Sie sind auf Seite 1von 16

Journal of Management and Governance 2: 1–16, 1998.

1
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

The Stock Market Reaction to Investment


Decisions: Evidence from Italy ?

EMANUELE BAJO1 , MARCO BIGELLI2 and SANDRO SANDRI2


1 Università Cattolica del Sacro Cuore, Milano, Italy; 2 Department of Applied Economic-Sciences,
University of Bologna, Piazza Scaravilli 1, 40126 Bologna, Italy; E-mail:
mbigelli@economia.unibo.it

Abstract. Based on a study of new investment announcements from 1989 to 1995 by Italian firms
listed on the Milan Stock Exchange, we find a positive stock price reaction to new investment deci-
sions. The stock price reaction is larger for joint venture announcements. The market response is
also larger for non-state owned companies and when the announcement is released in a period of
rising stock prices. The announced investment has no impact on the non-voting shares but increases
the voting shares’ market price through a significant revaluation of their vote-segment. We find
some evidence that new investments lead to management’s private benefits rather than towards firm
value. This is consistent with the typical Italian corporate governance structure, where a majority
shareholder safely controls a listed company while having only a fractional claim on the firm’s cash
flows.

1. Introduction
The competitive decline of U.S. firms versus Japanese firms in the 1980s has been
blamed on the lower U.S. rate of long term investments and R&D expenses. Porter
(1992a, 1992b) argues that the failure of the American capital investment system
is due to the so-called market myopia where the stock market rewards projects
with near-term cash flows and punishes investment in projects with long-term cash
flow profiles. In his opinion, the low rate of long-term investments in the U.S. can
be attributed to the failure of the corporate governance system, which is based
on institutional investors focused on short-run firm performance and with short-
lived ownership relations, as opposed to the long-lasting ownership links of the
Japanese Keiretsus. According to the market myopia hypothesis or institutional
investor hypothesis, management is unwilling to undertake long-term investments
in the fear that the consequent reduction in short-term earnings would depress
stock prices and make the firm a more attractive takeover target. The empiri-
? The present research has been financed by C.N.R. funds. We would like to thank Ugo Sini-
gaglia’s Dati Service Snc for having kindly offered some market data. We also thank Prof. Vikas
Mehrotra, the participants at the AFFI 1997 and EFMA 1997 meetings and two anonymous referees
of this Journal for their precious hints and useful comments.
2 EMANUELE BAJO ET AL.

cal implication of the market myopia theory is a negative market reaction at the
announcements of long-term investment decisions.
However, according to the shareholder value maximization hypothesis (Fama
and Jensen 1985), if the stock market is not myopic but efficient and managers try
to maximize firm’s value by financing positive NPV projects, the expected market
reaction to the announcement of a new investment should be positive. A positive
price-reaction is predicted also by the interpretation of Stein’s (1988) manager-
ial myopia. In order to signal their firms’ quality in an asymmetric information
setting, managers of profitable firms must maintain high earnings and dividends
and therefore forego positive NPV projects. Since only those projects with large
positive NPVs would be financed, the announcements of new investments should
be accompanied by positive abnormal returns in the firm’s stock price.
According to the Highly Competitive Market Hypothesis (Woolridge 1988),
a neutral market reaction to the announcements of new investments would be
expected if the products and factors markets are highly competitive and priced
so as strategic investment decision with positive NPV would be rare.
Extant empirical evidence from the U.S. has rejected the market myopia hypo-
thesis and found significant positive two-days cumulative abnormal returns of
about +1% at the announcements of new investments made by US companies
(McConnel and Muscarella 1985; Woolridge 1988; Woolridge and Snow 1990).
The stronger positive reaction found by Woolridge (1988), Woolridge and Snow
(1990) and Chan, Martin and Kensinger (1990) for investments in intangibles, such
as R&D, further rejects the market myopia hypothesis.
The Italian low rate of long-term investment and R&D spending is even less
than the American rate.1 However, this can not be attributed to any sort of market
myopia argument for at least two reasons. First, the Italian ownership structure
is completely different from the U.S. public company model: it is characterized
by the presence of a majority shareholder (often a family or a voting trust) and
institutional investors play a minor role. Secondly, the aggregate comparatively
lower level of long term investment and R&D spending can not be explained by a
hypothetical investor preference for short-run performance, as only a small number
of Italian companies are exchange-listed. However, in the light of the general trend
to broaden the Italian stock market and make it more representative of the country’s
industrial economy, it would be interesting to know the market reaction to the
announcement of new investments, as it could influence the aggregate propensity
to invest. Additional investigations are stimulated by the peculiar Italian corporate
governance structure, where the contemporary use of pyramidal groups and non-
voting shares allows the ultimate majority shareholder to have a safe control of the
company with a claim on the firm’s cash flows often below 5–10% (Brioschi et
al., 1989). In fact, by getting 100% of the private benefits and only 5–10% of the
increase in firm’s value, a rational majority shareholder should undertake invest-
ments that maximize his private benefits rather than firm value. Thus, a test of the
market reaction to the announcements of new investments made by Italian listed
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 3

