In these last few months of shifting focus to Latam Credit, I devoted a material amount of time to the transition from the mindset of a single name equity investor to the midset of a macro asset allocator having views on asset classes such as credit spreads, interest rates, and equity indexes (mainly S&P 500 and Bovespa) as well as equity sub-index views (like having a view on cyclicals versus non-cyclicals) to understand where investment is flowing.

Credit investing and rates investing are sensitive to investor relative return perceptions between asset classes. At the end of the day most bonds get paid back, yet there is still volatility in the asst class as funds flow into and out of credit and rates. Why is this?

The results of search for the answer to this why are surprising in that the skill-set and way of thinking at the index level of investing is in many ways different than what is requires for investing at the single-name level.

A quick example is earnings misses/beats. Single name investing many time is about knowing which company is better at its business than its competitors. So at the single name level it could be that a business’s competition is better and customers shifted to a competing product, and so the company lost sales and therefore missed earnings. In this case, though, at the index level sales likely stayed the same, and the earnings merely moved to someone else’s bottom line in the index. Who earned them only changed the the total earned in aggregate didn’t. So the index earnings don’t move in this case, and if index P/E doesn’t move, index price stays the same as well.

Inside the index one will see one company rally and one company sell-off but an index investor won’t feel anything. This concept is captured in the average correlation statistics of the S&P 500 (here shown as of July 28, 2011 from a Goldman Sachs report by David Kostin):

When pairwise correlation is very low, as it was in the mid 90s and mid 2000′s, the prevailing stories are idiosyncratic – stories about winners and losers within a sector and between sectors (cyclicals versus non cyclicals, etc).

In todays environment, pairwise correlation is actually quite high by historical standards. This can be thought as prices of all companies (to exaggerate a little bit) are rising and falling together.

This is interesting, because in general, as shown below (also from GS/Kostin), index earnings are slow to change (and they are only reported in 4 two month periods per year effectively).

Given that index earnings don’t change quickly, one might expect index price volatility to be low and therefore in high pair-wise correlation times, one might expect single name volatility to be low. This is not the case. Index volatility is not explained by earnings beats/misses at the index level.

So what’s going on then? The short answer is Index investing is 10% forecasting index earnings which is more or less easy given that they don’t change much/quickly day-to-day as shown above and 90% forecasting Index Price-to-Earnings ratio, which itself can be broken down into price-to-book and ROE forecasts.

Price to Earnings = Price to Book x Book / Earnings

Return on Equity = Earnings / Book, therefore:

Price to Earnings = Price to Book / ROE

This can be counterintuitive, but the implication is that improvements in ROE should lower Price to Earnings if Price to Book doesn’t change.

So bottom line is that Earnings don’t change much, so forecasting index levels becomes of game of forecasting price to earnings and/or price to book and ROE of the index.

The punchline is that what drives Price-to-Earnings many times are factors outside of the equities markets and many analyst equity models. One extremely important factor to watch and understand is 10 year US Treasury rates. They are extremely correlated with price to earnings of the S&P 500 as noted below (here shown as of August 3, 2011 from a UBS report by Jonathan Golub):

 So today the stories are not about individual winners and losers but about the merits of owning equities as opposed to rates.

There are very deep implications. From 2009, at least, to be a US Treasuries trader is to be an Equities Index Trader. Broadly people taking views on the S&P 500 index level are taking views on US Treasury 10 Year rates because the link between the two is that the same factor that drives 10 Year rates drives price-to-earnings ratios.

So to tie everything together, Index Earnings don’t change quickly. Most of the volatility in the index is explained by changes in Index Price-to-Earnings. So Index models (to forecast the Index Price) are really Index Price-to-Earnings models.

Furthermore, given the very high correlation, any equity index model has to consider the interest rate effect on price to earnings or at least have a reality check on whether the forecasted price-to-earnings level of the model checks witch what is likely in the Interest Rate market (particularly 10 Year Treasuries).

So the bottom line is that as growth expectations have been revised down, 10 year Treasuries have rallied, and this has pressured Price-to-Earnings of the S&P 500.

When will this stop? Probably when growth expectations are all repriced, which should finish in the next week or two. Keep an eye on GDP revisions and 10 Year Treasury interest rate action and you’ll get a sense of when it’s safe to be long equities again or when the 10 Year Treasury is at risk of selling off.

Abraço, Arthur

São Paulo Value