An Evaluation of the Value Relevance of Consolidated versus Unconsolidated Accounting Information: Evidence from Quoted Spanish Firms

AN EVALUATION OF THE VALUE RELEVANCE OFCONSOLIDATED VERSUS INCONSOLIDATED ACCOUNTING

INFORMATION: EVIDENCE FROM QUOTED SPANISHFIRMS

Christina Abad, Amalia Garcia-Borbolla, Neil Garrod, JoaquinaLaffarga, Manuel Larran and Juan Manuel Pinero.

Working Paper 2000/1

Department of Accounting and FinanceUniversity of Glasgow

Working Paper Series

The Department should like to acknowledge the help and support of the Wards Trust Fund set upin 1980 after the untimely death of James Cusator Wards.

Published 2000 by the Department of Accounting and Finance, University of Glasgow, GlasgowG12 8LE

ISBN 0 85261 637 6

1

AN EVALUATION OF THE VALUE RELEVANCE OF CONSOLIDATED VERSUS

UNCONSOLIDATED ACCOUNTING INFORMATION: EVIDENCE FROM QUOTED

SPANISH FIRMS.

Cristina Abad1, Amalia García-Borbolla2, Neil Garrod3, Joaquina Laffarga1, Manuel

Larrán2, and Juan Manuel Piñero2

1 Universidad de Sevilla.Address: Facultad de Ciencias Económicas y Empresariales. Departamento de Contabilidad y EconomíaFinanciera. Avenida Ramón y Cajal Nº1 Sevilla, 41.018, Spain.

2 Universidad de Cádiz.Address: Facultad de Ciencias Económicas y Empresariales. Departamento de Economía de la Empresa.Duque de Nájera, 8. Cádiz, 11.002, Spain.

3 University of Glasgow and Universidad de CadizAddress: Department of Accounting and Finance, University of Glasgow, 65-71, Southpark Avenue,Glasgow G12 8LE, Scotland, U.K.

Cristina AbadPhone number: 00·34·95 455 76 08E-mail: [email protected]

Amalia García-Borbolla.Phone number: 00·34·956 01 54 24E-mail: [email protected]

Neil GarrodPhone number: 00·44·141·330·5426E-mail: [email protected]

Joaquina Laffarga.Phone number: 00·34·95 455 76 09E-mail: [email protected]

Manuel Larrán.Phone number: 00·34·956 01 53 68E-mail: [email protected]

Juan Manuel Piñero.Phone number: 00·34·956 01 54 34.E-mail: [email protected]

Acknowledgement: Financial support for this paper under research project PB95-1254-C02-01, financed by the Dirección General de Enseñanza Superior is gratefully

2

acknowledged.

3

AN EVALUATION OF THE VALUE RELEVANCE OF CONSOLIDATED VERSUS

UNCONSOLIDATED ACCOUNTING INFORMATION:EVIDENCE FROM QUOTED

SPANISH FIRMS.

ABSTRACT:

In this paper we investigate the value-relevance of consolidated versus parent company

accounting information. In particular we investigate the value relevance of the minority interest

components of net total assets and earnings as currently reported and under the full entity

approach to consolidated reporting. An Edwards-Bell-Ohlson valuation framework is used to

generate results. By this means we cast light on the suitability of accounting regulation being

developed based upon the entity or parent company theories of consolidation. We carry out the

analysis in the Spanish context and the sample contains 474 observations of non-financial firms

quoted in the Madrid Stock Exchange for the period 1991-1997. The results from this analysis not

only have domestic relevance but provide guidance of a more international nature relating to the

impact of group definition, concepts of control and the most value relevant method of

consolidated disclosure. The results show that, from a valuation perspective, consolidated

information dominates unconsolidated, or parent company, information. However, neither the

currently reported minority interest components of net total assets and earnings, nor their values

under the full equity method of consolidation, are found to be value relevant. These results raise

the question of whether group definitions based on the equity theory of consolidation are the most

useful to investors.

4

AN EVALUATION OF THE VALUE RELEVANCE OF CONSOLIDATED VERSUS

UNCONSOLIDATED ACCOUNTING INFORMATION: EVIDENCE FROM QUOTED

SPANISH FIRMS.

1. INTRODUCTION

In this paper we investigate the value-relevance of consolidated versus parent company

accounting information. In particular we investigate the value relevance of the minority interest

components of net total assets and earnings as currently reported and under the full entity

approach to consolidated reporting, in the context of the Madrid Stock Exchange. The valuation

issues regarding consolidation are straightforward: do the parent company accounts or the

consolidated accounts, as currently prescribed, provide the most value relevant information for

pricing parent company shares? This issue has not been directly addressed previously for Spanish

companies. The value relevance of minority interest components of net total assets and earnings

and the choice of method to estimate these values, however, also have more international

ramifications and interest. Increasingly, accounting regulators are depending upon entity concepts

of consolidation to define the extent of ownership control (Financial Accounting Standards Board

(FASB), 1999; International Accounting Standards Committee (IASC), 1994) and, thus, the

definition of the reporting group. The research reported in this paper casts light on the suitability

of accounting regulation being developed upon the entity or parent company theories of

consolidation.

Under the parent company concept, the consolidated annual accounts are considered an extension

of the individual annual accounts of the parent company. Thus, the consolidated information is

considered complementary to the information disclosed in the parent company accounts. The

focus of this theory is that the interest of the parent company in the subsidiary companies is purely

financial in nature. Whilst not explicitly referred to in extant regulation, it is, by convention, the

parent company theory which underpins much of existing accounting practice world wide

(Beckman, 1995).

