Transcript Why do companies go public?

Why do companies go public?
•
•
1.
2.
3.
4.
5.
Motivation: 1. Why do some large firms stay private? IPO seems
to be a choice, not a necessary stage! 2. US data can not test it,
but Italy can!
Results:
Industry M/B and sales, followed by growth rate and profitability
have strongest positive effect on the likelihood of going public.
Investment, leverage, profitability, interest rate, and credit
concentration tend to decrease after listing.
Independent firms are more likely to go public after high sales,
high industry M/B, high profitability, high growth and high
investment, and to reduce their leverage, investment, profitability,
interest rate, and credit concentration after IPO.
Carve-outs are more likely to go public after high industry M/B and
high profitability only, and to reduce their profitability after IPO.
Controlling shareholders of independent firms do not divest
significantly after IPO but carve-outs do. But in both cases
controlling shareholders experience abnormal turnovers in three
years after IPO.
Why do companies go public?
•
1.
2.
3.
4.
5.
Conclusions:
Evidence is inconsistent with high growth opportunity
hypothesis. Firms do not go public to finance future
investment and growth.
Independent firms go public to rebalance their balance
sheet accounts and capital structure after high growth
and investment. Carve-outs choose best listing time to
maximize proceeds. Carve-outs are driven by financial
rather than real factors.
Evidence on bank credit suggests stronger bargaining
power and increased information for borrowers after
IPO.
Evidence on changes in ownership of controlling group
does not support the portfolio diversification hypothesis.
Evidence on reallocation of control suggests IPO as a
stage in the eventual sale of a company.
Why do companies go public?
•
1.
2.
3.
4.
5.
(1)
(2)
(3)
Data:
All firms must satisfy the minimum listing requirements, which may
change during sampling period. Once listed, the firm is taken out of
sample.
For carve-outs, the fixed flotation cost may be born (sunk cost) by their
parent firms, ie, no economies of scale. Besides before carve-outs, they
may have access to bank credit through their parent firms. Therefore they
must be separately considered.
After screening, only 40 independent firms and 29 carve-outs get listed
during 1982 and 1992.
Table 1 reports summary statistics. Compared to firms eligible to go
public but did not, IPO firms appear to be larger, older, more leveraged,
have lower loan rate, have higher investment, borrow from more banks,
and rely more on external financing .
Why are IPO firms in Italy or other Continental European countries much
larger and older than US firms?
high agency cost due to lack of enforcement of minority property rights
implicit fixed cost of a higher visibility to the tax and legal authorities
absence of VC or liquid stock market dedicated to small cap firms
hypothesis
What firms choose IPO
Adverse selection and moral Old and big firms
hazard cost
Fixed flotation cost
Big, independent firms
Confidentiality cost
Non high tech firms
Overcome borrowing
constraint (growth
opportunity)
High leverage, growth,
investment, industry
M/B
Best way for initial owners
to maximize proceeds from
eventual sales of the firm
Consequence after IPO
Low equity retention by initial
owners after IPO will have
worse performance
High investment, less leverage,
less or no change in payout
Higher turnover by initial owner
Diversification
Risky firms
Initial owners sell shares
Increase liquidity
Big firms
Diffuse ownership
Stock market monitoring
High investment
Use stock-based incentive,
high investment
More investor recognition
Diffuse ownership
Increase bargaining power
with banks
High interest rate, high
credit concentration
Lower interest rate, lower credit
concentration
Window of opportunity
(timing / overvaluation)
High industry M/B
Underperformance, no
increase in investment
Why do companies go public?
• Methodology: probit model
Pr(IPOi,t=1) = F(α1SIZEi,t-1 + α2CAPEXi,t-1 + α3GROWTHi,t + α4ROAi,t-1 +
α5LEVERAGEi,t-1 + α6MTBi,t + α7RCCi,t-1 + α8HERFINDAHLi,t-1 + ∑
γtYEARt)
F(z) is the standard normal cumulative distribution function (cdf),
F(z) →0 as z → -∞, F(z) →1 as z →+∞ and is non-linear,
F(z) = ∫-∞z φ(v) dv
where φ(z) is the standard normal density, φ(z) = (2π)-0.5exp(-z2/2)
The partial effect of SIZE on the probability of IPO is obtained from
partial derivative: ∂F(z)/∂SIZE = α1[dF(z)/dz] where dF(z)/dz = φ(z)
1. F is the cdf of a continuous random variable, φ is a probability
density function. F is a strictly increasing cdf, and so φ(z) >0 for all z.
Therefore the partial effect always has the same sign as α1
2. However g(z) depends on the level of each variable! So there is no
easy answer for the scale of any partial effect! So p. 43 is incorrect!
3. But relative partial effect is certain. The ratio of the partial effects for
SIZE and ROA is α1/α4
4. The largest partial effect of SIZE occurs when z=0. When z=o, φ(z)
reaches the maximum = (2π)-0.5 = 0.4 or so.
Why do companies go public?
•
Tax law makes IPO easier in year 1984 through 1986.
Expect the coefficients of these three year dummies to
be positive.
• Some unobservable firm-specific effect may be
correlated with regressors. For example, entrepreneurs
of traditional business are likely to resist IPO for the
cultural bias and these firms are more likely to have low
M/B. Therefore also estimate the linear probability
model with firm-specific effects.
• Why report standard error for each coefficient?
• Why is the SIZE insignificant for carve-outs?
1. Fixed flotation cost is partly sunk for subsidiaries
2. Size may proxy reputation
Why do companies go public?
• Yit = α + ∑j=03βjIPOi,t-j + β4IPOi,t-n + ∑j=03 γjQUOTi,t-j + ui + dt + eit
ui is firm-specific effect and dt is calendar year specific effect.
• Include QUOT to capture the effect of meeting the listing
requirement. This will correct the sample selection bias: more
profitable firms will be qualified for listing.
• To deal with the endogenous selection bias: firms that went public
have chosen to do so, should run the two-stage regression. Should
first estimate equation 1 and then include this estimated probability
of listing into equation 2.
• Also include lagged values of dependent variable and other lagged
dependent variables into equation 2.
• What are the meanings of β0 through β4?
Why do companies go public?
•
1.
2.
3.
4.
5.
Why do ROA declines after IPO?
Statistical coincidence - mean reversion: so include the
first lag of ROA and ROA in the year before IPO into
the equation 2 but results remain the same
Window dressing: public firms tend to inflate asset and
profit but private firms tend to deflate asset and profit.
High inflation rate in Italy makes government
encourage firms to step up the book value of asset.
Cash raised in IPO is temporarily invested in interest
earning financial assets
Moral hazard and adverse selection: change of the
incumbent’s stake at IPO is positively related to
change in post IPO ROA
Time the IPO
Why do companies go public?
•
1.
2.
3.
•
•
•
Why do bank credit interest rate falls after IPO?
Improvement in credit quality: leverage reduced and firms become
safer. So include ROA, Leverage, and Size to control for risk.
Results remain the same!
More information about borrowers
More outside financing options to weaken bank bargaining power
Initial owners divest very little of their holdings. This is true even if
we factor in how much the new equity raised at IPO or in
subsequent years was purchased by initial owners.
In 40.6% of cases, control group does not sell its equity and
demands new funds from outside investors. In another 40.6% of
cases, control group divests and does not raise new equity.
In the three years after the IPO, control group sells out the
controlling stake to an outsider in 13.6% of cases. This may be
due to (1) ease of transferring control of a public firm, (2) bad post
IPO performance , or (3) IPO as a step for controlling group to sell
the firm eventually.