Revisiting Directed Brokerage: Still No Free Lunch

grorrp COMMENTARY 57

DECEMBER 1998

RET'ISTTING DIRECTED BROI(ERAGE: STTLL

NCD

FREE LUNGH

A

7996 Plexus study of one clrent's drrected and non-dtrected trades concluded that whrle drrected trades had lower tmpact and commisslon cosrs, trmrng and opporlunity costs outwelghed the savings. More importanLly, drrected lradlng discourages lradrng rn raptdly movrng sltuattons. leadtng to significant underperformance relatlve to non-dtrected trades. Thls follow-up study of more clients and a longer time frame, comes to slmrlar conclusrons that [7J dlrected trades show no cost advantage when adjusted for expected costs, ancl [2J rndrrectly emphaslzrng lower cost trades leads to lower returns than reallzed throuqh non-dlrected trades.

How many sponsors would agree that

higher commissions can lead to

NYSE Rule 122 specifically prohibits a member firm from trading a stock with

many, we suspect. Sponsors who cap-

"...CgptUring 3( Of COtnSince the trader's mission cost the directture commissions iob is to control 'information broker at a time.

forfundadministra- ing sponsors 10.3( tion implicitly be- perform&nce..." lieve that a

higher

commission buys nothing of value. Consequently. rnanagers continue to

feel pressure to direct commissions to designated brokers for sponsor use. The burden ofproof that commissions

can be a tool, not just a cost, has fallen on the manager. Managers intuit that direction alfects investment perfomance, but seldom possess suf-

of ,

flow

to brokers in

a

way that secures best execution for all accounts, seqllencing of directed trades naturally results.

When prices move rapidly, timing becomes a priority. Accounts that direct trades to specific firms lose timing advantage whrle recapturing cornrnissions. How are we to assure this tradeoff is in the client's best interest?

ficient information to demonstrate that they are using commissions effectively. Measuring the true cost of direction requires looking beyond the obvious costs of commissions and direct impact. It needs to include timing costs

incurred while a directing broker seeks or awaits liquidity, as well as foregone returns when liquidity fails to develop. These costs capture the effects of disruptions to a manager's normal execution strategies. While proponents of direction argue that direction should not disrupt a manager's process, we show that differences resulting from the sequencing of trades can be significant.

The 1996 Plexus study of a growth manager's directed trading showed that:

o

section over time. The sample includes both large-cap and small-cap as well as value and growth strategies. ln 1Q98, the sample included $48 bil. of decisions of which 22% represented directed decisions.

Because the analysis covered

a

broader range of investment styles, the

effects of delays were reduced. Timing costs for directed trades fell, comparing more favorably to the non-directed trades. However, the benchmark costs for directed trades also fell. As a result, directed trades lagged their benchmark by 400%, compared to 8% for non-directed trades.

An interesting finding was that im-

Directed trades cost more than non-directed trades; trasted to 83 bp.

Plexus now has r"-rp to nine quarters of directed trading data for eight clients. This sample provides both a deeper pool for comparison and a good cross-

1

I

2 bp as con-

o

The ratio ofcosts to expected costs were more unfavorable for directed trades: 560 o ove r expectation compared to 40o over expectation.

o

Serious problems developed in experienced net returns, -3.l9oA for

directed trades vs. +0.470% for non-

directed. As a result, directed accounts realized lower returns for traded shares.

pact and commissions converged for the two groups of trades. Directed impact * commission costs rose from 25 to 2l bp, while non-directed costs fell from 39 to 33 bp. This is consistent with the steady drop in average commission rates due to increased use of Proprietary Trading Systems such as lnstinet and ITG/Posit. The question appears to be no longer the level of commissions; rather how the commissions are used.

Returns: tlre Bottorn Llne fon In the first study, we concluded

that the

most damaging aspect of directing brokerage was the 'adverse selection' incured when directed brokers fail to access liquidity in last moving situations. Because our first study covered a growth/ momentum-oriented manager, missed trades led to a large difference in retums" Adding more styles dampens the differential. but the perfonnace gap persists.

The significant difference for plan sponsors again occurred in the realized rates of return frorn trading. Market-

adjusted 30-day retllrns

bp

trades averaged

for non-directed-10 trades.

for

directed

versus +23 bp

The News!

On a cumulative basis, non-directed trades retumed +212 bp versus a loss of 87 bp for directed trades. Consistent with theory, directed trades out-

performed in only one period -- a quafter when average decision returns were negatlve.

Momentum-style managers and small cap managers appear to need all the broker flexibility and cooperation they can muster. Value managers, however, find time delay can be an ally, and directing may result in serendipitious delays. The exceptions are worth knowing to a sponsor who wants to effectively direct commissions.

Gonclusions

Take a look at our new Website at

www.plexusgroup.com. Find out

more about Plexus. study past commentaries, and soon access Plexus Group services online.

The "Plexus Group Sixth Conference Summary" is now arrailablelWrite to us for a free copy,

0r e-mail your request to info@plex usgroup.com.

Upcoming

Speeches

ary 25-26. New york: "New Technologies iunu

The table below summarizes the re- The lesson is that effective direction of sults. The benefits of recaptured com- trading requires careful planning. Plexus missions, assumed to be 3l per share, Group recommends: are added to gains which arise because smaller directed trades have lower com-

tWayne lhagner)

"Measuring Best Execution" (Steve Glass)

1' Direct only a modest portion of

mission and impact costs. oft'setting commissions(10-30%)basedonboth and capitalization' For these gains is the finding that directed manager style

their benchmark (of what they should cost). Finally, the performance of directed trades suffers when they are delayed or trades are more costly compared to

example' a large-cap value manager

can tolerate more directed trades than a small-cap value or large-cap growth

CBOE, Risk Management January 27 -30, Scottsdale: "Managing Transaction Costs" (lVayne Wagner)

manager'

too 2. Choose directed brokers in consultaquickly. The performance penalty tion with the manaser. abandoned because prices moved amounts to almost 15d per share.

3.

Use steo-outs as an alternative

Adding all the components together, the study shows that capturing 3p of com- 4. Consider whether the seruices purchased are essential to manager mission cost tlle directing sponsors performance. 10.31 per share of performance relative to non-directing accounts. 5. Install a monitoring program to evaluate executions and perfotmance.

len-Oiree+cC

TRADING PERFORMANCE

i+aCe+ i

0 0c

E

U

x'ro trodino

cos

t

over oenchm o'k

nderoerformonce ot oirecied trodes due to delovs

ond missed trodes

nHl;. 3 0c

3 0(

2C

40 0c

l6

4

50

l6

2C

il7e

l0

3c

-4

5C

l4

56

plexuggroue 11150 W. Olympic Blvd., #900

Los Angeles. CA 90064

PH: 310.3 12.55A5 FAX: 310.312.5506

,

.

Reprint any portion of this material with credit given to:

2C

8C

Also available at rvrvw.p lexus group. c om.

@

Plexus Croup, 1998