The Maps Are Being Redrawn

Building COMMENTARY

Betler

9rorrP, Perf ormernce

83

MARCH

2OO5

THE MAPS ARE BEING REDRAWN' "The world's biggest iceberg has crashed into an ice tongue protruding out of Antarctica, breaking a portion off, the European Space Agency said."

"The B-15A iceberg, about the size of Luxembourg, in the past month hit the Drygalski ice tongue, a feature marked on most maps of Antarctica, and broke a five-kilometer- (three-mile-) long section off, ESA said on its Web site, using satellite pictures to illustrate the damage." "Maps of Antarctica need to be amended," ESA said." wvwv.bloomberg.com, April 1 Bth 2005

The force behind this movement is pressure to reduce institutional trading costs. This pressure has three driving forces: (1) modeling, derived from academic

work, (2) motivation, derived from

pressing institutional needs, and (3) methodology, to define, implement and validate improvements to trading practices.

Modeling Jack Treynor and Andre Perold deserve credit for the implementation shortfall (iceberg) approach. This is the key insight, because it casts costs in the framework of a cost:benefit model. The old paradigm considered costs as something to be minimized. In effect, all other factors other than broker behavior were assumed to be fixed and unresponsive to trading strateEy and tactics. By linking trading costs to performance, Perold recast the problem

from one of cost control to one of prudent spending to achieve the investment goals. In addition, by creating an all-inclusive cost measure, the door was opened to the

Plexus Group's symbol for the hidden trading costs has been the icebero. which leads us to use the above geographical alteration as an analogy to what's happening with the 'T'tz* **t'*s***re? ##r??ss?gy ti*xreir; :xf"rtrpgf U.S. stock exchanges. In one week the NYSE announced a merger with ,S,tr#,€ #rpfl$ tr*S#*.#$s:psfd#*f arfff ARCA and the NASDAQ a purchase

*

FdY',$#

r**rx.#*

of the lnstinet system. What has caused this sudden flurry of exchange

#{#c:g;ra#pfuy

mergers? We see two goals; further reducing the cost of trading through the use of better technology along with accessing new markets. We believe this is the next logical step in response to forces that have been active for decades

*f fft* #tr#fldd *f

fFa*

cs?w*s$setgr"

discovery of the large but heretofore undiscovered hidden costs of trading: the iceberg. With this new l\,4any thanks to Peter Haynes of TD Newcrest Securities, who brought the iceberg news item to our attention and first developed the analogy to the chanqes in the market.

paradigm, it became possible to engineer changes to improve trading results.

Motivation The CFA Institute picked up on this same idea when they defined Best Execution as the trading process most likely

to maximize the value of client portfolios.2 Once an investment management firm truly internalizes the connection between costs and performance, high-cost trading impediments become intolerable. In recent years, institutional investors, led by -i./anguard and Fidelity, have intensified pressure on the exchanges to develop more cost-effective solutions. The irresistible iceberg of institutional trading needs broke the defense of the entrenched interests.

enabled us to hold costs constant in the face of massive growth in assets under management. Assuredly, both the investing environment and the market structure are much different than they were in the 70's and 80's.

Changes in the brokerage industry have led to much lower liquidity costs as long as prices remain stable. But as stability wavers, costs rise. Indeed, when adjusted for momentum, the cost reduction shrinks. One well known investment manager shared with us an intuition that trading costs haven't fallen at all -- they've just moved from one bucket (fixed commissions) to other buckets (impact and delay.) There's some truth to that insight, according to the Plexus data. We have found, for example, that the costs of trading small cap stocks haven't changed much at all, although the proportion of total trading in this category has dropped significantly since the Internet heyday. 200

Measu rement 100

In 1975, Cuneo and Wagner published an estimate of round trip costs of 2.6% for information motivated trades. This was in line with Demsetz's '1963 * the first we are

----r--+ -200 -100

aware of - estimate and Krause and Stoll's 1972 estimate. In Plexus' very first depiciion of an iceberg in 1992, the round trip costs were measured real change over 30 years.

-1

at2.B'h no :c .E

Turning to the present, current Plexus client data shows all-in round trip costs at an all time low of 1"5%. So it appears at first blush that costs have been cut roughly in

: .F

e

00

-200

.cl

&oo

200

INDMS $ev SCGS

-300

half since the turn of the millennium, with the big reductions starting to appear in 2003 as the effects of

-400

400

300

LCG

*

Return (bp)

i

decimalization began to kick in. Not so fast! As recently as 200'1, at the end of the Internet frenzy, costs peaked out at 3.1% round trip, much higher than current costs and higher than the cosl estimates of the 1960's and '70's. What this shows is that economic/market prospects and

"._*_*{ There's an important reason why this should be true, one

that harkens back to the heart of the cost:benefit

conditions play a major role in ,.,.,,,,,'-,:.. . ,,:.: determining the magnitude of transaction costs. ln contrast, 1,,,,, 1, ;,: .: ..,,... : : : the recent low costs are partly due to the low volatility/low ...:, :.,,... , : . : : :., .. return exoectations of recent .:

...,',:

.':.'.,.,..:.'::::ti:',.