Under the entity approach, the group is considered to be the dominant economic unit and thus the

5

consolidated annual accounts are considered to be the most suitable format for providing

information about the financial situation of the parent. The focus of this theory is of the operating

unit being the group rather than the parent (Moonitz, 1951).

Until recently, the reporting of consolidated financial statements has been the exception rather

than the rule. It was uncommon within the countries of the European Union (EU) until the

enactment of the Seventh Company Law Directive (Seventh Directive) and its incorporation into

national legislation. In many emerging economies consolidated financial statements are either

relatively recent requirements or non existent. In the United States (US), where consolidated

statements have been required since 1959 (AICPA, 1959), the FASB has embarked upon a broad

multi-phase project on consolidations and related matters at the urging of their constituents,

including the Securities and Exchange Commission (Harris et.al., 1997). This has resulted in an

exposure draft (FASB, 1999) in which the definition of control incorporated in the FASB´s

proposed Statement of 1995 (FASB, 1995) is revised and clarified.

The publication of the exposure draft underscores a remaining controversy surrounding the

relevance and nature of consolidation, even within the US where it has been “indigenous to

American financial reporting” (Moonitz, 1951, p.10): what is a group and how is it defined? In

the 1995 proposed Statement it was required that a controlling entity (parent) should consolidate

all entities that it controls (subsidiaries). In the 1999 exposure draft this is not changed but

revisions are suggested which clarify the proposed definition of control of an entity. The definition

incorporated in the 1999 exposure draft is that “control involves decision making ability that is

not shared with others” (FASB, 1999, para 6). This differs from the earlier proposal where the

focus was on decision making powers relating to another entity’s individual assets. Thus,

although not explicit, the exposure draft moves closer to the entity concept of consolidation and

away from the traditional parent company approach. However, this movement is purely in terms

of group definition and still leaves much of the accounting practice fully influenced by the parent

company theory. For example, goodwill is calculated on the price paid for the internally owned

proportion of the subsidiary only and excludes the minority interest’s proportion of the

subsidiary’s revalued assets.

6

We provide evidence on the suitability of such a move by investigating the valuation relevance of

standard disclosure under the two pure theories of consolidation and the current practice which

uses entity definitions of a group and parent company accounting. We do this in two main stages.

The first issue we address is whether consolidated disclosure of any kind provides value relevant

information in excess of parent company disclosure. The answer to this question is neither

obvious nor unequivocal (Chambers, 1968). Investment in subsidiaries is reported in the parent

company accounts at cost. Assuming historic cost disclosure is itself value relevant (and there is a

raft of literature supporting such a view since the seminal work of Ball and Brown, 1968) then

there is no obvious reason why consolidated disclosure, which disaggregates the investment into

its constituents assets and liabilities, should provide increased value relevant disclosure. Indeed,

the concealment of financial distress or underlying solvency in the parent and the potential

influence of exchange rate fluctuations from subsidiaries may well lead to less value relevant

disclosure (Cea, 1992).

On the other hand, there is evidence that assets employed in different lines of business generate

earnings streams which are valued differently (Emmanuel and Garrod, 1992). Thus it might be

expected that the market value of a parent company will not be influenced solely by the specific

assets and liabilities of the parent company itself. In the development of its activities, the parent

company controls the assets and liabilities of its subsidiary companies (whether or not it has 100%

control of them). This control will play an essential role in the determination of future earnings of

the parent and, consequently, of its dividends. Thus information relating to the asset and liability

structure of the parent company’s investment, as revealed in the consolidated accounts, could lead

to the consolidated accounts being more value relevant than the parent company accounts alone.

Evidence regarding this issue is scant, largely due to US companies not disclosing parent company

accounts alone. Whilst consolidated and unconsolidated data are disclosed in several countries,

published work is limited. Darrough and Harris (1991) examine the effects of consolidation in

Japan and find little evidence of incremental information content or value relevance of

consolidated data. They conclude, however, that these results cannot be generalised due to the

7

unique institutional environment and inter-firm ownership relations in Japan. A study in which

data on German companies are used is presented by Harris et. al. (1997). They provide weak

evidence that consolidated financial statements are more value relevant than unconsolidated

financial statements. However, this is not a consistent finding over all the sample years and the

flexibility afforded German companies in the GAAP which can be applied to consolidated

accounts is claimed to influence the results. Our results provide further evidence on this issue.

Although not a direct valuation study, Pellens and Linnhoff (1993) investigate the changes in

certain financial ratios between parent and consolidated financial statements in Germany.

Although only a small study, of 28 companies/groups, the results indicate a significant difference

between the two statements with debt being higher and profitability lower in the consolidated

accounts. Similar conclusions are drawn by Lambert and Zimmer (1996) from a different sample

of German companies. Studies using Spanish data produce very similar results (see, e.g. Blasco,

1995; Larrán, 1996; Abad et al., 1997). They show that solvency (but not liquidity) and

profitability ratios calculated from consolidated statements tend to be worse than the equivalent

ratios calculated from the parent company statements.

The second issue which we address is which type of consolidated accounts are the most value

relevant: those based on the entity or parent company approach? Whilst there are accounting

differences which flow from the choice of parent company or entity approach, these largely centre

around the size of recorded goodwill and minority interest, which net off against each other. The

primary difference between the two approaches rests on the view taken regarding the nature of

the minority interest itself. Under the parent company approach the minority interest is perceived

as a liability on the parent company’s claim on subsidiary net assets and only the net assets owned

by the parent will contribute to future earnings and dividends. Under the entity concept, however,

the minority interest is seen as an alternative source of equity for the parent’s total net asset

holding and the totality of the assets and liabilities of the group will contribute to the generation of

a particular level of earnings for the parent company, since these are the assets and liabilities that

it controls. Thus, when it comes to making valuations about the future dividend stream of the

parent company under the entity concept, the totality of the groups net assets should be

8

considered (i.e. including minority interest) whilst under the parent company approach these net

assets should be excluded from the valuation process. Under current reporting practice in Spain,

and most other large market economies, the method of calculating minority interest is based on

the parent company theory of consolidation. It excludes the minority interest share of any asset

revaluation in the subsidiary and any goodwill from the purchase. Consequently we carry out a re-

estimation of minority interest using a full entity theory approach which we use in our empirical

tests.

