Three times a day, Acadian Asset Manage-ment calculates the relative attractiveness and expected return of 40,000 stocks around the world. “You may notice we like numbers,” Churchill Franklin says with a laugh. The Acadian CEO is sitting in a conference room at the Boston headquarters of the quantitative investment firm for an interview in early May.
Across the table, Chief Investment Officer John Chisholm says that, in essence, what Acadian does is look for inefficiencies in the pricing of securities. “Investors make certain systematic behavioral errors in the way they make investment decisions,” he says.
Take value. Companies that are cheap in terms of price to earnings tend to generate higher returns over time than expensive companies. One reason: Investors tend to extrapolate from a stretch of strong earnings growth and bid up a company’s price, even when continued growth to support such valuations is improbable. Errors such as that can be tied to certain characteristics of stocks. “‘Factors’ is the quant term for them,” Chisholm says. “We measure empirically what’s the payoff associated with that error.”
Acadian then uses its models to predict, given the current market and macroeconomic environment, what the return associated with that characteristic will be in the near future. “Then we build portfolios that maximize the exposure to those particular factors or inefficiencies in the marketplace,” he says.
Acadian’s approach has worked: Its $8.5 billion Global All-Country Equity strategy, for example, returned 1 percentage point more than its benchmark net of fees from inception in 2003 to June 30. Acadian, which was founded in 1986, manages about $75 billion for institutions in more than 40 strategies. The firm’s $19 billion emerging markets equity strategy, its largest, returned an average of 7.1% a year net of fees from January 1, 1994, through June 30. By comparison, the fund’s benchmark, which since 2001 has been the MSCI Emerging Markets Index (Net), gained an average of 5.1% annually over the period.
To explain Acadian’s approach, Chisholm logs in on a keyboard connected to a large monitor on the wall. “We have a tool, think of it as a website, that the research and portfolio management team uses to understand what all the models are saying,” he says. Acadian has about 300 employees working in offices in London, Singapore, Sydney and Tokyo, as well as Boston. The firm is part of New York-listed OM Asset Management, which is majority-owned by London-based Old Mutual. “Let me just pull up a stock example,” Chisholm says.
Microsoft is suggested. Chisholm types away at the keyboard. The screen then displays what looks like an elaborate spreadsheet. “So, currently, Microsoft’s really mediocre,” he says, scanning a table of numbers on the screen. He points to some valuation numbers, which are shown in blue. “It doesn’t look horrible on a valuation perspective,” Chisholm says, pointing to a set of raw data on the right side of the screen. “These are the traditional valuation metrics in this top group here,” he says. Among the six pieces of data displayed are Microsoft’s current price-to-book ratio and price-to-trailing-12-month-earnings ratio, which on this day is 19.5 times.
Those six measures are related, Chisholm says: If a company is cheap in terms of one metric, it will typically be cheap for the others. “So we combined them into what we call an aggregate factor,” he says. Acadian labels that factor “price to intrinsic value,” and it’s shown as a raw value, score, weight and contribution in blue in a table on the left side of the screen.
The raw value is essentially an average of the six metrics. The score compares the raw value with those of the company’s peers. “A score of zero would mean that the company is neutral on that measure relative to its peer group,” he says. A score of 1 would mean that a company is 1 standard deviation more attractive than its peers. Microsoft’s score is 0.16.
Acadian’s secret sauce is how it weights the factors for a given company, Chisholm says. “So we have a stock factor attribution system,” he says. He opens another spreadsheet that tracks the entire universe of stocks and points to a column.
The firm ranked all of the 40,000 companies by the value characteristic relative to their peers at the end of March, Chisholm says. “We run a regre
ssion of that end-of-March ranking against April’s index-relative return for each of those companies,” he says. “That regression coefficient there is 51.2.”
That means that globally on average a company that was 1 standard deviation cheaper than its peers in March would have outperformed them in April by a 0.512 percentage point. “That tells us what’s the payoff to that characteristic in this peer group,” he says. Acadian takes the historical factor-return values and feeds them into yet another model. “It’s kind of the forward-looking piece: How do we expect these different characteristics to do in the future?” he says.
For each company, the firm sums the forecast payoffs from all of the factors to get a single expected return.
Acadian then inputs the data for the companies in a given strategy into an optimizer that trades off their attractiveness against transaction costs and risk. “That’s ultimately how we determine what to buy and sell in portfolios we manage,” he says. “There’s a lot of engineering that underlies the process.”
