Semiconductor Index (SOX) Analysis Updated

ARA Research
ARA (April’16)- © 2016 -
ARA Research
Philadelphia Semiconductor
Sector (SOX) Index Analysis
A.R. Alvarez
April 2016

Semiconductors
ARA (April’16)- © 2016 - 2
ARA Research
SOX Index Analysis – What’s New
 Data Updated to March 31, 2016
 SOX Index vs. U.S. Recessions 13 -17
 Correlation to NASDAQ / S&P500 20 - 21
 Adding Recession Filter to SOX 38 - 41
Momentum Investing

Semiconductors
ARA (April’16)- © 2016 - 3
ARA Research
SOX Index Analysis Outline
 SOX Index History 4 - 12
 SOX Index vs. U.S. Recessions 13 - 17
 SOX Index vs. U.S. Benchmarks 18 - 26
 SOX Index Momentum Analysis 27 - 41
 Appendix: SOX Background 42 - 50

Semiconductors
ARA (April’16)- © 2016 - 4
ARA Research
SOX Index History
ARA ResearchARA Research

Semiconductors
ARA (April’16)- © 2016 - 5
ARA Research
SOX Index Companies
 SOX Established Dec 1993; Options Trading
Started Sept 1994; Split 2:1 July 1995
 Top 30 Semi Companies By Market Cap
 Of 16 Companies on Index Sept 1994:
 Eight Still on Index (ADI, AMAT, CY, IDTI, INTC, MU,
TER, TXN)
 Five Were Bought (IRF, LSI, MOT, NVLS, NSM);
Acquiring Company on Index
 Three Dropped Off (AMD, LSCC, VLSI)
 Index Weight Limited to ~8%; Limits Impact
of Larger Market Cap Companies (Broadcom
/ Avago a Temporary Exception)
 Market vs. Index Modified Market Cap
 Aggregate MC ~$750B vs. Index MMC~$450B
 Max Company MC: Intel $151B vs. ~$35B Index MMC;
Similar to TI Which Has MC of ~$59B)
 Median Index MMC ~$15B, Market Cap ~$11B (Maxim)
 Min Semi Company Market Cap Si Labs ~$1.9B
Note: Green Highlights companies added since Oct’15 (replace Atmel, KLA-Tencor, and Sandisk)
Ticker Name
Weight
(%)
Index
MMC
($B)
Market
Cap
($B)
ADI Analog Devices, Inc. 4.0 $18 $18
AMAT Applied Materials, Inc. 4.0 $18 $24
ARMH ARM Holdings plc 1.4 $6.2 $19
ASML ASML Holding NV 1.8 $8.2 $41
AVGO Broadcom (Avago) 11.6 $52 $59
CAVM Cavium, Inc. 1.0 $4.7 $3.2
CY Cypress Semiconductor 0.9 $4.2 $2.6
IDTI Integrated Device Tech 0.9 $4.0 $2.7
INFN Infinera Corporation 0.7 $3.1 $2.1
INTC Intel Corporation 7.8 $35 $151
LLTC Linear Technology Corp 3.6 $16 $11
LRCX Lam Research Corp 4.1 $18 $13
MCHP Microchip Technology Inc. 3.3 $15 $11
MPWR Monolithic Power Systems 0.8 $3.8 $2.5
MRVL Marvell Technology Group 1.7 $7.5 $5.1
MSCC Microsemi Corporation 1.4 $6.2 $4.1
MU Micron Technology, Inc. 3.6 $16 $11
MXIM Maxim Integrated Products 3.4 $15 $11
NVDA NVIDIA Corporation 4.2 $19 $20
NXPI NXP Semiconductors NV 4.3 $19 $29
ON ON Semiconductor 1.3 $6.0 $4.1
QCOM Qualcomm Incorporated 7.4 $33 $77
QRVO Qorvo, Inc. 2.1 $9.3 $5.9
SLAB Silicon Labs 0.5 $2.4 $1.9
SUNE SunEdison, Inc. 0.0 $0.2 $0.2
SWKS Skyworks Solutions Inc. 3.9 $18 $14
TER Teradyne Inc. 1.4 $6.2 $4.2
TSM Taiwan Semiconductor Mfg 7.2 $32 $130
TXN Texas Instruments Inc. 8.0 $36 $59
XLNX Xilinx Inc. 3.5 $16 $12
Market Close April'20 2016 1.0 $450 $747

Semiconductors
ARA (April’16)- © 2016 - 6
ARA Research
SOX Index Employs Algorithm to Limit Contribution of Largest Market
Cap Companies; Balances Bi-Modal Nature of Market Cap
SOX Company Market Caps
Market Cap of Companies in SOX ($B)
Data as of April 20, 2016

