customer retention in the cover headline here age...

Month 2015

Cover Headline Here (Title Case)Cover subhead here (sentence case)

Q1 2019

Customer Retention in the Age of Electronic Trading

Executive SummaryBy mining our historical data on dealer-client relationships, Greenwich Associates found that they are becoming significantly less stable over time and that the relative stability advantage held by the bulge-bracket firms is eroding. We believe this is due to the increasingly sophisticated ways in which the buy side is scrutinizing their executions, using tools originally designed to prove best execution and now deploying them for trading optimization.

This has significant lessons for the sell side. First, for the top firms, the depth of client relationships is less important than the breadth of the client base. And all firms have a strong interest in reducing the volatility of their customer market share. It is likely broker-dealers will adopt their clients’ methods but in reverse—using their greater volumes to model client behavior, enabling them to intervene early.

METHODOLOGY

This report draws on several sets of Greenwich Associates data as well as 15 interviews conducted by senior analysts with asset managers in the United States and Canada. Respondents were asked about their current trade analytics processes as well as their expectations for the future evolution of trade analytics. The primary data set is derived from the annual Greenwich Associates U.S. Equity Investors study—in 2018, there were 275 respondents. The particular data set used for this analysis examines the pairwise broker-dealer client relationships from 2013–2014 and compares them to 2017–2018. Over 2,200 pairwise relationships are included in this data set.

CONTENTS

2 Executive Summary

3 Introduction

4 The Erosion of Client Loyalty Quantified

6 Lessons to Learn

6 TCA: From Cursory Glance to Electron Microscope

9 The Relationship is in the Details

11 Conclusion

2 | GREENWICH ASSOCIATES

Ken Monahan is a senioranalyst on the MarketStructure and Technologyteam, specializing in electronic trading, regulation and fixed-income markets.

WHILE THE AGGREGATE MARKET POSITIONS OF THE TOP BROKER-DEALERS REMAIN STABLE, THE MEDIAN CLIENT REALLOCATES ITS FLOW AMONG ITS BROKER-DEALERS SIGNIFICANTLY

BETWEEN 2014 AND 2018, THE MEDIAN SIZE OF CLIENTS’ BROKER-DEALER VOLUME REALLOCATIONS IN U.S. EQUITIES INCREASED BY

50%


IntroductionFor over 40 years, Greenwich Associates has been interviewing the buy side about the service they receive from their brokers in equity markets. One of the most popular uses of this data is the Greenwich Share Leaders. What league tables are to investment banking, Greenwich Share Leaders are to sales and trading. Wall Street is a very competitive place and keeping tabs on who’s on top, who’s up and who’s down is in the blood. What’s more, these are businesses with high fixed and low marginal costs, so these Share Leaders are almost certainly the profit leaders. People love to watch the money.

People actually running these businesses, however, not only want to know who is making the money but also, and more importantly to them, how the money is made. How do the Share Leaders do it? To try to understand this, Greenwich Associates has analyzed 2,200 pairwise broker-client relationships across two time periods. What lies beneath the surface is remarkable: The stability at the top of the Greenwich Share Leaders belies substantial instability at the level of individual broker-client relationships.

Note: Based on responses from 300 buy-side institutions in 2017 and 275 in 2018, weighted by commission spend of accounts. 1Greenwich Associates Trading Share represents a broker’s relative importance to the buy-side institutions within the Greenwich Associates universe. Scores are based upon the amount of business conducted with each respondent and the size of each responding institution based on commission spend with the sell-side community. 2Equity Trading encompasses sales trading, execution, sector trading, and electronic trading. Top five leading brokers are cited including ties in 2017 and top three in 2018. Source: Greenwich Associates 2017 and 2018 U.S. Equity Investors Studies

Greenwich Share Leaders — 2017 and 2018

J.P. Morgan

Morgan Stanley

Goldman Sachs

Bank of America Merrill Lynch

Credit Suisse

Broker

8.5%

8.2%

7.9%

7.5%

6.2%

Trading Share

1T

1T

1T

4

5

Statistical Rank

2017 U.S. Equity Trading Share1, 2

J.P. Morgan

Morgan Stanley

Goldman Sachs

Bank of America Merrill Lynch

Broker

9.0%

8.1%

7.4%

7.1%

Trading Share

1

2

3T

3T

Statistical Rank

2018 U.S. Equity Trading Share1, 2

The stability at the top of the Greenwich Share Leaders belies substantial instability at the level of individual broker-client relationships.


