The R&D Tax Credit Aspects of Artificially Intelligent Hedge Funds
Hedge-Funds
Today, thousands of hedge funds are in
existence, collectively managing over $1 trillion in assets.
These funds, which are unregulated and only available to
qualified investors, pursue myriad different trading
strategies including but not limited to long/short equity,
market neutral, event driven, macro, and distressed securities
investing. Each fund employs and adheres to its own unique
investment philosophy and process; however, every fund can
generally be classified as being either fundamental or
quantitative in style.
In the “quant” fund realm, many funds are scaling up
investments in “machine learning”, a subset of artificial
intelligence (AI) that is attracting considerable attention in
the money management industry.
Hedge funds developing and refining their own proprietary
machine learning technology should be aware of federal and
state Research and Development Tax Credits that are available
as a result of their innovation.
The R&D Tax Credit
Enacted in 1981, the Federal Research and
Development (R&D) Tax Credit allows a credit of up to 13
percent of eligible spending for new and improved products and
processes. Qualified research must meet the following four
criteria:
- New or improved products,
processes, or software
- Technological in nature
- Elimination of uncertainty
- Process of experimentation
Eligible costs include employee wages, cost of supplies, cost
of testing, contract research expenses, and costs associated
with developing a patent. On December 18, 2015 President Obama
signed the bill making the R&D Tax Credit permanent.
Beginning in 2016, the R&D credit can be used to offset
Alternative Minimum tax and startup businesses can utilize the
credit against payroll taxes.
Rise of the Machines
While it’s true that computerized
investment approaches have been around for decades, machine
learning, a form of artificial intelligence where dynamic
algorithms comb through immense sets of data seeking patterns
and predictive outcomes, expands upon traditional algorithmic
trading models which are often rules-based and static in
nature.
For example, a standard algorithmic investment model might
short the 30 best performing stocks in a generic index over a
specified time frame while simultaneously snatching-up the 30
worst performers over the same period, hoping to benefit from
a market reversal. Such a model, while systematic and complex,
is also rigid and fails to account for changes in things
beyond company fundamentals and market technicals.
By contrast, many quant funds in recent years have migrated
toward artificial intelligence where machines can scrutinize
large amounts of financial and non-financial data in the blink
of an eye. Specifically, several quant shops today employ a
type of machine learning called deep learning, which trains
machines to recognize patterns in big data by incorporating
human-like observations of pictures, words and audio . This
capability, in turn, enables machines to be able to recognize
potential actionable investment ideas often well before any
human might be able to do the same.
Consider, for instance, the hypothetical AI powered hedge fund
that looks at weather patterns and satellite images of crops
in order to make a bet on commodities . In this regard, an
artificially intelligent machine is able to identify, analyze
and adapt to real-time changes in the natural environment in
ways that traditional statistical-based analysis cannot
account for.
The Pioneers
The technology involved in creating
artificially intelligent machines is cutting-edge. These
machines do not just collect data and process information;
they draw inferences, answer questions, recommend actions and
display data in visual layouts with state-of-the-art features.
Described below is a summary of some of the key players in the
hedge fund industry that have embraced machine learning as a
natural extension of their quantitative investment DNA.
Bridgewater Associates
Bridgewater Associates, the world’s largest hedge fund manager
with approximately $150 billion under management, is turning
towards artificial intelligence. In March 2015 the firm
started a new artificial-intelligence unit with about half a
dozen people led by David Ferrucci, the all-star engineer from
International Business Machines Corp that developed the Watson
computer to beat human players on “Jeopardy” during his time
there.
The unit is tasked
with creating algorithms to make predictions based on
historical data and probabilities. But the technology
(which is still rather confidential) is far more innovative
than its ancestors from the late 1990’s who used quantitative
analysis to predict similar trends. The difference is
that the latest technology has self-learning capabilities, can
communicate with humans and is capable of analyzing far more
complex and voluminous data.