Semiconductors
ARA (April’16)- © 2016 - 7
ARA Research
SOX Index Cycles
SOX Index Has Exhibited 8 Cycles (Defined as ±30%) Since Inception
Through March 31, 2016

Semiconductors
ARA (April’16)- © 2016 - 8
ARA Research
SOX Index Cycle Min-Max
The Median “Up” Cycle was 123% vs. a Median “Down” Cycle of -57%, But The
Distribution Was Bi-Modal, Consistent with the SOX’s Relatively High Volatility
Median Min Max
Up 123% 50% 496%
Down -57% -69% -22%
Neutral 3% -4% 9%

Semiconductors
ARA (April’16)- © 2016 - 9
ARA Research
SOX Index Cycle Lengths
The Median “Up” Cycle Length was 19 Mths vs. a Median “Down”
Cycle Length of 10 Mths
Median Min Max
Up 19 6 30
Down 10 7 17
Neutral 14 11 17

Semiconductors
ARA (April’16)- © 2016 - 10
ARA Research
SOX Index vs. Semi Revenue
SOX Has Historically Served as Leading Indicator for Semi Revenue

Semiconductors
ARA (April’16)- © 2016 - 11
ARA Research
SOX / Semi Sales Ratio
Avg Std Dev
1995 - 1999 26 9
2000 - 2004 39 16
2001 - 2009 20 4
2010 - 2014 18 3
1995 - 2016 25 12
SOX / Semi Sales Ratio ~20 Over Past 10+ Years

Semiconductors
ARA (April’16)- © 2016 - 12
ARA Research
SOX Index Summary
 Eight “Up-Down” and 2 “Sideways” Cycles; Currently in Early Phase of a
“Down” or “Sideways” Cycle; Down ~10% From Previous Peak
 Peaked During “Internet” Bubble of 1999 & 2000; Still Trades at ~50% of
2000 Peak, Even Though Has Had a Steady Rise From 2008 Financial Crises
 Fairly Extreme Movements Either Up (>100%) or Down (> -75%) Over
Relatively Short Periods
 Median Up-Cycle Gain (123%) Was ~2X The Median Down-Cycle
 Median Up-Cycle Length (~20 Mths) Was ~2X Median Down- Cycle, So
Relative Slope (~6.5% Per Month) Was Same For Both Portions
 Setting Aside Two Neutral Cycles, Only One Cycle Was Less Than ±30%;
The Down-Cycle From June to Dec 2011 Which Only Dropped ~22%
 SOX Index Tends to Lead Semi Revenue By a Few Months on Down-Cycles,
But is Coincident With Semi Revenue Increases on Up-Cycles
 The Most Recent Cycle is an Exception Where Semi Revenue Declined Well
Before the SOX Index Started to Drop (Due to Heavy 2015 M&A ?)

Semiconductors
ARA (April’16)- © 2016 - 13
ARA Research
SOX Index:
U.S. Recessions

Semiconductors
ARA (April’16)- © 2016 - 14
ARA Research
SOX Index vs. U.S. Recession
NBER* Recessions
in Grey
SOX Index Has Exhibited Significant Drops Outside of Recession Periods
* National Bureau of Economic Research

Semiconductors
ARA (April’16)- © 2016 - 15
ARA Research
SOX Index Off-Peak Chart
NBER Recessions
in Grey
SOX “Off-Peak” Chart Since Index Inception Highlights 2000 Falloff

Semiconductors
ARA (April’16)- © 2016 - 16
ARA Research
NBER Recessions
in Grey
SOX Index Off-Peak Since 2001
SOX “Off-Peak” Chart Since 2001 Eliminates the Impact of the 2000 Run-
Up & More Clearly Shows Current Index Dynamics

Semiconductors
ARA (April’16)- © 2016 - 17
ARA Research
SOX vs. Recession Summary
 SOX Has Experienced Significant Drops Outside of Recession Boundaries
 8 Cycles with >30% Drop Identified Since Index Inception
 Index Period Encompassed 2 Recessions
 2001 Recession: The SOX Exhibited Behavior Specific to Dynamics of Its
Market Segment (Technology)
 ~ 1 Year Before Recession Start the SOX Had Major Drop; Also
Exhibited By Tech Heavy NASDAQ
 Precipitated by Telecom & Dot Com ‘Bubble’
 As Percentage of Peak, SOX Lost More Value Prior to Recession Than
After Start of Recession; Not Typical of Other Indices/Recessions
 2008 Recession: The SOX Behaved in a Manner Similar to Other Major US
Indices; i.e. If Total Drop From Previous Recent Peak = 100%:
 ~33% Drop Before Recession Start
 ~67% Drop After Recession Start
 Recovery Starts Well Before Recession End
 Given SOX’s High Correlation to Major US Indices (Pg. 20 & 21), It is Likely
to Exhibit Behavior More Like 2008 than 2001 in Next Recession