companies can not be considered as a test of the market myopia for two differ-
ent reasons. First, both because of the minor role played by institutional investors
(which are usually supposed to be the cause of myopic markets) and because an
eventual temporary fall in market prices due to the announcement of a long term
investment would not trigger any threat of a takeover, when the ownership structure
is characterized by the presence of a majority shareholder. Secondly, a negative
market reaction to new investment may be explained by negative NPV projects
rather than by myopic investors. This could be true especially in the context of
the Italian corporate governance structure, where a negative NPV project could
anyway be undertaken if it leads to an increase of the private benefits pertaining to
the majority shareholder.
This paper carries out the first empirical study of the market reaction to the
announcements of new investments on the Italian stock market. Similarly to the
above cited US empirical studies, we find a weakly significant (P-value = 0,057)
general positive reaction (+0.61%) to the announcements of new investments,
which is stronger and more significant for the announcements of joint ventures
(+1.26%), for non-state-owned firms (+0.77%) and when the new investment is
announced in a period of rising stock prices. Besides, by separating the market
reaction on the vote-segment of the voting shares, we find some evidence that new
investments are made to maximize majority shareholder’s private benefits rather
than firm’s value.
The paper is organized as follows. Section 2 describes the data collection pro-
cedure, the event study methodology, the profitability index, the separation of the
market reaction between the vote and the investment-segment, and the definition
of some independent variables used in the cross-sectional analysis of the observed
abnormal returns. Section 3 analyzes the statistical results while Section 4 presents
our conclusions.

2. The Empirical Research Methodology


2.1. THE SAMPLE
Our sample consists of new investment announcements in the 1989–1995 period
by firms listed on the Milan Stock Exchange. These announcements meet the
following selection criteria:
a) the first announcement of a new investment is correctly identified in the “Il Sole
24 Ore” CD-ROMs;2 the new investment is the only information announced
or, in case of other contemporary news releases,3 the declared investment size
must be greater than 20% of the market value of the firm’s outstanding equity;
b) the firm’s voting shares must be listed for at least 130 trading days before
the announcement date and its stock price series available either on the
“Metastock” database4 or in the “Il Sole 24 Ore” online web-database.
4 EMANUELE BAJO ET AL.

Table I. Distribution of the observed announcements of new investments across


industries and types of new investment

Industry # obs. Capital Joint R&D Other


expend. ventures

Automotive 10 5 2 0 3
Chemical 1 1 0 0 0
Computers 5 1 4 0 0
Energy 4 3 1 0 0
Pharmaceuticals 1 0 1 0 0
Financial 6 4 2 0 0
Food 2 1 0 0 1
Glass frames 1 0 0 0 1
Home appliances 1 0 1 0 0
Mechanical manufacturing 4 0 4 0 0
Paper 1 1 0 0 0
Retailing 1 1 0 0 0
Iron and Steel 1 1 0 0 0
Telecommunications 12 9 3 0 0
Textile 7 3 4 0 0
Tourism 3 3 0 0 0
Transportation 5 3 1 0 1
Wires and rubber 5 0 3 1 1

Total 70 36 26 1 7

The selection criterion under b) allows us to exclude those observations where


the announcement of the new investment may have only partially determined the
security’s return, given the relatively small size of the investment and the presence
of some other relevant information.
The final sample consists of 70 announcements of new investments made by
27 firms. Of these, 61 have no other contemporary news releases and 57 have
information on the investment size. The distribution of the new investments across
industries and types of investment is reported in Table I. 5

2.2. THE EVENT STUDY

We use the standard “market model” to estimate abnormal returns around the
announcement dates:

Ri,t = αi + βi Rm,t + ei,t ,


THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 5

where Ri,t is the company’s i return on day t; αi and βi are the market-model
parameters estimated in the time window starting on day –130 through day –31
before the “Sole 24 Ore” newspaper announcement date (day 0).
Rm,t is the BCI (Banca Commerciale Italiana) value-weighted market index
return on day t.
We also calculate cumulative abnormal returns (CARs) for several windows
both around the announcement date (two, three and five-day CARs) and from day
–30 to day +30 around the announcement date. Since the three-days CAR (–1, 0,
+1) appeared to be the most significant and representative of the market reaction,
we focus on the results for the three-days event window while also giving some
information on the other CARs.6

2.3. THE NPV INDEX


In order to have an estimate of the new-investments’ average profitability and make
some cross-sectional comparisons, following Huson, Morck, Smith and Yu (1996)
we have appositely constructed a profitability index (NPVindex).
The abnormal return in response to the announcement of a new investment is
determined both by the project profitability or net present value and by firm size,
as measured by the market value of equity. In fact, consider a project with an initial
cost of $100 and a net present value of $10. A firm with a market capitalization
worth $1000 would experience an abnormal return equal to +1% (+$10/$1000)
while a firm with a market capitalization of $10.000 would see its stock price rise
by only 0.1% ($10/$10.000). In order to eliminate such size effect we first compute
the market expectation of the project’s net present value as:
X
E(NP V ) = CARI · VEi ·
i

where CARi is the three-days CAR of the class i of shares (voting, non-voting or
privileged shares) and VEi is the market value of the firm’s class i equity as of two
days before the announcement date.
To get a measure of the projects’ relative profitability, we then scale such NPV
estimate by its initial cost (when available in the announcement) and defined the
ratio as the NPVindex, that is:
E(NP V )
NP V index = .
I0
In the above example, the NPVindex of the project is now equal to 10% for both
the two firms.
6 EMANUELE BAJO ET AL.

2.4. THE SEPARATE MARKET REACTION ON THE VOTE AND THE


INVESTMENT- SEGMENT

Italian voting shares quote at a large premium over non-voting shares, (on average
about 80%, Zingales 1994), despite lower dividends.7 The high voting premium can
be explained by both large private benefits from the control of an Italian company
and from the possibility of expropriating minority shareholders and non-voting
shareholder rights (Zingales 1994; Bigelli and Caprio 1996).8 Therefore, we split
the pre-announcement market price of the voting shares (Pv ) into two parts, the
investment-segment (Inv) and the vote-segment (Vote):

Pv = I nv + V ote.

We define the investment-segment as the price of the non-voting shares (Pnv ) minus
a discount for taking into consideration the lower dividend stream for the voting
shareholders:

I nv = Pnv. − discount.

We compute the appropriate discount for each dual-class firm in the sample as
the present value at the pre-announcement date of a perpetuity given by the higher
dividend pertaining to the non-voting share either by law (2% of the par value of the
share) or by the company’s charter. The present value of the perpetuity is obtained
using as a discounting rate the current 12-months T-bill gross rate plus a 5% risk
premium. Consider the CAR for the voting shares (CARv ) as a weighted CAR of
the market reactions on the vote-segment (CARvot e) and the investment-segment
(CARinv ):
I nv V ote
CARv = CARinv. + CARvot e · .
Pv Pv
The CAR of the investment-segment (CARinv ) can be considered equal to the CAR
of the non voting shares, since the discount (the difference between the investment-
segment and the value of a non-voting share) can be treated as a constant, as its
value depends on interest rates and the dividend surplus pertaining to a non-voting
share by law or by the company’s charter.9 The CAR of the vote-segment can there-
fore be derived by observing the CAR of the voting shares and by substituting the
CAR on the investment-segment with the CAR observed on the non-voting shares,
as follows:
Pv I nv
CARvot e = CARv · − CARnv · .
V ote V ote
New investments could be financed to maximize management’s private benefits
rather than firm’s value (Jensen 1986). This risk is probably higher in the context of
the Italian corporate governance structure, where the majority shareholder controls
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 7

the company while having only a small claim of the firm’s cash flows. If new
investments have zero net present value but increase the majority shareholder’s
private benefits, we should observe a zero CAR on the non-voting shares and a
positive CAR on the voting shares, as higher expected private benefits would be
reflected in a more valuable vote-segment and a higher price of the voting shares.

