New York – The Wall Street Letter featured a story this week about how research directors are attempting to value research. Buy-side clients have been pushing sell-side research providers to better estimate the proportion of commissions that relate to prop research. The key factors in putting a value on research cited were:
- "how accurate the analysts buy/sell calls are"
- "how often clients want to talk to different analysts"
- "how the information is different from rival firms"
- 'how the salespeople and traders view analysts"
The topic was covered at the recent SIFMA conference and the most notable quote was provided by Lara Warner, head of US equity at Credit Suisse—"I find it pretty ironic that its' our job to value other industries and other companies, but we can't even value our own industry".
In this missive we discuss the various value metrics and their relationship to the theory of valuation.
In valuation exercises, a lot of time is spent discussing DCF (discounted cash flow) models. With these models the value is assessed to be the present value of a stream of cash flows. On the surface this seems simple enough. But the reality is that the choice of the discount rate entails choices that relate to the individuals' choices between risk and return.
Compared to the WSL article, we have reordered their list according to our assessment of estimation difficulty. The first point relates to the estimation of the synthetic returns that would be generated by the analyst's recommendations had an investor acted upon every recommendation. This has been followed for some time now and is an essential part of the analysis of the value or research to the client. Another relatively simple metric relates to the second point. Here it is again relatively simple to devise a metric around the number and/or time spent by analysts in dialogue with clients. For example, one could calculate the number of calls/per client/per analyst.
The third point relates to the differentiation of the research ideas provided. This metric is probably less difficult to estimate than it is to utilize effectively, because we all know that it is not valuable to provide different ideas that are wrong. Perhaps a metric that combines the buy/sell accuracy with research idea differentiation is a possibility. One could envision a metric that combines the performance of an analyst with the analyst's recommendation disparity index. (We are working on it)
The fourth point supports the usage of the broker vote in
determination of value. Some see this as
a popularity contest that favors analysts that are constantly marketing their
research to the buy-side. However, this approach, though flawed, gets to the
heart of valuation theory. According to microeconomic theory, valuation is aggregation
of individual choices that relates to each person’s preferences. In this sense
the real value of research is a combination of price and the satisfaction
derived from the consumption of the research. To my taste, cutting the price of fried
scorpions will have no impact on my consumption.
So in valuing sell-side research, we are back to
microeconomic consumer theory as a basis for valuation. Like it or not--and for most of us, its
not—all finance theory eventually depends upon consumer theory.
There is one other way to value a service; establish a
market and trade it. The closest we can get to this is through the
implementation of CCAs (CSAs) between the sell-side and the buy-side—or more
generally, the unbundling of research cost from transactions costs. This is a
frightening prospect for the sell-side, since unbundling of services almost
always reduces the pricing of those services. In a fully unbundled model value maximization
to the entire market would occur when the marginal value of research was equal
to the marginal cost to the buy-side. For the sell-side this is perhaps even
scarier than pervious assertion. Fortunately for the sell-side, there is really
no way to estimate this equation.
In the valuation approach, it seems the broker
vote model is the best we can do on the consumer preference side. The other
metrics are fairly well developed, but still need work.
Posted at 11:19 am by Thomas Hutchinson
Permalink