Semiconductors
ARA (April’16)- © 2016 - 18
ARA Research
SOX Index
Benchmarks

Semiconductors
ARA (April’16)- © 2016 - 19
ARA Research
SOX Index Behavior “Mimics” S&P 500 and NASDAQ – How Well?
SOX Index vs. S&P500 & NASDAQ

Semiconductors
ARA (April’16)- © 2016 - 20
ARA Research
SOX Correlation to NASDAQ
1994 - 2016 2001 - 2016 2006 - 2016
0.59 0.48 0.79
Regression Coefficient (R2)
During Past 10 Years SOX & NASDAQ Have Become Highly Correlated

Semiconductors
ARA (April’16)- © 2016 - 21
ARA Research
SOX Correlation to S&P500
1994 - 2016 2001 - 2016 2006 - 2016
0.5 0.58 0.89
During Past 10 Years SOX & S&P500 Have Become Highly Correlated

Semiconductors
ARA (April’16)- © 2016 - 22
ARA Research
$1,000 Becomes*:
Benchmark Performance: B&H
End Mar'16
S&P $4,560
Russell $4,378
DJI $4,783
NASDAQ $6,578
SOX $5,690
* Since SOX Inception

Semiconductors
ARA (April’16)- © 2016 - 23
ARA Research
SOX Index Relative Performance
vs. NASDAQ: SOX Performance Generally Worse in any Given 5
Year Period; Worse Overall
vs. S&P500: SOX Performance Better in “Good” 5 Year Periods
& Worse in “Bad” 5 Year; Slightly Better Overall
Similar Performance For All 5 Indices in Last 5 Year Period
CAGR
CAGR S&P500 Dow R2000 NASDAQ SOX
1995 - 1999 25% 26% 14% 38% 36%
2000 - 2004 -3.5% 1% 5% -12% -11%
2005 - 2009 -1.2% 2.4% -0.2% 1.4% -3.0%
2010 - 2014 13% 14% 13% 16% 14%
1995 - 2016 7.2% 7.3% 7.2% 9.3% 7.5%
Note: Through End March’16

Semiconductors
ARA (April’16)- © 2016 - 24
ARA Research
vs. NASDAQ: SOX Annualized Volatility ~50% Higher,
Decreasing Modestly (From 1.9X to 1.4X) Over Time
vs. S&P500: SOX Annualized Volatility ~2x Higher,
Decreasing Significantly (From 2.5X to 1.6X) Over Time
Annualized Volatility
Volatility S&P500 Dow R2000 NASDAQ SOX
1995 - 1999 16% 16% 14% 21% 40%
2000 - 2004 20% 20% 23% 35% 57%
2005 - 2009 24% 22% 30% 25% 32%
2010 - 2014 16% 15% 22% 18% 25%
1995 - 2016 19% 18% 22% 25% 39%
Note: Through End March’16

Semiconductors
ARA (April’16)- © 2016 - 25
ARA Research
NASDAQ, S&P 500 and Dow Indices Historical Risk Adjusted
Return (CAGR/Volatility) All ~2X Better Than SOX
For the Most Recent 5 Years (Since March 2016) The Difference
in Risk Adjusted Return Has Narrowed Such That the NASDAQ,
S&P 500 and Dow Indices Are ‘Only’ 1.5 Better Than SOX
CAGR /Annualized Volatility
CAGR/Volatility S&P500 Dow R2000 NASDAQ SOX
1995 - 1999 1.6 1.7 1.0 1.8 0.9
2000 - 2004 -0.2 0.0 0.2 -0.3 -0.2
2005 - 2009 -0.1 0.1 0.0 0.1 -0.1
2010 - 2014 0.8 1.0 0.6 0.9 0.6
1995 - 2016 0.4 0.4 0.3 0.4 0.2

Semiconductors
ARA (April’16)- © 2016 - 26
ARA Research
Indices Comparison Summary
 Over The Past 10 Years The SOX Has Exhibited Relatively High Correlation
to the NASDAQ (R2 ~0.8) and S&P500 (R2 ~0.9); Prior to That Not So Much
 For The Period 1995 – March 2016, The NASDAQ, SOX, and S&P500 CAGR
was 9.3%, 7.5%, & 7.2% With a Volatility of 25%, 39%, & 19% Respectively
 SOX Relative Risk (CAGR/Volatility) Was Worse Than NASDAQ or S&P500
 Individual 5 Year Periods For Most Part Exhibit Similar Relative Results
 $1,000 Invested in Each of the 5 Indices Since May 1994 Grow to ~$6,600,
~$5,700, ~$4,400 to ~$4,800 (Excluding Dividends) March 2016 For
NASDAQ, SOX, DJI / Russell 2000/S&P500 Respectively
 Whereas The Four Benchmark Indices are Currently Above or Close to
Their All Time High, The SOX is at ~50% of It’s 2000 Peak
 If One Were Unlucky Enough to Have Invested at or Near The SOX Peak in
2000, They Would Still Have Lost ~50% of The Investment as of March 2016
 The SOX Has Exceeded Its Pre-2008 Financial Crises ‘Peak’, But It Took 6
Years to Recover