To test for valuation differences between parent company accounts and consolidated accounts

prepared under the entity and parent company approaches we utilise a valuation model based on

the Edwards-Bell-Ohlson (EBO) model.

In the next section we discuss international diversity surrounding the reporting of minority

interest. In addition, we review relevant literature from financial economics that provides a

theoretical framework within which consolidated disclosure based on the entity and the parent

company approaches can be assessed. In section 3 we develop our theoretical model, research

method, hypotheses and sample. In section 4 the results of our empirical tests are reported and

our conclusions are to be found in section 5.

International Diversity in Categorising Minority Interest.

Whilst accounting procedures relating to consolidation still primarily reflect a parent company

approach (Beckman, 1995), US regulators, since the early 1990s, have been intimating a

preference for the entity theory of consolidation in the definition of a subsidiary which needs to be

consolidated and in the categorisation of minority interest. In its 1991 Discussion Memorandum

(FASB, 1991) the FASB expresses its preference for the concept of the economic unit. They

argue that control of all subsidiary activity is held by a single management team, that of the parent

company, and, thus, the concept of the economic unit best represents the totality of the assets and

liabilities under the control of the parent company.

For its part, the Financial Accounting Standards Committee of the American Accounting

9

Association (1994), in reply to this Memorandum, emphasises the fact that the choice between

one or other approach for the preparation of the consolidated information should be consistent

with the objective of improving the accounting information provided to users. They agree with the

FASB in supporting the use of the concept of the economic unit as the basis for the consolidation,

since, it is claimed, the relevant entity, on which information helpful to users should be published,

is the economic unit controlled by the managers of the parent company. In a later report, they

confirm this view (AAA, 1996) and state that the basis for consolidation should be the concept of

the economic unit, and support the inclusion of the minority interest as part of the equity shown in

the consolidated balance sheet.

In the UK, the Accounting Standards Board (ASB) in Financial Reporting Standard (FRS) 2:

“Accounting for Subsidiary Undertakings”, issued in 1992 (ASB, 1992), state that the purpose of

the consolidated financial statements is to reflect the economic unit which conducts its activities

under the control of the parent company. It also establishes that in the consolidated balance sheet

the minority interest is shown as part of the capital and reserves of the group, as is the case under

the entity concept of consolidation.

Thus regulators in both countries, the UK and US, which have disclosed consolidated accounts

for the longest time (Parker, 1984) are in favour of generating consolidated accounts in line with

the entity concept. This philosophy filtered through to the Seventh Directive which requires

consolidated disclosure throughout the countries of the Union. Although harmonisation is a clear

goal of European Directives they are all worded in such a way that national interpretations can be

placed upon them in the national regulations through which they are enacted. This is certainly true

of the Seventh Directive. Whilst the Directive requires consolidation and the use of accounting

procedures familiar in the UK and US, the earlier research on German consolidated disclosure

highlights the flexibility afforded in national interpretations of the Directive. A similar situation is

revealed in Spain.

The Spanish regulations specify that “given the diverse approaches that may be adopted towards

consolidation, particularly regarding the technical rules covering it, the following regulations have

10

been drafted according to the criterion that the consolidated accounts constitute a prolongation of

those of the parent or dominant company” (RD 1815/1991 translation). In other words, under

Spanish legislation, the consolidated annual accounts are considered to be complementary

information to the individual accounts, so that the presentation of the former does not represent

an exemption from the presentation of the latter. The Spanish regulators are making a clear

statement that the parent company theory rather than the entity theory of consolidated reporting,

underlies their regulation.

This view is further corroborated by the way in which minority interest is required to be reported.

In Article 45.2 of the Law of Mercantile Adaptation to the European Community Standard (Ley

19/89) it states that the consolidated balance sheet will show separately the amount corresponding

to the shareholders or members outside the group, which will appear in a specific section suitably

designated. Thus, as established in the Standards for the Formulation of Consolidated Annual

Accounts (RD 1815/1991), the minority interests are located separately from the capital and

reserves of the group and as an intermediate category between these and the long-term liabilities.

The Spanish regulations are in line with those advocated by the IASC. In International

Accounting Standard (IAS) 27 the entity theory definition of a group is adopted in that the

concept of control is not associated solely with a majority holding in the subsidiary’s equity.

Control can also exist when the parent owns one half or less of the voting power in the subsidiary

(IASC, 1994, paragraph 12). However, in IAS 22, the treatment of minority interest reflects the

parent company theory, given that minority interest does not include the part of goodwill which

does not belong to the parent company ((IASC, 1993, paragraph 40). Also, minority interest

should be presented in the consolidated balance sheet separately from liabilities and the parent

shareholders' equity, and the minority interest claim on earnings should also be separately

presented. (IASC, 1993, paragraph 26)

Assumptions regarding, the nature of the minority shareholders or minority interests are critical

when deciding on the appropriate method of consolidation. The Financial Economics literature

includes a number of papers which address the nature of minority interest from a theoretical

11

viewpoint. In summarising these, Beckman (1995), identifies two fundamental types of group

restructuring transactions which result in minority interest: transactions by which one company

takes over another either with the initial intention of obtaining 100% participation in the target

company, or simply with the intention of obtaining sufficient to gain control of the target

company; or, transactions by which a parent company sells part of its participation in a wholly

owned subsidiary company.

