Dave Berguer

H.J. Turtle

CROSS-SECTIONAL PERFORMANCE
AND INVESTOR SENTIMENT IN A
MULTIPLE RISK FACTOR MODEL

Objective
Are opaque securities more sensitive to

measures of market sentiment?

Is ex-ante known investor sentiment related

to marginal performance of opaque and

translucent securities?

Investor sentiment and firmcharacteristic data

Sentiment is considered broadly as general

optimism os pessimism towards future stock

returns;
Sentiment is measured using the monthly

sentiment index of Baker and Wurgler;

Measuring attributes of
sentiment-prone stocks

Measuring attributes of
sentiment-prone stocks
The regression model used was:

For j = 1, 2, ..., N;
For t = 1, 2, ..., N;
Where:
N = number of cross-sectional observations;
T = number of time series;
Rj,t = excess return of asset j during period t;
Rm,t = excess market return during period t;

Measuring attributes of
sentiment-prone stocks
Based on j,sent, stocks are grouped into 10

portfolios;
A typical firm in the high sentiment

sensitivity grouping is expected to display

volatile returns, a small equity base, low
earnings, low dividends, high distress risk and
have a relatively intangible assets.

Measuring attributes of
sentiment-prone stocks
Results in table 2 support the hypothesis that

firms with high sensitivity to investor sentiment

tend to be relatively opaque;

Table 2 document a strong relation between the

firms that we estimate to have the highest

sensitivity to investor sentiment, and the opaque
firm characteristics that Baker and Wurgler
(2006) hypothesize, after controlling for market
risk.

Measuring attributes of
sentiment-prone stocks
To provide robustness results regarding the

relation between firm characteristics and

sentiment sensitivities documented in table
2, the authors expand their model to control
for multiple risk sources.

Measuring attributes of
sentiment-prone stocks
Augmented regression model:

Where:
Rsmb, Rhml, and Rmom, represent the small minus
big, high minus low, and momentum risk factors,
respectively.
Risk factor data comes from Ken Frenchs data

library.

Measuring attributes of
sentiment-prone stocks
Again, results strongly support the

hypothesis that sentiment-prone stocks

display opaque firm characteristics;
However, the difference in sample averages

of firm characteristics across portfolios are

dampened when additional risk factors are
considered.

Measuring attributes of
sentiment-prone stocks
Table 2 and 3 corroborates the hypothesis

that sentiment-prone stock portfolios are

small, intangible and volatile;
To verify that sentiment-sensitivities exhibit

similar patterns across firm characteristics

with ex ante available information, eq 1 and 2
are reestimated using roling windows.

Measuring attributes of
sentiment-prone stocks
As expected, younger firms, and firms with

larger root mean square error, are more

opaque, and more sensitive to sentiment;
Results are congruent with prior analysis,

except for BM in table 5 that loses its

significance.

Performance conditional on
investor sentiment

Performance conditional on
investor sentiment
Analysis to this point documents a robust

relation between opacity and sentiment;

Analysis is now shifted to whether ex ante

known sentiment result in positive portfolio

performance.

Performance conditional on
investor sentiment
Subsequent analysis considers expected

marginal performance during period t, given

only information available in t-1;
Conditional alphas provide the marginal

performance of a given sentiment portfolio

for a given level of systematic risk.

Performance conditional on
investor sentiment
Conditional alpha is estimated directly from

the folowing unconditional regression:

Where the conditioning information instrument,

Sent, is known at the beginning of each

investment interval.

Performance conditional on
investor sentiment
Previous analysis indicates strong relation

between sentiment sensitivities and firm

characteristics;
Firm characteristics of previous analysis are

now used to form 10 portfolios based on a

firms ranking of the specific characteristic at
that point in time;

Performance conditional on
investor sentiment
Given a known investor sentiment realization,

the conditional alpha may be written as:

The resultant conditional alpha measures

marginal performance from the conditional
regression of the portfolio return against the risk
factors where excess returns for all portfolios
and factors are linearly related to the underlying

information instruments.

Table 6
Suggests the expected contrarian nature of

sentiment as a conditioning variable for

opaque firms;

Table 7
We observe large variation in marginal

performance for opaque firms across levels of

investor sentiment, with little variation in
marginal performance for translucent firms
across the same levels of investor sentiment.

Conditional alpha across

volatility portfolios
Variation in conditional alpha, across portfolios, for a given level of
sentiment, for a given portfolio

Variation in the conditional alpha, across

portfolios, for a given sentiment level

Conditional alpha across

size portfolios

Conditional alpha across age

portfolios

Conditional alpha across

earnings portfolios

Conditional alpha across

dividend portfolios

Conditional alpha across

prop, plant & eqp portfolios

Conditional alpha across R&D

portfolios

Conditional alpha across B/M

portfolios

Conditional alpha across

sales growth portfolios

Table 8
Robustness test
Expands the model to account for other risk

factors;
Results are robust.

Conclusion
Most sentiment-prone stocks tend to exhibit

opaque characteristics (volatile, small, young

and intangible);
Opaque portfolios offer the greatest marginal
performance when sentiment levels are
lowest;
Opaque portfolios exhibit much greater
variation in conditional alpha estimates
across levels of investor sentiment relative to
translucent portfolios.