Signal From Noise
But this 50,000-foot view can only tell us so much. A classification system must also break down its universe of stocks into more manageable and useful categories. If we use the technology sector as an example, securities would first be sorted into broad buckets—hardware manufacturers, software developers, etc.—and then further organized into smaller segments, like semiconductors, telecommunications equipment, computer services and so on.
All three classification systems choose to divvy up their stock universes via four levels of increasingly detailed categorization. (See Figure 2.)
Notice that even though GICS includes fewer companies in its universe than both TRBC and ICB, it offers more granularity at deeper sorting levels. At Level 3, it provides 16 more categories than its nearest competitor, TRBC, and at Level 4, it offers 30 additional classifications.
Generally speaking, higher granularity is a good thing—but only up to a point. No fourth-level sector funds currently exist, making GICS’ increased granularity here mostly an academic exercise. That’s not to say such funds won’t ever exist—particularly as ETF providers search for the first-mover advantage in ever-narrower slices of the market—but at some point, sector ETFs become too specific to attract the volume, liquidity or diversification needed to remain investable. Can the market really support a footwear ETF, or a soft-drinks fund?
A classification system’s sorting scheme must strike a balance between practicality and granularity. With the most stocks and the most granularity at Level 2, and middle-of-the-road granularity at Levels 3 and 4, we believe TRBC offers the best selection compromise of the three systems.
Balancing Act: Sorting Methodology
Knowing how many buckets stocks fall into isn’t enough to judge a classification system; you also have to know why they’re directed to those buckets in the first place. Sorting methodology is the secret sauce of sector classification, and it’s often the smallest details that will create the grossest differences between sectors from different systems.
In order to determine the true nature of a company’s business—and therefore which bucket it should fall into—all three classification systems start with the same metric: revenue. For TRBC and GICS,1 if a company derives 60 percent of its revenue from a single business segment, then it’s lumped into that relevant category; ICB uses a similar approach, lowering the cutoff to 51 percent. Either way, these limits ensure that companies are organized according to their truly dominant business segment, which only a major shift in their revenue stream could change.
The process gets trickier, of course, when a company operates more than one line of business, which many, if not most large companies do these days. Should a firm’s revenue result from several business lines, the three classification systems each adopt different proprietary approaches to assess that company’s core business.
Under the ICB system, if a company possesses two or more business lines, indexers turn to audited publicly available accounting information and directors’ reports. ICB, however, reserves the right to classify stocks either by the industrial process they employ or by their end products, a decision not always transparent and obvious to an end investor.