 |
|
Our investment process is bottom-up and utterly disciplined, creating value-oriented, sector-neutral, diversified, fully invested portfolios.
|
|
|
The foundation of our work is a universe of seasoned, suitable, liquid, U.S.-listed stocks. We evaluate companies relative to their industry peers using three categories (or pillars) of attractiveness: value, management, and momentum. Overarching this multifactor valuation is a portfolio construction process that optimizes the tradeoff between expected return and multi-faceted risk. Actual, realized trading costs are used to guide us toward efficient implementation. The integration of research, portfolio management, and trading is essential to our decision-making process.
|
|
|
|
|
+ |
INVESTMENT UNIVERSE
|
|
|
We begin with every stock that trades on a major U.S. exchange (comparable to the Dow Jones Wilshire 5000, about 6,500 names). After screening, we end up with an investment universe of 2,000 or so companies that are seasoned (at least three years of operating history), suitable (no bankruptcies, ADRs, gold stocks, REITs, or funds), and liquid (minimum of $700,000 average daily trading volume). Our strategies draw on all or parts of this universe:
|
|
|
|
• |
Top Cap targets the largest 200 securities by market cap.
|
|
• |
Large Cap targets securities among the largest 350 or 1,000 caps, depending on the benchmark.
|
|
• |
Mid Cap targets stocks within the realm of the Russell Midcap Index (caps 200 - 1,000).
|
|
• |
Small Cap eliminates companies with market capitalizations greater than the Russell 2500 Index.
|
|
• |
130/30 strategies purchase or sell short any company in the target capitalization range.
|
|
• |
Dollar-Neutral Long/Short strategies purchase or sell short any company in our entire investment universe.
|
|
|
+ |
PEER-GROUP CLASSIFICATION
|
|
|
We divide our universe into 13 sectors and 34 industries that reflect statistical and fundamental economic relationships among stocks. Our research shows multi-factor valuation to be most productive when relative value is assessed on a peer-group basis.
|
|
|
| Sectors | Industries |
| | Capital goods | producer goods, aerospace, construction | | Consumer discretionary | media, retail, leisure, apparel | | Consumer durables | motor vehicles, household durables |
| Consumer staples
|
food, household products, tobacco |
| Energy |
oil & gas, energy equipment & services |
| Financial
|
banks, miscellaneous finance, insurance, real estate |
| Healthcare
|
pharmaceuticals, equipment, providers, biotech |
| Materials |
chemicals, paper, metals & mining |
| Services
|
business services |
| Technology |
software, hardware, semiconductors, comm equip, electrical |
| Telecommunications |
telecommunications |
| Transportation |
transportation |
| Utilities |
utilities |
|
|
|
|
|
Note: Our sectors have evolved from nine in 1984 (based on Wilshire
Associates' work) to 10 in 1994, 11 from 1995 through 2000, and now 13 (based on GICS, Global Industry Classification Standard). In broad terms,
we differ from the 10 GICS sectors by our carve-out of capital goods, services, transportation, and consumer durables
from GICS’ industrials and consumer discretionary sectors.
|
|
|
In addition to aiding security analysis, peer-group classification also contributes to portfolio diversification.
|
|
+ |
MULTIFACTOR VALUATION
|
|
|
We evaluate securities relative to their peers, based on three categories of attractiveness:
|
|
|
VALUE — We favor asset-rich companies with higher earnings compared to price. We examine price-driven measures at various points on the balance sheet and income statement: ratios based on book value, sales, a composite of operating earnings, and our own concoction of forecasted earnings. The lower the multiple — in other words, the higher the fundamental "yield" — the higher the expected return.
|
|
|
MANAGEMENT — We use numerous measures to gauge the effectiveness and outlook of management: return on operating assets (looking back up to five years, with emphasis on the more recent past), earnings quality (as opposed to earnings quantity, which we assess in value, above), forecasted long-term earnings growth, the trend in sales compared to assets, net share repurchase, and insider trading (purchases receive greater emphasis than sales).
|
|
|
MOMENTUM — We measure momentum by examining earnings revision, relative price and volume action, price stability, earnings surprise, and short interest. Earnings revision is our own formulation of the trend in analysts' expectations about company earnings one year hence. Relative price and volume action are measured over the preceding year.
