1

 G. William Schwert
 University of Rochester
 Simon School of Business
 http://schwert.ssb.rochester.edu/FMA2012.htm

2

 Journal of Finance 65 (April 2010) 425465
 with Michelle Lowry and Micah Officer
 Interesting blend of time series and cross sectional modeling issues
 Research question is motivated by the apparent difficulty that issuing
firms and underwriters have in setting IPO prices anywhere near the
subsequent secondary market price (i.e., IPO underpricing)

3


4


5


6


7


8

 IR_{i} = b_{0}
+ b_{1} Rank_{i} +
b_{2}
Log(Shares_{i}) + b_{3} Tech_{i}
 + b_{4} VC_{i} +
b_{5}
NYSE_{i} + b_{6} NASDAQ_{i}
 + b_{7}
Log(Firm Age_{i} + 1) + b_{8} Price
Update_{i} + e_{i}.
(1)
 Log(s^{2}(e_{i})) =
g_{0} +
g_{1}
Rank_{i} + g_{2} Log(Shares_{i})
 + g_{3}
Tech_{i} + g_{4} VC_{i} +
g_{5}
NYSE_{i} + g_{6} NASDAQ_{i}
 + g_{7}
Log(Firm Age_{i} + 1) + g_{8} Price
Update_{i}_{
}(2)

9


10

 This a little unusual, since the IPO returns are for different
securities and they are not equally spaced through time
 Effectively, we are treating these observations as coming from the “IPO
return process,” which we assume is stationary
 As you will see, this seems to work pretty well . . .

11

 IR_{i} = b_{0}
+ b_{1} Rank_{i} +
b_{2}
Log(Shares_{i}) + b_{3} Tech_{i}
 + b_{4} VC_{i} +
b_{5}
NYSE_{i} + b_{6} NASDAQ_{i}
 + b_{7}
Log(Firm Age_{i} + 1) + b_{8} Price
Update_{i}
 + [(1θL)/(1fL)] e_{i}
 = .948, θ = .905 => low,
but persistent autocorrelations of returns
 LjungBox(20) drops from 2,848 to 129

12

 Log(s^{2}(e_{i})) =
g_{0} +
g_{1}
Rank_{i} + g_{2} Log(Shares_{i})
 + g_{3}
Tech_{i} + g_{4} VC_{i} +
g_{5}
NYSE_{i} + g_{6} NASDAQ_{i}
 + g_{7}
Log(Firm Age_{i} + 1) + g_{8} Price
Update_{i}
 EGARCH model:
 log(s^{2}_{t}) = w + a log[e_{i1}^{2}/s^{2}(e_{i1})]
+ d log(s^{2}_{t1})
 Var(e_{i}) = s^{2}_{t }∙ s^{2}(e_{i})

13

 ARCH intercept w = .025
 ARCH coefficient a = .016
 GARCH coefficient d = .984
 Very persistent time series volatility
 LjungBox(20) for autocorrelations drops to 57
 LjungBox(20) for autocorrelations of squared residuals drops to 67
(from 317 for ARMA model)

14

 Evidence is consistent with timevarying information asymmetry story
 But the extreme persistence of IRs and volatility, given the
characteristics of the offering, suggests that there are important
aspects of uncertainty about the valuation of IPOs that are simply hard
to predict
 Suggests alternative methods for selling IPOs are worth considering –
e.g., IPO auctions . . .

15

 The general approach of focusing on uncertainty has many possible
applications in corporate finance as well as in capital markets areas
 Modeling uncertainty as a function of firm/deal characteristics gives a
richer set of tools to look at information asymmetry and other similar
questions

16

 Finally, modeling dispersion using both time series and cross sectional
tools allows for better inference
 In much the same way that Mitch Petersen’s paper on the importance of
clustering in calculating standard errors for crosssectional models
used in corporate finance has become “stateoftheart,” correctly using
WLS or MLE leads to much more reliable inferences for the “mean
equation”
