Feb 22, 2021
The Composer Manifesto
The Great Stagnation
Asset management hasn’t changed all that much over the last few decades. Despite all the headlines about consolidation, disruption, pressure on fees, and so on, the asset management industry today doesn’t look too different from a decade ago. Full-service wealth managers still charge high fees while offering dubious value. The last real product innovation was the Exchange Traded Fund (ETF), hailing back to 1993.
The industry’s most recent change is for brokers simply to lower prices, with commissions dropping to zero in many cases. A lack of pricing power, however, is more an indication of commoditization than anything else — when you haven’t done anything different, the only way to stand out is to charge less. The biggest asset managers’ web interfaces are underwhelming, to say the least. Pros and retail investors alike end up managing their portfolios in messy Excel spreadsheets — if they haven’t given up keeping tabs (pun intended) altogether!
Why hasn’t software changed everything already? There’s likely no single, all-encompassing answer. But a likely culprit is that, to date, fintech software attempts have been too practical, lacking the boldness and radical product thinking we’ve seen in other domains. Instead of radical reform, they’ve focused on incremental change, automating a single step of a process here or removing a layer of friction there. Bolder action is needed. To break out of the current stasis, the entire infrastructure — human and machine — that underpins asset management needs to be rethought and rebuilt, starting with a blank slate.
Composer, as it stands today, is a no-code tool for building automated investing strategies. Yet everything we do today is in service of our long-term mission to completely rebuild asset management with software.The rest of this piece details what we think the future will look like and how we fit into it.
The Great Decoupling
Holding a variety of assets across different strategies is pretty normal these days — and if modern portfolio theory is correct, it’s also a good thing, thanks to the benefits of diversification. But in order to invest in different types of assets today, these assets often need to be maintained in multiple accounts. This situation leads to two sources of entropy:
- The need to execute buy and sell orders in multiple accounts to go from current positions to target allocations.
- A growing number of siloed data sources to pull from, wrangle and aggregate to get an idea of consolidated portfolio performance.
Both of these challenges are pretty daunting, and frankly they will just get worse as entropy increases.
The idea that we will use one single monolithic provider for all data, portfolio management, and trade execution seems more and more nonsensical by the day. The proliferation of new types of alternative assets, particularly crypto and DeFi, shows no sign of slowing down, and these new asset types don’t play well with traditional brokers.
The entropy generated from new data sources is even more pronounced. On top of the need to wrangle data coming from different account and asset types, there will be a growing number of external data sources — whether company data scraped from 10-K filings, sentiment analysis of social networks, signals derived from time series data, or new data sources we can’t even anticipate today.
To adapt to these and future changes, Composer’s architecture decouples data, logic, and trade execution. Specifically, Composer is a single logic layer that takes an arbitrary number of data providers as input and sends instructions as output to an arbitrary number of brokerages, exchanges, and/or custodians.
The shape of things to come
Once logic is cleanly decoupled from data and execution, it’s possible to create a light-weight container to share the most intellectually valuable and creative part of investing — the investing strategy. An investing strategy, after all, is a set of decisions that rely on data as input.
This might sound like it’s solely applicable to quant investing, but it’s not. Even the most old school fundamental investor follows this process. They might pore over financial statements with a magnifying glass in hand, highlighting items with a pen, but they are nonetheless ingesting data to make decisions: Is this company undervalued? Should I allocate money to this asset? If so, how much money?
Recognizing that the logical decisions that comprise any type of investing strategy are the most valuable link in the chain (and the thing that drives the generous compensation of fund managers), it makes sense to containerize strategies. At Composer, we call this strategy container a Symphony.
Granting a license to dream, the platonic ideal of a Symphony will have certain key properties: interoperable, immutable, versioned, shareable, and monetizable.
Interoperable
If the investing strategy is the thing of value, and the brokerage is largely a commodity, people shouldn’t be forced to use any one brokerage. This means the Symphony needs to easily interoperate with many different brokerages, much like how Stripe makes it easy to integrate with many different payment processors. Interoperation would entail both pulling account-level data needed to make sense of the aggregate holdings of a portfolio from across accounts and sending order instructions. There has been some progress on the former “read” case, with much less progress on the latter “write” case — although both could use improvement.
There’s machine-machine interoperability, as discussed above, and then there’s machine-human interoperability (i.e., “good user experience” ). In the end, humans are the ones to create, inspect, and modify Symphonies, whether they know how to code or not. Composer is aligned with the philosophy that software will augment humans and improve their decision making capabilities, rather than replace them altogether. This philosophy dates all the way back to the very originator of computer science, Ada Lovelace. As Walter Issacson once explained in an interview:
I think in 20 years, you’re going to have ubiquitous computing, meaning they’re going to be so intimately connected to us, we won’t even notice that our watch is a computer or our eyeglasses are a computer. And I think you’ll have natural language so that you can just talk to your computer. You won’t have to type in, you’ll say things like, did the Yankees win? And on your eyeglasses, on your watch, it will happen, because I think things get more and more personal in the digital age. They get more intimately and humanly connected to us. And that takes us back to Ada Lovelace, who really believed that it wasn’t our machines that were going to become artificial intelligence that would go off and do things on their own, but that the trajectory of invention and innovation would be to connect humanity to our machines better, to make our machines much more personal, so they really become part of our lives instead of this quest for an artificial intelligence or a singularity, where machines will get to think without us and won’t need us anymore.
Given this belief, the platonic ideal must have an accessible no-code interface in addition to a well-designed API to allow developers to easily extend Symphonies. Which brings us to our next desired property.