|
|
|
Each component of attractiveness varies in influence, determined by a stock's industry, size, and growth characteristics. Forecasted-earnings yield, for example, is a better indicator of value for pharmaceutical companies than for producers of hardware technology; estimate revision is more powerful among small-cap issues than large-cap; and price-to-book and price-to-sales illustrate how growth comes into play (both are more powerful predictors of future stock prices among relatively lower-growth stocks).
|
|
|
We derive an all-in excess expected return for each company. All-in means we boil our multi-factor valuation down to a single number; excess suggests above (or below) the level of the market; expected indicates forward-looking. Each stock's summary measure of attractiveness carries across all of our strategies.
|
|
+ |
PORTFOLIO CONSTRUCTION
|
|
|
Portfolios are fully invested, sector-neutral, and optimally diversified. We make controlled, stock-specific bets across many names in each of our 13 sectors. Sector weights and position sizes are driven by the target benchmark. In our absolute-return strategy, where there is no optimization benchmark, sector weights are driven by our attempt to maximize expected return while minimizing total volatility.
|
|
|
Minimum position size, at cost, is 0.3%. Maximum position size is determined by a security's weight in the benchmark and varies in proportion to the size of the company. We buy no more than 1.2% (at cost) over the benchmark weight for the largest companies in our investment universe. For a non-benchmark name (or a non-benchmark strategy), 1.2% is the absolute maximum weight (up to 3.0% for long/short concentrated). Portfolios hold an average of 80 to 200 names (per side for long/short), depending upon the strategy. Generally, the smaller the targeted capitalization of a strategy, the more names held.
|
|
|
Portfolio rebalancing is systematic and disciplined. Purchases and sales are driven by changes in valuations from the ongoing evaluation of our investment universe, tempered by anticipated transaction costs. Portfolios remain fully invested at all times.
|
|
|
"To hell with a balanced portfolio. I want you to sell my Fenwick Chemical and sell it now."
|
|
+ |
OPTIMAL DIVERSIFICATION
|
|
|
Individual security weights are driven by combining our notion of future profits and our risk model, the ultimate process by which we create portfolios. As discussed, we employ commonsense rules to reduce risk — lots of stocks, sector-neutrality (13 sectors), and controlled and limited stock-specific bets. In addition, we consider five other factors:
|
|
|
INDUSTRIES — Within sectors, 34 industry groups come into play (exposure is controlled but is not neutral).
|
|
|
FUNDAMENTAL CHARACTERISTICS — Sector-by-sector capitalization size, growth, value, and market sensitivity are used to measure and temper fundamental risk.
|
|
|
INTEREST-RATE SENSITIVITY — Stocks' sensitivity to changes in the 10-year Treasury rate matter.
|
|
|
RETURN CLUSTER GROUPS — Diversification among statistical groups reflecting long-term price movements is our aim.
|
|
|
COVARIANCE — Individual stock-by-stock covariance is considered on a sector basis, measured by residual returns over the prior 500 trading days after extracting all of the influence of the other risk and diversification factors.
|
|
|
The most important determinant of a security’s weight (including "negative" weights on the short side) is our forecast of future profits. Although we seek to squeeze out every penny of excess return, we also seek to balance multi-faceted risk.
|
|
+ |
IMPLEMENTATION
|
|
|
Cost-effective portfolio implementation is a vital part of our investment process. Our trading is based strictly on “best execution” and aims to minimize total transaction costs. We measure total transaction costs as implementation shortfall — the difference between valuation price and execution price, including commissions (which average between 1¢ and 2¢ per share), dealer spreads, market impact, and opportunity costs.
|
|
|
We capture actual trading results in our transaction-cost model to guide our selection of trading methods and venues. We utilize blind-strike package trading, ECNs, electronic crossing, as well as traditional agency broker/dealers. Our transaction-cost model also informs our rebalancing effort: realistic, empirically derived, stock-specific estimates of cost help us assess purchases and sales. Annual turnover ranges between 75% and 150% (per side for long/short). Soft dollars are never used.
|
|
+ |
ART, SCIENCE, AND ENGINEERING
|
|
|
It is said that investment management is neither art nor science, but more a matter of engineering. In that spirit (and with tongue firmly planted in cheek), we end this discussion of our disciplined investment process with a flow chart:
|
|
|
|