2.5. INDEPENDENT VARIABLES

We use the following independent variables to explain the cross-sectional variation


among the abnormal returns observed at the announcements of new investments:

The investment relative size (Relsize): this variable is the ratio of the new invest-
ment size (when available) on the market value of the firm’s equity as of two days
before the announcement date. In case of joint ventures announcements, the invest-
ment size refers to the capital invested in the joint venture by the announcing firm.
As in Woolrdidge and Snow (1990) and Chan, Martin and Kensinger (1990), we
check if the market response to a new investment is sensitive to its size, as it could
be expected that larger investments would have a greater impact on firm value.

The company size (Firmsize): as a proxy for firm size we take the natural log of
the market value of the firm’s combined equity (voting, non-voting and privileged
shares) as of two days before the public announcement of the new investment. We
decided to control for size since it could be positively correlated with liquidity.

The company Market-to-book ratio (M/B-Ratio): it’s our proxy of Tobin’s Q, com-
puted as the ratio of the pre-announcement firm’s market capitalization over the
company’s book value as of the end of the fiscal year preceding the announcement.
Being a proxy of the firm’s performance, the higher the market-to-book ratio, the
better the management ability in previous investments, and the investors’ reaction
to the announcement of new investments.

A dummy for the type of new investment (Dtype): as found by Wooldridge and
Snow (1990), this dummy should highlight a possible different reaction for the two
main types of investment announcements: capital expenditures (Dtype = 0) and
joint ventures (Dtype = 1).

A dummy for state-owned companies (Dstateown): we set this variable equal to 1


when the company is directly or indirectly state-owned. This way, we can test for
a different market reaction to new investments for state-owned companies.

A dummy for a rising stock market period (Dbullmkt): we set such dummy variable
equal to 1 in the rising stock market period lasting from January 2nd 1993 till Octo-
ber 6th 1993. According to Choe, Masulis and Nanda (1993), in the expansionary
8 EMANUELE BAJO ET AL.

Table II. Percentage three days CARs for the whole sample and for the sub-samples partitioned
by types of new investment and ownership. P-values for T-test (equal means) and Wilkoxon
signed-rank Test (equal medians)

n Mean Median Min Max %CAR>0

Whole sample 70 +0.61% +0.30% –5.23% 10.08% 55.17%


(0.057) (0.403)

Panel A
Capital expenditures 36 +0.11% 0.19% –5.23% 6.47% 55.50%
(0.793) (0.618)

Joint ventures 26 +1.26% +0.63% –2.90% 10.08% 57.69%


(0.039) (0.557)

Difference –1.15% –0.44%


(0.095) (0.609)

Panel B
Privately-owned companies 53 +0.77% +0.36% –3.75% 10.08% 56.60%
(0.036) (0.410)

State-owned companies 17 +0.07% +0.15% –5.23% 5.08% 52.94%


(0.907) (1.000)

Difference +0.70% +0.21%


(0.341) (0.782)

phases of the economic cycle there are more investment opportunities with positive
net present values. If their thesis holds, the announcement of a new investment in
the middle of a rising stock market preceding a new expansionary economic cycle
should be more welcomed by investors.

A dummy for multi-year investments (Dyears): as in Woolridge and Snow (1990),


by setting this variable equal to 1 for those investments planned for more than one
year, we verify if there is a different market reaction between short and long term
investments.

3. Results
3.1. THE MARKET REACTION TO NEW INVESTMENTS

The announcements of new investments by Italian-listed companies are accom-


panied by a positive and weakly significant (P-value = 0.057) average abnormal-
return, equal to +0.61% for the whole sample of 70 observations in the three
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 9

days around the announcement date (Table II). When we split the overall sample
into the two subsets of capital expenditure and joint venture announcements,10 we
observe an average insignificant three-days CAR equal to +0.11% for new capital
expenditures and a significant positive CAR equal to +1.26% for new joint venture
announcements. Although the two CARs differ only at a 10% probability level, the
larger price-effect of the joint venture announcements can be viewed as supportive
of the synergy hypothesis and confirms the results found by McConnel and Nantell
(1985), Woolridge (1988) and Woolridge and Snow (1990). In Panel B of Table II,
we partition the whole sample into new investments announced by privately-owned
companies and state-owned companies. The announcement of new investments
generates a significant positive CAR equal to +0.77% for privately-owned com-
panies, and an insignificant market reaction (CAR = +0.07%) for state-owned
companies.
As far as the results of the other event windows, the market reaction in the
five-days around the announcement date (–2, +2) has always the same sign of the
three-days CAR but it is generally less significant, which suggests that the stock
market has been fairly efficient in promptly adjusting stock prices to the new
information released. Furthermore, the overall lack of significance of two more
CARs measures computed on two event-windows before the announcement date
(respectively from –30 till –1, and –7 till –1) seem to exclude an anticipation of the
event by some informed traders. The reported absence of insider trading (relatively
unusual on the Italian stock market of the studied period) is probably due to the
fact that the new information was not expected to affect stock prices so much as to
give rise to a relevant trading profit. The weak abnormal returns observed around
the announcement support such interpretation.