The Erosion of Client Loyalty QuantifiedGreenwich Associates has historical data on the broker-dealer allocations of thousands of clients. To study this, we examined 1,100 broker-client relationships between 2013 and 2014 and another 1,100 in 2017–2018. We looked at each client’s volume allocation across all of its execution broker-dealers and measured the change in share allocation per broker-dealer year over year, in absolute-value terms. That is, if a firm changed its allocation to a given broker-dealer from 10% to 15%, that’s an increase of 50% in allocation-share terms. This represents a normalized measure of the stability of broker-client relationships. We also divided the broker-dealers in our sample into “bulge bracket” and “non-bulge bracket” groups. This enables us to look at the role that scale plays in stabilizing client relationships. The results were remarkable.

The best way to get a handle on the data is to look at it graphically. A number of changes occur in the shape and the center point of the distributions of the total sample from ‘13–’14 to ‘17–’18. The mean and variance of the ’13–‘14 distribution are dominated by outliers, a small number of very large reallocations, so it is better to focus on the medians, the center of gravity of the distribution. The ’17–‘18 distribution is less influenced by outliers but the center of gravity shifts significantly, from 22% to 36%. What is happening here is that extreme changes in market share are becoming less common, but the most common changes in market share are becoming more extreme: The median allocation to any given broker-dealer changes by a more than a third year over year.

These effects are still more interesting when examined on a segmented basis. What we see in the ‘13–’14 period is that the distribution of reallocations among the bulge bracket is both significantly lower as well as much more stable. The median change for the bulge bracket is 21% versus 28% for the non-bulge. This seems intuitively correct, as the bulge bracket firms, by definition, touch the buy side at more points and so have deeper, and presumably stickier, relationships with them.

ABSOLUTE VALUE OF CHANGE IN ALLOCATION PER DEALER—CLIENT RELATIONSHIP—TOTAL MARKET

Fre

qu

ency

% Reallocation10%

Note: Based on interviews with 316 U.S. traders in 2014 and 275 in 2018.Sources: Greenwich Associates 2018 North American and 2014 U.S. Equity Investor Studies

2017–20182013–2014

0%

10%

20%

30%

40%

20% 30% 40% 50% 60% 70% 80% 90%

2% 1%

100%

2% 1%

More

7%2%

6%

13%

5% 7%2% 4%

16% 16%

9%

17%

26%

13%

19%

7%

14%

2013–2014median

2017–2018median

22% 36%

11%

The median allocation to any given broker-dealer changes by more than a third year over year.


Looking at the relative position of the two segments in the ‘17—’18 period is very instructive. The distribution of the non-bulge bracket firms contracts notably as the standard deviation and the mean both decline substantially. However, the center of gravity for both groups increases significantly: The bulge-bracket median jumps from 21% to 35%, and the non-bulge moves from 28% to 39%. Across the industry, dealer-client relationships are much less stable. What is perhaps the most significant finding of this analysis is that the stability advantage enjoyed by the bulge bracket firms has almost evaporated in just four years.

It is interesting that while this is clear evidence of substantial reallocations among broker-dealers on a client-by-client basis, it is completely opaque to anyone looking only at the rankings. Between 2017 and 2018, there was no change at all in the rank ordering of the Greenwich Share Leaders, even though the median client relationship was reallocating a third of its volume. What can account for the top-line stability given all the churn beneath the surface? What can be learned from this?

BULGE BRACKET VS NON-BULGE BRACKET CLIENT REALLOCATIONS—2013–2014

Fre

qu

ency

10%

Non-bulge bracketBulge bracket

0%

10%

20%

30%

20% 40% 50% 60% 70% 80% 90% 100% More

Note: Based on interviews with 316 U.S. traders.Sources: Greenwich Associates 2014 U.S. Equity Investor Study

30%

1% 2% 1%3%

5%

12%

7%5% 4% 5%

2% 2%

15% 16%

9%7%

29%

18%

5%

9%

% Reallocation

Bulgebracketmedian

Non-bulgebracketmedian

21% 28%

20%20%

BULGE BRACKET VS NON-BULGE BRACKET CLIENT REALLOCATIONS—2017–2018

Fre

qu

ency

% Reallocation10%

Non-bulge bracketBulge bracket

0%

10%

20%

30%

20% 30% 50% 60% 70% 80% 90%

2%

100% More

Note: Based on interviews with 275 U.S. traders.Sources: Greenwich Associates 2018 North American Equity Investor Study