In the first of the cases, when the initial bid is for 100% ownership, it may happen that the

optimum decision of any particular individual shareholder could be not to accept the offer, but

rather free-ride on the increased profits generated by the new management team from the

combination of businesses (Grossman and Hart, 1980). In this situation minority interests can be

characterised as members who maintain their participation in the consolidated entity after the

take-over; therefore they are investors who share in the profits derived from the acquisition and

new management . This is consistent with the economic unit concept.

If the initial bid is only intended to gain a controlling share of the target then there exists an

optimal proportion of target company shares to be acquired. In this case, there exists empirical

evidence (Bradley, 1980) indicating that the value of the shares acquired, once the offer has been

completed, is not from their proportional claims to the net cash flows of the target firm but from

the control over all the resources that the shares confer. As the entity approach provides a more

faithful representation of the whole of the assets and liabilities under the control of the parent

company, it may be inferred that accounting information prepared on the basis of the entity theory

will show a higher correlation with the market value of the parent company than that which might

be derived from the parent company theory.

Sale transactions which lead to the creation of minority interests often occur when the subsidiary

company has high growth potential and is in need of external financing (Schipper and Smith,

1986). The parent company will be prepared to accept the costs of an outsider having a minority

interest since the parent may obtain sufficient profits from the availability of separate financing to

the subsidiary. The work of Schipper and Smith (1986) and Nanda (1991) indicate that equity

12

carveouts, as they call them, are simple equivalents to parent company equity issues.

Thus in all three cases identified in the literature, minority interest can be seen to arise within a an

entity concept. The minority interest is clearly an equity substitute and thus economic theory all

points towards the entity concept of consolidation.

In summary, two fundamental arguments may be offered in support of the entity theory of

consolidation. Firstly, the minority interests are seen as shareholders of the whole group or

consolidated entity. Secondly, the control of 100% of the shares of the subsidiary company is in

the hands of a single management team (that of the parent company), even though, as a

consequence of the transactions, minority interests arise. For this reason, 100% of the market

value of the subsidiary should be included in the consolidation. In addition if stock prices reflect

the present value of the future dividend stream and group assets and liabilities, under the control

and management of the parent company, all contribute to the generation of group earnings and,

therefore, of future dividends, then group reporting under the entity concept will be more

informative to the shareholders of the parent company.

3. Research Method, Hypotheses and Sample

The value relevance of the different forms of consolidated data is evaluated by means of a

valuation model with share price (firm value) being modelled as a function of purely accounting

variables, namely earnings and book value of equity. The formulation of share price as a function

of the present book value plus the discounted value of future abnormal earnings has been known

for some time. This valuation model was originally proposed by Preinreich in 1938 and later used

by Edwards and Bell (1961), Edey (1962) and Peasnell (1982), but has been reintroduced and

popularised in the accounting literature by the work of Ohlson (1995) and Feltham and Ohlson

(1995).

The valuation model takes as its starting pount the dividend discount model.

)r+(1

]d[ E = P ii+tt

=1it ∑∞ (1)

13

Where:

Pt = stock price at time t

dt = dividends at time t

r = discount rate, and

Et[] = expectations operand at time t

Application of the clean surplus accounting assumption, which can be expressed as:

tt1tt debvbv −=− − (2)

Where:

et represents the total value of ordinary shareholder earnings for period t

bvt represents the value of ordinary shareholder equity, at the end of accounting period t.

all other variables are as previously defined

leads to:

)r+(1

]e[ E + bv = P i

ai+tt

1=itt ∑∞ (3)

Where:

ea t = abnormal earnings at time t = (et – rbvt-1)

all other variables are as previously defined

Simplifying, and yet realistic, assumptions can then be placed on the expectations operators which

lead to the following formulation (Rees, 1997).

t2t10t ebvP α+α+α= (4)

As has been well documented in previous research using similar models (e.g. Collins et.al. 1997;

Rees, 1997; Garrod and Rees, 1998) this particular model has certain advantages over traditional

returns models for exploring issues related to value relevance. For our purposes it has the

additional property of lending itself to decomposition so that the two main explanatory variables,

earnings and book value of equity, can be subdivided into component parts. We take advantage of

this in this study by decomposing earnings and book value of equity into parent company,

consolidated, reported minority interest and full entity method minority interest sections. By

14

adopting this approach, the nature of least squares regressions means that the t-statistic on each of

the component parts of earnings or book value of equity provides a direct measure of the

incremental value of each component to share price (Jennings, 1990).

We decompose equation (4) by utilising the following set of variables (valued at the year-end):

Pit: Market price per share of the ordinary shares of parent company i at time t.

eit: Earnings per share of parent company i at time t.

bvit: Book value of equity per share of parent company i at time t.

epit: Consolidated earnings per share under the parent company theory for group i attime t

bvpit: Consolidated book value per share under the parent company theory for group i attime t.

erit: Consolidated earnings per share incorporating reported minority interest share ofearnings for group i at time t.

bvrit: Consolidated book value per share incorporating reported minority interest forgroup i at time t.

eeit: Consolidated earnings per share under the entity theory for group i at time t.

bveit: Consolidated book value per share under the entity theory for group i at time t.

Parent company’s earnings (eit) and parent company’s book value (bvit) are obtained from the

individual accounts of the parent company.

Consolidated earnings under the parent company theory (epit) corresponds to the profit

attributable to the parent company in the consolidated profit and loss account. Consolidated

earnings incorporating reported minority interest share of earnings (erit) is calculated by adding the

profit attributable to the minority interests, as recorded in the consolidated profit and loss

account, to epit. The consolidated book value of equity under the parent company theory (bvpit)

corresponds to the equity in the consolidated balance sheet. The, consolidated book value of

equity incorporating reported minority interest (bvrit) is calculated by adding the minority

interests, as recorded in the consolidated balance sheet, to bvpit.