3.2. THE CROSS - SECTIONAL VARIATION OF THE MARKET REACTION TO NEW


INVESTMENTS

From the results of the three cross-sectional regression models reported in Table III,
only the dummy variable Bullmkt always results highly significant. This indicates
that new investments are considered more profitable and welcomed more favorably
by investors when announced in a period of rising stock prices, as sustained by
Choe, Masulis and Nanda (1993). Among the other dummy variables, only Dtype
results weakly significant in some models (1 and 2), therefore confirming the more
favorable reaction to joint venture announcements. Finally, the lack of significance
for the regression coefficients of Firmsize, Relsize and M/B-ratio implies that the
stock price-effect at the announcement of new investments does not seem to be
influenced by firm size, the relative size of the new investment, and the company’s
market-to-book ratio.
The regression results of the other CAR measures confirm the results from the
univariate analysis, that is, a market reaction concentrated around the announce-
ment date and not anticipated. In fact, when using the five-days CAR as the
10 EMANUELE BAJO ET AL.

Table III. OLS cross-sectional regressions of the three-days CARs on a set of potential
explanatory variables. (Values for student-T in brackets and P-value for the White test’s
Chi-squared)

Model 1 2 3
(n = 62) (n = 62) (n = 49)

Constant –0.015 –0.018 –0.018


(–1.439) (–1.513) (–0.938)

Dtype 0.012 0.013 0.010


(1.841)∗ (1.868)∗ (1.397)

Dstateown –0.011 –0.009 –0.014


(–1.298) (–1.146) (–1.544)

Dyears 0.015 0.015 0.015


(1.655) (1.703) (1.734)

Dbullmkt 0.025 0.026 0.031


(2.669)∗∗∗ (2.688)∗∗∗ (3.419)∗∗∗

M/B-ratio 0.001 0.001


(0.523) (0.587)

Firmsize 0.000
(0.038)

Relsize 0.000
(1.392)

R2 0.214 0.218 0.366


Adjusted R2 0.159 0.148 0.257
White test 0.578 0.556 0.387

dependent variable, the signs of the explanatory variables do not change: Bullmkt
is still always significant, but all the models R-squares are slightly lower.11 When
we used the two CAR measures computed over the pre-announcement period, the
regressions’ R-squares drop substantially and the models are not significant.

3.3. WHICH IS THE NET PRESENT VALUE OF NEW INVESTMENTS ?

In order to determine the net present value of new investments we use the prof-
itability index (NPVindex) explained above. In the light of what we say in the
next paragraph, we have to point out that the value of the expected NPV used in
the index is measured as the abnormal change in the firm’s market capitalization,
therefore including any appreciation of the vote-segment of the voting shares. Some
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 11
Table IV. Measures of the profitability index (NPVindex) in the sub-samples
of announcements with declared investment size partitioned by an investment
relative size (Relsize) greater than a given percentage of the firm’s market
capitalization before the announcement. P-values for T-test (equal means)

Cutoff n Mean Median Std. Dev. Min. Max.

Relsize > 0 57 0.647 0.002 3.680 –10.910 19.952


(0.189)

Relsize > 5% 44 0.041 0.001 0.140 –0.311 0.429


(0.059)

Relsize > 25% 32 0.021 0.001 0.056 –0.042 0.236


(0.044)

