We write algorithms that take observational data from satellites, pixels, and turn them into real-time insights about our world. The applications for this technology extend from farming to government and, as you might expect, finance. Weve got hedge funds knocking on our door almost every day, but our ambitions will take us far beyond the business of investing.
We use artificial intelligence, via evolutionary algorithms, to explain and predict complex ecosystems, such as patient outcomes in health care. Our platform does a good job of predicting financial markets. Still, were not interested in being a financial services company in the long run. Were interested in explaining complexity and predicting successful outcomes in a range of domains.
We model trust among entities in a network by examining social connections, online interactions and backgrounds. Our ability to predict trust- and creditworthiness will have a huge impact on the peer-to-peer industry thats developing, such as in ride sharing or apartment renting. And, of course, itll also create a massive opportunity in financial services, but thats not our focus.
Will the financial services industry soon be challenged by technology entrepreneurs with little initial or no exclusive interest in the investment business? The three abridged and anonymized quotes above, from real people running real technology companies, would seem to suggest that, yes, thats increasingly probable. Which is not to say that companies arent emerging from Silicon Valley to target bloated segments of the financial services industry (that is, pretty much all of the financial services industry) they definitely are. But weve noticed something rather more profound in the past year: The hot technologies being developed today will offer unparalleled insight into the complex world around us, and the applications to the entire domain of finance and investing are countless.
One example: The ascendance of nonbiological intelligence means computing systems will learn and process many types of inputs far faster than even the most-expert individuals. Once experts partner with the systems, these man-machine teams will become extremely competent at rules-based goal seeking. The days of using scarce computing resources to model complex systems backcasting, calibrating, validating and eventually forecasting are nearly over. Massive and scalable parallel systems are now available for rent by the second, and the implications of having access to almost limitless computing environments to attack the largest data sets imaginable augur a paradigm shift in the discovery and communication of relevant data patterns. Machine-learning science and technologies are increasingly agnostic to the internal mechanisms of models. Rather, the scale of data that can be evaluated and processed in real time on massive grid networks will allow systems to inform their operators as to what the key variables are, instead of being restricted by top-down architectures.
In short, a growing number of computing systems and technologies will empower people, organizations, networks and information in transformative ways. Service industries will be particularly affected, as they often require human, labor-intensive analytics and networking to scale. But if technologies can help people network and analyze faster and better, some of the companies in the industries that provide these very services will face an existential challenge. As with the rise of computing and the Internet, we expect new technologies in the coming decade to challenge service industries, such as finance, in ways that few people today appreciate.
What does all this mean for institutional investors? Well, it means that computer systems may soon do for them directly what asset managers have been doing as third parties. It means the private club of hedge funds, which you once happily paid 2 and 20 to join, wont be as exclusive as it used to be. It means the mythical black box that some asset managers use to drive outperformance and reinforce their own market power will be democratized. It means we will see a gradual emancipation of allocators from high-cost intermediaries and a dramatic move on the part of institutional investors toward much more flexible, and affordable, technology companies.
We think technology has reached a tipping point in finance. Up until now advances have served to empower private sector intermediaries, which traditionally have been relied upon to allocate capital to its highest use, at the expense of the big asset owners. Indeed, academic research shows quite clearly that smart financial intermediaries have used technology in the past few decades to obtain higher rents and reinforce their competitive advantage within uncompetitive markets. This undoubtedly results from the fact that technology often came with opacity and complexity, which was (and is) a recipe for high fees and compounding economies of scale flowing back to intermediaries. Indeed, to access the most-powerful technologies, institutional investors and others have had to work with the most-expensive gatekeepers.
At any rate, we believe Silicon Valley will help the community of institutional investors turn the tables. Ultimately, technology will help them streamline and strengthen operations, manage and distribute knowledge, access unique (and heretofore expensive) markets and level the playing field with the private financial services industry. As institutional investors adopt innovative technologies, we expect the very nature of financial intermediation to evolve. And that should be universally positive for our financial and, indeed, our capitalist system. For finance to function effectively, pensions, endowments, sovereign funds and other long-term institutional investors need to behave like highly sophisticated financial consumers. But to do that, they will need help from aligned intermediaries providing new and powerful technologies.
