How Innovation and Technological Savvy Would Help Institutional Investors

The role of institutional investors in modern society often goes underappreciated. In fact, the viability of many socio-economic systems is premised on institutional investors succeeding in their missions. Sovereign wealth funds help stabilize the macro-economy and currency prices, finance critical infrastructure, and invest for citizens’ long-term future. Pension schemes provide transfer mechanisms that permit billions of people to have financial security in later life. And endowments and foundations fund scientific research and education, as well as sustain the arts.

To execute these crucial functions, institutional investors must remain competitive within today’s global financial ecosystem. And this means adopting advanced technologies, such as machine learning and sophisticated data-management platforms. Yet most institutional investors (hereafter “Investors”) have historically struggled not just to embrace new technology; they are often unable to innovate in general. The reason: Investors were rarely – if ever – set up with innovation as a central design principle or priority. Instead, many Investors have been initially configured in ways that resist change, rather than drive it. For example, prudent person rules and fiduciary obligations often put the brakes on innovative ideas, as do market- and peer-based benchmarks, which push Investors to hug indices and converge with industry norms. Consequently, technology is treated as peripheral (or at best non-focal) by most Investors and isn’t used to develop long-term competitive advantages.

Our research – which has entailed dozens of interviews and hundreds of hours of direct dialogue with Investors – highlights a path that could resolve both of these problems at once: By reorienting their organizations around technology, rather than being mere users of it, Investors can ensure deeper innovation across the entire process of investment decision-making. Below, we describe our findings on the chief obstacles to such technological reorientation, as well as means for removing them. Many of these insights are transferable to other types of organizations that may seek to further develop their own capacities for innovation. We then depict what a future of deeper innovation, enabled by technological reorientation, could mean for institutional investors.

We first note that simply improving access to better technology won’t suffice for most Investors. Investors have a long history of outsourcing technology and innovation by relying on third parties to supply them, often indirectly, inadequately, and at excessive cost. For example, many Investors claim their main access to advanced technology is through external asset managers (e.g., hedge funds). Reliance on external managers, however, creates problems of transparency and control. External managers with truly cutting-edge technology are usually unwilling to expose its details, which makes it nearly impossible for Investors to assess whether the high fees those managers charge for technological “prowess” are justified. Such secrecy also encourages external managers with inferior technology to follow suit and reap higher fees.

More broadly, whenever Investors rely on external managers for technological access, they lose control over what can be done with it. Investors (heavily) subsidize external managers’ technology, but can’t use it themselves to, for example, run their own risk analyses). These subsidies strengthen Investors’ addiction to external solutions by widening the gap between outside managers’ technological capabilities and their own. Investors need to reverse this growing disadvantage.

The main barriers to doing so stem from Investors’ data quality and organizational norms. For most of history, improving investment technology has meant increasing the speed at which data reaches investment decision-makers: from using carrier pigeons, to telegraph stock-tickers, to internet relays. Yet speed confers only fleeting advantages and shifts emphasis to exploiting short-term arbitrages. Institutional investors operate over long horizons and should instead look to build durable competitive edges, which is why they should focus upon data quality (i.e., granularity, volume, and accuracy). Nevertheless, Investors find it hard to maintain suitable data quality due to a benefit-cost disconnect: They concentrate too narrowly on what benefits a given dataset can provide (e.g., assessing it only for single applications), while not considering the overall costs to properly clean, maintain, and analyze it.

On the organizational side, Investors’ obstacles are threefold. First, the diffusion of data and technology across their organizations is impeded by fragmented communication and lack of visibility across data and technology resources contained in different pockets of the organization. For example, many datasets reside on local memory-drives accessible only by single users and are shared only point-to-point – as spreadsheet attachments to emails. Second, technology “transformation projects” are generally too large and top-down in structure. Such big change efforts are prone to overruns in both time and cost, and foster fatigue and noncompliance among personnel. Third, Investors tend to maintain an isolated perspective on technology: When appraising their own technological capabilities or seeking new solutions, they often do not look widely enough and fixate on current trends and status quos among their peers. Together, these factors stand in the way of Investors being able to reorient their organizations around technology.

Our research indicates that these obstacles can all be greatly mitigated or even eliminated. There are four keys to doing so, which entail: 1) more comprehensive data-quality evaluations; 2) dynamic culture; 3) empowered communication; and 4) creative cooperation. Concerns about data quality can be dealt with through enhanced data-governance systems based on sharper tools for more holistically appraising the quality of data as well as valuing its proper use. For instance, best-in-class Investors maintain well-designed data hierarchies that dictate what levels of quality are necessary for data to be used for specific types of investment decisions. These Investors also are acutely attuned to how organizational culture and day-to-day behavioral patterns affect data exchanges and upkeep by personnel and structure their data-management protocols accordingly.

