Analysts, Industries, and Price Momentum

1
Analysts, Industries, and Price Momentum

• Leslie Boni, University of New Mexico Kent L.
  Womack, Tuck School at Dartmouth

2
(No Transcript)
3
Premise of the Paper

• Analysts are industry experts, they follow one
  (or sometimes two) industries
• ? If industry-level analysis is their main
  expertise, the proper test of their skill is
  within-industry comparisons
• ? So, we do industry-based analysis

4
Our Research Questions

• Can and do analysts identify winning and losing
  stocks within their industry specializations?
• Why is intra-industry analysis better, if it is?
• How do analysts recommendation changes
  temporally relate to price momentum and industry
  momentum? i.e. Do they predict or follow
  industry momentum?

5
Data

• CRSP returns and market capitalizations for
  companies traded on NYSE, AMEX, and Nasdaq
• Industry categories SP/Morgan Stanley Global
  Industry Classification Standards (GICS) codes
  (via Compustat)
• 59 industries, closest to how the Street breaks
  down coverage
• I/B/E/S (now part of Thomson, First Call)
  collects rating changes from virtually all
  brokerage firms, we use US stock exchanges
• 186,217 recommendation changes from 433
  brokerages (including subs) on 7,980 companies
• Time period Jan. 1996 Nov 2002
• There are 12,395 companies with CRSP returns and
  GICS codes
• So 4,415 of these are neglected firms, not
  followed

6
Table 1A of Recommendation Upgrades/Downgrades
7
Table 1B Recommendation Upgrades/Downgrades3-Day
Rets
8
Table 2 Brokerage Firm Characteristics
9
Table 3 Analyst Coverage in GICS Industries
10
Question 1 Can and do analysts identify
winning and losing stocks within their industry
specializations?

• Research Design Self-financing portfolios
• We compare a consensus-level strategy . . .
• Consensus Level Strategy 1 Buy Best
  consensus level stock, Short worst consensus
  level stock portfolios equally weighted by
  industry
• . . . with two recommendation-change strategies
  
• ? Define a net AgChange measure (Upgrades)
  - (Downgrades) in month
• Changes Strategies Buy All net upgraded
  stocks, Short all net downgraded stocks
• 1. Equally weight portfolios by industry
• 2. Equally weight all stocks

11
Table 4 Returns from Recommendation Portfolios
Bottom Line Levels are virtually worthless, but
Changes are quite valuable, even starting, on
average, one-half month after the Rec Change
(since they are evenly spaced over previous
month).
12
Question 2 Why Does Industry Analysis
Help?Table 6The Value of Industry
Diversification
13
Table 4 and Table 5, col. 5 The Consistency of
Returns from Industry Portfolios
14
Table 4 Returns from Recommendation Portfolios
(continued)
Note All portfolios load positively on
momentum factor, but especially non-industry
weighted ones.
15
Table 4Value Weighted Returns from
Recommendation Portfolios
16
Fig 2 Dilution of long-short return as we
constrain market cap of stocks recommended
17
Figure 1Compare to vanilla momentum strategy
18
Industry Rec Information Measures

• We construct two monthly aggregate measures of
  industry recommendation information.
• The original AgChange for each stock, aggregated
  for each industry a basic measure of
  positive/negative information flow from analysts
• AgChanges/Analyst Coverage in the stock
• reduces weighting of stock in industry measure
  when more analysts . . .

19
Question 3 Cross Industry Issues/Questions

• Do industry returns exhibit price momentum when
  industries are classified using the GICS
  classifications?
• Does analyst recommendation information,
  aggregated by industry, exhibit month-to-month
  momentum?
• Do industries that are net upgraded (downgraded)
  have better (worse) returns than other industries
  in the month in which the recommendation changes
  occur?
• Do analysts chase industry price momentum?
• Does analyst recommendation information have
  value for picking future hot and cold industries?

20
Table 8Serial Correlation Industry Returns
Rec Measures
21
Table 9Cross Serial Correlation Industry Rets
vs. Rec Measures
22
Table 10Panel A
23
Table 10Panel B
24
Conclusions and Future Work

• Conclusions
• Analysts are willing and able to signal winning
  and losing stocks within their industries. Key
  signal upgrades and downgrades
• Our (not optimized) Recommendation-change
  Strategy shows
• Returns 1.2 per mo., 14 per yr.
• With returns MUCH less volatile than popular
  price momentum strategies, and they avoid the
  negative January effect
• Industry-rotation strategies of buying upgraded
  stocks in hot industries and shorting
  downgraded stocks in cold industries produce
  slightly higher returns (1.4 per month) but
  with greater risk than the industry-neutral
  strategies.
• Policy Implications
• Comparisons of research firm performance that
  fail to account for the degree and substantial
  benefit of industry diversification may lead to
  incorrect inferences