Semiconductors
ARA (April’16)- © 2016 - 27
ARA Research
SOX Index
Momentum Analysis

Semiconductors
ARA (April’16)- © 2016 - 28
ARA Research
SOX Index Momentum
 Objective:
 Determine if Simple Momentum Rules Can Be Utilized on the SOX Index to
Obtain a Better Return Than a “Buy & Hold” Strategy
 Two Distinct Momentum Rules Utilized
 First Momentum Rule (Momentum Always “On”):
 Conventional Use of Momentum
 Buy on Positive Momentum: “Current” Price > Lookback Period Price
 Sell on Negative Momentum: “Current” Price < Lookback Period Price
 Second Momentum Rule (Momentum “On” Only During Recession):
 Growth Trend Timing: http://www.philosophicaleconomics.com/2016/01/gtt/
 Use Conference Board Co-Incident Composite Indicator (or Your Favorite
Recession Indicator(s) ) to Determine Whether In or Out of Recession
 Outside Recession: Buy on Positive Momentum or Stay Invested (B&H)
 Inside Recession: Apply First Momentum Rule

Semiconductors
ARA (April’16)- © 2016 - 29
ARA Research
Momentum Calculations
 Daily and End of Month Market Close Calculations Utilized
 Lookback Periods:
 Absolute Performance: 3, 6, and 12 Month
 Absolute Performance With Non-Zero (2.5%) Cutoff: 6 Months
 Moving Average: 200 Day Simple Moving Average and Various Exponentially
Weighted Moving Averages
 Buy-Sell Decisions Made Once Per Mth; Last Trading Day of
the Mth
 Tax Scenarios (Deferred vs. Instant):
1) Taxed at End of Holding Period (Retirement Account)s; Also True for B&H
2) Taxed at 33% on a Trade Gain, No Tax Benefit on Trade Loss
 Fees/Spread: Low Cost ($10) Trades, Spread Not Taken Into
Account
 Invested Return and Maximum Draw-Down Calculated

Semiconductors
ARA (April’16)- © 2016 - 30
ARA Research
Momentum Summary: 1st Rule
 A Momentum-Based Trading Rule Can Yield Better Results Than B&H
 Lookback Period Matters – A Lot: Only 200 Day SMA and 3 Mth Lookbacks
Provided Greater Returns Than B&H in a Taxed Account
 Taxes Matter: In a Deferred Tax Account Momentum-Based Trading Yields
Better Results Than B&H in All Cases Except for 12 Mth Lookback
 No Momentum Rule Will Outperform B&H All of The Time; There Can be
Significant Periods (Years) of Underperformance
 3 Month Lookback Period Gave Best Overall Performance Over 20 Yr
Period, But in a Taxed Account Underperformed B&H for First 14 Years
 200 Day Simple Moving Avg Outperformed B&H Through 2013, But Has
Underperformed Since & is Now ~ Break-even in a Taxed Account
B&H Taxed No Tax Taxed No Tax Taxed No Tax Taxed No Tax Taxed No Tax
$1,000 Becomes $4,142 $5,103 $10,147 $2,650 $4,289 $1,860 $2,279 $4,276 $6,877 $3,685 $5,611
$ Trades (B&S) n/a 58 58 44 44 26 26 38 38 42 42
Avg Trades/Year n/a 2.9 2.9 2.2 2.2 1.3 1.3 1.9 1.9 2.1 2
Max Draw Down 73% 60% 50% 57% 44% 65% 60% 50% 42% 57% 44%
Median Draw Down 52% 37% 29% 42% 38% 39% 39% 35% 30% 41% 36%
3 Mth Lookback 6 Mth Lookback 12 Mth Lookback 200 Day SMA 6M LB / 2.5% CritSOX Results Through
March 31, 2016

Semiconductors
ARA (April’16)- © 2016 - 31
ARA Research
SOX B&H vs. Mtm 3 Mth Lookback
B&H Taxed No Tax
$1,000 Becomes $4,142 $5,103 $10,147
$ Trades (B&S) n/a 58 58
Avg Trades/Year n/a 2.9 2.9
Max Draw Down 73% 60% 50%
Median Draw Down 52% 37% 29%
3 Mth Lookback

Semiconductors
ARA (April’16)- © 2016 - 32
ARA Research
B&H Taxed No Tax
$1,000 Becomes $4,142 $2,650 $4,289
6 Mth Lookback