Relsize > 50% 26 0.008 0.000 0.032 –0.042 0.147


(0.229)

precautions must be taken for such an index to have representative values. In fact,
suppose the measured abnormal return is equal to +1% and the investment size is
equal only to 0.1% of the market value of the outstanding equity, which is worth
$100. According to the NPVindex the project would have a NPV equal to 1000% of
its initial cost [E(NPV) = $1], which is obviously unrealistic. This happens because
the measured abnormal return is likely to be only partially generated by the capital
expenditure announcement, considering its small size. This is the reason why our
sample selection procedure excluded those announcements accompanied by some
other relevant information if the investment size was not highly relevant (above
20% of the market value of the firm’s equity). At the same time, we determine
the NPVindex on different subsets of our sample by excluding those investments
whose relative size is lower than a given percentage of the firm’s market capital-
ization before the announcement (Table IV). As it can be seen from the minimum
and maximum values in the first raw of Table IV, when small investments are not
excluded the NPVindex can assume meaningless values. By cutting off investments
with relative size lower than 5% of the firm’ market capitalization, the NPVindex
starts giving more realistic values, since the NPV is equal to 4.1% of the initial
investment, on average, and varies from –31.1% to +42.9%. When we use the 25%
relative size cutoff, the average profitability index reduces to about +2%, though it
is still weakly significant. When we consider only the largest investment projects
(relative size greater than 50% of the firm’s capitalization), the profitability index
is not different from zero (+0.8%), varies in the narrowest range (from –4.2%
to +14.7%) and presents the lowest standard deviation. These results support the
Highly Competitive Market Hypothesis, since they can be interpreted as if the larger
is the investment (and the better is its NPV estimate) the lower is the likelihood
12 EMANUELE BAJO ET AL.

Table V. Comparisons of the three-days CARs for the voting and non-voting shares and
for the vote and investment-segment of the voting share. Comparisons are made on the
sub-sample of 27 observations with both voting and non-voting shares listed and with
a voting premium (Pv − Pnv ) greater than 20% of the non-voting share market price.
P-values for T-test (equal means) and Wilkoxon signed-rank test (equal medians)

Mean Median Min Max

Premium% 69.58% 74.48% 24.92% 139.53%


Vote% 32.99% 36.95% 11.95% 54.07%

Panel A
CARv +0.58% +0.36% –2.90% +5.88%
(0.175) (0.701)

CARnv –0.16% –0.13% –6.57% +9.79%


(0.346) (0.701)

Difference +0.73% +0.79% –11.19% +7.48%


(0.071) (0.248)

Panel B
CARvot e +2.52% +1.83% –9.42% +28.60%
(0.088) (0.122)

CARinv –0.16% –0.13% –6.57% +9.79%


(0.346) (0.701)

Difference +2.68% +1.68% –19.21% +28.73%


(0.067) (0.248)

that the project has a return well in excess of the cost of capital. We also run sev-
eral cross-sectional regression of the NPVindex on the above defined independent
variables without reporting significant results.

3.4. FIRM ’ S VALUE VERSUS PRIVATE BENEFITS MAXIMIZATION

The huge average spread between the Italian voting and non-voting shares allows
us to separate the stock market reaction to the announcement of a new investments
into expected net present value from the project and from changes in expected
private benefits. In fact, since the non-voting share market price reflects only the
present value of expected cash flows, its price variation at the announcement of a
new investment should reflect the expected net present value from the project. On
the other hand, as above explained, the voting share market price can be thought
as made of two parts: a vote-segment and an investment-segment. While the
investment-segment incorporates the expected net present value from the project,
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 13

changes in the expected private benefits would be reflected in a different value


of the vote-segment. By looking at a possible different reaction between the two
segments of a voting share we can infer if the new project is expected to increase
private benefits or firm’s value.
In a subset of 27 observations with both the voting and the non-voting shares
listed and with a voting premium (Pv − Pnv ) greater than 20%,12 we used the above
explained methodology to separate the vote-segment part of the voting shares and
derive its three-days CAR. In the selected subset, voting shares were quoting at an
average premium of about 70% on the non-voting share market price and this cor-
responded to a vote-segment which was explaining almost 30% of the voting share
market price. From Panel A of Table V we can see that the market reaction to the
announcement of a new investment13 has been slightly negative for the non-voting
shares (–0.16%) and positive for the voting shares (+0.58%). Though the signed-
rank test for the difference (+0.73%) is significant only at the 10% probability
level, this represents a first piece of evidence that new investments favor voting
shareholders through an increase of the expected private benefits incorporated in
the voting premium. When we separate the market reaction on the vote and on
the investment-segment, the above difference gets larger and more significant.14
The effect of new investment announcements is an increase in the expected private
benefit equal to +2.52% (as measured by CARvot e) and a slight decrease in the
expected cash flow equal to –0.16% (as measured by CARinv ). The signed-rank
test on the difference (+2.68%) is significant at the 7% probability level and robust:
it remains under the 10% probability level when using different minimum voting
premium cutoffs in the sub-sample selection (15% and 25%). The results seem
therefore to offer some evidence that Italian listed companies may finance new
investments to maximize the majority shareholder’s private benefits rather than firm
value.

