US sector analysis (part 2): No benefit from excluding financials

In this second US sector analysis post on total return data since 1993 we try to answer the question that started all of this in the first place: do investors benefit on a risk-adjusted basis by excluding financials from their portfolios?

To answer this question we have to make some basic assumptions about an alternative portfolio strategy. Before we go into the details the conclusion is pretty clear: investors do not get any significant additional positive risk-adjusted returns on their portfolios by excluding financials.

Excluding financials does not add risk-adjusted value
In our previous sector analysis part 1 we showed that financials have made bad risk-adjusted returns since 1993 comparatively. So what would happen if financials were excluded from a market portfolio (defined as a portfolio with 9 out of the 10 GICS sectors with respective market cap weighted returns over time) and compared that to the market portfolio (S&P 500 Total Return Index)? Our guess was better returns and less volatility.

Below you see the results with the market portfolio ex. financials having annualised returns of 8.5 percent since 1993 compared to 8.0 percent for the S&P 500; so an annualised active premium of 0.5 percent. The two other portfolios are equally weighted portfolios of 1) market portfolio excluding financials and IT, 2) an equal weighted portfolio comprised only of the energy, consumer staples and health care sectors - the three best sectors based on the Information Ratio described in our part 1 analysis. We will get back to those two portfolios later.

Interestingly the market portfolio ex. financials has only slightly less annualised volatility (15.0% vs. 15.5%) compared to S&P 500 and thus only slightly better Sharpe Ratio (0.24 vs. 0.21).

Based on the Information Ratio our market portfolio ex. financials has 0.22 compared to of course zero for the S&P 500 as we use that index as our benchmark.

Some would say that sounds okay. Is it significant? No. A two sample t-test of the market portfolio ex. financials' and S&P 500 mean reveals a very low statistic of only 0.0787 (p-value = 0.9373). In other words we cannot reject the hypothesis that the difference in means is equal to zero. Thus investors do not benefit from excluding financials when you compare to a a market portfolio holding all other sectors with their market weights. On a quick note: the two other portfolios in the former chart did not have significant alpha either and this is despite the equal weighted porfolio consisting of energy, consumer staples and health care having an annuaslised return of 11.2 percent and thus annualised active premium (alpha) of 3.2 percent. Is that not worth reflecting upon?

To get an intuitive idea of why the alpha (the difference in returns between the two portfolios) is not significant see the chart below. It shows you the monthly excess returns (alpha) and the mean value (the light blue dotted line) which is barely above zero.

So what has to be done instead to beat the market?
This question we will answer next week in our final analysis where we will test Warren Buffet's performance against the S&P 500 and draw some conclusions about winner strategies.