Bad Beta, Good Beta By John Y. Campbell And Tuomo Vuolteenaho Presentation by Michael-Paul James Cash flow and discount rate betas estimates stock market risk factors more efficiently than CAPM over time. ○ Cash flow news ■ Stock returns covariance with cash flows news ○ Discount rate news ■ Stock returns covariance with discount rate news ● “Bad” cash flow beta (risk) demands higher premiums than “good” discount rate beta (risk). ● Value and small stocks have higher cash flow betas than growth and large stocks on average. ● High average returns on value and small stocks are appropriate compensation for risk, not an unrealized benefit to ownership. ● Overweighting small and value stocks benefit low risk aversion equity investors ● Underweighting small and value stocks benefit high risk aversion equity investors ● Model offers strong explanatory power in the cross section of asset returns with theoretical values ● ICAPM outperforms the CAPM in empirical research

1. Bad Beta, Good Beta
By John Y. Campbell And Tuomo Vuolteenaho
Presentation by Michael-Paul James
1

2. Table of contents
2
Michael-Paul James
Introduction
story, questions, context, issues,
literature
01
News
How Cash-Flow News and
Discount-Rate News Move the
Market, Return-Decomposition
Framework, VAR State Variables and
Estimation
02
Measure Betas
Measuring Cash-Flow and
Discount-Rate Betas
03
Pricing Betas
Pricing Cash-Flow and
Discount-Rate Betas, An
Intertemporal Asset Pricing Model,
Empirical Estimates of Risk Premia,
Additional Robustness Checks,
Loose Ends and Future Directions
04
Conclusions
05

3. Introduction
01
story, questions, context, issues, literature
3
Michael-Paul James

4. Measuring Stock Market Risks
4
Michael-Paul James
● CAPM
○ Beta with the market portfolio of all invested wealth measures a
stock’s risk, not other characteristics should influence a rational
investors required return.
○ CAPM fails after 1960 in empirical tests
■ Small and value stocks delivered higher average returns then
their beta suggests.
■ High past betas produce similar returns to low past betas
controlling for size
■ Empirical biases favors small and value stocks with low betas.

5. Measuring Stock Market Risks
5
Michael-Paul James
● Two components of the market portfolio
○ News about future cash flows
■ Bad news: Wealth decreases and investment opportunities
remain the same
○ News about changes in the discount rate or cost of capital
■ Bad news: Wealth decreases but future investment
opportunities improve
○ Assumption
■ Investors demand higher premiums on assets that covary with
cash flow news than covary with discount rates which are
compensated with higher potential future returns

6. Measuring Stock Market Risks
6
Michael-Paul James
● Division of news
○ Cash flow beta
○ Discount rate beta
● Merton’s (1973) ICAPM
○ Discount rate beta equals the variance of the market return
○ Cash flow beta equals γ (relative risk aversion) times the variance of
the market return
■ If γ > 1 then the risk averse investor demand a higher premium
for cash flow risk
● Labeling
○ Bad cash flow beta (high risk premium) versus (relatively) good
discount rate beta (low risk premium)

7. Measuring Stock Market Risks
7
Michael-Paul James
● Improvements to the model
○ After 1963, the two beta model (ICAPM) improves predictive power
over CAPM
■ Possibly explains:
● Why growth stocks with high (good) betas have lower
average returns.
● Why value stocks with high (bad) betas have high average
returns.

8. Measuring Stock Market Risks
8
Michael-Paul James
● Campbell’s (1993) improvement to ICAPM
○ Solves discrete time empirical version
○ Assumes asset returns are homoskedastic
■ Variance of the residual (error term) is constant.
○ Investor has recursive preferences
■ Describe trade-offs between current and future consumption
using with the certainty equivalent of next period's utility
(single index for future)

9. Measuring Stock Market Risks
9
Michael-Paul James
● Paper improvements
○ Small stock value spread to predict aggregate stock returns.
■ Empirically high returns to small growth stocks forecast low
returns on the aggregate stock market
■ Hypothetically due to negative correlation with innovations to
investment opportunities

