The
Journal Interview
Gary P. Brinson, CFA
is the founder, Chairman and Chief Investment Officer of UBS Brinson,
a Chicago-based money manager with approximately $158 billion in institutional
asset portfolios under management, as well as $122 billion in UBS Private
Banking mutual fund assets. For its clients, which include the world's
largest corporations and government pension plans, the firm invests
across the global financial markets. In his 30-year career, Mr. Brinson
also has authored books and lectured at university level. He is a former
Executive Committee member and past Chairman of the Institute of Chartered
Financial Analysts, a current member of the Editorial Advisory Board
of The Journal of Portfolio Management and a trustee of the Research
Foundation of the Institute of chartered Financial Analysts. He is on
the Board of the Economic Club of Chicago and was named 1991's Outstanding
Financial Executive by the Financial Management Association.
David Spaulding: Please tell me a little bit
about your background.
Gary P. Brinson:
When I was an undergraduate, at Seattle University, in the early
1960s I had an interest in accounting that then shifted to Finance and
Investments. I went on to graduate school, at Washington State University,
and stayed with that emphasis. When I completed graduate school in 1969,
I left for the East Coast to work at Travelers Insurance Company, which
is where I began my career in investment management. This is actually
my 30th year in the business.
DS:
How did you end up forming Brinson Partners?
GPB: It was a series of steps: From 1969 to 1979 I was at Travelers.
Then I came here to Chicago in 1979, as the chief investment officer
for what was then the First National Bank of Chicago, where we setup
an asset management business. Then in 1989, due to the capital means
of the bank and with my motivation and that of my partners, we decided
that it would be best to do a management buyout. That is when we created
Brinson Partners.
DS: So, your firm is only 10 years old.
GPB:
The people really date back to 1981 in terms of when we were all working
together at First Chicago. Although we changed the name of the firm
to Brinson Partners in 1989, when we did the buyout, really the investment
process, the investment people have been working together since 1981.
DS:
Here’s a little bit of my background: In the mid-80s I had responsibility
for the production of performance reports. Back then we were doing them
on a monthly basis for internal use and actually only annually for our
clients. I’m curious, has your time frame for performance information
shortened?
GPB:
Well, I think you obviously have different constituencies for performance
reporting. I would say the typical client need for performance reporting
tends to be quarterly and in some cases monthly. But in the formal sense
usually the reports are done on a quarterly basis. From my standpoint,
for internal purposes we monitor some categories of portfolios with
daily reporting, the tracking of performance numbers are then followed
with a thorough analysis of performance and attribution on a monthly
basis.
DS:
So you get daily performance on occasion today?
GPB: Yes.
DS: Is
that growing in importance for you?
GPB: I wouldn’t
say it’s growing in importance. But what it does provide is a monitoring
mechanism by which we can follow various asset class categories for
anything transpiring that might look out of bounds, with respect to
the risk budget and risk management that we have assigned to that particular
asset class or category.
DS:
How about for the clients? Do you envision them wanting to see numbers
on a more frequent basis than monthly?
GPB:
From my own standpoint, no. I don’t think the client wants to get into
the job of doing the monitoring of the risk management and the risk
budget. The client needs to be sensitive to the capabilities of the
manager when he hires him or her. But it seems to me that it is a) a
bit overwhelming, and b) not very efficient for the client to try and
take over that duty within a time frame shorter than a month. Furthermore,
you can argue that even monthly data is too frequent.
DS:
As far as reporting, do you find your clients wanting more than just
performance numbers? Is it typical for you to provide them with attribution
and risk information as well?
GPB:
It really differs quite a bit by type of client. Some clients are very
interested in getting down into the attribution and looking at the decomposition
of returns and what’s going on. Then there are other clients that have
absolutely no interest in that. These clients want just to see the basic
performance numbers and then we really are reporting to those expressed
client needs.
DS:
Let’s shift gears a little bit. What’s your view of the AIMR-PPSSM?
GPB:
I think we are happy with it. It brings a leveling of the playing field
so-to-speak. It brings a certain level of respectability and integrity
to the reporting process. So, from an industry standpoint, I say it’s
quite healthy; and it is intentionally keeping away the abuses that
might happen.
DS:
Now there are a new set of standards that are coming out – GIPSSM
– and, given the
fact that your company is very global, do you anticipate complying with
these standards?
GPB:
We are planning to comply with them. I suspect, given the scope of this,
it will take us a bit of time to get the relevant information pulled
together and position it within that set of standards. But that’s still
in our plans.
DS:
Do you see yourself as an anomaly compared with your peers, when it
comes to performance measurement, attribution knowledge, and expectations?
I know that you’ve done research and have written articles on attribution;
and I’d say that you are quite knowledgeable about that. But do you
think that your peers are as sophisticated or as comfortable with those
concepts?
GPB:
It is hard to generalize. I think that in any endeavor like this there
is sort of a distribution of awareness of what these issues are. I don’t
track that very closely. Certainly there is a subset of the peer group
that spends a lot of time and intellectual energy as we do in this subject
area. I think it is equally true that there are people who – for whatever
reason – don’t want to get involved in those kinds of statistics.