The values for eeit and bveit are estimated in the following way. Both minority interest and

goodwill need to be increased by the proportion of subsidiaries owned by external parties. We

15

estimate this proportion in two separate ways:

X1 = (BVC – BVP) + (IP – IC) . (Method 1)

MI + (BVC – BVP) + (IP – IC)

X2 = IP – GWC . (Method 2)

MI + IP - GWC

Where:

MI = reported minority interest in the consolidated accounts

BVC = the book value of equity in the consolidated accounts

BVP = the book value of equity in the parent company accounts

IP = the value of investments in associates and subsidiaries reported in the `parent company accounts

IC = the value of investments in associates reported in the consolidated accounts

GWC = the goodwill recorded in the consolidated accounts

Two methods are used as both contain approximations due to the nature of Spanish consolidation

methods. Under these, in the consolidated accounts, the investment in associated companies is

split between book value investment and goodwill, the latter being assimilated within the goodwill

figure relating to the subsidiaries. Thus in method 1, IC is an underestimate of the parent’s

investment in subsidiaries and X1 will be an overestimate of the average proportion of subsidiaries

owned by the parent. This problem does not occur in method 2 as IC is not used, but this second

method suffers from the fact that the reported goodwill figure in the accounts has been amortised

to some extent whilst the theoretically correct figure to be used in the formula is the goodwill at

the date of purchase. Thus goodwill will be underestimated and X2 will also be an overestimate of

the average proportion of subsidiaries owned by the parent1. As neither of these problems can be

resolved, given the disclosure in the annual accounts, data generated using both methods are

tested in the empirical analysis.

The impact of moving to the full entity method is to increase goodwill, and minority interest,

1 This is only the case if goodwill is positive. The reverse is true if goodwill is negative. This occurs in a number of

16

which will result in an increase in the amortisation expense of goodwill2 which will, in turn, reduce

group earnings by ((1-X)/X) of the existing goodwill amortisation figure. The minority interest

figure will increase by (MI/X – the fall in consolidated earnings).

Taking equation 4 and replacing it by the four components of earnings and book value of equity

we generate our testable model.

The null form hypotheses that we are testing are:

H01: There is no incremental value relevant information contained in the data disclosed in the

consolidated accounts over those disclosed in the parent company accounts.

H02: There is no incremental value relevant information contained in the minority interest figure

disclosed under current regulation.

H03: There is no incremental value relevant information contained in the entity form of

consolidated accounts over the data disclosed under current regulations.

If null hypothesis H01 is to be rejected then we should identify a significant coefficient on any or

all of α2, α3, α4, α6, α7 or α8.

If null hypothesis H02 is to be rejected then we should identify a significant coefficient on either α3

or α7.

If null hypothesis H03 is to be rejected then we should identify a significant coefficient on either α4

or α8.

The sample used corresponds to non-financial companies quoted on the Madrid Stock Exchange.

The information is obtained from the “Auditorías de Sociedades Emisoras” Data Base compiled

by the Comisión Nacional del Mercado de Valores and Extel Financial Company Analysis Service.

We have taken the accounting information from the first, and the stock market information from

the sample companies.2 In general, only positive goodwill can be amortised.

.......(5).................... )ee()ee()ee(e

)bvbv()bvbv()bvbv(bvP

eitrit8ritfit7fitit6it5

eitrit4ritfit3fitit2it10it

−α+−α+−α+α

+−α+−α+−α+α+α=

17

the second, specifically, market capitalisation, the number of shares issued, price of the individual

share, and dividends paid.

The data are collected from 1991, the year when the publication of consolidated accounts was

made obligatory in Spain, up to and including 1997. After excluding cases with missing data, 506

useable cases remain. As is normal in these types of study outliers, defined as the cases lying in the

top and bottom 1% of the three basic variables (price, earnings and book value of equity), are

excluded. This results in a final sample of 474 observations. The time distribution of the final

sample is: 76 in financial year 1991; 73 in 1992; 73 in 1993; 68 1994; 65 in 1995, 59 in 1996 and

60 in 1997.

4 Results

Descriptive statistics and a correlation matrix for the model variables are to be found in Tables 1

and 2 respectively. The results of the regression for the pooled sample are to be found in Table 3.

The results when X1 is used as the proxy for the average proportion of subsidiaries owned by the

parent are recorded in Panel A and those using X2 in Panel B.

The first thing to note is that the coefficients on the book value of equity (bvit) and earnings (eit)

of the parent company accounts are significant and in line with previous studies, using both

Spanish (Giner and Rees, 1999) and other data (e.g. Collins et.al., 1997; Garrod and Rees, 1998;

Harris et al., 1997). This provides additional confidence in the results and the statistical

significance of the models as a whole (F-test value of 101.87 using X1 and of 101.65 using X2,

both significant at the 0.000001 level)

In relation to the first of the hypotheses, it can be observed that, the coefficients on both (bvpit –

bvit) and (epit – eit) are positive (1.3416 and 5.7425 respectively when using X1 and 1.3947 and

5.7846 respectively when using X2 ) and significant (book value at the 0.0001 level and earnings

at the 0.05 level in both regressions) in both models, indicating that the consolidated disclosure is

more value relevant than the parent company disclosure alone. The t-statistic on the book value

variable is almost double that on the earnings variable, consistent with the view that the

18

revaluation of assets and recognition of goodwill provide a better proxy for future earnings and

cash flow than the simple incorporation of the parent’s share of subsidiary and associates retained

earnings

The finding that the consolidated information has more value relevance than that of the individual

parent company is consistent with the results of the study by Larrán and Rees (1999). Interviews

with Spanish financial analysts; revealed that valuations of the parent company were based on

group rather than the individual annual accounts, unless the parent company’s activities were

highly differentiated from the rest of the group’s.