10. News
02
How Cash-Flow News and Discount-Rate News Move the Market
Return-Decomposition Framework
VAR State Variables and Estimation
10
Michael-Paul James

11. ● Loglinear approximate decomposition of returns
○ Accounting identity: stock returns are associated with future cash
flow or future discount rate news
○ Expected increase in cash flows increases future capital gains
○ Expected increase in discount rate decreases future capital gains
○
● rt+1
log stock return ● dt+1
log dividend paid
● Δ one period change ● Et
rational expectation
● ρ discount coefficient ● NCF
cash flow news
● NDR
discount rate news
Return Decomposition Framework
11
Michael-Paul James

12. ● First order VAR (vector autoregression) model & linear functions of
news
● zt+1
m-by-1 state vector with rt+1
as its first element
● Γ m-by-1 vector & m-by-m matrix of constant parameters:
● ut+1
iid m-by-1 vector of shocks
● λ ≡ ρΓ(I-ρΓ)-1
maps news to VAR shocks
● (I-ρΓ)-1
Increases weight on persistant news variables
Return Decomposition Framework
12
Michael-Paul James

13. VAR state variables and estimation
13
Michael-Paul James
● Four state variables
○ re
m,t+1
: Excess market return
■ Log (value weighted index / treasury bills)
○ TY: Yield spread between long term and short term bonds
■ 10 year constant maturity taxable bond yield / short term taxable notes
■ Tracks business cycle
○ PE: Market’s smoothed price-earnings ratio
■ Log (S&P 500 price index / 10-year moving average of S&P earnings)
■ High PE = low long run expected returns (constant earnings growth)
○ VS: Small-stock value spread
■ Log (book-to-market ratios of small value and small growth)
■ ICAPM motivated: more sensitive to changes in discount rates, equity
market, and financial conditions.

14. Table
1:
Descriptive
Statistics
Of
The
VAR
State
Variables
14
Notes: The table shows the descriptive statistics of the VAR state variables estimated from the full sample period 1928:12–2001:12, 877 monthly
data points. rM e is the excess log return on the CRSP value-weighted index. TY is the term yield spread in percentage points, measured as the
yield difference between ten-year constant-maturity taxable bonds and short-term taxable notes. PE is the log ratio of the S&P 500’s price to the
S&P 500’s ten-year moving average of earnings. VS is the small-stock value-spread, the difference in the log book-to-market ratios of small value
and small growth stocks. The small-value and small-growth portfolios are two of the six elementary portfolios constructed by Davis et al. (2000).
“Stdev.” denotes standard deviation and “Autocorr.” the first-order autocorrelation of the series.
VS spread negatively
correlated to excess
returns
TABLE 1: Descriptive Statistics Of The VAR State Variables
Variable Mean Median Stdev. Min Max Autocorr.
re
M
0.004 0.009 0.056 -0.344 0.322 0.108
TY 0.629 0.55 0.643 -1.35 2.72 0.906
PE 2.868 2.852 0.374 1.501 3.891 0.992
VS 1.653 1.522 0.374 1.192 2.713 0.992
Correlations re
M,t+1
TYt+1
PEt+1
VSt+1
re
M,t+1
1 0.071 -0.006 -0.03
TYt+1
0.071 1 -0.253 0.423
PEt+1
-0.006 -0.253 1 -0.32
VSt+1
-0.03 0.423 -0.32 1
re
M,t
0.103 0.065 0.07 -0.031
TYt
0.07 0.906 -0.248 0.42
PEt
-0.09 -0.263 0.992 -0.318
VSt
-0.025 0.425 -0.322 0.992
Higher PE negatively
correlated to excess
returns
Higher yield spread
positively correlated to
excess returns
TY, PE, and VS highly
correlated to previous
period, re
is not

