Algorithmic Trading Survey - Hedge Fund 2023
Algorithmic Trading Survey - Hedge Fund 2023
Algorithmic Trading Survey - Hedge Fund 2023
N UT R | V H
E YE ]D ]
THE
2023
ALGORITHMIC
TRADING
SURVEY
HEDGE FUNDS
With special analysis
from Datos Insights
Issue 76 // thetradenews.com // 61
Issue76
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
W
hile the hedge is clear that market challenges are productivity, ease of use, speed,
fund industry far from over, making the results and cost. The final two categories,
was unable from The TRADE’s Algorithmic speed and cost, replaced execution
to escape the Trading Survey and the factors consistency and customer support
rampant recession fears and impacting the use of algos and in the top five highest rated aspects
market volatility that echoed the perceptions of quality and of service for 2022.
throughout global markets in execution all the more timely. Of the fifteen categories assessed
2022, there are some silver linings. The overall score of hedge in the survey, 10 saw year-over-
According to the Barclay Hedge fund respondents to the 2023 year decreases in their overall
Fund Index, hedge funds lost Algorithmic Trading Survey ratings (Figure 1). The largest
8.59% in 2022 which may not declined to 5.74 from 2022’s drop was observed in execution
seem great except when compared score of 5.80, marking the lowest consulting which decreased 30
to the S&P 500 Total Return overall score by hedge fund basis points to 5.45, followed by
Index which fell 18.14%. In fact, respondents since 2019’s survey customer support which decreased
2022 marked the first time in (5.72). Respondents were based 20 basis points to 5.82. Providers
over 14 years where hedge funds primarily in the UK (43%), Europe are well aware that while access,
outperformed the S&P 500 in a (31%), North America (12%), speed and ease of use are table
calendar year. Overall industry and rest of world (11%), with a stakes, the relationship is the
performance however was once handful coming from APAC (4%). differentiator. Rather than a
again truly nonlinear, with the The most impactful features reflection of providers pulling
largest hedge funds and those of algorithms for hedge fund back on the personal services and
deploying macro strategies coming respondents differed some from high touch they are providing
out ahead of their counterparts. last year and were identified as clients, we believe there are two
As we move further into 2023, it dark pool access, increased trader factors at play driving this notable
62 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
ALGO 2023
Figure 1: Rating of algo performance ALGO 2022
ALGO 2021
Aspect of service Weighted average score
5.89
Ease of use
6.00
6.06
Speed 5.87
5.90
5.93
5.85
Cost 5.86
5.79
5.84
Execution consistency
5.96
5.76
5.82
Customer support 6.02
6.04
5.79
Reduce market impact 5.89
5.76
5.75
Flexibility/sophistication of SOR 5.68
5.78
5.69
Anonymity 5.68
5.85
5.62
Order routing logic/analysis 5.57
5.75
5.59
Price improvement 5.56
5.55
5.57
Algo monitoring capabilities 5.70
5.70
5.47
Customisation features 5.61
5.67
5.45
Execution consulting 5.75
5.60
5.20 5.30 5.40 5.50 5.60 5.70 5.80 5.90 6.00 6.10
Issue 76 // thetradenews.com // 63
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
decrease. First, the talent wars Figure 2: Reasons for using algorithms (% of responses)
created a high degree of turnover Feature 2023 2022 2021
and loss of expertise that has
Ease of use 12.03 12.06 11.21
played out across industries. The
loss of both key contacts and the Reduce market impact 10.87 11.58 10.78
expertise they developed around Increased trader productivity 9.33 10.24 8.68
specific solutions is something we
Consistency of execution performance 8.83 10.42 10.29
are seeing more broadly. Second,
buy-side desks are becoming Higher speed, lower latency 8.44 6.99 8.28
increasingly sophisticated and Customisation capabilities 8.37 6.37 8.47
their expectations for the support
Better prices (price improvement) 8.25 7.36 9.27
providers can offer is likely in
many cases running ahead of the Lower commission rates 7.67 8.17 7.02
capabilities of those providers. Greater anonymity 7.48 7.55 6.90
Although ranking in the bottom
Flexibility and sophistication of SOR 6.28 5.02 5.91
half of the service aspects assessed
in the survey, the largest increases Algo monitoring capabilities 6.05 7.50 7.18
were seen in flexibility and Data on venue/order routing logic or analysis 5.17 4.46 4.16
sophistication of SOR (+.07) and
Results match pre-trade estimates 1.23 2.28 1.85
order routing logic/analysis (+.05).