DS: This
may seem like an odd question, but it actually came up when I was at
a conference recently. If past performance is not an indication of future
results, what then is the value of performance measurement?
GPB: Performance
information is necessary and valuable for reporting and providing an
audit trail for actual results. I do not believe that past performance
is a valuable guide for forecasting what will happen in the future.
DS: Given
your involvement with the global market, I am curious as to how much
more complex you feel it is to deal with performance measurement and
attribution on a global basis versus domestic?
GPB:
Clearly you’re bringing in the new dimension of multi-currency; so any
time you are dealing with another variable, it makes the process more
complex. Yet, it is fairly straight forward. At least I would say today
it’s very straight forward. A couple of our people, Denis Karnosky and
Brian Singer, wrote an ICFA (Institute of Chartered Financial Analysts)
research foundation monograph1 on how to correct performance attribution
in a multi-currency environment. The reason we did that was that up
until then we had seen things from clients and consultants that, to
be blunt about it, were just done wrong. That is to say, the methodology
was wrong. It can be complicated. If people do think it through, they
still can come up with erroneous attribution terms. I think today it
is generally accepted that people understand how to treat the currency
factor. Global attribution does bring in that extra dimension. But beyond
dealing with these global portfolios and attribution, it really is just
an extension of what you would be doing in the comparative US portfolio.
DS:
What additional complexities does emerging market investing introduce
into performance measurement?
GPB: The
data is noisy. As you have suggested, the benchmarks haven’t really
solidified. I suppose it just requires us to cope with information,
price and benchmark information, that is not of the same quality or
caliber that we are used to dealing with.
DS:
On page 21 of your book,2 which I know was written prior to the Euro,
you mentioned the absence of a single security. I also recognized that
the Euro is not per se a single security, but it’s pretty close to that
in Europe. So what impact has this had on trading costs and risks?
GPB: I don’t
think it changes risk very much because now you are simplifying things.
You’re dealing with the Euro instead of dealing with separate currencies.
So it simplifies the process of currency management and currency hedging
and probably lowers the transaction costs of doing that. But I don’t
know that it would have any effect on risk management, other than that
you could argue that it is now easier to surround those countries around
the Euro with the single currency dealings, with cross-correlations
between French Francs and German Marks, for example.
DS:
You mentioned that transaction costs are something that your firm monitors
to try to enhance your performance. Please explain.
GPB:
It is. We subscribe to some services from outside vendors, as well as
our own internal transaction cost monitoring process. Transaction costs
are the enemy of investment performance and anything we can do to minimize
the impact of those transaction costs, to improve the performance of
the portfolios, is of interest to us.
DS: Getting
back to reporting and performance attribution for a moment, how well
do you think your clients understand that concept? It has been my opinion
that a number of people feel they need to have attribution done as well.
This rise in interest may just be because it is a new buzzword and,
because of this, many really don’t comprehend it. I am just curious
whether you think your clients have a good grasp of it?
GPB: I
think it is hard to generalize. I think some of our clients are very
sophisticated and have a very good grasp of it and, more importantly,
have a good conceptual understanding. On the other end of the spectrum,
there would be others who are not as familiar with it, although I see
this, with the passage of time, becoming an area of growing interest.
DS:
Should AIMR include standards for attribution? If so, what areas should
be covered?
GPB:
I would be very opposed to that. I think that when you talk about attribution
you are talking about diagnosis. It is very important for AIMR to be
involved with the integrity of the performance number. The thing is
when you start getting clinical diagnostics of performance, there are
a myriad ways that one can construct diagnostics and ways that people
care about their wants. I think it would be a big mistake for AIMR to
come in and try to say that there is just a single type of diagnostic
to be applied to performance numbers. So I would draw a strong distinction
between the integrity of the numbers, which is what AIMR is addressing,
and people taking those numbers and applying various diagnostics.
DS: Can
we expect any changes to be taking place in performance attribution
as the rules have been defined?
GPB: I
think that they’ve largely been defined. But when you start getting
into statistical analysis, you never want to say that there won’t be
further breakthroughs and further perspectives that may come from your
own journal in terms of the way people think about – and the way they
take – the data and display it, diagnose it and provide some further
insight. On the one hand, I would say quite a bit has been achieved.
I would be very reluctant to say that we are not going to see further
innovation and insight in this area.
DS:
Let’s talk about this for a moment. What risk measures are important
to you?
GPB: The
two basic risk measures are measuring the absolute risk of the portfolio
in terms of it’s volatility, and the relative risk in terms of its tracking
error, since the benchmark has been established with the client. Increasingly
we find that the clients are relatively sophisticated; and I think this
is a very good trend toward prescribing with the manager the expected
amount of tracking error in a certain style and a certain manager’s
capability that the client can anticipate. That diminishes then the
element of surprise and the after-the-fact statement from the client
as to, “Well, gee, I didn’t understand your returns could swing as much
as they are swinging.” So I suspect we are going to see a lot more in
terms of this notion of tracking error, risk budgeting, and those types
of elements.