15. Table
2:
VAR
Parameter
Estimates
15
Notes: The table shows the OLS parameter
estimates for a first-order VAR model including
a constant, the log excess market return (re
M
),
term yield spread (TY), price-earnings ratio
(PE), and small-stock value spread (VS). Each
set of three rows corresponds to a different
dependent variable. The first five columns
report coefficients on the five explanatory
variables, and the remaining columns show R2
and F statistics. OLS standard errors are in
square brackets and bootstrap standard errors
in parentheses. Bootstrap standard errors are
computed from 2,500 simulated realizations.
The table also reports the correlation matrix of
the shocks with shock standard deviations on
the diagonal, labeled “corr/std.” Sample period
for the dependent variables is 1929:1–2001:12,
876 monthly data points.
TY, PE, and VS autoregressive
coefficient of 0.879, 0.994, &
0.991 respectively
TABLE 2: VAR Parameter Estimates
Constant re
M,t
TYt
PEt
VSt
R2
% F
re
M,t
0.062 0.094 0.006 -0.014 -0.013 2.57 5.34
[0.020] [0.033] [0.003] [0.005] [0.006]
(0.026) (0.034) (0.003) (0.007) (0.008)
TYt+1
0.046 0.046 0.879 -0.036 0.082 82.41 1.02x103
[0.097] [0.165] [0.016] [0.026] [0.028]
(0.012) (0.170) (0.017) (0.031) (0.036)
PEt+1
0.019 0.519 0.002 0.994 -0.003 99.06 2.29x104
[0.013] [0.022] [0.002] [0.004] [0.004]
(0.017) (0.022) (0.002) (0.004) (0.005)
VSt+1
0.014 -0.005 0.002 0.000 0.991 98.4 1.34x104
[0.017] [0.029] [0.003] [0.005] [0.005]
(0.024) (0.028) (0.003) (0.006) (0.008)
corr/std re
M,t+1
TYt+1
PEt+1
VSt+1
re
M,t+1
0.055 0.018 0.777 -0.052
(0.003) (0.048) (0.018) (0.052)
TYt+1
0.018 0.268 0.018 -0.012
(0.048) (0.013) (0.039) (0.034)
PEt+1
0.777 0.018 0.036 -0.086
(0.018) (0.039) (0.002) (0.045)
VSt+1
-0.052 -0.012 -0.086 0.047
(0.052) (0.034) (0.045) (0.003)
VS some predictability on TY
Re
some predictability on PE
Excess returns: momentum
evident, TY, PE, and VS offers
predictability
PE & VS highly persistent

16. Table
3:
Cash-Flow
and
Discount-Rate
News
For
The
Market
Portfolio
16
Notes: The table shows the properties of cash-flow news (NCF
) and discount-rate news (NDR
) implied by the VAR model of Table 2. The upper-left
section of the table shows the covariance matrix of the news terms. The upper-right section shows the correlation matrix of the news terms
with standard deviations on the diagonal. The lower-left section shows the correlation of shocks to individual state variables with the news
terms. The lower-right section shows the functions (e1' + e1'λ, e1'λ) that map the state-variable shocks to cash-flow and discount-rate news. We
define λ ≡ ρΓ(I - ρΓ)-1
, where is the estimated VAR transition matrix from Table 2 and ρ is set to 0.95 per annum. re
M
is the excess log return on the
CRSP value-weighted index. TY is the term yield spread. PE is the log ratio of the S&P 500’s price to the S&P 500’s ten-year moving average of
earnings. VS is the small-stock value-spread, the difference in log book-to-markets of value and growth stocks. Bootstrap standard errors (in
parentheses) are computed from 2,500 simulated realizations.
TABLE 3: Cash-Flow and Discount-Rate News For The Market Portfolio
News covariance NCF
NDR
News corr/std NCF
NDR
NCF
0.00064 0.00015 NCF
0.0252 0.114
(0.00022) (0.00037) (0.004) (0.232)
NDR
0.00015 0.00267 NDR
0.114 0.0517
(0.00037) (0.00070) (0.232) (0.007)
Shock correlations NCF
NDR
Functions NCF
NDR
re
M
shock 0.352 -0.890 re
M
shock 0.602 -0.398
(0.224) (0.036) (0.060) (0.060)
TY shock 0.128 0.042 TY shock 0.011 0.011
(0.134) (0.081) (0.013) (0.013)
PE shock -0.204 -0.925 PE shock -0.883 -0.883
(0.238) (0.039) (0.104) (0.104)
VS shock -0.493 -0.186 VS shock -0.283 -0.283
(0.243) (0.152) (0.160) (0.160)
DR SD of 5%
CF SD of 2.5%
DR & CF uncorrelated
at 0.114
re
and PE negatively
correlated with NDR
re
and PE weakly
correlated with NCF

