IPOs. We omitted firms paying dividends or firms with a stock price below
$10 a share. On October 10, the stub value for Stratos was −$1.66 a share,
and the synthetic short price constructed using six-month options was 24
percent below the actual price of Stratos (similar to the deviation seen for
Palm in table 4.6), or $5.89 below the actual price per share. For the twenty-
eight other firms, the average synthetic short price was only 3 percent below
the actual price, or 87 cents per share, easily explainable with bid/ask
spreads on options. The maximum deviation was 8 percent below the ac-
tual price, only a third of the deviation observed for Palm and Stratos. On
the basis of this evidence, the Palm and Stratos cases appear to present un-
usually large violations of put-call parity.
To conclude, in the case of Palm and Stratos, we have strong evidence
from options markets confirming that the new issues are overpriced, and no
one should buy them (at least without lending them out, which not everyone
can do in equilibrium) because cheaper alternatives are available. Although
shares in the parent are not perfect substitutes for shares in the subsidiary
(because of the risk of spin-off ), the synthetic shares are virtually identical.
Although not an exploitable arbitrage opportunity, this is a case of blatant
mispricing.
- What Causes Mispricing?
We hope to have convinced even the most jaded reader that the cases we are
studying are clear violations of the law of one price. Given that arbitrage can-
not correct the mispricing, why would anyone buy the overpriced security?
Why are some investors willing to buy shares in Palm when there are cheaper
alternatives available in the market, either by buying the parent or by buying
Palm synthetically in the options market? In this section we investigate this
question, first by asking a simple question: Who buys the expensive sub-
sidiary shares, and how long do they hold them? We then look at IPO day re-
turns for evidence on how these investors affect prices of the parent.
A. Investor CharacteristicsColumns 1 and 2 of table 4.8 display volume data for both parents and
subsidiaries in our six cases with negative stubs. We show turnover for
the first twenty days of trading, defined as average daily volume divided by
shares outstanding (for parents) or by total shares sold to the public (for
the IPO). The turnover measure does not include the first day of trading it-
self. All twelve stocks trade on NASDAQ. Since NASDAQ is a dealer mar-
ket, reported volume includes dealer trades, and the turnover caused by
trades between actual investors is approximately half the turnover reported
in table 4.8.
MISPRICING IN TECH STOCK CARVE-OUTS 159