The second hypothesis relates to whether the minority interest portion of book value of equity or

group earnings is value relevant. When using the X1 proxy for the average proportion of

subsidiaries owned by the parent, we see that earnings variable is insignificant (t-statistics of

0.8082) whilst the book value variable is just significant at the 5% level (t-statistics of 2.0112).

However, when the X2 proxy for the average proportion of subsidiaries owned by the parent is

used then neither of the variables is significant (t-statistics of –0.8656 and 0.8506). Further

investigation of the results using X1 uncovered the fact that , there are two extreme values (more

than 12 standard deviations greater than the mean) of this variable and if these two observations

are eliminated then the t-statistic on the book value variable falls to 1.8218 which is significant

only at the 6.9% level. Thus we discover no support for the value relevance of the minority

interest component of earnings and only very weak support for the value relevance of the minority

interest component of book value of equity.

The final hypothesis relates to the use of full equity theory values for minority interest in both

book value and earnings. Neither of the variables ((bvet – bvrt) and (eet – ert)), using either X1 or X2

to generate the revised earnings and book value estimates, are significant. This is consistent with

the full entity theory minority interest share of earnings and net total assets providing no

additional valuation relevant information over that currently disclosed.

In summary, our results indicate the following answers to our two research questions. Firstly,

19

consolidated reports do provide additional value relevant information to that provided in parent

company accounts. In other words, the market does value the retained earnings of subsidiaries

and associated companies which are revealed in the consolidated accounts. Secondly, whilst we

have identified some weak evidence in support of the value relevance of the reported minority

interest share of net total assets, we find none for the reported minority interest in earnings and

none for either measure if calculated using the full equity theory of consolidation. The

insignificance of the minority interest share of earnings is understandable as the parent company

shareholders have no claim over these earnings. However, the weak support we identify for the

reported minority interest share of net total assets is consistent with parent company shareholders

being aware of this alternative from of financing for the net total assets over which they have

control and taking it into account (to a very limited extent) when forecasting future earnings.

6. CONCLUDING REMARKS

The issue of whether to report consolidated accounts and, if so, how is one of considerable

topical interest. In general there is much regulatory support for requiring consolidated disclosure

but there is less unanimity on the appropriate method of reporting to adopt. The main area of

contention is the classification of minority interest: should it be reported with equity, with debt or

in a separate section somewhere between the two? In both the USA and the UK, the minority

interest is considered as part of the capital of the group. Recent pronouncements from the FASB

indicate that the US regulators wish to move ever closer to the entity concept of consolidation.

There has been a variety of responses to this issue from EU countries following the enactment of

the 7th. EU Company Law Directive, which requires consolidated disclosure. Spain is one of those

countries which require disclosure which underscores the distinctiveness of minority interest by

reporting it separately to the capital of the group. The results we report in this paper support this

view.

We find overwhelming evidence that the consolidated accounts provide incremental value relevant

information over the disclosure of the parent company alone. However, we find no (or, at best,

sparse) evidence that minority interest as currently reported nor as would be reported under the

full equity method of consolidation provide any additional value relevant information. These

20

results raise questions about the suitability of moving even closer to a full equity definition of a

reporting group, as it would appear is the intention of the FASB. Whilst the concept of control

may be an attractive one from a theoretical perspective of the group, the results reported above

are consistent with investor usefulness not being well served by such a move.

21

REFERENCES

A.A.A. Financial Accounting Standards Committee, 1994, “Response to the FASB DiscussionMemorandum ‘Consolidation Policy and Procedures’”, Accounting Horizons, Vol.8 No.2, pp.120-125.

A.A.A. Financial Accounting Standards Committee, 1996, “Response to the FASB ExposureDraft, ‘Proposed Statement of Financial Accounting Standards- Consolidated FinancialStatements: Policy and Procedures’”, Accounting Horizons, Vol.10 No.3, pp. 182-185.

Abad, C., B. Garcia-Barbella, J. Laffarga, M. Larran and J. Piñero, 1977, “Análisis de lasDiferencias entre la Información Contable de las Cuentas Individuales y las Consolidadas”,Actualidad Financiera , November, pp. 27-47.

Accounting Standards Board, 1992, “Financial Reporting Standard 2: ‘Accounting for SubsidiaryUndertakings’”.

AICPA, 1959, Accounting Research Bulletin No. 51: Consolidated Financial Statements, NewYork: AICPA.

Ball, R. and P. Brown, 1968, “An Empirical Evaluation of Accounting Numbers”, Journal ofAccounting Research, Autumn, pp. 159-178.

Beckman, J.K., 1995, “The Economic Unit Approach to Consolidated Financial Statements:Support from the Financial Economics Literature”, Journal of Accounting Literature , No. 14,pp. 1-23.

Blasco, M.P.,1995, “El análisis de los Estados Contables Consolidados. Una AproximaciónConceptual y Empírica”, Doctoral Thesis, Universidad de Zaragoza.

Bradley, M., 1980, “Interfirm Tender Offers and the Market for Corporate Control”, Journal ofBusiness , October, pp. 345-376.

Cea, J.L.,1992, “Algunas Anotaciones sobre la Imagen Fiel y sobre el Concepto de CuentasAnuales Consolidadas de los Grupos de Sociedades”, Revista de Estudios Financieros , No. 108,pp. 65-81.

Chambers, R., 1968, Consolidated Statements are not Really Necessary, Australian Accountant,February, p. 90.

