Seeking Alpha
Profile| Send Message|
( followers)  

by Maz Jadallah

Great article entitled"Rob Arnott's Magic Indexing Formula" by Shawn Tully at Fortune magazine recently. The article describes Arnott's investment research approach and how his FTSE RAFI US1000 index, which is offered by Charles Schwab as a mutual fund and by Invesco Powershare as an ETF (NYSEARCA:PRF), has outperformed the S&P500 index by 2.4 percentage points per year since inception (12/30/2005). Impressive. For those of you unfamiliar with Rob Arnott and his RAF indexing approach, it is definitely worth a look.

Arnott's approach as described in the article basically boils down to one simple premise, market cap weighted indexes tend to over weight expensive growth stocks with potentially very adverse results. To illustrate this dynamic, the article points out an astonishing fact that Arnott discovered early in his career; from 2000 and 2002, the average stock in the Russell 3000 increased by 20% while the index fell by 20%. To avoid this systemic flaw, Arnott's index weights stocks based on their “economic footprint” instead of their market cap. The approach takes into account a company's revenues, cash flows, book value and dividends to derive its “economic footprint” and then weights the stock accordingly. The result, termed “the value effect”, is that the index tends to overweight companies trading at cheap prices and underweight companies trading at high multiples – which I'm sure we can all agree is a good way to approach stock selection.

We thought it would be worthwhile to compare the performance and holdings of one of our index clones against that of Arnott's PRF strategy. Our AlphaClone Hedge Fund Index [AC/HFI] Top Holding clone is perhaps the most representative clone for hedge funds in our universe as it invests quarterly in the largest holding from each hedge fund in our database. Holdings are weighted in proportion to the number of holders (i.e. if six managers hold stock x and three managers hold stock y, then stock x would have twice the weight as stock y). The results are interesting.

While our clone outperforms PRF by a significant amount (see table below), both strategies trounce the broader market. As of 3/31/2010, Arnott's PRF index returned 3.04% annualized since inception (12/30/2005) while AC/HFI returned 4.1% annualized over the same period.

AC/HFI Backtest vs. PRF – Annualized Return Since Inception

Ann. Return Inception*

Growth $10K

AC/HFI

4.10%

$11,901

PRF (Index)

3.04%

$11,446

PRF (NYSE:NAV)

2.52%

$11,199

S&P500TR

0.59%

$10,257

Russell 1000

0.80%

$10,346

* Results are from 12/30/2005 to 3/31/2010. Source: PRF results are from PowerShares. AC/HFI results are from AlphaClone.

It's also worth noting that roughly two thirds of the 128 unique holdings in AC/HFI's current portfolio also appear as a current holding in Arnott's PRF portfolio. Perhaps even more interesting is that for stocks that appear in both strategies, there is broad agreement on which stocks to overweight (see table below).

AC/HFI vs. PRF: Top 20 Common Holdings & Relative Weights

Ticker

Name

Weight Ratio*

Weight Ratio*

AAPL

Apple Inc

8.46

5.08

JPM

JPMorgan Chase Co

4.23

15.68

ACL

Alcon Inc

3.52

0.31

XTO

XTO Energy Inc

2.82

1.27

BAC

Bank of America Corp

2.82

21.99

CCI

Crown Castle Inter..

2.11

0.34

GOOG

Google Inc Class A

2.11

2.88

TGT

Target Cp

2.11

3.45

PFE

Pfizer Inc

2.11

13.32

XOM

Exxon Mobil Corp

2.11

28.84

TDG

Transdigm Group Inc

1.41

0.09

CLF

Cleveland-Cliffs ...

1.41

0.23

VMED

Virgin Media Inc

1.41

0.73

MCK

McKesson Corp

1.41

3.53

CVS

CVS Caremark Corp

1.41

4.02

IBM

International Bus...

1.41

8.2

WFC

Wells Fargo & Co

1.41

13.81

VZ

Verizon Comm

1.41

15.57

C

Citigroup Inc

1.41

16.45

Average Weight

0.70%

0.10%

* “Weight Ratio” is the holding's actual weight in each strategy divided by the strategy's equal weighted allocation.

That is quite a bit of overlap, but what about the 47 stocks in AlphaClone's portfolio that don't show up in PRF? Well if PRF picks 1000 stocks based on their relative value, then the 47 AC/HFI stocks that don't appear in PRF must be more growth oriented stocks. The table below lists the ten stocks that have the highest allocation in AlphaClone's portfolio but which also do not appear in PRF. As predicted, six of the ten stocks have higher P/E ratios than their sector or sub-sector but there are a couple of surprises as well; Priceline PCLN and Wet Seal WTSLA both have P/E ratios well below their sub-sectors. In fact, Wet Seal has the lowest P/E ratio of any other company listed in the Apparel Stores sub-sector. Why don't they appear in Arnott's index? More on that later.

Top 10 AC/HFI Stocks Not In PRF

Ticker

Name

P/E Stock

P/E Sector*

CIT

CIT Group Inc

23.7

14.6

GLD

SPDR Gold

51.2

n/a

SPY

SPDR S&P 500

n/a

n/a

GMCR

Green Mountain Co.

52.1

17.5

CISG

Cninsure Inc

27.2

23

GGP

General Growth Pr...

n/a

14.6

NFLX

Netflix Inc

52.6

23.1

PCLN

Priceline.com Inc

17.8

36.2

WTSLA

Wet Seal Inc Class A

4.1

19.5

MJN

Mead Johnson Nutr.

25.3

17.5

* P/E of sub-sector is used unless unavailable, then P/E of the sector used.

Digging still deeper, we compared the current sector allocations for each of the two strategies. The table below outlines the current sector allocations for each [click to enlarge].

Ac:hfi vs prf sectors

The two strategies agree on quite a bit:

  • Both strategies are the most bullish on the Financials sector, allocating roughly twice as much to it as they do to the next highest allocated sector.

  • Four of the top five sectors in each strategy are the same; Financials, information Technology, Consumer Discretionary and Energy. Combined, the four sectors account for 64% of AlphaClone's total portfolio vs 55% of Arnott's.

  • Both strategies rank Consumer Discretionary third among all sectors.

But there are also some notable differences:

  • AlphaClone's strategy seems to show a lot more conviction as measured by the spread between the highest allocated and lowest allocated sector; 28 percentage points for AlphaClone vs. 17 percentage points for PRF.

  • AlphaClone allocates nearly twice as much to the Materials sector and only half as much to the Industrials and Consumer Stapes sectors. AlphaClone's allocation to Utilities is also only one tenth that of PRF albeit both allocations are relatively low.

In summary, both the PRF and AC/HFI strategies overlap significantly across sector and individual stock allocations, and both outperform the market by a significant amount. But what gives AC/HFI the edge in performance? In our humble opinion it simply boils down to active stock selection (AC/HFI) vs. passive stock screening (PRF) .

We are huge fans of Rob Arnott and Research Affiliates. In many ways, the firm serves as a role model for AlphaClone and what we aspire to become. Their ideas are simple, powerful, intuitive and they've been able to build a formidable research and asset management business around that simplicity.

Source: Selecting Stocks Based on the 'Value Effect': AlphaClone vs. PRF