28.01.2014 Views

Download - IndexUniverse.com

Download - IndexUniverse.com

Download - IndexUniverse.com

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

the net asset value of a fund and one or more benchmark<br />

indexes as a framework for analysis of an actively managed<br />

fund’s performance.<br />

Net Tracking Error As A Framework<br />

For Fund Performance Evaluation<br />

This section illustrates one way to organize your examination<br />

and evaluation of transaction costs and a number of<br />

other major cost and value-added elements that determine<br />

fund performance. The example is an actively managed<br />

mutual fund (to illustrate the breadth of the analytical possibilities),<br />

but most of the hidden cost and value-added<br />

elements apply to index and actively managed ETFs just as<br />

well. After all, even if an index fund is passively managed,<br />

“passive” should not mean “mindless.” Transaction costs<br />

offset at least some of any value added by both active and<br />

passive investment processes. Transaction costs associated<br />

with index <strong>com</strong>position changes are a dead-weight drag on<br />

the performance of any portfolio that replicates an index. In<br />

most cases, the <strong>com</strong>position-change transaction costs are<br />

embedded in the performance of the index, but that does not<br />

mean we can’t estimate them.<br />

In my youth, I was a great fan of the late Don Herbert’s<br />

“Mr. Wizard” science TV show. While childhood memories<br />

are sometimes inaccurate, I recall having heard from time<br />

to time, as Mr. Wizard and his youthful apprentices donned<br />

safety glasses, an admonition something like, “Don’t try<br />

this at home.” While that may not have been the precise<br />

warning Mr. Wizard used, it is probably an appropriate<br />

warning for anyone who might try to assemble the data<br />

described in this section from the fund databases available<br />

today. While most of the information described here can be<br />

developed from SEC filings, the data assembly and calculations<br />

are beyond what most advisers, let alone individual<br />

investors, have the resources to undertake. The purpose of<br />

this discussion is to illustrate what should soon be possible<br />

and to offer a preview of how better data on mutual funds<br />

and ETFs that should be<strong>com</strong>e available over the next few<br />

years can raise the level of analytical discourse and improve<br />

the fund selection process. Improvements in data availability<br />

should make this kind of fund analysis almost routine<br />

within a few years.<br />

As indicated at the beginning of the first article in this<br />

series, most fund rating systems explicitly recognize the<br />

limitations of past performance as a predictor of future<br />

investment results—and then proceed to focus on just such<br />

performance <strong>com</strong>parisons. The overwhelming focus of fund<br />

service ratings is on <strong>com</strong>parisons of a fund’s historic performance<br />

to peer group performance. One problem with past<br />

performance is that it is a “noisy” measure, at best, of the<br />

value added or subtracted by a fund’s investment process.<br />

Effective fund evaluation will attempt to break down the<br />

<strong>com</strong>ponents of performance to separate the wheat (valueadded<br />

elements and costs) from the chaff (noise).<br />

I propose to use the net tracking error difference calculation<br />

as an organizing framework to incorporate all the favorable<br />

and unfavorable elements affecting fund performance.<br />

The fund manager’s objective should be to achieve the best<br />

possible performance for investors, not the smallest possible<br />

tracking error relative to a flawed benchmark—or relative to<br />

any benchmark, for that matter. The objective of the manager<br />

of any fund—indexed or active—should be to maximize<br />

positive tracking error relative to an appropriate benchmark,<br />

subject to risk constraints that are appropriate for the fund. 3<br />

Calculating tracking error relative to several benchmarks lets<br />

us use the multiple <strong>com</strong>parisons to increase our understanding<br />

of why a fund’s performance has been good or poor. A<br />

large positive tracking error in fund performance is almost<br />

certainly more desirable than a negative tracking error (or<br />

than no tracking error at all), but any useful analysis is much<br />

more <strong>com</strong>plex than that statement suggests. 4<br />

To make it most useful, net tracking error should be<br />

viewed as a summary measure of the positive and negative<br />

causal elements of value added—or of poor performance.<br />

Calculating the individual elements of cost and performance<br />

and displaying them as offsetting <strong>com</strong>ponents of a net<br />

tracking error calculation can provide useful insight into the<br />

interaction of the determinants of performance. Some of<br />

these calculations require methodology that goes far beyond<br />

what fund services can offer today, but this kind of analysis<br />

can enlighten investors in ways that make development and<br />

application of this methodology inevitable as the available<br />

data improves. Of course, there will always be random elements<br />

that limit the value of even the best analysis. Noise<br />

that is not subject to an unequivocal explanation can be a<br />

sizable <strong>com</strong>ponent of any fund evaluation. While noise limits<br />

the usefulness of the tracking error framework, we are looking<br />

for the causes of performance and are asking appropriate<br />

questions. That is a substantial improvement over what is<br />

being done by most fund evaluators today.<br />

There is no reason to calculate tracking error only relative<br />

to the template index of an index fund or only to a benchmark<br />

for a market segment similar to the fund’s portfolio. Tracking<br />

error for both indexed and actively managed funds measured<br />

relative to several indexes can highlight important index and<br />

fund characteristics. It can reveal elements of index transparency<br />

costs, the quality of the selections in a fund based on<br />

a quantitative security selection process or the astuteness<br />

of the stock picks of a traditional active manager. Multiple<br />

tracking error calculations can reveal that a poorly performing<br />

benchmark is being used as an index fund template. Tracking<br />

error measured relative to <strong>com</strong>petitive funds, particularly<br />

with the <strong>com</strong>parative or <strong>com</strong>prehensive tracking error divided<br />

into operating costs, trading costs and other elements, can<br />

highlight features that a skilled fund analyst or a determined<br />

do-it-yourself investor can use to improve fund selection.<br />

When the net tracking error is broken down into security<br />

selection value added, operating costs, transaction costs and<br />

other measures that incorporate and highlight cost elements<br />

and the effect of any risks accepted and management decisions<br />

made, the result is a rich tapestry that reveals important<br />

characteristics of the fund and its investment process that an<br />

adviser will want to understand. Harking back to my earlier<br />

warning, this analysis is not something for the average do-ityourself<br />

investor or even an adviser to undertake at the present<br />

time. However, advisers need to prepare their thinking<br />

52<br />

January/February 2010

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!