of mispricing. So negative stubs should be considered the extreme cases
of unambiguous mispricing. For the potential negative stubs, we gathered
information that was available in real time to construct the estimated
ratio. In all but one case the uncertainty about the final ratio appears to be
small.^2
Of our eighteen firms, nine clearly had positive stubs. We classify three
stubs as marginally negative. These were cases in which we observed small
negative stubs on one or two days only or the correct ratio is sometimes un-
clear because of changing numbers of shares. For these cases we think that
a reasonable person would not be convinced that the stub was negative
given all available information. None of the three marginal cases involves a
negative stub at or near the IPO date.
We identify six cases of unambiguously negative stubs: UBID, Retek, PF-
SWeb, Xpedior, Palm, and Stratos Lightwave.^3 All six are technology
stocks. UBID is an online auction firm. Retek produces business-to-business
inventory software. PFSWeb provides transactions management services for
e-commerce. Xpedior is an e-business consulting firm. Stratos Lightwave is
an optical networking firm. Both the six parents and the six subsidiaries
trade on NASDAQ.
As shown in table 4.2, for the six cases with negative stubs, four were
negative at closing prices on the first day of trading, and the other two were
negative by two days after. For five of the cases, the stub was negative for at
least two months, with a maximum of 187 trading days for Stratos. For
one case, Xpedior, the stub was negative for only two days before turning
positive again. Xpedior’s minimum stub also had a fairly small magnitude
of only −19 percent of the parent company’s value, unlike the other five,
which had minimum stubs of −39 to −137 percent of the parent’s value.
Thus Xpedior is a much weaker case in terms of the persistence and magni-
tude of the mispricing.
Table 4.2 shows the magnitude of the mispricing in a variety of ways.
Perhaps the most relevant is the market value of the shares trading in the
subsidiary. This number (which uses the number of publicly trading shares,
not the number of outstanding shares) is at its peak, $2.5 billion, for Palm,
meaning that investors worth $2.5 billion thought it was better to own
Palm than to own 3Com.
138 LAMONT AND THALER
(^2) In one case, Retek, there appears to have been substantial uncertainty about the final ratio
since the parent’s number of shares was somewhat volatile. Retek’s parent ultimately decided
to accelerate the vesting of the options held by insiders.
(^3) In these six cases, we calculated the estimated distribution ratio (prior to the actual distri-
bution ratio announcement) as follows. For UBID and Retek, we used the ratio from the CRSP
shares outstanding. For Palm and Stratos, we always used the ratios provided by Spin-off Ad-
visors. For PFSWeb, we used the ratio provided by Spin-off Advisors until March 2000 and
then used the CRSP shares. For Xpedior, we used the ratio provided by the company web page
(in real time).