Collins, D., E. Maydew and I. Weiss, 1997, Changes in the Value-Relevance of Earnings andBook Values over the Past Forty Years, Journal of Accounting and Economics, Vol. 24, pp. 39-67.

Darrough, M.N. and T.S. Harris, 1991, Do Management Forecasts of Earnings Affect Stock

22

Prices in Japan?,in Ziemba W.T., W. Bailey and Y. Hamao (eds), Japanese Financial MarketsResearch, Amsterdam North Holland, pp. 197-229.

Edey, H., 1962, “Business valuation, Goodwill and the Super-profit Method”, in Studies inAccounting Theory, W. Baxter and S. Davidson (eds.), London:, Sweet and Maxwell.

Edwards, J. and P. Bell, 1961, The Theory and Measurement of Business Income., University ofCalifornia Press.

Emmanuel, C.R. and N. Garrod, 1992, Segment Reporting: International Issues and Evidence;Prentice-Hall, Hemel Hempstead.

Financial Accounting Standards Board, 1991, “Discussion Memorandum: ‘Consolidation Policyand Procedures’”.

Financial Accounting Standards Board, 1995, Exposure Draft: “Proposed Statement of FinancialAccounting Standards - Consolidated Financial Statements: Policy and Procedures” .

Financial Accounting Standards Board, 1999, Exposure Draft (revised): “Proposed Statement ofFinancial Accounting Standards - Consolidated Financial Statements: Purpose and Policy” .

Feltham, G. and J. Ohlson, 1995, “Valuation and Clean Surplus Accounting for Operating andFinancial Activities”, Contemporary Accounting Research , Vol. 11 No. 2, pp. 689-731.

Garrod, N. and W. Rees, 1998, “International Diversification and Firm Value”, Journal ofBusiness Finance and Accounting, Vol. 25, Nos. 9 and 10, pp. 1255 -1281.

Giner, B. and W. Rees, 1999, “A Valuation Based Analysis of the Accounting Reforms”, Journalof Management and Governance, Vol. 3, No. 1, pp. 1-18.

Grossman, S. and O. Hart, 1980, Takeover bids, the Free-rider Problem and the Theory of theCorporation, Bell Journal of Economics, Spring, p. 42-64.

Harris, T.S., M.H. Lang and H.P. Moller,1997, “Unconsolidated versus Consolidated accountingMeasures: Empirical Evidence from Listed German Companies”, 20th Annual Congress of theEuropean Accounting Association.

International Accounting Standards Committee, 1993, IAS 22(Revised). Business Combinations.

International Accounting Standards Committee, 1994, IAS 27(Revised). Consolidated FinancialStatements and Accounting for Investments in Subsidiaries.

Jennings, R., 1990, A Note on Interpreting Incremental Information Content, The AccountingReview, October, p. 925-932.

23

Lambert, C. and I. Zimmer, 1996, “The Effect of New Consolidation Accounting Rules onCorporate Structures and Financial Statements”, 19th Annual Congress of the EuropeanAccounting Association

Larrán, M., 1996, “Análisis Financiero de grupos de empresas: un estudio empírico del período91-94”, ponencia presentada al VII Encuentro de Profesores Universitarios de Contabilidad,ASEPUC.

Larrán, M. and W. Rees, 1999, “Tecnicas, Recursos Informativos y Practicas Seguidas por losAnalistas Financieros: Un Estudio Empirico”, Monografia de Investigacion, Instituto Español deAnelistas Financeros, Madrid.

Ley 19/89 de 25 de Julio, de Reforma Parcial y adaptación de la Legislación Mercantil a lasDirectivas de la CEE en materia de Sociedades (BOE 27/7/89).

Moonitz, M., 1951, The Entity Theory of Consolidated Statements, Foundation Press.

Nanda, V. 1991, On the Good News in Equity Carve-Outs, The Journal of Finance, December,pp. 1717-1737.

Ohlson, J., “Earnings, Book Values and Dividends in Equity Valuation”, ContemporaryAccounting Research (1995), Vol. 11 No.2, pp. 661-87.

Parker, R. H. 1984, Explaining National Differences in Consolidated Accounts, In Lee, T.A. andR.H. Parker (eds), The Evolution of Corporate Financial Reporting, Garland Press.

Peasnell, K., 1982, “Some Formal Connections between Economic Values and Yields andAccounting Numbers”, Journal of Business Finance and Accounting , Vol. 9, No.1, pp. 361-81.

Pellens, B. and V. Linnhoff, 1993, Financial Analysis of Group Accounts in Germany: AnEmpirical Study, S. J. Gray., A.G. Genenberg and P.D. Gordon, pp. 110-120.

Rees, W., 1997, “The Impact of Dividends, Debt and Investment on Valuation Models”, Journalof Business, Finance and Accounting, Vol. 24, No. (7)&(8), pp. 1111-1140.

Schipper, K. and A. Smith, 1986, A Comparison of Equity Carveouts and Seasoned EquityOfferings: Share price Effects and Corporate Restructuring, Journal of Financial Economics,Vol. 15, pp. 153-186.