As opposed to following static instructions, programs learn as
markets change and adjust to new and forthcoming
information. The programs are effective, advocates say,
because they can crunch huge amounts of data in short periods,
"learn" what works, and adjust their strategies on the
fly. In contrast, the typical quantitative approach may
employ a single strategy or even a combination of strategies
at once, but may not move between them or modify them based on
what the program determines works best.
Renaissance Technologies
Renaissance Technologies, the East Setauket, NY fund with
headquarters near Stony Brook University and operations in
Manhattan, has been developing innovative self-learning
financial analysis technology as well. The company was
founded in 1982 by President and CEO James H. Simmons and has
approximately $65 billion in assets under management, as of
October 2015. It has consistently been a leader in
cutting-edge financial management technology around the world,
employing artificial intelligence to make sense of patterns
that are not easily decipherable to the human mind.
Two Sigma Investments
Two Sigma is a systematic hedge fund based in New York that
uses machine learning and distributed computing as part of its
investment strategy. The company is run by John Overdeck
and David Siegel and has $35 billion in assets under
management. Two Sigma believes that artificial intelligence
and machine learning offer measurable benefits over human
fundamental due diligence and traditional quant methods by
supplementing statistics with real-world information.
MAN AHL
Man Group’s AHL fund out of the United Kingdom, which has over
$19 billion under management, has tested and used machine
learning strategies with demonstrable success. In fact, a
particular machine learning strategy helped AHL move from
negative to positive returns when it was first introduced in
August of 2015. Although it’s still early in the machine
learning life cycle, Man Group has resolved to commit more
funding to AI than to any other venture – all but affirming
its massive appeal and potential.
The Impact of Cloud Computing
While large and powerful hedge funds have
always had the means to invest in and devote resources to AI,
cloud computing has made it feasible for start-ups and
entrepreneurs to help hedge funds create intricate computer
models. Indeed, the pooling of many computers into one
powerful entity, accessible via the Internet, has allowed
things to be done at unprecedented cost and scale .
For example, Braxton McKee, founder of Ufora – a data science
engineering company based in New York, has been able to
leverage the powers of cloud technology to create his own
models at a nearly negligible cost. Additionally, Domino Data
Lab, founded by a few Bridgewater alums and based in San
Francisco, enables quantitative hedge fund analysts to test
new ideas, collaborate with other data enthusiasts, and
seamlessly produce new trading models, all through access to
the cloud .
There is no doubt that cloud computing is a major
transformational force in the IT world. In addition to
enhancing IT productivity, it serves as an increasingly
important platform for innovation amongst hedge funds seeking
to capitalize on the use of artificial intelligence.
Challenges Ahead
Several notable funds, as summarized
previously, now employ some form of artificial intelligence to
predict trends and to better understand markets and human
behavior. Nonetheless, as any good quantitative analyst will
agree, models are only as good as the assumptions underpinning
them. Improvements in computing power and big data
analysis have opened the door for major innovations in the
field of artificial intelligence; however, such advancements
have also given rise to the concern of extraneous information,
or “noise”, distorting the accuracy of predictive models. For
this reason, some doubters believe that machine learning will
quickly become an out-moded practice in the quantitative
investing domain.
In addition to being susceptible to noise, skeptics also
contend that successful AI trading strategies will eventually
be copied by other funds . As a consequence of widespread
replication, the asymmetric information edge originally
captured by the innovating fund will evaporate and
subsequently be fully priced into the market.
Artificially intelligent hedge funds are thus well-served to
experiment with a plethora of different machine learning
technologies and trading strategies, and should engage an
R&D tax credit service provider to benefit from each new
iteration and development.
Conclusion
Automated, artificial intelligence-based trading systems are
beginning to capture the attention of global hedge funds.
Artificial intelligence and, more specifically, machine
learning, can analyze large swaths of data with great speed
and adapt to new information the way old, static models
cannot. Federal and state R&D Tax credits are available to
stimulate the efforts in AI hedge fund innovation and should
be taken advantage of, accordingly.