, AI can be used to analyze affiliate decisions to improve their product and platform design; identify and correct data issues; and analyze unstructured ESG information to improve reporting and portfolio allocation. It can also contribute significantly to improving ESG reporting and target tracking.
However, there are still challenges in analyzing the many data available, and choosing one measure over another can have a significant impact on the outcome. Ultimately, a comprehensive investment process should avoid placing too much reliance on a single measure. In addition, the costs of managing alternative datasets must also be considered: not only the costs of acquiring the data, but also the investment required to store and integrate these large datasets. Overall, the consensus is that the integration of ESG into investment approaches will become deeper and the ability to use robust data will play a major role in this process. AI can not only help extract relevant information from existing data sources, but it also offers exciting opportunities to create new ones.
Conclusion
In other words, AI will increasingly infiltrate plan operations and the impact will not be limited to investment decision making, "There are many ways to apply AI and machine learning to make operational processes and frameworks more efficient.
In ESG investing, it is starting to be used and tested, although many limitations remain, and it may be some time before practical, standardized applications are in place.
If ESG investing is about considering the material opportunities and risks of sustainable decision-making, AI has both considerable benefits and risks to watch out for. In short, while giving ESG investing the opportunity to develop and expand, AI may itself pose an ESG risk to companies that wish to embark on this path.