Simple Sector Adjustment For Value Investing

|
Includes: ADT, AES, CTL, IR, KEY, MET, MRO, MUR, OI, PRU, R, RIG, THC, VAL
by: Investing Discipline
Summary

Simple, quantitative value strategies are one of the easiest ways of achieving superior returns over the long term.

However, value strategies that are too simple tend to ignore important factors such as valuations relative to sector average.

Today, we are taking our strategy one step further to show how sector adjustments can enhance returns and reduce volatility.

Introduction

In my previous articles, I have shown that very simple, long only, equal weighted strategies can outperform their underlying indices by a wide margin. In short, these strategies are based on one or more valuation multiples such as Price to Book, Price to Equity, Price to Cash-flow, or their composite z-scores.

While these strategies have historically outperformed the S&P 500 in both absolute and risk-adjusted terms, returns have generally been more volatile compared to the index. One of the reasons for volatility is sector concentration, a natural outcome of blindly sorting stocks on valuations without looking at their multiples relative to their sector's average. I will now show how a very simple sector adjustment process can further enhance returns while significantly reducing volatility and downside risk of the basic value strategies.

The basic strategy

All data used in the analysis are drawn from historical lists of S&P 500 index constituents from 1989 Q3. For each quarter end, I sort all S&P 500 stocks by their Price to Cash Flow ratio (PC) and split them into ten deciles. PC is lagged by one quarter to ensure that only observable financial information is used to build portfolios, as companies are required to disclose their quarterly results within three months. For example, portfolio constructed at the end of Q2 2011 will be based on financial results for Q1 in that year. Each of the ten portfolios consists of an equally-weighted basket of stocks in each decile.

Sector-adjusted strategy

The above strategy obviously disregards sectors completely. For the sector-adjusted strategy, I first calculate z-scores from the Price to cash flow data for all 500 stocks at each quarter end. Second, I calculate z-scores for each sector at each quarter end, again based on the observable data in order to mitigate the look-ahead bias. Third, I calculate the average of the two z-scores, rank all constituents by the combined z-score and split them into ten deciles. Again, each of the ten portfolios consists of an equally-weighted basket of stocks in each decile. This technique applies equal significance to stocks that are cheap relative to the entire index and stocks that are cheap relative to their peers.

Strategies compared

Let us first take a look at cumulative returns for each of the ten deciles for both strategies. A trained eye will notice that the sector-adjusted strategy (in the right panel) has a slightly better behaved, orderly spread range of decile returns. More importantly, the "cheap" deciles (including the cheapest one) for the sector adjusted strategy exhibits less volatility and significantly milder drawdowns relative to its unadjusted counterpart.

The following chart shows the cheapest deciles of each strategy on the same panel. The chart confirms the key message and reveals that the maximum drawdown was only 43.99% for the sector adjusted strategy, compared to 61% for the unadjusted one.

Importantly, the maximum drawdown figure of 43.99% is lower than 45.8% for the S&P 500 and the returns are significantly better, as shown in the table below. Adjusting for sector-relative valuations increases the excess returns from 17.9% to 18.4% and reduces volatility from 24.9% to 19.3%, resulting in a superior sharpe ratio of as much as 0.95, on an annualized basis. Downside risk measures are also generally lower, with the only exception of the Modified Expected Shortfall.

Table 1: Returns and Volatility for selected deciles

PC

PC (sector adjusted)

S&P 500

Annualized Return

17.88

18.36

9.55

Annualized Std Dev

24.94

19.32

15.83

Annualized Sharpe (Rf=0%)

0.72

0.95

0.60

Table 2: Downside Risk for selected deciles

PC

PC (sector adjusted)

S&P 500

Semi Deviation

8.22

6.89

6.01

Gain Deviation

9.84

6.74

4.74

Loss Deviation

7.50

5.82

5.93

Downside Deviation (MAR=3.33%)

6.28

5.04

5.19

Downside Deviation (Rf=0%)

5.94

4.70

4.84

Downside Deviation (0%)

5.94

4.70

4.84

Maximum Drawdown

61.00

43.99

45.80

Historical VaR (95%)

-11.41

-11.33

-13.32

Historical ES (95%)

-20.19

-15.75

-15.82

Modified VaR (95%)

-11.47

-10.88

-11.47

Modified ES (95%)

-12.29

-16.16

-16.26

Example

I have re-run the process for structuring the basic PC and sector adjusted PC portfolio based on the currently available data. Below are the top 10 holdings the cheapest long-only deciles for each strategy:

Table 3: Top ten long ideas (PC)

Ticker

Name

Sector

PC

PC (S&P 500)

PC (sector)

AES

AES

Utilities

3.26

14.41

6.41

CTL

Centurylink

Telecommunications

3.27

14.41

6.09

KEY

Keycorp

Financials

2.03

14.41

13.83

MUR

Murphy Oil

Oil & Gas

2.45

14.41

6.87

PRU

Prudential Finl.

Financials

2.09

14.41

13.83

MRO

Marathon Oil

Oil & Gas

3.36

14.41

6.87

RIG

Transocean

Oil & Gas

2.03

14.41

6.87

ESV

Ensco Class A

Oil & Gas

0.00

14.41

6.87

IR

Ingersoll-Rand

Industrials

0.13

14.41

14.29

THC

Tenet Healthcare

Healthcare

2.20

14.41

20.50

Table 4: Top ten long ideas (sector-adjusted PC)

Ticker

Name

Sector

PC

PC (S&P 500)

PC (sector)

AES

AES

Utilities

3.26

14.41

6.41

KEY

Keycorp

Financials

2.03

14.41

13.83

MET

Metlife

Financials

3.75

14.41

13.83

OI

Owens Illinois New

Industrials

4.30

14.41

14.29

R

Ryder System

Industrials

3.53

14.41

14.29

PRU

Prudential Finl.

Financials

2.09

14.41

13.83

RIG

Transocean

Oil & Gas

2.03

14.41

6.87

ESV

Ensco Class A

Oil & Gas

0.00

14.41

6.87

IR

Ingersoll-Rand

Industrials

0.13

14.41

14.29

ADT

ADT

Industrials

4.21

14.41

14.29

It should be clear that the basic unadjusted strategy is currently overweighting Oil and Gas, the cheapest sector at the moment. The chart below shows that the Sector Adjusted strategy generally exhibits a better balance, even though it seems to be overweighting financials at the moment.

Conclusion

In summary, z-scoring is a very simple technique that can be applied to a range of problems. Today, it helped us effectively tackle the problem of stock selection based on two different factors: cheapness relative to an average stock and cheapness relative to an average stock in the sector. Adjusting for sector valuations has historically helped reduce volatility and enhance returns by a significant margin.

Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.