40%

1% 0%2% 1% 2%

13% 14%

6%9%

5% 4%

19%

12%

19%16%

11%15%

11% 10%14% 15%

Bulgebracketmedian

Non-bulgebracketmedian

35% 39%


Lessons to LearnThe lesson for the bulge bracket is the vital importance of scale. The continued dominance of the Greenwich Share Leaders shows that their leadership is driven as much by the breadth of their client base as by the depth of the service. For the large firms with broad client bases, many of the individual reallocations cancel one another out, dampening its aggregate effect. It’s also a warning, however, that the breadth of service provision, and the relationship it has supported, is losing its power over allocations, relative to firms that are focused only on execution.

The implications for non-bulge bracket firms are also significant. The decrease in loyalty means that the chances to win business away from the larger firms are increasing. On the other hand, since they generally have fewer client relationships, those firms that are not able to scale up their breadth of relationships to offset the increasing restlessness among the clients face significant risks. They have a powerful interest in both onboarding new clients and in significantly reducing the degree to which their own clients reallocate their flow. For these firms, improvements in service provision are a matter of life and death.

TCA: From Cursory Glance to Electron MicroscopeFor firms looking to either defend themselves against reduced client loyalty or to lever it to capture share, it’s important to know what is driving the reduction. This is a complicated task, as the data is very descriptive, but significantly less prescriptive. Still, some likely drivers stand out more than others.

Greenwich Associates data shows the expanded use and improved quality of electronic execution is a likely factor. Over the same period that we see a major convergence between the bulge and the non-bulge bracket firms, we see a significant advance in measures of customer satisfaction with electronic trading for non-bulge bracket firms. We ask our research participants to rank the degree to which their e-trading needs are being met by their broker-dealers. Two things have happened: First, the standards have gotten higher—the level of satisfaction that earned the top spot in 2014 doesn’t make the top five in 2018. In addition, non-bulge bracket firms now dominate the rankings. Given the resource disparity between the two groups of firms, these results likely represent a significant increase in investment by the non-bulge bracket firms.

The depth of relation-ship associated with breadth of service is losing its power over allocations, relative to firms that are focused only on execution.

The decrease in loyalty means that the chances to win business away from the larger firms are increasing.


This is also borne out by recent Greenwich Associates data on the buy side, which demonstrates the increasing sophistication with which they are evaluating their sell-side broker-dealers. Transaction cost analysis has evolved from being a check-the-box exercise to show compliance with best-execution requirements, to being central to the trading process. Indeed in 2018, “evaluating trading effectiveness against benchmarks” overtook “oversight and reporting” as the most common use to which TCA is put. Additionally, over half the firms in our study use it to measure trading effectiveness.

This is apparent in the qualitative interviews we conducted for this study as well. Sophisticated buy-side firms are faced with a serious challenge of proliferating algorithms. A large, real-money buy-side firm might have 20 broker-dealers , each offering a suite of 10 algos. Such a firm is in the simultaneously enviable and difficult position of having to evaluate 200 algos. What it might do is choose the style (aggressor, VWAP, etc.) of algos they use most often, and then use an algo wheel to randomize the implementation of any given firm’s algo when that specific style is called for. They then use their TCA plan to evaluate the executions.

GREENWICH ASSOCIATES ELECTRONIC TRADING SATISFACTION RANKINGS—U.S. EQUITIES

Firm Name Point Score Rank Firm Name Point Score Rank

Non-bulge

Non-bulge

Non-bulge

Non-bulge

Non-bulge

Bulge

Bulge

Non-bulge

Bulge

Non-bulge

60

58

58

56

56

55

55

54

53

53

1

2T

2T

4T

4T

6T

6T

7

8T

8T

Non-bulge

Non-bulge

Non-bulge

Non-bulge

Non-bulge

Bulge

Non-bulge

Non-bulge

Non-bulge

Non-bulge

69

65

63

62

61

61

60

60

59

58

1

2

3

4

5T

5T

7T

7T

9

10

Note: Based on interviews with 316 U.S. traders in 2014 and 275 in 2018. Point scores are calculated based on the quality of a broker-dealer’s electronic trading service and capabilities. Respondents are asked to rank their broker-dealers on a 1–5 scale, where 5 is worth 100 points, 4 is worth 50 points, 3 is worth 25 points, 2 is worth 12.5 points and a 1 is worth 0 points. Reported Point Scores show the average score for each broker-dealer.Sources: Greenwich Associates 2014 and 2018 U.S. Equity Investor Studies

2014 2018

Two things have happened: First, the standards have gotten higher—the level of satisfaction that earned the top spot in 2014 doesn’t make the top five in 2018. In addition, non-bulge bracket firms now dominate the rankings.