Semiconductors
ARA (April’16)- © 2016 - 33
ARA Research
B&H Taxed No Tax
$1,000 Becomes $4,142 $1,860 $2,279
12 Mth Lookback

Semiconductors
ARA (April’16)- © 2016 - 34
ARA Research
SOX Index & 200 Day SMA

Semiconductors
ARA (April’16)- © 2016 - 35
ARA Research
B&H vs. 200 Day SMA Lookback
B&H Taxed No Tax
$1,000 Becomes $4,142 $4,276 $6,877
200 Day SMA

Semiconductors
ARA (April’16)- © 2016 - 36
ARA Research
Impact of Non-Zero Threshold
Criteria: 6 Mth Look Price ≥ 2.5% Higher
Than Current Price to “Buy” Through March 31, 2016
B&H Taxed No Tax
$1,000 Becomes $4,142 $3,685 $5,611
6 Mth LB / 2.5% Criteria

Semiconductors
ARA (April’16)- © 2016 - 38
ARA Research
Where Are We So Far?
 Compared to “Buy & Hold” Investing No Really Good
Alternatives Using Simple Momentum Rules on SOX
 Results Are Highly Dependent on Lookback Period; Only
One of The Four Lookback Periods Beat B&H in a Taxed
Account With Any Degree of Margin
 The Best Performing Lookback Period (3 Mths), While
Producing a Good End Result, Under-Performed “Buy &
Hold” For The First 14 Years of The Index
Is There a Way To Adjust Momentum in a Simple
Manner That Significantly Improves The Results?

Semiconductors
ARA (April’16)- © 2016 - 39
ARA Research
Momentum Summary: 2nd Rule
 Applying Recession Filter (Use Mtm Rule Only Within Recession Limits)
Significantly Improves Performance Over Using Mtm Rule at All Times
 Required Very Few (Only 6 Trades; 3 Buy-Sells) Trades Over 20+ Years
 Does Not Provide as Good Overall Return as Mtm 3 Mth Lookback, But Does
Outperform B&H in Any Given Month For Almost The Entire 20+ Year Period
 As Noted Previously The Mtm 3 Mth Lookback Method Underperformed B&H
For Most of the First ~14 Years of the SOX (Pg. 37)
 Much More Likely to Stay With an Investment Strategy When It Out-Performs
Alternative Strategies in Most Years
B&H Taxed No Tax Taxed No Tax Taxed No Tax
$1,000 Becomes $4,142 $5,103 $10,147 $4,276 $6,877 $6,917 $8,590
$ Trades (B&S) n/a 58 58 38 38 6 6
Avg Trades/Year n/a 2.9 2.9 1.9 1.9 0.3 0.3
Max Draw Down 73% 60% 50% 50% 42% 66% 66%
Median Draw Down 52% 37% 29% 35% 30% 52% 42%
SMA - Recession3 Mth Lookback 200 Day SMASOX Results Through
March 31, 2016

Semiconductors
ARA (April’16)- © 2016 - 40
ARA Research
Momentum Rule Only in Recession
Criteria: 200 Day SMA Applied Only During Recessionary Periods
B&H Taxed No Tax
$1,000 Becomes $4,142 $6,917 $8,590
200 Day SMA - RecessionSOX Results Through
March 31, 2016

Semiconductors
ARA (April’16)- © 2016 - 41
ARA Research
Momentum: Always vs. Recession
Criteria: 200 Day SMA – Res = Momentum Rule Applied Only During Recessions
200 Day SMA – Momentum Rule Applied Entire Time
B&H Always Recessions
$1,000 Becomes $4,142 $4,276 $6,917
200 Day SMA
SOX Results Through
March 31, 2016

Semiconductors
ARA (April’16)- © 2016 - 43
ARA Research
WW Semiconductor Sales
Growth Rate Drop From Historical 15% CAGR to 6% Since 2000
Given Industry Size & Penetration Don’t Expect Return to 15% CAGR Unless
Global GDP Exceeds 4 – 4.5% Consistently
“Semiconductor Industry Primer 2015”
D. Wong (Wells Fargo) Oct’15

Semiconductors
ARA (April’16)- © 2016 - 44
ARA Research
NASDAQ Correlation to S&P500
1994 - 2016 2001 - 2016 2006 - 2016
0.90 0.94 0.93