In fact, some respondents made
it a point to note that they want (+2.00%), higher speed/lower rates. Additionally, firms in this
to see additional features such latency (+1.46%), and flexibility demographic, who typically have
as real-time routing and venue and sophistication of SOR larger positions to move, continued
analysis tools, automatic merging (+1.26%). to rank greater anonymity among
of similar orders and increased When it comes to firm size, the their top five reasons for using
conditional trading functionality. results are generally consistent algorithms. Anonymity was ranked
Respondent’s reasons for using across respondents, although ninth most important by all hedge
algorithms are presented in Figure there were a few exceptions. In fund respondents overall.
2 as a percentage of respondents contrast to the broader survey The 2023 survey reveals that
for years 2021-2023. The top results, firms with over $50B in hedge funds on average, decreased
reasons for using algorithms AUM responding to the survey the number of algo providers
as indicated by hedge fund tended to put less emphasis on they use (Figure 3). This marks
respondents remain unchanged higher speed/lower latency and a reversal from last year’s survey
from last year. They are ease customisation capabilities and results that showed firms of
of use (12.03%), reduce market instead increased their response all sizes either increasing or
impact (10.87%), increased rate in areas like algo monitoring maintaining the number of algo
trader productivity (9.33%), capabilities and lower commission providers used. The decline in
and consistency of execution
performance (8.83%). Despite
remaining the primary reasons Figure 3: Average number of providers used by AuM (USD billions)
for using algorithms, all four of Feature 2023 2022 2021
these categories did experience a
decline in their percentage rates More than $50 billion 5.17 3.94 3.94
compared to the prior year. In $10 - 50 billion 5.00 5.52 4.00
fact, consistency of execution $1 - 10 billion 4.89 5.39 3.90
performance had the largest
year-over-year decline of any $0.5 - 1 billion 4.60 3.89 3.36
category (-1.60%). The largest $.025- 0.5 billion 1.67 2.60 2.20
increases in reasons for using Up to $0.25 billion 2.25 3.33 1.80
algorithms however were seen
in customisation capabilities Not answered 3.71 4.00 3.78
64 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
ALGO 2023
Figure 4: Number of providers used (% of responses) ALGO 2022
ALGO 2021
32.14
1-2 34.96
35.82
19.05
3-4 19.51
23.13
48.81
>5 45.53
41.04
0 10 20 30 40 50
the number of providers reflects this trend is likely multi-faceted. those that answered >80% (-6.99).
a long-term trend disrupted by First, while the survey did not This means of course that the
the 2020-2021 pandemic in which ask about prime brokerage brackets representing between 0
firms are looking to consolidate relationships, decisions around and 40% of portfolio value traded
relationships. This year’s data the use of different providers is via algorithms experienced year-
once again shows some anomalies likely correlated to both prime over-year increases in percentage
however, that indicate firms broker relationships and the use of of respondents.
are continuing to evaluate new disparate strategies. More broadly The drop in the percentage of
providers and their solutions. The there are two conflicting business respondents that use algorithms
increased sophistication of the imperatives that are driving the to trade the majority of their
buy-side trading desk will continue addition and consolidation of value was even more dramatic
to evolve, and that evolution providers. On the one hand, funds when it came to the largest firms
creates opportunities and are occasionally driven to diversify responding to this year’s survey.
competitive threats for providers. vendor exposure and capabilities The results show that almost 49%
When AUM is removed from the by engaging in additional provider of respondents from firms with
equation, the number of providers relationships. On the other over $50B in AUM used algos to
used by hedge funds continues to hand, with providers generally trade the majority of their value
show a barbell like distribution, improving their offerings over the in 2023 compared to 75% the year
with little change from prior years years, there have been efforts to before. This could be a result of
(Figure 4). Most respondents fall consolidate relationships that have hedge funds shifting their holdings
into the camp of using either one occurred on both sides. increasingly toward asset classes
to two providers (32%) or more According to this year’s survey, that are not typically traded via
than five (49%). The remaining less than half (45%) of hedge algos. In addition, as we see funds
19% of respondents use a number fund respondents say they use move toward multi-asset trading
in between. Although it is perhaps algorithms to trade the majority of strategies, survey respondents now
slow moving it is interesting to their portfolio value (Figure 5), a include entries for instruments
note that over the last few years, notable decrease from the almost outside of just equities, such as
the gap between the percentage 66% that claimed to do so the year fixed income, FX, and ETF’s,
of respondents that use one to before. It’s therefore unsurprising which they typical use algos for
two providers and those that that when broken down into less often than they do equities.