The authors of this article straddle the worlds of academia, technology and finance. Ashby is executive and research director at the Global Projects Center (GPC) at Stanford University and a senior adviser to the Office of the Chief Investment Officer of the Regents at the University of California. Daniel is CEO of Kensho Technologies, a Google Ventures and Goldman Sachs Groupbacked financial technology company that builds analytics platforms for the financial industry. Daniel also runs the financial technology research project at Stanfords GPC. We are both passionate about the role technology will play in financial services, and we have dedicated much of our professional lives to this topic. We are both entrepreneurs, investors, advisers, researchers and writers, which means that listing all our potential conflicts would more or less consume the remainder of this article. For that reason, weve decided not to use the names of any commercial entities from which we could hope to derive even a possibility of economic benefit. This will allow us to use insights from our own experiences without (we hope) being accused of talking our own book.
Our purpose in this article is not to sound an alarm on Wall Street about the threat technology presents to its business even if thats a by-product of our writing it. Rather, we want institutional investors and Silicon Valley to better understand how they might work together in the years ahead. In short, this is not a story about the demise of global financial centers as much as its a story about the rise of a virtual financial center. And we believe this virtual financial center will be extremely powerful, providing institutional investors with an entirely new place to operate and achieve their objectives.
The job of an investor is to take money and turn it into more money. Put another way, the product that all institutional investors create is the same: They make returns. Thats all investors really do. And to achieve this, they all use, for the most part, the same inputs: To their initial stock of money they add a healthy amount of human capital, a dash of informational advantage and a dose of process. Persistent outperformance requires an investment organization to apply high-caliber people and efficient processes in creative ways to develop proprietary sources of information and, ultimately, knowledge. Its this knowledge that allows investors to generate outperformance.
In thinking about the future of institutional investment, especially how technology will change the way investors do business, its useful to revert to these three inputs people, process and information and how they can come together to create knowledge. Whatever your unique approach to investing may be, its component parts likely fall into one of these categories. People refers to the talented individuals who drive returns for investment organizations. Talent is critical in the investment business, which is why skilled investors have sufficient leverage to get paid more than skilled professionals in any other industry. Process refers to the decision-making inputs required by an organization to execute on its strategic plans and achieve its long-term objectives. Process also refers to governance, which is a critical factor for success, as the board has the ability to give the organization the resources it needs to achieve its objectives. Information refers to the insights and, ultimately, knowledge used by investors to make decisions. This may refer to the algorithms running in a black box, to a highly cultivated network of trusted individuals or to the theoretical models in a finance textbook. Its important to note that not all information is of equal importance, and informational advantages are crucial. Its for this reason that some investors are willing to push their information gathering to the boundaries of what is legal.
In our view, these three inputs and the way in which they are combined offer a framework for considering the key factors that drive success or failure among investors. These inputs are often of different qualities and combined in different ways to achieve similar return objectives. For example, the endowment model of institutional investment is based largely on informational advantages; the best endowments are adept at leveraging their networks to identify and access top managers to drive high performance. The Canadian model of institutional investment is based largely on process and human resources; institutions that follow it pay competitive salaries and manage assets internally, reducing fee overhang and improving their ability to think creatively about portfolio construction. The traditional model of institutional investment, which outsources all of the assets to external service providers, assumes that people and information are best procured outside the confines of an institutional investment organization.
As University of Oxford professor Gordon Clark suggests, the key differences among the varying investment models often stem from a simple decision on whether to make or buy the key inputs (or a subset of inputs) to produce investment returns. If you have a robust governance framework that values investments in data-processing and knowledge-acquisition infrastructure and has the ability to pay top talent, you may choose to make most of your money on your own. If not, you may choose to take a hybrid approach. Indeed, the question underpinning the different models of institutional investment is ultimately whether the processes, people or information should be developed within an organization or sourced on the market.
But what this question fails to consider and what nearly all large institutional investors perhaps do not appreciate is the extent to which technology will alter the way in which these three inputs interact. Indeed, what the existing analyses of present and future models for investment did not realize was how much people, process and information could be disrupted through technological innovation. In the sections that follow, we will review some of the technological disruptions looming in each of these categories.
The largest institutional investors in the world are in most cases public in nature and located in cities far from major financial centers. Although hiring talent can be a challenge in any location, it is made all the more difficult by the fact that these public funds are limited by both compensation and geography. Many investors need to fill public sector jobs in cities like Edmonton, Juneau and Sacramento with people who can compete in and with the private sector. This isnt easy, especially with yawning salary differentials.
The conventional finance wisdom, however, has it that the highest-paid investors are the best at their jobs. If an investment organization wants to get the best returns, this logic goes, it has to be willing to pay the highest salaries. This is why public pensions and sovereign funds, according to intermediaries, should outsource to intermediaries.
