How Good is Your Investment Sales Team?
In a recent article we describe how some asset managers are transforming their product distribution using artificial intelligence. We call this distribution analytics. The transformation requires overcoming three key challenges. Here, we consider the second challenge: sales performance evaluation.
Much has been written on separating luck from skill in investment management. But, how can we tell if the sales team is doing a good job? We can of course simply look at commissions, but that doesn’t seem fully satisfactory. In his book Principles, Ray Dalio advises us to, “Pay more attention to the swing than the shot,” meaning that we should focus on the process more than on the outcome.
For instance, imagine you’re a sales associate at Bridgewater Associates. It’s April 2020, COVID-19 is raging and your flagship fund just lost 20%. Dalio admits that he was “blindsided” by the pandemic. You may not be able to attract any inflows at all in Q2. In fact, outflows are more likely. However, what you do and say to clients over the coming quarter can still make a big difference. How should your firm evaluate your performance in Q2? Surely not just by looking at your commissions.
Asset flows into an investment product are driven by a mix of factors:
● Sales/relationship strength
● Marketing/brand strength
● Product performance
● Luck
Many asset managers struggle to separate these factors. And, it’s a high-stakes struggle. Those who focus on outcomes such as commissions or Assets Under Management (AUM) find it difficult to hold teams accountable. The sales team ends up complaining that the prospects they get from marketing are poor. Marketing starts complaining product performance is not competitive enough. Meanwhile, portfolio managers begin to complain they are misunderstood by the market.
By sorting out these influences, we help our clients evaluate which parts of the business are working and which aren’t. They are then able to course-correct and make improvements.
We begin with the balance sheet equation: Ending AUM = Beginning AUM + Investment Return + Asset Flows. For now, let’s ignore distributions and non-organic growth.
On the left side of the chart, we break down a product’s total return into three components: market, category, and product return. Let’s take a concrete example using PIMCO’s Active Bond Exchange Traded Fund (ETF) (Ticker: BOND) as of July 13, 2020:
From these figures, we calculate the “Category vs. Market Return” as -0.71%. Since this is negative, Core-Plus was not the place to be in the bond market in 2020. On the other hand, the “Product vs. Category Return” is +0.17%, indicating this PIMCO portfolio management team did well within the confines of their mandate. PIMCO’s executive management should probably evaluate this team’s performance using “Product vs. Category Return” rather than “Category vs. Market Return.” After all, PIMCO is paying this team to form the best possible Core-Plus portfolio, not to pick winning categories.
We perform a similar analysis on asset flows (the right side of the chart), but we cannot compare them directly as we did with investment returns, because they are at different scales.
It helps to think in terms of market share:
● Category vs. Market Flows: In this fact set, 10% of the bond market was allocated to the Core-Plus category at the beginning of the period. If its market share had remained constant, the Core-Plus category would have suffered 10% of the market’s outflows or $4,418 million. It actually did better than that, so its “Category vs. Market Flows” are positive: -2,345 — (-4,418) = $2,073 million.
● Product vs. Category Flows: The ETF captured 0.30% of the Core-Plus category at the beginning of the period. If its share had remained constant, the ETF would have suffered 0.30% of the category outflows or approximately $7 million. It actually had inflows of $507 million, so its “Product vs. Category Flows” were 507 — (-7) = $514 million.
Here is a summary of our analysis for PIMCO’s ETF for the period of Jan. 1 — July 12, 2020:
The goal is to attribute each of these to a different team. Of course, no team is an island, but this framework helps with some useful distinctions.
Returns are relatively easier to attribute:
● Portfolio managers are most responsible for the “Product vs. Category Return.”
● Executive leaders who set the firm’s product lineup are most responsible for the “Category vs. Market Return” metric. The better job they do at entering winning categories and exiting lagging ones, the higher this metric goes.
Flows are more difficult:
● Sales is most responsible for the “Product vs. Category Flows” metric, but portfolio managers influence it as well. Since many investors chase performance, past returns will influence current flows.
● Marketing is most responsible for the “Category vs. Market Flows” metric, because they are tasked with translating the firm’s product lineup into an attractive brand. However, firm leadership influences this as well. Categories with good past performance are easier to sell. To use a poker metaphor, firm leadership deals the hand that marketing must play.
To isolate sales from product performance, we use the following regression:
Product vs. Category Flows in Current Period = β * Product vs. Category Returns in Past Period + α
In this equation β is the regression coefficient and α is a measure of the value added by the sales team, similar to α in a Capital Asset Pricing Model (CAPM). Put another way, α is the actual flows vs. those that would be expected given historical product performance.
Following the same logic, we isolate marketing from category performance using this regression:
Category vs. Market Flows in Current Period = β * Category vs. Market Returns in Past Period + α
The equations above are simple regressions with one factor: performance in a past period, say the prior 12 months. In practice, we expand them to include
● Multiple past periods
● Other past performance measures, e.g., volatility, drawdown, etc.
● More flexible model forms, supporting non-linear relationships
As we add factors and flexibility, we fit the data better and make the α a purer measure of sales and marketing skill, respectively. This would be similar to the various extensions of CAPM for returns, making α a purer measure of investment skill. Following that literature, we use several tests to ensure we do not overfit the data.
Notes:
● Much of the framework and analysis are indebted to Dr. Jan Jaap Hazenberg’s “A New Framework for Analyzing Market Share Dynamics among Fund Families.” Dr. Hazenberg uses relative flows and AUM-weighted returns to decompose market share changes. We present a simplified version that replaces relative flows with dollar flows and weighted returns with simple returns. We would like to thank Dr. Hazenberg for his help in reviewing his framework and findings.
● In analyzing the PIMCO ETF’s flows, we use the following sources:
o ETF flows are from ETFdb.com through July 13, 2020. ();
o Bond market flows are from Baird through May 2020. ();
o Historical ETF Net Asset Value is from PIMCO’s semi-annual report as of Dec. 31, 2019.
o Bond market size is from SIFMA. We show corporate debt outstanding as of Q4 2019.
o Category flows and AUM are placeholders used to illustrate this calculation. The real figures are available from a variety of sources, such as Lipper, the Investment Company Institute (ICI), Broadridge, and MarketMetrics.