4. Conclusions
We study the effect of new investment announcements by companies listed on
the Milan Stock Exchange. Our results show an average positive market reaction
to the new investment announcements. The three-day market-model cumulative
abnormal return for the overall sample is significant and equals +0.61% (P value =
0.057). However, as found in some US studies (Woolridge 1988; Woolridge and
Snow 1990), the market response to the announcements of new joint ventures
(+1.26%) is significantly larger than for new capital expenditures (+0.11%). A
more favorable reaction is also found for privately-owned companies (+0.77%),
than for government-owned companies (+0.07%), and when the new investment is
announced in a period of rising stock prices.
When the expected net present value is computed as the change in the firm’s
market capitalization around the announcement date, the NPVs from the new
projects are estimated at about +2% of the initial cost and weakly significant. In
14 EMANUELE BAJO ET AL.

accordance with the Highly Competitive Market Hypothesis, the values observed
for the constructed profitability index seem to indicate that the larger the size of
the investment relative to the market value of the firm’s equity, the lower the
profitability index (i.e. the ratio of the expected net present value on its initial
cost).
The announcement of the new investments has a different effect on voting and
non-voting shares. In the subset of observations where both classes of shares were
listed, the voting shares experienced a three-days CAR equal to +0.58% versus a
–0.16% of the non-voting shares. By separating the market reaction on the voting
shares between the vote-segment and the investment-segment, we find that the
announcements of new investments are considered to have zero or negative net
present values, as the investment-segment barely varies, while they significantly
increase the expected private benefits of the majority shareholder, as indicated by
the positive three-days CAR on the vote-segment (+2.52%). In other words, the
Italian corporate governance structure, where a majority shareholder safely con-
trols a listed company while having only a small claim on the firm’s cash flows,
seems to favor investments which maximize management’s private benefits rather
than overall firm value.

Notes
1 In 1995 Italian investments in R&D were only equal to 1.3% of its GDP, versus a percentage equal
to 2.4% for US and almost 3% for Japan (Data from the OCSE-World Economic Forum).
2 “Il Sole 24 Ore” is the most widespread Italian financial newspaper.
3 Typically information on sales or earnings. The information on how the new investment is going
to be financed is rarely released and is not as important as for the US market. In fact, while the
announcement of a future equity issue would generally depress stock prices in the US setting, active
insiders and the common signal of an implicit increase of the dividend yield lead to a more dispersed
and generally favorable reaction to the announcements of Italian equity rights issues (Bigelli 1998).
Moreover, the sign and size of the reaction have been found to depend on the terms of the offering,
which are not known at the investment decision date.
4 The metastock database is the stock price database of the eponymous software for technical
analysis.
5 Seven observations could not be classified as capital expenditures, joint ventures or investment
in R&D as they were either the development of a new product or a combination of the other three
classified types of investments.
6 A correct interpretation of the results depends, therefore, on the implied assumptions that the stock
market is efficient in translating the new information on investment decision into stock prices.
7 Non-voting shares are entitled by law (Law 216/1974) to a minimum dividend equal to 5% of par
value and to an additional dividend of 2% of par value in excess to the dividend paid to the voting
shares.
8 As examples of private benefits, different from the classical perquisites, we can quote the following
possibilities for the majority shareholder: paying itself “special dividends” (as called by Shleifer
and Vishny 1997) by exploiting the business relationships with the other companies they control;
selling controlled firms over their value when sold to companies at a lower stage of the pyramidal
group (as highlighted by the “Stet case-study” shown by Zingales 1994) and vice versa, when sold
to a controlling company of the group; doing insider trading (which was legal till 1991 and then
THE STOCK MARKET REACTION TO INVESTMENT DECISIONS 15