Semiconductors
ARA (April’16)- © 2016 - 45
ARA Research
SOX Index Summary
 PHLX Semiconductor Sector Index is a Modified Capitalization Weighted
Index Made Up of the Major Companies in the Semiconductor Sector
 Two Versions
 Price Return Index (SOX): Calculated Without Regard To Cash Dividends on Index
Securities (Initiated Dec 1st 1993 at Base Value = $100)
 Total Return Index (XSOX): Reinvests Cash Dividends on Ex-Date.; Synchronized to
Value of The Price Return Index at The Close on Dec 22, 2009
 Both Indexes Reflect Extraordinary Cash Distributions
 Index Eligibility (Main Points)
 Be Listed on NASDAQ Stock Market or New York Stock Exchange
 Have Minimum Market Capitalization of $100 Million
 Have Traded at Least 1.5 Million Shares in Each of Previous Six Months
 Index Evaluation
 Index Securities Evaluated Every Sept; Additions/Deletions Effective 3rd Friday of Sept
 Securities Meeting Eligibility Criteria as of End of July Ranked By Market Capitalization
 Top 30 Securities By Market Capitalization Included in Index

Semiconductors
ARA (April’16)- © 2016 - 46
ARA Research
SOX Index Calculation
 Value of Index Equals Aggregate Value of Index Share Weights
Step 1) Calculate Total Market Value of Index = Sum of Market Value of Each Index Component = The
Last Sale Price Multiplied By Number of Shares Outstanding
Step 2) Calculate Weight of Each Index Component = Its Market Value Divided By Total Market Value
of Index
Step 3) Employ Capitalization Weighting Adjustment/Modification Algorithm to Limit Contribution of
Largest Market Cap Companies to 8% or Less (See Rebalancing Section)
Step 4) Calculate Aggregate “Modified” Capitalization Weighted Market Value = Sum of Each Index
Component's Weight and Last-sale Price
Step 5) Divide Aggregate Weighted Market Value By Divisor, Which Scales Aggregate Value to Lower
Value More Suitable For Reporting
 Index Balancing
 Index Employs a Modified Market Capitalization Weighting Methodology
 Each Qtr, Index Rebalanced Such That Max Weight of Any Security Does Not Exceed 8% and No
More Than 5 Securities Are at That Cap
 Excess Weight of Any Capped Security is Distributed Proportionally Across Remaining Index
Securities
 If After Redistribution, Any of The 5 Highest Ranked Index Securities are Below 8%, These
Securities are Not Capped
 Next, Any Remaining Index Securities in Excess of 4% are Capped at 4% and The Excess Weight
is Redistributed Proportionally Across Remaining Index Securities
 Process is Repeated, As Necessary, To Derive The Final Weights