Table 1 Summary statistics for variables used in the regression model

bvit bvpit - bvit bvrit - bvpit bveit – bvrit

(using X1)bveit – bvrit

(using X2)eit epit - eit erit - epit eeit – erit

(using X1)eeit – erit

(using X2)PRICE

Mean 2.1620 0.2157 0.2020 0.2059 0.1422 0.1429 0.0332 0.0136 -0.0038 -0.0028 3.222219 Median 1.7046 0.0385 0.0296 0.0007 0.0007 0.1089 0.0114 0.0001 -0.0000 -0.0000 2.300000 Maximum 13.3926 3.7294 4.5642 19.3865 6.2919 1.0724 1.2333 0.4894 0.0215 0.0397 18.90000 Minimum -0.1343 -1.0949 0.0000 -0.0085 -0.0986 -1.0837 -0.3172 -0.1758 -0.2620 -0.1230 0.100000 Std. Dev. 1.6415 0.5301 0.5792 1.3160 0.6777 0.2682 0.1090 0.0579 0.0189 0.0123 3.137025 Skewness 1.4955 2.6552 5.1039 11.6172 6.7947 0.0087 3.7849 4.7828 -8.3952 -5.5919 1.937494 Kurtosis 7.4926 12.5951 31.5481 155.8762 52.5916 6.0214 36.6185 33.9439 92.9312 43.7843 7.309077

Observations

474 474 474 474 474 474 474 474 474 474 474

Pit: Market price per share of the ordinary shares of parent company i at time t.eit: Earnings per share of parent company i at time t.bvit: Book value of equity per share of parent company i at time t.epit: Consolidated earnings per share under the parent company theory for group i at

time tbvpit: Consolidated book value per share under the parent company theory for group i at

time t.erit: Consolidated earnings per share incorporating reported minority interest share of

earnings for group i at time t.bvrit: Consolidated book value per share incorporating reported minority interest for

group i at time t.eeit: Consolidated earnings per share under the entity theory for group i at time t.bveit: Consolidated book value per share under the entity theory for group i at time t.

1

Table 2 Correlation matrix for variables used in the regression model

bvit bvpit - bvit bvrit - bvpit bveit – bvrit

(using X1)bveit – bvrit

(using X2)eit epit - eit erit - epit eeit – erit

(using X1)eeit – erit

(using X2)PRICE

bvit 1.0000

bvpit - bvit 0.2722 1.0000

bvrit - bvpit 0.3229 0.4522 1.0000

bveit – bvrit

(using X1) 0.2205 0.2723 0.5988 1.0000

bveit – bvrit

(using X2) 0.2637 0.3261 0.9000 0.6304 1.0000

eit 0.5993 0.2544 0.2075 0.1455 0.1267 1.0000

epit - eit 0.1706 0.5298 0.2840 0.1015 0.2365 0.1136 1.0000

erit - epit 0.2570 0.3953 0.8027 0.3834 0.7479 0.2275 0.4061 1.0000

eeit – erit

(using X1)-0.1871 -0.1757 -0.5617 -0.3525 -0.3871 -0.1310 -0.0584 -0.3950 1.0000

eeit – erit

(using X2)-0.2164 -0.2212 -0.6893 -0.3041 -0.5458 -0.1270 -0.1261 -0.5371 0.9215 1.0000

PRICE 0.5825 0.5425 0.4139 0.2588 0.3313 0.6394 0.4241 0.3800 -0.1718 -0.2295 1.0000

Pit: Market price per share of the ordinary shares of parent company i at time t.eit: Earnings per share of parent company i at time t.bvit: Book value of equity per share of parent company i at time t.epit: Consolidated earnings per share under the parent company theory for group i at time tbvpit: Consolidated book value per share under the parent company theory for group i at time t.erit: Consolidated earnings per share incorporating reported minority interest share of earnings for group i at time t.bvrit: Consolidated book value per share incorporating reported minority interest for group i at time t.eeit: Consolidated earnings per share under the entity theory for group i at time t.bveit: Consolidated book value per share under the entity theory for group i at time t.

3

Table 3 Results of regression

using 474 observations taken from Spanish companies for the years 1991-1997

Model using X1 as the proxy for theaverage proportion of subsidiaries

owned by the parent

Model using X2 as the proxy for theaverage proportion of subsidiaries

owned by the parentVariable Coefficient t-Statistic Prob. Coefficient t-Statistic Prob.

InterceptC 1.0885 5.9272 0.0000 1.1002 5.9179 0.0000bvit 0.3838 3.2433 0.0013 0.3817 3.2148 0.0014

bvpit - bvit 1.3416 4.4512 0.0000 1.3947 4.4109 0.0000bvrit - bvpit 0.9540 2.0112 0.0449 0.5373 0.8506 0.3955bveit – bvrit 0.0171 0.3498 0.7267 0.3166 0.7958 0.4265

eit 4.9352 6.2949 0.0000 4.9654 6.2685 0.0000epit - eit 5.7425 2.4988 0.0128 5.7846 2.5066 0.0125erit - epit -3.4168 -0.8082 0.4194 -3.5192 -0.8656 0.3872eeit – erit 8.1407 5.0716 0.1091 4.0697 0.3715 0.7104

Adjusted R2 0.6305 0.6300F-statistic 101.8709 0.0000 101.6528 0.0000

Figures in bold, italic and underlined at significant at the 0.0001 level.Figures in bold and italic are significant at the 0.01 level.Figures in italic are significant at the 0.05 level.

Pit: Market price per share of the ordinary shares of parent company i at time t.eit: Earnings per share of parent company i at time t.bvit: Book value of equity per share of parent company i at time t.epit: Consolidated earnings per share under the parent company theory for group i attime tbvpit: Consolidated book value per share under the parent company theory for group i attime t.erit: Consolidated earnings per share incorporating reported minority interest share ofearnings for group i at time t.bvrit: Consolidated book value per share incorporating reported minority interest forgroup i at time t.eeit: Consolidated earnings per share under the entity theory for group i at time t.bveit: Consolidated book value per share under the entity theory for group i at time t

)ee()ee()ee(e

)bvbv()bvbv()bvbv(bvP

eitrit8ritpit7pitit6it5

eitrit4ritpit3pitit2it10it

−α+−α+−α+α

+−α+−α+−α+α+α=

4