Modular and Extensible
As with any new, bold software innovation, the Symphony won’t be static nor perfect at conception. Instead, there will be an active community of developers and enthusiasts to extend the functionality of a Symphony in ways the initial creators could never dream of. This might mean new core functionality in the Symphony itself or new external integrations.
While extensions can take many forms, they should all be modular and composable. That is, each new piece of functionality should lead to a combinatorial explosion of possibilities. Much like a choir isn’t the mere addition of soloists, but rather the harmonies the choir can perform as a group, each new element added to the Composer ecosystem should harmonize with the rest.
Sharable and communicable
In addition to making it easy for humans to interact with software, humans need to be able to interact with each other — that is, they need to be able to communicate. Right now, communicating about an investing strategy is a disaster. The main mediums available are Excel sheets, code, and long-form natural language. All of these mediums are deeply flawed; they’re either too ambiguous, or they take far too long and too much cognitive effort to parse. Code, which is the most rigorously testable and unambiguous medium, is completely uninterpretable to anyone who can’t code. Natural language is the most accessible, but the reader may easily interpret a strategy differently from what the original author intended to convey. Excel offers a middle ground, but it has repeatedly been the source of grave errors and mistakes, with more complex workbooks relying on a dizzying chain of formulas and manually inputted data.
A clear visual representation of a Symphony is essential for quick, easy, and unambiguous communication. A shared visual language allows for the dissemination and discussion of ideas. Many of the best investing strategies have historically been the product of highly collaborative teams. Anything that facilitates Symphony collaboration should pay dividends.
Immutable and versioned
Just as engineers version their code with Github, designers version their designs in Figma, and writers version their work in Google Docs, it should be easy to version Symphonies. As with any creative knowledge work, it’s helpful to track changes made over time, enriched with comments and metadata that indicate why a change was made. If there’s a mistake, it should be easy to roll back to a previous version. A related, equally desirable property is the ability to “fork” different versions of a Symphony and run them in parallel for purposes of either experimentation or diversification.
Connected to versioning, there should be an immutable record of a Symphony. This engenders trust for all relevant parties — whether users, creators, editors or distributors. If someone is the originator of a Symphony, they want to ensure intellectual credit stays with them. If a strategy is the result of modifications by others, they will want some form of credit. Discerning end users will want to know the provenance trail of a Symphony as part of their diligence process. Just as DeFi and (at least to some extent) fiat-denominated assets have protections against counterfeiting and debasement, so should the Symphony that “sits on top” of those same assets.
Monetizable
Right now, the best way to make money in asset management is, well, to work in asset management. But being an asset manager isn’t easy — nor is becoming one. Even sophisticated private investors eschew launching funds to invest in public markets, as the setup costs and regulatory burden prove to be too much of a headache.
The future will be different, and the lines between professional and retail investors will blur. Credentials will matter less and performance will matter more. And physical location will continue to dwindle in importance. Even traditional asset managers will need to welcome these changes to keep up, as more forward-thinking and capital-efficient firms embrace the new reality.
All the properties discussed so far make it easier for anyone, anywhere to make a living working on the most enjoyable and intellectually satisfying part of asset management: strategy creation. And there will be more than one way to contribute and earn income, depending on one’s skill set. Everyone from a quant, to a fundamental stock picker, to a developer of ancillary apps and services will have a chance to shine — and to be financially rewarded.
The Symphony is also poised to disrupt the funds of funds industry. Currently, funds that invest in other fund managers charge ludicrously high fees and are inaccessible to anyone who isn’t either part of an institution or ultra-high-net-worth. The Symphony will provide new avenues for people who are skilled at selecting, blending, and curating others’ strategies to be rewarded for their taste and acumen, without needing to charge exorbitant fees for doing so.
Investing as a creative endeavour.
In some ways, the most radical change needed to evolve asset management isn’t even technical. It’s a cultural change, challenging long-held beliefs about what asset management should look like. What if the experience of managing investing strategies not only didn’t suck, but actually delighted? This only sounds like naive, foolish optimism before considering that tools in fields like design, film, coding, and so on are already fun, stimulating, and creative.
Before the advent of personal computers with graphical user interfaces in the 1980’s, computers were the domain of niche hobbyists, and before that, large institutions. Computers weren’t supposed to be easy to use, let alone fun, for average people. They were large, complicated, computing devices operated by engineers and scientists.
The introduction of the microprocessor in the 1970’s made home computing a physical possibility — but it took decades for non-engineers to widely adopt them for other forms of creative work. Apple rightfully gets a ton of credit for this cultural shift — even if they didn’t build most of the core technologies that made the shift possible. Rather, it was a shift in perspective that was very radical and non-obvious at the time. Steve Jobs was guided by a different set of values than the rest of the industry, the subject of countless books and articles. One value was especially odd: the quest for aesthetic perfection.
How does rehashing this widely known piece of history relate to asset management? Renaissance Technologies, the most successful hedge fund of all time in terms of performance, might be the one firm that could credibly claim to have “solved” the market as a purely scientific exercise. And yet, Jim Simons, its founder, has attributed his firm’s success in part to the following principle:
Be guided by beauty. I really mean that. I think pretty much everything I’ve done has had an aesthetic component, at least to me. Well we might think that “building a company that’s trading bonds, what’s so aesthetic about that?” What’s aesthetic about it is doing it right. Getting the right kind of people, and approaching the problem, and doing it right. And if we feel that we’re the first one to do it right, that’s a terrific feeling, and it’s a beautiful thing to do something right…”be guided by beauty” is not such a bad principle.
The software we build at Composer — from the outermost layers of our visual design all the way to our backend architecture — is guided by a search for beauty of the kind Simons describes. We can’t wait to see the beautiful things you will build.
