Selection & universe

Selection comes before allocation and systems are needed to assist in this selection. Why systems? First, because there are many types of selection; manager, asset, region, risk, profile, strategy, or other constraints like liquidity. Second; systems are better equipped to handle a deluge of information. Regarding universe; selection is made from it. The universe is large. Do we know our universe? Do we understand it? Or is the universe that matters just about popularity rather than a real character?

What is statistical character of a group? Example; what is the weekly performance ranking of Indian auto sector among the 30 sector indices in India on a scale of 0 to 100 percentile; near 0, near 50 or near 100 percentile? If you were buying an auto stock would your selection change if the answer is different? It should matter to you even if your investing style is momentum, reversion or just trading. It does not harm to trade in the direction of the larger trend.

Whenever you buy or sell, you rank your selection. This ranking could be conscious or subconscious, or have an opaque or transparent frame. Most of us rely on ranking heuristics, simple thumb rules, example behavioural finance talks about how high volume and high price as attractors. Many times entries and exits may not involve analysing quarterly or annual performances. It may not be a common practice for an investor to look at a sector performance before buying a sector component. If you are not looking or analysing the respective universe, the good news is that you are not alone. However, the bad news is that your portfolio runs the risk of underperformance.

Let's substantiate our hypothesis that defining a universe could influence your selection and hence your portfolio performance. Early January 2013 in a group of top 10 equity indices made of Nikkei, Hang Seng, CAC, Sensex, DAX, FTSE, STOXX 50, Bovespa, S&P 500 and Shanghai Composite, it was Nikkei that showcased top relative outperformance for the fourth quarter of 2012. The respective index topped the ranking list. FTSE on the other hand ranked in the middle among the 10 global indices.

However, when we looked at the absolute performance of the top 10 global indices among a larger group of 1000 equity indices, the results were starkly different. Nikkei was now cheaper, ranked in the bottom 20 per cent of the new group while FTSE was ranked higher at 40 percentile, more expensive than Nikkei. Just moving from relative to absolute rankings and from a smaller to a larger group had inverted the valuation perspective. A different universe could indicate a different approach to selection. As we near the first half of 2013, Nikkei fared better than FTSE 100. So if we filter out our selections through multiple universes, it could lead to a better selection and hence higher possibility of outperforming peers and respective benchmarks.

How does one do this? Extend and define your universe and look at the bigger picture. Look at a sector if you are selecting a stock, look at a group of sectors if you want to allocate in a sector, look at regional markets if investing locally, and look intermarket if you are allocating in multiple assets. How will it help? A confluence of confirmations improves signal quality. If something is a buy across groups, it makes the signal a better buy, which means more return per unit of risk.

Sensex 30 and Nifty 50 are two different groups, one has 30 stocks and the other has 50. The common stocks also come with different weightages. We ran a query on the Orpheus passive extreme reversion style for Sensex and Nifty to find which stocks were less than 50 percentile on a quarterly performance ranking, were common in both groups and had a positive seasonality. The selections from this sub-universe were SBI, NTPC and L&T.

The author is CMT, and Founder Orpheus CAPITALS, a global alternative research firm