not severely repressed as in other countries); expropriating non-voting shareholders through rights
offerings (as shown in Bigelli and Caprio 1996) and others.
9 We are also implicitly assuming that the two stock categories have the same liquidity so that the
price adjustments to the new information take place contemporaneously. However, Zingales (1994)
reports that when liquidity is measured by both the number of shares traded and the total value of
transactions voting shares result more liquid, while when we look at the number of shares traded
over the number of shares outstanding non-voting shares result more liquid, as large blocks of voting
shares are never traded.
10 As explained above when commenting Table I, eight observations could not be classified either
as capital expenditures or joint ventures and could not be treated as a separate homogeneous sub-
sample.
11 Respectively equal for the three models to 0.170, 0.172 and 0.336.
12 We need to take only those observations with a minimum voting premium to avoid generating
unreasonable results. In fact, let consider an example where voting shares quote at a premium of
only 1% on the non-voting shares market price and define the vote-segment simply as the difference
between the voting and the non-voting share market prices. In this case, a comparable market reaction
on the two classes of shares, equal to a +1% CAR on the non-voting shares and a +1.1% CAR on the
voting share would correspond to a +10% CAR on the vote-segment and make us conclude that the
investment is made to maximize private benefits.
13 Only the three-days CAR have been considered since from the previous analysis this time win-
dow appeared to be that which has best captured the market reaction to the announcement of new
investments.
14 Though the medians are still not significant, probably because of the small sample size.

References
Bigelli, M. (1998). The Quasi-Split Effect, Active Insiders and The Italian Market Reaction to Equity
Rights Issues. European Financial Management 4: 185–206.
Bigelli, M. & Caprio, L. (1996). Equity Rights Issues and Differential Returns Between the Voting
and Non-Voting Classes of Shares: Evidence from Italy. Working paper presented at the AFFI
Conference 1996 and at the EFA Conference 1996.
Brioschi, F., Buzzacchi, L. & Colombo, M. G. (1989). Gruppi industriali e mercati finanziari.
Harvard Espansione 45: 119–123.
Chan, S. H., Martin, J. D. & Kensinger, J. W. (1990). Corporate Research and Development
Expenditures and Share Value. Journal of Financial Economics 26: 255–276.
Choe, H., Masulis, R. W. & Nanda, V. (1993). Common Stock Offerings Across the Business Cycle.
Journal of Empirical Finance 1: 3–31.
Fama, E. F. & Jensen, M. (1985). Organizational Forms and Investment Decisions. Journal of
Financial Economics 14: 101–119.
Huson, M., Morck, R., Smith, G. & Yu, W. (1996). Cross-Sectional Variation in the Stock Market’s
Reaction to Long Term Corporate Investment Decisions. Working paper presented at the NFA
Conference 1996.
Jensen, M. C. (1986). Agency Costs of Free Cash Flows, Corporate Finance and Takeovers. American
Economic Review 76: 323–329.
McConnell, J. J. & Nantell, T. J. (1985). Corporate Combinations and Common Stock Returns: The
case of Joint Ventures. Journal of Finance 40: 519–536.
McConnel, J. J. & Muscarella, C. J. (1985). Corporate Capital Expenditure Decisions and the Market
Value of a Firm. Journal of Financial Economics 14: 399–422.
Porter, M. E. (1992a). Capital Choices: Changing the Way America Invests in Industry. Council on
Competitiveness and Harvard Business School.
16 EMANUELE BAJO ET AL.

Porter, M. E. (1992b). Capital Disadvantage: America’s Failing Capital Investment System. Harvard
Business Review, September–October: 65–82.
Shleifer, A. & Vishny, R. (1997). A Survey of Corporate Governance. The Journal of Finance 52:
737–783.
Stein, J. C. (1988). Efficient Capital Markets, Inefficient firms: A Model of Myopic Corporate
Behavior. Quarterly Journal of Economics November: 655–669.
Woolridge, J. R. (1988). Competitive Decline and Corporate Restructuring: Is a Myopic Stock
Market to Blame? Journal of Applied Corporate Finance: 26–36.
Woolridge, J. R. & Snow, C. (1990). Stock Market Reaction to Strategic Investment Decisions.
Strategic Management Journal 11: 353–363.
Zingales, L. (1994). The Value of the Voting Right: A Study of the Milan Stock Exchange
Experience. Review of Financial Studies 7: 125–148.

Address for correspondence: Marco Bigelli, Università di Bologna, Dipartimento di Discipline


Economico-aziendali, Piazza Scaravilli 2, 40126 Bologna, Italy
E-mail: mbigelli@economia.unibo.it

Das könnte Ihnen auch gefallen