We bet we can find more hedge fund employees earning $500 million per year in the U.S. today than you can find pension fund employees earning $500,000 per year. You on? Actually, dont take that bet. Youll lose.
But this raises an important question: Are the hedge fund employees who set the stage for the 1,000-times differential in compensation 1,000 times smarter? Of course they arent. They arent superhumans. Many simply have supercomputers. The technologies that certain hedge funds have been using are, in fact, at least 1,000 times better. Indeed, some early hedge funds made significant investments in technology, and they continue to reap rewards from those investments today. The economies of scale those funds enjoyed served to reinforce their hegemony. This is still the situation on the ground today, for the most part, though more and more managers rely on or help shape financial services through communication and discovery technologies like social media.
Thanks to technologys ability to gather and analyze enormous amounts of data, investors that use technology are able to employ far fewer people to perform the same or better analytics. You can think of technology as an augmentation of capabilities or a brain extender; as with any other profession, the work flow an investor can accomplish is generally fixed by the limits of his or her experience, skill and intelligence. Discount what are for now unforeseeable events chance and so forth add a widely varying array of access to information (market indicators and economic data, however sophisticated or simple), and you have a basic outline of the capabilities of any individual financial analyst: one analyst, one brain, surrounded by the tools of the trade and possessed of differing abilities to use those tools.
But now imagine if that person were not just one person, if the hard limit that is a single brain was in fact a much softer limit than we thought. Imagine a dozen artificially intelligent versions of any analyst say, of you taking in information and parsing it, making sense of it exactly as you might do but on a massive scale. That information is sorted and then passed to you, the human investor, as a set of recommendations to act upon as you like.
The result is, essentially, the creation of digital clones of a good analyst or trader. The applications for machine learning and artificial intelligence will prove invaluable for institutional investing. Deriving real-time insight from the daily data deluge, scraping data from the web as well as compiling and analyzing more-conventional sources of financial information, could easily become an automated process.
These days its not difficult to imagine that bots might learn to make rapid, granular decisions about which stocks to buy, in keeping with any given investment style that is, in keeping with your investment style. These bots could be trained through machine learning to identify the signals you would identify, independent of any too-stringent, rule-bound conditions. Bots can learn, simulate, replicate and amplify the reach of that idiosyncratic perspective.
An asset might exhibit certain traits and behavior that would be impossible to reduce to a logical set of conditions, but you might nevertheless find it appealing. You can only look at one chart or read one news story at a time to draw your personality-driven inferences; bots can look at millions simultaneously and eventually will be able to do so from a perspective they learned directly from you. Bots would not make investment decisions independently but would pass them on to traders as a set of near perfectly engineered recommendations derived by functional digital replicas, or avatars, of their own brains.
In sum, though quants have had their day in the sun, artificially intelligent bots may soon put them out of business. And as a result, the complete reliance on talented investment professionals, one of the scarcest resources in the investment business today, will give way to even-more-talented man-machine teams. The implications for a sovereign fund in Doha or a pension fund in Auckland are significant.
We now take for granted the ability to zoom into a virtual map containing millions of data points with our fingertips and to have a system reveal to us relevant information in stages and layers as we tap and swipe phone numbers, photos, addresses and GPS coordinates. Voice-commanded real-time computation of dozens of alternative traffic routes, factoring in live satellite data on millions of vehicle movements, is just another part of everyday life in Silicon Valley a new way of understanding and engaging with the world, pioneered by technology and since spread outward to just about every city in the world. Similarly, networking online to find old friends or new jobs has become commonplace. The simple act of saying swipe left or swipe right will have significant meaning for many people reading this. We think its fair to say that our day-to-day lives have been completely revolutionized by technology in the past decade.
By contrast, the world of finance is still, by and large, living in the past, reliant on calcified processes and technologies. Investors have neither the interface granularity nor the processing speed needed to properly model and compute big data. This limitation spawned the era of the quant and the data scientist, who still use complex programming languages for statistical modeling inside hedge funds and other intermediaries. But as noted earlier, the quants life span is running out as these processes move from Wall Street (or rather Connecticut) to Silicon Valley and Boston or, in theory, to any other city with the right technology infrastructure. More companies are looking to sell access to data and insight platforms on a license model rather than offer asset management services using those platforms with a hefty fee. The quotes from corporate executives at the top of this article give some understanding of the kinds of partnerships that many investment organizations will be forging with technology companies in the years ahead.