,..

:::';,::

,:

:1,:i''lii'lti;r-1,{.i{il'.i?l'::if':l$

,1:',.a,.,.::":',a:::;::l .;';,..,.1:.;.',::

::..,:::ill{;}:

.

vears.

Another way of looking at this is to say that changes in market structure and improved trading practices have To make the point more simply, we've abbreviated the CFA Institute definition, which can be found in the Trade l\4anagement Guidelines on www.cfainstitute.org. We should also point out that the CFA Inst;tute endorses neither the implementation shortfall method nor any other measurement 2

ro

"r"rt"ii".

.n"rnrll method.

When we segregate

the Plexus data by manager style, we notice that "growth"

and "momentum" managers consistently have higher 3 ln the early 1990's, I met Harold Demsetz at a social event and praised his

work as seminal to my own thinking. He said I was the first person who ever mentioned the oaoer to him.

costs of trading than "value" or "contrarian" managers. In Commentary 55 of May 1998, we measured trading costs

Thus the changes in the markets and the coming titanic struggle between NYSE/ARCA and NASDAQ/Instinet will change the geography of the world of investing. Trading of equities will merge with trading of bonds, options and

for large cap growth managers at 2.5 times the cost of large cap value managers. Small cap growth managers spent twice the cost of small cap value managers.

futures, opening new types of strategies that are too difficult to execute now. We will redraw the map, and investors will be the winners.

The chart, reproduced from Commentary 55, graphs trading costs by style against 30 day market adjusted returns and draws two interesting conclusions: (1) the dots representing costs vs. returns lie very close to a straight line; and (2) the costs approached the experienced returns for each style. i.e. costs go up in lock step

What will come of this in the future? Clearly both the NYSE and NASDAQ are gathering resources for a titanic

clash. lt's hard to see who, if anyone, will ultimately dominate. Any predictions are highly speculative, but our guess would be a duopolistic standoff with both

with the information value.

systems surviving. We also predict continuing budding of new trading technologies aimed at niche markets. Keeps everybody on their toes.

We explained ". . . the value of the research dictates the acceptable cost as managers filter out ideas with greater expected cost, or stop trading when costs exceed idea value. Either wav. manaqers overall are able to calibrate

Investors will be the clear winners, but the manner

in

which investors win may be a little more indirect than simply from reduced trade costs. The real gains will come from the nncnino rrn of new investment strategies.

As the traditional barriers

to

progress retreat, exchanges will move on to the real work to be done. The key to exchange improvement for institutional investors lies, as always, in the enhancement of liquidity and the strengthening of confidentiality. The major challenge for the new consolidations will be to figure out how information technology and high speed connectivity can be used to promote these commanding institutional

costs to research value." To summarize, in times of high expectations such as the internet boom, transaction prices go up; in low expectation

times like 2002-2004, they go down. What we see here is the returns expectations translated into an upper bound

on the cost of liquidity. This relationship of costs to expectations seems to be a pointer to market efficiency, and goes a long way to explain why active management

needs. However, connectivity won't solve all trading problems. Ticking off orders against the book is fine when volatility is low and the book is deep. However, high tech

net of costs looks like a zero-sum oame.

solutions can traumatize when volatility is high, momentum is strong and the book is thin. In those

Lower Cosf Markets

situations hightouch markets show their merit, and findthe-liquiciity services becorne more highly valued than blazing speed. lf a headlong rush to high tech connectivity destroys the high-touch capabilities of the market, investors could be hurt.

It would be incorrect to conclude that the level of exchange costs doesn't matter; it matters a great deal. When costs come down due to enhancements in market functionality, more ideas become actionable. Institutional

investors can profit from ideas that have smaller

of institutional new boom drives

Finally, don't expect the average cost

expectations, to the benefit of investors. This is very evident in the world of the high-turnover hedge funds, who are able to trade for tinv returns at even tinier trade

trading

to remain low when a

expectations skyward. Transaction costs will still be topdown driven by value expectations, not cost structures.

costs.

Plexus

News

Put this on your calendar! The Plexus Group 9th Client Conference is scheduled for September 18 - 20, 2005 at Fairmont Turnberry Is/e Resort & Ctub

*,

the

(http://www.fairmont.com/turnberryisle/)

;/. :l: 7: ,:,