17. Figure
1:
Cash-Flow
and
Discount-Rate
Recessions
17
Notes: This figure plots the cash-flow
news and negative of discount-rate news,
smoothed with a trailing exponentially
weighted moving average. The decay
parameter is set to 0.08 per month, and
the smoothed news series are generated
as MAt
(N) = 0.08Nt
(1 0.08)MAt-1
(N). The
dotted vertical lines denote NBER
business-cycle troughs. The sample
period is 1929:1–2001:12.
Dotted lines are troughs of a
recession
Profitability recession
Valuation recession
Mixed recession

18. Measure Betas
03
Measuring Cash-Flow and Discount-Rate Betas
18
Michael-Paul James

19. Measuring the betas separately
19
Michael-Paul James
● Cash flow beta
● Discount rate beta
● Total market beta
● Value stocks have higher cash flow (ROE: Return on Equity) betas than
growth stocks

20. Table
4:
Cash-Flow
and
Discount-Rate
Betas
In
The
Early
Sample
20
Notes: The table shows the estimated cash-flow (βCF
) and discount-rate betas (βDR
) for the 25 ME- and BE/ME-sorted portfolios and 20
risk-sorted portfolios. “Growth” denotes the lowest BE/ME, “value” the highest BE/ME, “small” the lowest ME, and “large” the highest ME stocks.
bVS
, bTY
, and brM
are past return-loadings on value-spread shock, term-yield shock, and market-return shock. “Diff.” is the difference between the
extreme cells. Standard errors [in brackets] are conditional on the estimated news series. Estimates are for the 1929:1–1963:6 period.
TABLE 4: Cash-Flow and Discount-Rate Betas In The Early Sample
βCF
Growth 2 3 4 Value Diff.
Small 0.53 [0.11] 0.46 [0.09] 0.40 [0.08] 0.42 [0.07] 0.49 [0.08] -0.04 [0.07]
2 0.30 [0.06] 0.34 [0.06] 0.36 [0.06] 0.38 [0.06] 0.45 [0.08] 0.16 [0.04]
3 0.30 [0.06] 0.28 [0.05] 0.31 [0.06] 0.35 [0.06] 0.47 [0.08] 0.18 [0.04]
4 0.20 [0.05] 0.26 [0.05] 0.31 [0.05] 0.35 [0.07] 0.50 [0.09] 0.30 [0.05]
Large 0.20 [0.05] 0.19 [0.05] 0.28 [0.06] 0.33 [0.07] 0.40 [0.09] 0.19 [0.06]
Diff. -0.33 [0.09] -0.26 [0.06] -0.12 [0.05] -0.09 [0.04] -0.10 [0.04]
βDR
Growth 2 3 4 Value Diff.
Small 1.32 [0.18] 1.46 [0.19] 1.32 [0.15] 1.27 [0.14] 1.27 [0.15] -0.06 [0.15]
2 1.04 [0.11] 1.15 [0.11] 1.09 [0.11] 1.25 [0.11] 1.25 [0.13] 0.21 [0.08]
3 1.13 [0.10] 1.01 [0.08] 1.08 [0.09] 1.05 [0.10] 1.27 [0.12] 0.14 [0.06]
4 0.87 [0.07] 0.97 [0.08] 0.97 [0.09] 1.06 [0.10] 1.36 [0.13] 0.49 [0.10]
Large 0.88 [0.07] 0.82 [0.07] 0.87 [0.08] 1.06 [0.09] 1.18 [0.12] 0.31 [0.10]
Diff. -0.45 [0.15] -0.64 [0.15] -0.43 [0.10] -0.21 [0.08] -0.08 [0.10]
βCF
Lo brM
2 3 4 Hi brM
Diff.
Lo bVS
0.21 [0.04] 0.25 [0.05] 0.31 [0.06] 0.37 [0.07] 0.45 [0.09] 0.25 [0.05]
Hi bVS
0.15 [0.03] 0.19 [0.04] 0.25 [0.06] 0.28 [0.06] 0.37 [0.08] 0.22 [0.05]
Lo bTY
0.18 [0.04] 0.21 [0.05] 0.26 [0.06] 0.31 [0.07] 0.41 [0.08] 0.23 [0.04]
Hi bTY
0.16 [0.04] 0.21 [0.04] 0.27 [0.05] 0.32 [0.06] 0.40 [0.08] 0.24 [0.05]
βDR
Lo brM
2 3 4 Hi brM
Diff.
Lo bVS
0.73 [0.06] 0.87 [0.07] 1.04 [0.09] 1.20 [0.11] 1.46 [0.13] 0.73 [0.09]
Hi bVS
0.64 [0.05] 0.75 [0.07] 0.96 [0.08] 1.09 [0.09] 1.30 [0.11] 0.66 [0.08]
Lo bTY
0.73 [0.06] 0.85 [0.07] 1.00 [0.09] 1.17 [0.10] 1.38 [0.12] 0.64 [0.08]
Hi bTY
0.65 [0.06] 0.76 [0.06] 0.88 [0.08] 1.09 [0.10] 1.34 [0.12] 0.69 [0.09]
Average of value CF betas
less growth: 0.156
Average of value DR betas
less growth: 0.218
Average of small DR betas
less large: 0.18
Average of small DR betas
less large: 0.366
Small β higher than large
Value β higher than growth
Small and value stocks
were riskier from 1929-63
High past market betas
induce high CF & DR beta