Sourcing investment opportunities also remains antiquated. Remarkably, the kinds of matchmaking tools that can connect you with your future spouse have not yet changed the way you source and screen investment opportunities. But fear not the big-data machine-learning platform that matches you with hot dates will soon be matching you with hot deals. In fact, we know of several existing companies, targeting different alternative-asset classes, that are working to match investors to investments based on hundreds of inputs. These companies, which are recruiting top engineers from social networking companies, are building matchmaking and correlation engines that will connect you with the best deals for you. They will assess your characteristics as an investor and your networks and match those opportunities that youve indicated you might be interested in (via predetermined signals) but that you can most rapidly de-risk (thanks to your unique qualities). One of these companies is even building technologies to construct optimized syndicates of different investor groups that will coinvest in a deal, maximizing the value brought to entrepreneurs by a community of different investors and increasing the likelihood of success.
But how many investment bankers realize this is happening? How many venture capitalists understand whats coming to their comfy niche? In our view, those companies and individuals that made their money in financial services because they sat at the intersection of networks (see brokers, bankers and some asset managers) should be nervous. A powerful matchmaking engine that can thoughtfully understand investors and entrepreneurs and can assess, using massive data sets and parallel processing, the potential for successful partnerships will give even the top venture capitalists a run for their money especially because the fees charged by the technology platform will be a tiny fraction of what the venture capital firms charge.
In sum, we are close to a future in which institutional investors will no longer have to rely on third-party managers to assess the morass of signals and heaps of data to make informed investment decisions. The financial services companies in global money centers many of which have accumulated talent and technology on a massive scale to inform their own investment decisions will be challenged by technology companies that have unparalleled analytics capabilities to inform everybodys investment decisions. And these technology companies will form the basis of a virtual pool of resources that will bring most if not all of the professional capabilities of financial centers sourcing, screening, conducting due diligence, structuring, syndicating, trading and monitoring, among other tasks to the fingertips of investors around the world. All they will require is access to a connected device. And the technological systems that will make all this possible live in what we are calling the virtual financial center.
We spend a lot of our time trying to help investors think creatively about how they do things. We do this for two key reasons. First, we believe that the best investors accept financial markets as constantly changing ecosystems in which good ideas are ephemeral and there are rewards for spotting new opportunities early and acting in an entrepreneurial manner. Second, we believe new approaches will be almost by definition less competitive than traditional approaches, meaning that creative investors can reduce the fees and costs associated with their investment execution. In short, we think it pays to constantly think creatively about the process of investing. Most investors dont seem to agree. In reality, institutional investment organizations are often allergic to innovation, enjoying monopoly control over their own asset bases. They prefer the company of a good herd, and the individuals in charge tend to focus on managing political and career risk instead of investment risks.
One of the main reasons for this behavior is that most investment organizations are flying blind. The theoretical models they use to build their portfolios, which are based on unrealistic assumptions like rational actors and efficient markets, are often powerless to explain let alone predict crises and often follow one another right into the middle of them. In fact, the majority of financial models derive from work done in the 1960s to the 80s, when pencil-paper tractability was prized precisely because of a lack of computational power. Worse, such investors recognize all too well that they live in a very dangerous world, but their business, risk and information technology systems are obsolete or lacking in critical functionality, redundancy and security. Boards of directors are told repeatedly by consultants and intermediaries that effective investment management in financial markets is enormously demanding in terms of talent and resourcing, that this is a business for professionals and that No disrespect, yall, but pension fund employees arent professionals.
In this context, its reasonable that institutional investors rush into the arms of costly intermediaries and the mainstream financial service providers. Its understandable that they spend their time constructing diversified portfolios of risks rather than digging into individual assessments of opportunities. Its understandable that they hug benchmarks and match peers; buy products, not assets; invest in managers, not companies; and manage according to expected returns, not risks. They optimize ratios rather than focusing on key variables, and they use diversification as a crutch rather than as a tool. And over time these mechanisms, which were meant to simplify investors lives, have become exactly the things that complicate them. The models and abstractions the products and managers are not substitutes for real-world understanding and investing.
Can technology help? No doubt. Boards and managers will soon have systems that can explain and predict markets. They will have the ability to slice and dice their own portfolios in real time. They will finally have powerful tools to plot a course that meets their own unique needs, recognizing who they are, where they are and where they can go in an idiosyncratic world of constraints and challenges. The era of big data will reduce our reliance on top-down models by allowing explanatory data to emerge into coherent explanations and predictions, without draining human resources or relying too heavily on external managers and consultants. Again, an asset might exhibit certain traits and behavior that would be impossible to reduce to a logical set of conditions, but data-mining tools will help make sense of it all.