Semiconductors
ARA (April’16)- © 2016 - 47
ARA Research
SOX Index Companies 2013 - 15
Ticker Name
Weight
(%)
Index
MMC
($B)
Market
Cap
($B)
Ticker Name
Weight
(%)
Index
MMC
($B)
Ticker Name
Weight
(%)
Index
MMC
($B)
ADI Analog Devices, Inc. 4.1 $18.4 $20.0 ADI Analog Devices, Inc. 3.9 $19.8 ADI Analog Devices, Inc. 3.8 $9
AMAT Applied Materials, Inc. 3.7 $16.3 $19.7 ALTR Altera (Intel) 2.7 $13.5 ALTR Altera (Intel) 3.3 $7
ARMH ARM Holdings plc 1.8 $8.0 $23.1 AMAT Applied Materials, Inc. 6.2 $31.4 AMAT Applied Materials, Inc. 6.3 $14
ASML ASML Holding NV 1.8 $8.0 $40.3 ARMH ARM Holdings plc 1.4 $6.9 AMD Advanced Micro Devices 0.7 $2
ATML Atmel Corporation 1.0 $4.5 $3.6 ASML ASML Holding NV 2.0 $10.3 ARMH ARM Holdings plc 1.8 $4
AVGO Avago Technologies 7.2 $32.3 $35.7 ATML Atmel Corporation 0.8 $4.0 ASML ASML Holding NV 3.0 $7
CAVM Cavium, Inc. 1.1 $5.0 $4.0 AVGO Avago Technologies 4.2 $21.6 ATML Atmel Corporation 1.0 $2
CY Cypress Semiconductor 0.9 $4.1 $3.4 BRCM Broadcom (Avago) 4.3 $21.8 AVGO Avago Technologies 3.4 $8
INFN Infinera Corporation 0.7 $3.3 $2.4 CREE Cree 0.9 $4.8 BRCM Broadcom (Avago) 3.8 $9
INTC Intel Corporation 8.7 $39.1 $166.0 FSL Freescale (NXP) 1.5 $7.7 CREE Cree 2.2 $5
KLAC KLA-Tencor Corporation 2.9 $12.9 $10.2 INTC Intel Corporation 7.8 $40.0 FSL Freescale (NXP) 1.2 $3
LLTC Linear Technology Corp 3.1 $13.9 $11.3 KLAC KLA-Tencor Corporation 3.3 $16.7 INTC Intel Corporation 8.0 $18
LRCX Lam Research Corporation 3.2 $14.5 $11.8 LLTC Linear Technology Corp 2.6 $13.0 KLAC KLA-Tencor Corporation 3.3 $8
MCHP Microchip Technology Inc. 2.9 $13.0 $10.4 LRCX Lam Research Corporation 3.1 $16.1 LLTC Linear Technology Corp 2.9 $7
MRVL Marvell Technology Group 1.3 $6.0 $5.0 MCHP Microchip Technology Inc. 2.2 $11.0 LRCX Lam Research Corporation 2.7 $6
MSCC Microsemi Corporation 1.0 $4.3 $3.5 MRVL Marvell Technology Group 1.7 $8.7 LSI LSI CORP 1.4 $3
MU Micron Technology, Inc. 3.7 $16.5 $18.6 MU Micron Technology, Inc. 8.1 $41.4 MCHP Microchip Technology Inc. 2.6 $6
MXIM Maxim Integrated Products 3.1 $14.1 $10.8 MXIM Maxim Integrated Products 2.1 $10.6 MRVL Marvell Technology Group 1.8 $4
NVDA NVIDIA Corporation 4.3 $19.4 $15.4 NVDA NVIDIA Corporation 2.6 $13.5 MU Micron Technology, Inc. 4.7 $11
NXPI NXP Semiconductors NV 4.2 $18.8 $22.3 NXPI NXP Semiconductors NV 4.0 $20.6 MXIM Maxim Integrated Products 2.6 $6
ON ON Semiconductor 1.3 $5.7 $4.7 ON ON Semiconductor 0.9 $4.7 NVDA NVIDIA Corporation 2.7 $6
QCOM Qualcomm Incorporated 7.7 $34.7 $95.4 QCOM Qualcomm Incorporated 8.3 $42.3 NXPI NXP Semiconductors NV 3.2 $7
QRVO Qorvo, Inc. 2.1 $9.4 $7.7 RFMD RF Micro Devices (Quorvo) 0.9 $4.8 ON ON Semiconductor 1.0 $2
SNDK SanDisk Corp. 4.5 $19.9 $16.0 SNDK SanDisk Corp. 3.9 $19.7 QCOM Qualcomm Incorporated 7.6 $17
SUNE SunEdison, Inc. 0.7 $3.1 $2.5 SUNE SunEdison, Inc. 1.3 $6.7 SNDK SanDisk Corp. 4.6 $10
SWKS Skyworks Solutions Inc. 3.4 $15.2 $15.8 SWKS Skyworks Solutions Inc. 2.8 $14.0 SWKS Skyworks Solutions Inc. 1.5 $3
TER Teradyne Inc. 1.1 $5.0 $4.0 TER Teradyne Inc. 0.9 $4.8 TER Teradyne Inc. 1.0 $2
TSM Taiwan Semiconductor Mfg 6.0 $26.7 $116.5 TSM Taiwan Semiconductor Mfg 4.2 $21.6 TSM Taiwan Semiconductor Mfg 5.9 $13
TXN Texas Instruments Inc. 8.8 $39.4 $60.5 TXN Texas Instruments Inc. 8.3 $42.3 TXN Texas Instruments Inc. 8.0 $18
XLNX Xilinx Inc. 3.5 $15.6 $48.7 XLNX Xilinx Inc. 3.0 $15.2 XLNX Xilinx Inc. 3.7 $8
Market Close Oct'22 2015 1.0 $447 $809 1.0 $509 1.0 $227
October 2015 October 2014 October 2013