21. Table
5:
Cash-Flow
and
Discount-Rate
Betas
In
The
Modern
Sample
21
Notes: The table shows the estimated cash-flow (βCF
) and discount-rate betas (βDR
) for the 25 ME- and BE/ME-sorted portfolios and 20
risk-sorted portfolios. “Growth” denotes the lowest BE/ME, “value” the highest BE/ME, “small” the lowest ME, and “large” the highest ME stocks.
bVS
, bTY
, and brM
are past return-loadings on value-spread shock, term-yield shock, and market-return shock. “Diff.” is the difference between the
extreme cells. Standard errors [in brackets] are conditional on the estimated news series. Estimates are for the 1963:7–2001:12 period.
TABLE 5: Cash-Flow and Discount-Rate Betas In The Modern Sample
βCF
Growth 2 3 4 Value Diff.
Small 0.06 [0.07] 0.07 [0.06] 0.09 [0.05] 0.09 [0.04] 0.13 [0.04] 0.07 [0.04]
2 0.04 [0.06] 0.08 [0.05] 0.10 [0.04] 0.11 [0.04] 0.12 [0.04] 0.09 [0.03]
3 0.03 [0.05] 0.09 [0.04] 0.11 [0.04] 0.12 [0.03] 0.13 [0.04] 0.09 [0.04]
4 0.03 [0.05] 0.10 [0.04] 0.11 [0.03] 0.11 [0.03] 0.13 [0.04] 0.10 [0.04]
Large 0.03 [0.04] 0.08 [0.03] 0.09 [0.03] 0.11 [0.03] 0.11 [0.03] 0.09 [0.03]
Diff. -0.03 [0.05] 0.02 [0.05] -0.01 [0.04] 0.02 [0.04] -0.01 [0.04]
βDR
Growth 2 3 4 Value Diff.
Small 1.66 [0.13] 1.37 [0.11] 1.18 [0.10] 1.12 [0.09] 1.12 [0.10] -0.54 [0.08]
2 1.54 [0.11] 1.22 [0.09] 1.07 [0.08] 0.96 [0.08] 1.03 [0.09] -0.52 [0.08]
3 1.41 [0.10] 1.11 [0.08] 0.95 [0.08] 0.82 [0.07] 0.94 [0.09] -0.47 [0.09]
4 1.27 [0.09] 1.05 [0.08] 0.89 [0.07] 0.79 [0.07] 0.87 [0.08] -0.41 [0.09]
Large 1.00 [0.07] 0.87 [0.07] 0.74 [0.06] 0.63 [0.07] 0.68 [0.07] -0.33 [0.08]
Diff. -0.66 [0.12] -0.50 [0.11] -0.44 [0.10] -0.49 [0.09] -0.44 [0.08]
βCF
Lo brM
2 3 4 Hi brM
Diff.
Lo bVS
0.09 [0.03] 0.08 [0.03] 0.10 [0.04] 0.10 [0.04] 0.12 [0.05] 0.04 [0.04]
Hi bVS
0.06 [0.03] 0.06 [0.03] 0.07 [0.04] 0.05 [0.05] 0.06 [0.06] -0.01 [0.04]
Lo bTY
0.06 [0.03] 0.04 [0.03] 0.08 [0.04] 0.08 [0.04] 0.06 [0.06] 0.00 [0.04]
Hi bTY
0.09 [0.03] 0.07 [0.03] 0.09 [0.03] 0.08 [0.04] 0.10 [0.05] 0.00 [0.04]
βDR
Lo brM
2 3 4 Hi brM
Diff.
Lo bVS
0.57 [0.06] 0.77 [0.06] 0.88 [0.07] 1.12 [0.08] 1.40 [0.09] 0.82 [0.08]
Hi bVS
0.67 [0.06] 0.85 [0.07] 1.06 [0.07] 1.30 [0.09] 1.58 [0.11] 0.91 [0.11]
Lo bTY
0.73 [0.07] 0.86 [0.07] 1.05 [0.07] 1.23 [0.08] 1.60 [0.12] 0.87 [0.10]
Hi bTY
0.61 [0.06] 0.79 [0.06] 0.91 [0.06] 1.11 [0.07] 1.39 [0.09] 0.78 [0.08]
Average of value CF betas
less growth: 0.086
Average of value DR
betas less growth: -0.448
Average of small DR betas
less large: 0.004
Average of small DR betas
less large: 0.506
Small β higher than large
Value β higher than growth
Growth has higher β than
value but more good DR β
Growth: more cash?
Pre-1963 value debt load?
DR: investor sentiment?