But organizational culture is not just a given. To reorient around technology, Investors must work to promote a learning ethos that promotes experimentation with new technology in a controlled way (i.e., at a small scale that limits the consequences when some experiments fail). However, such experimentation and learning require resources, and Investors need to ensure that the right types and amounts of resources (especially time) are reserved for such purposes. There are many ways for Investors to ensure such resources are available. These include time budgets (like Google’s 20 percent concept, where employees get a fraction of their workweeks to pursue innovation projects outside the scope of their usual tasks), daily innovation tracking (a la kaizen principles popularized by Toyota), radical transparency (which exposes improvable areas and is heavily practiced by the mega-fund, Bridgewater), and Red Teams (which are units tasked with pinpointing weaknesses in organizational strategy and processes and ways to correct them).

Communication can be improved with keener understanding (and visibility to personnel) of where various data resides within the organization, as well as how it tends to flow across it. Such efforts to map organizational data should be paired with changes that make data management a more social process – for example by making data repositories more accessible inside the organization and allowing for collocation of user input (e.g., comments, use cases) with the data. Additionally, storage schemes that preserve more of the context in which data were gathered (e.g., by whom, for what purpose, along with what other data, through what procedures) can not only help enhance management of data quality, but also find ways to get new, better data.

Finally, Investors, should demand greater cooperation among themselves regarding data and technology. Many Investors compete very little with one another, as compared with how they must compete in the financial ecosystem at large. Thus, cooperatively sharing knowledge about technology could confer massive benefits to Investors as a group and greatly help in breaking their dependence on third parties for technology. And such sharing needn’t be limited to existing knowledge: Pooling resources to develop new technologies, data, and capabilities could greatly empower the institutional-investment community across the world.

These changes will not necessarily be easy, but the advantages to Investors, and the global socio-economy, could be enormous. Worldwide, Investors are stewards of more than $70 trillion in capital. A mere 10 basis points (0.1 percent) of that amount spent on technology would far surpass the research and development budgets of Google, Apple, Facebook, Amazon, Microsoft, and IBM combined. That’s some significant innovation power that’s going untapped. What sorts of new solutions could that capital unleash if it were to be repurposed for innovation?

To get the flavor of an answer, consider how some of the latest advances in the field of deep learning – which uses artificial neural networks to perform sophisticated inference tasks – could apply in institutional investing. In 2016, researchers Mount Sinai Hospital in New York built a deep-learning system for making clinical diagnoses: It was given no pre-programmed medical knowledge but learned to create complex representations of patients based on their medical records alone. It used these to diagnose over 90 diseases and accurately predict future health events for patients, days, weeks, and even months in advance of when expert human clinicians were able to do it. Imagine the benefits to Investors of having deep-learning to aid risk detection and alert human decision-makers to market disruptions in advance.

In 2015, researchers from Tufts University constructed an artificial-intelligence algorithm that solved a biological riddle that had confounded scientists for a century: how a type of flatworm is able to regenerate healthy copies of itself when sliced into segments. The system reverse-engineered the biophysical pathways responsible for that process and came up with a new theory of how these pathways allow regeneration. This theory was not only accurate but easily interpreted by human researchers. Imagine how useful it would be for Investors to be able to consult more true-to-reality theories of financial risk and market function (versus many of the leading economic theories that prevail today) and have technology that can build actionable models from them.

In 2017 Google researchers engineered an artificial-intelligence engine, AutoML, capable of constructing its own machine-learning systems. AutoML was shown to be capable of building programs that in many cases performed better than the machine-learning architectures designed by Google’s own top engineers. Imagine how much Investors’ technological sophistication might advance if, instead of hiring expensive in-house programmers, they were able to have inexpensive algorithms construct tools for them. Or, if instead of hiring fleets of quants to build trading and valuation models, Investors could let artificial-intelligence engines do the job.

The benefits of these possibilities to society should be significant, as they would significantly help Investors achieve their respective missions. But before any of them can become reality, Investors need to commit to reorienting themselves around technology.

This post comes to us from Ashby H. B. Monk and Dane Rook at Stanford University. It is based on their recent paper, “The Technological Investor: Deeper Innovation Through Reorientation,” available here.

Leave a Reply

Your email address will not be published. Required fields are marked *