Boards will finally have an effective risk management function based on real-world agents that will facilitate the development of an effective investment function. They will only empower managers and delegate authorities when they are confident that risk systems are functioning effectively. Investors need not be averse to complexity, but they need to be able to ensure that they fully appreciate the component risks of every investment and seek to diversify those risks at the total portfolio level. Technology will make this much easier than it is today.
Technology will also help overcome the limitations of existing governance models. It will provide a means of collaborating with peer organizations. It will help minimize errors and biases via checklists and diligence tools, all of which will increase efficiency: Investors will better understand their portfolios and be able to manage them more effectively. No longer will a pension fund have to beg for ten more people technology will make a single individual capable of doing the work of ten people. The only losers in this world will be the highest-cost intermediaries, whose own dominance was likely a function of technology and informational asymmetry (not to mention the poor governance of the asset owners).
Many institutional investors today are focused on how they can grow. The prospect of working with misaligned intermediaries in global financial centers has made pension and sovereign funds search for ways to increase their head counts, their systems and even their global footprints all in the name of reducing their reliance on the for-profit financial industry. We can think of five public investors with at least 700 bodies sitting in no less than four offices in global financial centers and this strikes us as potentially wasteful. No doubt this approach offers a more aligned access point to financial markets than other investment models, but has the pendulum swung too far? Ten years from now, will the community of institutional investors need massive global organizations to reach their potential and achieve their objectives?
We think the next generation of institutional investors will take a different and likely smarter approach. Rather than developing massive teams and physical outposts in global financial centers, future investors will likely stake their claims to the resources available in the virtual financial center. Silicon Valley will develop offerings that empower people, streamline processes and rapidly convert data and information to actionable knowledge. The 700 people in those extra offices may have to go elsewhere, as a core team of highly talented and technologically networked individuals back at headquarters may be capable of doing the same job.
A more inclusive revolution is coming to finance, and with it we expect to see a new, lean model of institutional investment rooted in dynamic capabilities and technological sophistication. The tech model, as weve been describing it to friends, will offer a path for all institutional investors to find aligned access points in financial markets. We see it as ideal for those funds that are looking for aligned access to markets but are unwilling to build global empires (as in the Canadian model) or prostrate themselves to the very agents they should be disciplining (as in the endowment model).
We expect to see in the coming decade a generation of investors empowered by world-class technology, focused on creatively combining people, process and information in new ways, some of which weve touched upon and whose appearance we hope to herald with this article.
The Canadian models reliance on people and process will diminish if technology makes talent less expensive and an organization even more responsive and rich in information for its decisions. The endowment model will no longer best suit institutional investors that can think through big data on their own and break away from their dependence on managers. The best information will go to those with the best technology, and the best technology will live in Silicon Valley and other tech-driven regions and cities a point so obvious its almost not worth making. It also seems likely that the shifts we outline will amplify one another as they interact, making the virtual financial center the home of the tech model a more powerful and empowering aggregation of services and knowledge than ever was empowered by world-class technology available in London, New York or Hong Kong.
To be sure, tech model investors will rely on technology to do the majority of their work and instead will focus their talents only on those areas where they can truly add value in the marketplace. These investors will use their own characteristics in a deliberate way to move into markets with minimal competition. (Recall that they will be working with extended brains with smarter approaches to research and analysis especially; that is, they may be acting more deliberatively as well as more deliberately.) For example, being a long-term investor offers opportunities that short-term investors dont have. Also, being a local trusted partner to companies and project developers can create unique and proprietary opportunities. Last, a large investor may be constrained in its ability to access top managers, pushing it into alternative access points for similar risk exposures. These are unique characteristics that future investors will want to consider in their own portfolio construction.
But to develop unique and innovative capabilities that leverage competitive advantages, institutional investors will need to expand their partners to include technology companies. This isnt as crazy as it might seem. Were already seeing some institutional investors look to technology companies as key partners, filling the roles that used to be played by consultants and asset managers.
Among other benefits, placing their trust in the virtual financial center will help institutional investors access the promise of machine learning and data mining. Theyll also begin to overcome the geographical restrictions that have kept them from attracting top talent. (It isnt clear that top talent will continue to work at traditional firms, as technology companies continue to attract young professionals who might formerly have gone to Wall Street.) In short, everywhere one looks, technology has begun to revise a relatively conservative, slow-moving industry.
Over the past 30 years, advances in technology have served to empower financial intermediaries. Going forward, technology may threaten their existence. As we saw in the quotes that opened this article, the final aims of the most-exciting technology companies may lie elsewhere. In the meantime, a new virtual financial center is rising.