Fama-French

1. CAPM, example 1, size

(a) Expected returns

(b) Betas

VW market EW market
2. Discount rates update

3. CAPM Example 2: industry portfolios

60
 1926−1979 1980−2010
20 14

12
15
Average Return 10
10 8

5 6

4
0
2

−5 0
−0.5 0 0.5 1 1.5 −0.5 0 0.5 1 1.5
Betas Betas

Figure 2: CAPM on Fama-French size portfolios, and , 10 and 30 year government bonds,
montlhy data 1926-2009. The diagonal line is the fit of a cross-sectional regression.

14

12
Durbl

10 Chems
MoneyShops
Oil
NoDur Manuf
8 Rm
E(R )

Cross sectional regression Telcm

e

Utils
Other
6

Cross section, no γ
4

Time series regression (through Rm, Rf)

0 Rf
0 0.5 1 1.5
β

4. FF: What about book/market sorted portfolios?

(a) Facts: There is a big spread in average returns. But market beta is a disaster.

61
5. Discount rates graphs

62
 Average returns and betas

0.8
E(r)

0.6

b x E(rmrf)

Average return
0.4

β x E(rmrf)

0.2

h x E(hml)

−0.2
Growth Value

Average returns and betas for Fama - French 10 B/M sorted portfolios. Monthly data
1963-2010.

Average returns and betas Factor analysis

0.6

0.8 0.4 Factor 1, 88.8% of var.

E(r)
Average return

0.6 0.2
b x E(rmrf)
0.4 0
β x E(rmrf)
0.2 −0.2
h x E(hml) Factor 2, 5.3% of var.
0 −0.4

−0.2
Growth Value Growth Value

B/M sorted portfolios, monthly data 1963-2010. Left panel: Average returns, market beta
× market premium, and two-factor betas times premiums. Right panel: eigenvectors of the
largest two eigenvalues in the covariance matrix of excess returns.

63
 1963−2009 1926−1963
1
V V
1.2 8 9
0.8
9
1 5 6

Average Return
8 7
0.6 7 32
6 0.8 G 4
5 342
0.4 0.6
G
0.4
0.2
0.2

0 Rf 0 Rf
0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5
Betas Betas
Value effect before and after 1963. Average returns on Fama - French 10 portfolios sorted
by book/market equity vs. CAPM betas. Monthly data. Source: Ken French’s website.

6. Fama-French solution:

(a) Run time series regressions that include additional factors (portfolios of stocks)
SMB, HML

Rtei = αi + bi Rtem + si SMBt + hi HMLt + εit ; t = 1, 2...T for each i = 1, 2..N.

(b) Look across stocks

E(Rei ) = αi + bi E(Rem ) + si E(SMB) + hi E(HML)

7. Fama-French paper: See Table 1

(a) FF factors

64
 HML

30% 40% 30%

Size (NYSE breaks)

S/L S/M S/H

SMB
B/L B/M B/H

Book/market (NYSE breaks)

HML = (S/H + B/H)/2 – (S/L+B/L)/2
SMB = (S/L +S/M+S/H)/3 – (B/L+B/M+B/H)/3

65
 0.4 0.5

Loading 1

Loading 2
0.2 0

0 −0.5
1 1
2 2
3 5 3 5
4 4 4 4
3 3
5 2 5 2
size 1 value size 1 value

0.5 1
Loading 3

Loading 4
0.5
0
0

−0.5 −0.5
1 1
2 2
3 5 3 5
4 4 4 4
3 3
5 2 5 2
size 1 value size 1 value

8.

66
 Empirical Asset Pricing Flowchart

Group stocks by some characteristic (size, B/M, past return, etc.)

Is there a spread in average returns? No

Yes

Really? Do the statistics right? Survivor/selection bias? Out of sample? No

Yes

Are high average returns explained by high market betas? Yes

No

Are high average returns explained by multifactor betas? Yes

No

Does a new multifactor model seem plausible, work?

Yes No Trade on it. Hope it lasts.

Fame and fortune Take up behavioral finance

9.
Portfolio
Mean
E(R)

Securities

Better weights?

Portfolio
1 2 3 4 5
Log(Book/market)

67