In 2018 “evaluating trading effectiveness against benchmarks” overtook “oversight and reporting” as the most common use to which TCA data is put.


The number of orders executed by buy-side shops is small—on average under five per day per billion dollars under management. So to properly evaluate their broker-dealer, it’s necessary to gather and aggregate execution data over time. This used to be difficult, but no more. The universal use of TCA has now made it possible for firms to collect trans-action data over time on a per-broker-dealer, indeed on a per-algo, basis. This data is then used to determine the execution quality and to reallocate their flow accordingly.

Greenwich Associates data also shows that the buy side is using increasingly sophisticated criteria to judge the sell side. For example, we asked which aspects of their execution, if improved, would have

USE OF TCA IN THE INVESTMENT PROCESS—U.S. EQUITY TRADERS

A post-trade review to analyzetrader/trade e�ectiveness

against benchmark(s)

Venue analysis

Alpha profiling

Oversight/Reporting

Trade surveillance

As a pre- and post-trade tool tomodel/forecast trades and to analyze

trader/trading e�ciency and e�ectiveness

A pre-trade modeling/forecasting practiceto anticipate the outcome of a trade

Real-time, in-trade analytics

To help determine trader compensation

92%73%

Note: Based on 50 respondents in 2018 and 44 in 2016.Sources: Greenwich Associates 2018 and 2016 Market Structure & Trading Technology Studies

20182016

86%84%

60%48%

48%N/A

46%41%

44%N/A

42%25%

32%30%

26%11%

EXECUTION QUALITY FACTORS WITH GREATEST IMPACTON PROFITABILITY

Fill rates 60%

Latency 27%

Cancellation rates 20%

Reject rates 13%

Cancellation speed/notification 13%

RFQ response time 13%

Note: Based on 15 responses.Sources: Greenwich Associates 2018 TCA to TQA Study

We asked which aspects of their execution, if improved, would have the greatest effect on their profitability. The No. 1 response was “fill rates,” followed by latency and cancellation rates.


the greatest effect on their profitability. The No. 1 response was “fill rates,” followed by latency and cancellation rates. These are remarkably technical aspects of execution. The fact that they are so well understood is evidence that the spread of TCA, and its integration into the trading process, is raising the bar for the sell side. It is also likely the reason behind the decreasing stability of the dealer-client flow allocations we see in the data—because clients armed with data can be more confident in their decision-making.

The Relationship is in the DetailsTo compete effectively, sell-side firms must, therefore, pay significant attention to improving their own technological infrastructure, so it can withstand the greater scrutiny the buy side is employing. They would do well to apply the same level of rigor to monitoring their own processes as the buy side now applies to the result. In reality, they must apply even more rigor. This is because while the buy side can execute in haste and transaction cost analyze at leisure, the sell side must execute in haste but cannot conduct a performance analysis at leisure. It has to be live and, ideally, predictive. Given the increasing proclivity of the buy side to reallocate their flow, issues have to be dealt with before they become large enough for the TCA engines to detect them.

Luckily for the sell side, because they sit at the center of a nexus of clients, their throughput is significantly larger. So just as the buy side can now make use of the profusion and advance of TCA tools, so too can the sell side make use of advances in data science and machine learning. Sell-side firms that have developed the capacity to monitor their message traffic across the order lifecycle can identify errors as they occur and seek to address them immediately. Take, for example, the main issue that clients say would enhance their profitability: increased fill rates. This makes sense, as firms in the TCA field have recently begun to focus on “delay” costs—the implicit costs of not transacting. A sell-side firm that is effectively monitoring its message flow can monitor its fill rates across venues. In the event of a decrease in fill rates at a particular venue, it can begin routing customer flow away from that venue and toward others where the fill rates are higher.