DS: Are
there any other measures of risk that you utilize?
GPB: There
are, but they are not related to performance. Obviously, in terms of
fixed income markets there are various measures of risk relating to
bond ratings and other types of risk that might be assessed with high
yield securities or people investing in bankrupt situations where you’re
dealing with yield, and balance sheet, income statements and such things
like this.
DS:
Are you satisfied with the way risk is being measured?
GPB:
I don’t know if there is a perfect way to capture risk, but I would
say generally speaking that I am satisfied with statistical tools that
are available and the developing tools that cascade off of them. I am
thinking when I am saying that, of things like VaR (Value at Risk) and
other types of tools that allow you to get into a portfolio and understand
the dynamics.
DS: As
you know there are a number of different risk measures that are offered
as proponents of downside risk. Perhaps one of the most recent ones
is the Modigliani Method, which you may have heard about it. Why do
you think there are so many available? If you are so comfortable with
your methods, why are there so many opinions about risk?
GPB:
I don’t think that there is anything wrong with it, as we ourselves
produce a number of different types of risk ratios or risk dimensions.
There is the Treynor Ratio. There is the Sharpe Ratio. There are all
kinds of different statistical techniques. The way we look at that is
when we are reporting on a particular portfolio, we will lay those all
out without saying that we prefer any one. Rather, we’re saying that
we’re looking at that field of measurement to see what the different
statistics can tell us about the performance and about risk.
DS:
In your book you encourage the use of market value weighted total return
indexes. This isn’t the universal approach and there is a huge universe
of competing benchmarks. How do you select the ones that you use?
GPB:
Well, the idea is largely worked out with the client and the consultant.
When a mandate is given it is usual, I would say quite typical, that
the client would prescribe the mandated benchmark by which he or she
is going to make an assessment of the manager. So long as that benchmark
is consistent with the style that we are bringing to that particular
mandate, we are quite happy to work with the client and the consultant
in using that as the tracking record.
DS:
Do you favor the use of a peer group as a benchmark?
GPB: What
I favor is the parallel measuring of performance against market standard
benchmarks and against peer groups. Peer groups though, as you know,
are a tricky area. You have to be very certain that the data is cleansed
and that the composite comparisons truly are pure comparisons. A good
example is in the fixed income area. Fixed Income mandates allow the
use of the low investment rate credits. Other fixed income mandates
don’t. So, if you are measuring two different bond managers; and you
are measuring them against the peer group; you have to be very careful
as to whether that peer group includes the use of below investment rate,
high yield securities or whether it contains strictly investment grade.
It is not always clear in terms of the pure data that one sees published.
It is not always clear, unless you really drill down into it, just how
cleansed that data is with respect to the strict definition of a peer
group.
DS: Given
the recent introduction of the Euro and the return of Hong Kong to China,
two events that came about after your book was published, can we expect
an update to your book?
GPB:
I haven’t talked to Roger Ibbotson about that for awhile. I’d say both
of us right now are pretty busy with what we are doing and don’t have
any item on the drawing board to do that in the near future. But as
you correctly point out, there is a lot of change in the world and maybe,
at some point, we may reconsider.
DS:
Getting back briefly to performance, I think you would probably agree
that we wouldn’t be able to do today what we do to performance without
the state of technology. Given that now we have the Internet, do you
think that it will have any role in performance measurement and or reporting?
GPB: I
think, just in general, we are liable to see the evolution of managers
reporting to clients with the ability of the client to access performance
data, as well as other portfolio data, through an Internet connection.
Where the client can come into the server on the manager side and look
at and receive the data in a more timely fashion and probably in a less
cumbersome way than it is currently done with paper and faxing, the
physical thing.
DS:
Do you have any closing comments regarding performance in general?
GPB: I
remain concerned that people who look at performance data don’t always
place it within the proper statistical context. What I mean by that
is that there is a lack of appreciation as to just how noisy the statistical
data is, and even though if one looks academically at the statistical
definition of how that data should be interpreted, most people will
try to draw inferences from that over periods of time that in many cases
are far too short for that to have any statistical viability. I suppose
it is just human nature that people want to try and rush to judgment
and to a conclusion about whether something is very good or very bad.
There seems to be a lack of patience and perhaps a lack of understanding
with the statistics themselves as to where that really does become a
statistically meaningful insight, as opposed to a number that is drawn
from such a small sample period that it is inappropriate to use.
DS: Thank
You.
ENDNOTES
1 Denis S. Karnosky
and Brian Singer, (1994), Global Asset Management and Performance Attribution,
Charlottesville, VA, The Research Foundation of the Institute of Chartered
Financial Analyst.
2 Roger G. Ibbotson
and Gary P. Brinson, (1993), Global Investing: The Professional’s Guide
to World Capital Markets, New York, NY, McGraw-Hill, Inc.
To
Receive a Trial Issue or to Begin Your Subscription to The Journal
of Performance Measurement
To
View a List of Previous Articles from The Journal of Performance
Measurement
To
View Submission Guidelines for The Journal of Performance Measurement