Semiconductors
ARA (April’16)- © 2016 - 48
ARA Research
SOX Index Sept vs. Oct 2010
 SOX Index Transitioned to
Modified Market Cap Algorithm
and 30 Companies Oct 2010
 Prior to That SOX Index Carrier
16 – 21 Companies, While the
iShares ETF (SOXX) Carried 52
Companies
 Since Oct 2010 Both The SOX
and SOXX Carried the Identical
30 Companies
 Blue = Companies Added
Oct’2010
 Weightings Shown are for SOXX
Ticker Name
Weight
(%)
Index
MMC
($B)
Ticker Name
ALTR Altera (Intel) 4.18 $7.4 ALTR Altera (Intel)
AMAT Applied Materials, Inc. 7.86 $13.9 AMAT Applied Materials, Inc.
AMD Advanced Micro Devices 3.25 $5.8 AMD Advanced Micro Devices
ATHR Atheros (Qcom) 1.45 $2.6 ATHR Atheros (Qcom)
AVGO Avago Technologies 3.88 $6.9 BRCM Broadcom (Avago)
BRCM Broadcom (Avago) 8.98 $15.9 CREE Cree
CREE Cree 3.16 $5.6 HITT Hittite Microwave (ADI)
CRUS Cirrus Logic 0.57 $1.0 INTC Intel Corp
HITT Hittite Microwave (ADI) 1.04 $1.8 KLAC KLA-Tencor Corp
INTC Intel Corp 7.48 $13.2 LLTC Linear Technology Corp
KLAC KLA-Tencor Corp 3.94 $7.0 LRCX Lam Research Corp
LLTC Linear Technology Corp 3.72 $6.6 MU Micron Technology
LRCX Lam Research Corp 3.74 $6.6 NSM National Semi (TI)
MKSI MKS Instruments 0.68 $1.2 NVLS Novellus Systems (Lam)
MRVL Marvell Technology Group 6.85 $12.1 SNDK SanDisk Corp.
MU Micron Technology, Inc. 4.22 $7.5 STM ST Microelectronics
NETL Netlogic Micro (BCOM) 1.26 $2.2 TER Teradyne
NSM National Semi (TI) 2.16 $3.8 TSM Taiwan Semiconductor Mfg
NVDA NVIDIA Corporation 4.26 $7.5 TXN Texas Instruments Inc.
NVLS Novellus Systems (Lam) 1.77 $3.1 WFR MEMC Electronics
POWI Power Integrations 0.63 $1.1 XLNX Xilinx Inc.
RBCN Rubicon Technology 0.35 $0.6
SNDK SanDisk Corp. 3.73 $6.6
STM ST Microelectronics 0.46 $0.8
TER Teradyne Inc. 1.34 $2.4
TSM Taiwan Semiconductor Mfg 3.83 $6.8
TXN Texas Instruments Inc. 8.48 $15.0
VECO Veeco Instruments 1.13 $2.0
WFR MEMC Electronics 1.92 $3.4
XLNX Xilinx Inc. 3.67 $6.5
100% $177
September 2010October 2010

Semiconductors
ARA (April’16)- © 2016 - 49
ARA Research
SOX Index Cycles Table
Median Min Max
Up 123% 50% 496%
Down -57% -69% -22%
Neutral 3% -17% -4%
Median Min Max
Up 19 6 30
Down 10 7 17
Neutral 14 11 17
Cycle Change Summary Cycle Length Summary
Cycle
Start
Cycle
End
Months
in Cycle
Cycle
Change
Cycle
Start
Cycle
End
Months
in Cycle
Cycle
Change
May-94 Sep-95 17 153% Feb-04 Aug-04 7 -65%
Oct-95 Jul-96 10 -52% Sep-05 Apr-06 20 50%
Aug-96 Sep-97 14 171% May-06 Sep-07 17 -4%
Oct-97 Aug-98 11 -50% Oct-07 Feb-09 17 -63%
Sep-98 Aug-00 24 496% Mar-09 May-11 27 133%
Sep-00 Sep-01 13 -69% Jun-11 Dec-11 7 -22%
Oct-01 Mar-02 6 72% Jan-12 Nov-12 11 9%
Apr-02 Jan-03 10 -57% Dec-12 May-15 30 93%
Feb-03 Jan-04 12 113% Jun-15 Mar-16 5 -9%

Semiconductors
ARA (April’16)- © 2016 - 50
ARA Research
Index Volatility
 Basic Definition:
 Standard Deviation of Historical Returns
 Basically a “Dispersion” Measure; Greater Dispersion (Volatility) Implies Greater Risk
 Upside & Downside Price Movements Are Equally Considered in the Volatility
Calculation; The Sortino Ratio Attempts to Capture Just Downside Volatility
 Calculation
Step 1) Calculate Daily Periodic Stock Return (Ri) = (Pi – Pi-1)/Pi ; Where Pi is Current
Day’s Closing Price and Pi-1 is Closing Price on Previous Day
 More Often Ri = Ln(Pi/Pi-1) is Used as it Better Represents a Continuously
Compounded Return; Approximately The Same For Most Short Intervals
 This was the Method Employed in This Study
Step 2) Decide on Time Period; Volatility for Both the Full 20 Year (1994 – 2014) as Well
as 5 Year Intervals Was Calculated for The 3 Indices Examined in This Study
Step 3) Calculate Daily Sample Variance = [ Sum(Pi-1 – Avg P)2 ] / (m-1); Where ‘m’ =
Number of Sample Data Points Over Which The Sum is Performed
Step 4) Calculate Annualized Sample Volatility = Sqrt (Sample Variance) x Sqrt (250);
250 Being The Number of Trading Days in a Typical Year
 Note: To Calculate Annualized Weekly or Monthly Volatility Multiply The
Corresponding Weekly or Monthly Volatility By Sqrt(52) or Sqrt(12) Respectively;
The Calculated Volatility Will Be Slightly Different in Each Case

Semiconductor Index (SOX) Analysis Updated

More Related Content

More from Tony Alvarez (10)

Recently uploaded (20)

Semiconductor Index (SOX) Analysis Updated