22. Pricing Betas
04
Pricing Cash-Flow and Discount-Rate Betas
An Intertemporal Asset Pricing Model
Empirical Estimates of Risk Premia
Additional Robustness Checks
Loose Ends and Future Directions
22
Michael-Paul James

23. Pricing CF and DR betas
23
Michael-Paul James
● Risk premium requirements under optimal portfolio strategy (p)
● Arranged: bad CF β has risk price increase by γ greater than good DR β
○ Log returns = CF portfolio risk + DR portfolio risk
○ Expected returns = CF market risk + DR market risk

24. Table
6:
Asset
Pricing
Tests
For
The
Early
Sample
24
Notes: The table shows premia estimated from the 1929:1–1963:6 sample for an unrestricted factor model, the two-beta ICAPM, and the CAPM.
The test assets are the 25 ME- and BE/ME-sorted portfolios and 20 risk-sorted portfolios. The second column per model constrains the
zero-beta rate (Rzb
) to equal the risk-free rate (Rrf
). Estimates are from a cross-sectional regression of average simple excess test-asset returns
(monthly in fractions) on an intercept and estimated cash-flow (βCF
) and discount-rate betas (βDR
). Standard errors and critical values [A] are
conditional on the estimated news series and (B) incorporating full estimation uncertainty of the news terms. The test rejects if the pricing
error is higher than the listed 5 percent critical value.
TABLE 6: Asset Pricing Tests For The Early Sample
Parameter Factor model Two-beta ICAPM CAPM
Rzb
less Rrf
(g0
) 0.0042 0 0.0023 0 0.0023 0
% per annum 4.98% 0% 2.76% 0% 2.74% 0%
Std. err. A [0.0032] N/A [0.0024] N/A [0.0028] N/A
Std. err. B (0.0029) N/A (0.0030) N/A (0.0028) N/A
βCF
premium (g1
) 0.0173 0.0069 0.0083 0.0148 0.0051 0.0067
% per annum 20.76% 8.22% 9.91% 17.80% 6.11% 8.00%
Std. err. A [0.0231] [0.221] [0.0167] [0.0175] [0.0046] [0.0034]
Std. err. B (0.0266) (0.0248) (0.0221) (0.0442) (0.0046) (0.0034)
βDR
premium (g2
) -0.0003 0.0066 0.0041 0.0041 0.0051 0.0067
% per annum -0.41% 7.93% 4.95% 4.95% 6.11% 8.00%
Std. err. A [0.0092] [0.0067] [0.0006] [0.0006] [0.0046] [0.0034]
Std. err. B (0.0088) (0.0071) (0.0006) (0.0006) (0.0046) (0.0034)
R2
48.08% 40.26% 45.85% 37.98% 44.52% 40.26%
Pricing error 0.0117 0.0126 0.0119 0.0133 0.0127 0.0126
5% critic. val. A [0.019] [0.024] [0.024] [0.033] [0.021] [0.027]
5% critic. val. B (0.019) (0.024) (0.031) (0.099) (0.021) (0.027)
Zero beta in excess of
treasury bill rate
Cash flow premium
Discount rate premium
Factor model, ICAPM,
CAPM with 38-48% R2
not
rejections at 5%

25. Figure
2:
Performance
Of
The
CAPM
And
ICAPM,
1929:1–1963:6
25
Notes: The four diagrams correspond to