The ability to harness message monitoring and to combine it with machine learning has the potential to take this one step further. Given the large data flows, it is possible to train machine-learning applications to look for message patterns that have, in the past, led to changes in fill rates. This would enable a sell-side firm to proactively begin rerouting order flow away from a venue at which it has a reasonable expectation

While the buy side can execute in haste and conduct TCA at leisure, the sell side must execute in haste but cannot conduct a performance analysis at leisure. It has to be live and, ideally, predictive. Given the increasing proclivity of the buy side to reallocate their flow, issues have to be dealt with before they become large enough for the TCA engines to detect them.


of near-term changes in fill rates. Utilization of machine learning for this kind of predictive issue resolution would be the Holy Grail of sell-side execution. It returns the broker-dealer to the status quo ante—the buy side’s increased propensity to reroute flow, driven by ever more sophisticated performance measurement, would be rendered impotent because execution issues would be addressed before they were capable of being detected by the most sophisticated tools.

What’s more, it’s possible that the sell side could reverse the rerouting tide still more by using message flows and machine learning to turn the lens back on their client’s activity. This is because just as someone looking only at Share Leaders will miss the competition below, a firm looking only at final executions will miss the execution components, fill rate, latency, and cancellation rates that drive client decisions. Better scrutiny could discern changes in client message traffic across the order lifecycle that might presage changes in how those clients allocate their flow to a particular broker-dealer. If such a link exists, and it can be modeled predictively, it would enable firms to take action to address the root causes of those shifts or to strengthen the relationship before the business moves elsewhere. Timely interventions by salespeople might reduce the level of volatility still further, bringing the technological revolution in trading full circle—by enhancing the importance of relationship management.

HIGH-FREQUENCY TRADING

It’s worth noting that in the same way that advances in buy-side TCA have pressed the sell side to improve their monitoring, improved capacity on the sell side to redirect their flows based on their real-time monitoring has made it necessary for liquidity providers to up their game as well. This is because they stand to lose access to order flow in the event that the metrics by which they are measured—fill rates for example—fall off.

In many ways, this issue is even more complex for trading firms. They have to monitor not only their own transaction data, but also reams of market data both on the markets on which they’re pricing as well as on markets that affect the pricing and risk models they are using. Proprietary trading firms have long been at the technological leading edge, and message monitoring has been something they’ve been doing as a matter of course. Today, they’re able to leverage external vendors in order to focus their technology resources on the more relevant ends of the business or to smooth their way into new asset classes as opportunities arise.


ConclusionWhile the rank ordering of the Greenwich Share Leaders has been remarkably stable, beneath the surface buy-side firms are increasingly willing to reallocate their order flow, and at the median, routinely change a given broker-dealer’s share by a third year over year. This is likely driven by an increase in the sophistication with which they are able to monitor broker-dealer performance and determine what execution factors affect their profitability.

This has important implications for the sell side. The larger firms are able to maintain their market position by spreading the risk resulting from the increased level of reallocated flow out among a large number of clients. Non-bulge bracket firms, which lack the cushioning effect of scale to protect them from reduced loyalty driven by sophistication, have had to focus on their technology stack—and there is evidence that they have. More to the point, the sell side can elide the greater scrutiny they are under by monitoring their own messages and using error detection and machine learning to address issues before they rise to the level of being detectable by the clients.

Cover Illustration: © iStockphoto/yongyuan

The data reported in this document reflect solely the views reported to Greenwich Associates by the research participants. Interviewees may be asked about their use of and demand for financial products and services and about investment practices in relevant financial markets. Greenwich Associates compiles the data received, conducts statistical analysis and reviews for presentation purposes in order to produce the final results. Unless otherwise indicated, any opinions or market observations made are strictly our own.

© 2019 Greenwich Associates, LLC. Javelin Strategy & Research is a division of Greenwich Associates. All rights reserved. No portion of these materials may be copied, reproduced, distributed or transmitted, electronically or otherwise, to external parties or publicly without the permission of Greenwich Associates, LLC. Greenwich Associates,® Competitive Challenges,® Greenwich Quality Index,® Greenwich ACCESS,™ Greenwich AIM™ and Greenwich Reports® are registered marks of Greenwich Associates, LLC. Greenwich Associates may also have rights in certain other marks used in these materials.

greenwich.com [email protected] Ph +1 203.625.5038 Doc ID 19-2016

customer retention in the cover headline here age...

Documents