use more than five has gotten smaller brackets one can see that In 2023, we see hedge funds
wider. In the last three years, the all bar one over the 50% threshold, decreasing their use of nearly
percentage of respondents using the 70-80% bracket, experienced every type of algorithm (Figure
one to two providers has slowly year-over-year declines. The 6), which is in direct contrast to
decreased while the percentage largest was seen in the percentage last year’s results but in line with
of those using more than five has of respondents that said they trade this year’s drop in overall algo
slowly increased. The reason for between 60-70% (-13.24), and performance ratings as well as the
Issue 76 // thetradenews.com // 65
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
marked decline in the percentage Figure 5: Algorithm usage by value traded (% of responses)
of portfolio value traded with Feature 2023 2022 2021
algos. The only exception to
Not answered 2.67 1.63 5.14
this is dark liquidity seeking
algos and “other”, which saw an 0 - 5% 14.67 6.50 3.70
increase of almost 5% and 2.3% 5 - 10 % 5.33 3.25 6.17
by respondents year-over-year.
10 - 20% 14.67 5.69 4.73
The algorithms with the largest
declines in usage by hedge funds 20 - 30% 6.67 4.07 9.05
were target close/auction (-14.9%), 30 - 40% 5.33 4.07 8.85
TWAP (-10%), and VWAP (-9.4%).
40 - 50% 5.33 8.94 8.44
While the movement away from
participation-based algorithms 50 - 60% 13.33 15.45 7.82
reflects the growing sophistication 60 - 70% 9.33 22.76 12.76
of the buy-side trading community,
70 - 80% 9.33 7.32 12.96
an overwhelming majority of
hedge fund respondents continue >80% 13.33 20.33 20.37
to use dark liquidity seeking
(84.5%), VWAP (70.2%), and %
Figure 6: Types of algorithms used (% of responses)
volume (69.1%) algorithms to
fulfill their trading needs. Feature 2023 2022 2021
We also see that amongst the Other 7.14 4.88 6.58
largest, and primarily hybrid
Implementation shortfall (basket) 15.48 17.89 10.91
traditional/hedge players, an
even distribution of Bloomberg TWAP 27.38 37.40 27.16
EMSX, FactSet Portware, and Implementation shortfall (single stock) 45.24 45.53 44.24
FlexTrade being used to access
Target Close/Auction Algos 45.24 60.16 no data
algorithms. Amongst those with
less than $50 billion in assets % Volume (participation) 69.05 69.11 56.38
under management however, VWAP 70.24 79.67 58.64
Bloomberg EMSX continues to be
Dark liquidity seeking 84.52 79.67 67.49
the dominant EMS platform.
66 // TheTRADE // Q2 2023
In association with
SAVE
THE SIX Swiss
DATE
Exchange
LEADERS IN
2023
TRADING
Celebrating excellence in trading
and execution since 2010
#leadersintrading
8 November, 2023,
The Savoy, London
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Berenberg
68 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Bernstein
Issue 76 // thetradenews.com // 69
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
70 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Citi
KEY STATS
Issue 76 // thetradenews.com // 71
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Goldman Sachs
72 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Jefferies
Issue 76 // thetradenews.com // 73
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
JP Morgan
74 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Kepler Cheuvreux
Issue 76 // thetradenews.com // 75
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Liquidnet
76 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Morgan Stanley
Issue 76 // thetradenews.com // 77
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Redburn
78 // TheTRADE // Q2 2023
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
UBS
Issue 76 // thetradenews.com // 79
[ A L G O R I T H M I C T R A D I N G S U R V E Y ]
Others
BARCLAYS RATINGS FOR ALGORITHMIC PERFORMANCE
80 // TheTRADE // Q2 2023