(clockwise from the top left) the ICAPM
with a free zero-beta rate, the ICAPM with
the zero-beta rate constrained to the
risk-free rate, the CAPM with a
constrained zero-beta rate, and the CAPM
with an unconstrained zero-beta rate. The
horizontal axes correspond to the
predicted average excess returns and
the vertical axes to the sample average
realized excess returns. The predicted
values are from regressions presented in
Table 6. Triangles denote the 25 ME and
BE/ME portfolios and asterisks the 20
risk-sorted portfolios.
* 20 risk sorted portfolios
Δ 24 Fama-French portfolios
Good CAPM due to bad CF
beta being constant fraction

26. Table
6:
Asset
Pricing
Tests
For
The
Early
Sample
26
Notes: The table shows premia estimated from the 1963:7–2001:12 sample for an unrestricted factor model, the two-beta ICAPM, and the CAPM.
The test assets are the 25 ME- and BE/ME-sorted portfolios and 20 risk-sorted portfolios. The second column per model constrains the zero-beta
rate (Rzb) to equal the risk-free rate (Rrf). Estimates are from a cross-sectional regression of average simple excess test-asset returns (monthly in
fractions) on an intercept and estimated cash-flow (βCF
) and discount-rate betas (βDR
). Standard errors and critical values [A] are conditional on
the estimated news series and (B) incorporating full estimation uncertainty of the news terms. The test rejects if the pricing error is higher than
the listed 5 percent critical value.
TABLE 6: Asset Pricing Tests For The Early Sample
Parameter Factor model Two-beta ICAPM CAPM
Rzb
less Rrf
(g0
) 0.0009 0 -0.0009 0 0.0069 0
% per annum 1.05% 0% -1.04% 0% 8.24% 0%
Std. err. A [0.0029] N/A [0.0031] N/A [0.0026] N/A
Std. err. B (0.0033) N/A (0.0031) N/A (0.0026) N/A
βCF
premium (g1
) 0.0529 0.0572 0.0575 0.0483 -0.0007 0.0051
% per annum 63.47% 68.59% 69.04% 57.92% -0.83% 6.10%
Std. err. A [0.0178] [0.0163] [0.0182] [0.0272] [0.0034] [0.0023]
Std. err. B (0.0325) (0.0444) (0.0262) (0.0423) (0.0034) (0.0023)
βDR
premium (g2
) 0.0007 0.0012 0.0020 0.0020 -0.0007 0.0051
% per annum 0.88% 1.44% 2.43% 2.43% -0.83% 6.10%
Std. err. A [0.0033] [0.0031] [0.0002] [0.0002] [0.0034] [0.0023]
Std. err. B (0.0085) (0.0099) (0.0002) (0.0002) (0.0034) (0.0023)
R2
52.10% 51.59% 49.26% 47.41% 3.10% -61.57%
Pricing error 0.0271 0.0269 0.0272 0.0275 0.0592 0.0875
5% critic. val. A [0.028] [0.042] [0.051] [0.314] [0.032] [0.046]
5% critic. val. B (0.030) (0.071) (0.051) (0.488) (0.032) (0.046)
Zero beta in excess of
treasury bill rate
Cash flow premium
Discount rate premium
Factor model, ICAPM, with
47-52% R2
not rejections
at 5%: CAPM fails

27. Figure
3:
Performance
Of
The
CAPM
And
ICAPM,
1963:7–2001:12
27
Notes: The four diagrams correspond to
(clockwise from the top left) the ICAPM
with a free zero-beta rate, the ICAPM with
the zero-beta rate constrained to the
risk-free rate, the CAPM with a
constrained zero-beta rate, and the CAPM
with an unconstrained zero-beta rate. The
horizontal axes correspond to the
predicted average excess returns and the
vertical axes to the sample average
realized excess returns. The predicted
values are from regressions presented in
Table 6. Triangles denote the 25 ME and
BE/ME portfolios and asterisks the 20
risk-sorted portfolios.
CAPM offer no predictability
under both the zero-beta
and risk-free rate

28. Robustness checks
28
Michael-Paul James
● Small sample bias
○ Negligible bias in VAR coefficient of returns on value spread (VS)
○ Little bias in estimated volatility of CF and DR news
○ Upward bias in negative CF and DR coefficients on VS
● Conditional pricing
○ Estimate covariences for each test asset with rolling 3-year window
○ Regress realized cross products on lagged rolling covariance
estimates and dummies in two pooled regressions
○ Regress realized simple excess returns on fitted conditional
covariances with restrictions

29. Robustness checks
29
Michael-Paul James
● Sensitivity to ρ
○ Based on a priori knowledge not estimations
○ Value of ρ has little impact on results except when zero-beta is
restricted to treasury-bill rate
● Data frequency
○ Repeated monthly tests with quarterly and annual data with
similar results in magnitude and direction.
● Sensitivity to additional state variables
○ Added treasury bill rate and log dividend price ratio with similar
results

30. Loose ends and future directions
30
Michael-Paul James
● Assumes constant variance for market returns, cash flow news, and
discount rate news.
● Assumes rational long term investor always holds a constant
proportion of assets in equities
○ No incentive to time the market
● No discussion of source of variation in the market’s discount rate
● No discussion of how CF and DR betas are linked to the underlying
cash flows of value and growth firms.
● Primary determinant of cost of capital is not market beta but cash flow
of a project.

31. Conclusions
05
31
Michael-Paul James

32. Summary
32
Michael-Paul James
● Cash flow and discount rate betas estimates stock market risk factors
more efficiently than CAPM over time.
○ Cash flow news
■ Stock returns covariance with cash flows news
○ Discount rate news
■ Stock returns covariance with discount rate news
● “Bad” cash flow beta (risk) demands higher premiums than “good”
discount rate beta (risk).
● Value and small stocks have higher cash flow betas than growth and
large stocks on average.

33. Summary
33
Michael-Paul James
● High average returns on value and small stocks are appropriate
compensation for risk, not an unrealized benefit to ownership.
● Overweighting small and value stocks benefit low risk aversion equity
investors
● Underweighting small and value stocks benefit high risk aversion
equity investors
● Model offers strong explanatory power in the cross section of asset
returns with theoretical values
● ICAPM outperforms the CAPM in empirical research

34. You are Amazing
Ask me all the questions you desire. I will do my best to answer honestly
and strive to grasp your intent and creativity.
34
Michael-Paul James