An exploration into the relation between corporate environmental reputation and financial performance:
 
UK evidence
 
Abstract
 
Nowadays, senior managers face the pressure to allocate corporate resources, especially on the social environmental issues. A large number of empirical studies have investigated the relationship between corporate environmental performance (CEP) and corporate financial performance (CFP). However, due to inconsistent results achieved, there is not a clear answer for the debate that whether strategic managers should add invests on corporate social responsibility or not. Methodological issues in these studies contribute to the mixed results. This study uses an improved methodology based on UK evidence, including better corporate environmental performance measure, to examine the relation between CEP and CFP. The results show that better prior financial performance leads to better environmental performance, which supports slack resources theory, while higher level of corporate environmental performance improves subsequent financial performance, which is consistent with good management theory. And accounting-based measures are found to be better than stock market-based measures to assess financial performance in this kind of research. Therefore, this dissertation gives useful directions for senior managers and further researches. Additional investment should be undertaken to achieve a higher level of financial performance.
 
Table of contents
 
1. Introduction…………………………………………………………………..…5-6
 
2. Literature Review
 
2.1 Samples Selection………………………………………………………….…...7-8#p#分頁標題#e#
 
2.2 Different environmental performance measures………………………………9-11
 
2.3 Different corporate financial performance measures…………………..…….11-14
 
2.4 Different control variables……………………………………………………14-15
 
2.5 Applied different theories…………………………………………………….15-16
 
2.6 The direction of causality………………………………………………………...16
 
2.7 Different analysis methods……………………………………………………….17
 
2.8 Possible research errors……………………………………………………….17-18
 
3. Theoretical Framework
 
3.1 Slack Resources Theory …………………………………………………………26
 
3.2 Good Management Theory …………………………………………………..26-27
 
3.3 Hypothesis…………………………………………………………………….27-28
 
4. Methodology
 
4.1 Sample……………………………………………………………………………29
 
4.2 Corporate environmental reputation measure………………………………...29-30
 
4.3 Corporate financial performance measures…………………………………..31-32
 
4.4 Control variables ……………………………………………………………..32-34
 
4.5 Model specification…………………………………………………………...35-36#p#分頁標題#e#
 
5. Analysis of Findings
 
5.1 Data Analysis…….…………………………………………………………........37
 
5.2 Results Analysis ….……………………………………………….……........38-45
 
5.3 Implications for Management…………………………………….…….........45-47
 
5.4 Limitations …….………………………………………………………….....48-51
 
5.5 Implications for further research…….………………………………………52-53
 
6. Conclusion……………………………………………………………………….54
 
7. Bibliography…………………………………………….………..………….55-59
 
Appendices
 
Appendix 1a………………………………………………………………………60-61
 
Appendix 1b………………………………………………………………………62-63
 
Appendix 2a………………………………………………………………………64-65
 
Appendix 2b………………………………………………………………………65-67
 
List of Tables
 
Table 1……………………………………………………………………………19-24
 
Table 2……………………………………………………………………………….36#p#分頁標題#e#
 
Table 3……………………………………………………………………………….38
 
Table 4……………………………………………………………………………….40
 
Table 5……………………………………………………………………………….41
 
Table 6……………………………………………………………………………….43
 
1. Introduction
 
The concept ‘corporate social responsibility’ (CSR) has been raised in the twentieth century. Nowadays, corporate social responsibility pays an important role in corporate management. It is essential for strategic managers to decide how to allocate the resources in order to maximise the profit. Therefore, managers face consistent pressure on scarce corporate resources allocation. (Waddock and Graves, 1997) Prahalad and Hamel (1994) pointed out that the pressure of resources allocation tends to come from the social issues of management rather than traditional strategic management. This means strategic decisions for resources allocation become more complex, because managers should consider both financial and social environmental outcome when making a decision. (Waddock and Graves, 1997) Mcwilliams and Siegel (2000) asserted that key stakeholders, including ‘customers, employees, suppliers, community groups, governments, and some shareholders’, encourage firms to invest more on programs related to corporate social responsibility. On the contrast, some studies (eg. Aupperle, Carroll and Hatfield, 1985; Ullmann, 1985) argued that CSR investment would incur a significant cost, which could reduce the profit.
 
There are a large number of literatures investigated the relation between corporate environmental performance and financial performance in the last forty years (Orlitzky, Schmidt and Rynes, 2003), due to the controversy discussed above. However, the results are inconsistent, for example, Salama (2005) suggested there was a significant positive relationship between CEP and CFP; Ingram and Frazier (1983) asserted CEP is negatively related to CFP; Barnett and Salomon (2006) argued there should be a curvilinear relationship; and Surroca, Tribo and Waddock (2010) found on direct link between CEP and CFP. Several reasons might contribute to mixed results, such as different samples, different CEP and CFP measures, and different control variables. Thus, it is difficult for senior managers to decision whether additional investment should be undertaken on CSR or not, and how much should be invested, due to mixed results in this research area.#p#分頁標題#e#
 
Therefore, it is meaningful and worthy to further examine the relation between corporate environmental performance and financial performance with more reasonable methodology, in order to achieve more valuable results. The key research question of this study is that whether there is a relationship between corporate environmental reputation (CER) and corporate financial performance (CFP), if so, in what direction the causation runs.
 
This study investigates the CER-CFP link based on UK evidence. Thus, 163 British companies are selected as the sample, while MAC survey is chosen as the CER measure. MAC survey is the ‘Community and Environmental Responsibility’ ratings of 239 British companies. (Salama, 2005) MAC survey is a valid proxy for CER, as it reflects different aspects of CER. In order to compare accounting-based measures and stock market-based measures, both of them are selected to assess the financial performance. Ordinary Least Squares (OLS) regression is the main analysis method in this study. The results should be able to answer the research question satisfactorily, while finding out which financial performance measure is better in the CEP-CFP relationship research. The results achieved in this study can give a direction for strategic managers when making decisions on resources allocation. Moreover, results also are an attempt to give directions for further researchers in this area.
 
In the following sections, details for how this research carries out and results achieved will be fully presented. In the Literature Review section, main literatures in this research area will be analysed and research directions for this study could be achieved; the theories applied in this study will be discussed in the Theoretical Framework Section; in the Methodology Section, details for how this research carries out will be presented; and in the Findings section, results achieved will be fully discussed. Finally, a conclusion of this study will be made.
 
2. Literature Review
 
Wood (1991) pointed out that there have been a large number of literatures that have investigated the relation between corporate environmental performance and financial performance. The contributions and limitations of these empirical research studies will be summarised in this section. The more recent literatures, which should have more valuation for further research in this area, get more coverage.
 
As mentioned earlier, the results of pervious researches for CSR-CFP relationship are inconsistent. (Orlitzky et al., 2003) Some studies discovered a positive relationship between environmental performance and financial performance, such as Waddock and Graves (1997), Stanwick and Stanwick (1998), Salama (2005). Some suggested a negative relationship, such as Ingram and Frazier (1983), Freedman and Jaggi (1982). Some asserted that there should be a curvilinear relationship, such as Barnett and Salomon (2006). The others argued that there is no relationship, such as Mcwilliams and Siegel (2000), Surroca, Tribo and Waddock (2010). To sum up, there are eight main reasons for mixed results: (1) Samples selection (2) Different corporate environmental performance measures; (3) Different corporate financial performance measures; (4) Different control variables; (5) Applied different theories; (6) The direction of causality; (7) Different analysis methods; (8) Possible research errors.#p#分頁標題#e#
 
2.1 Samples Selection
 
The first issue addressed in this part is the industries examined in the studies. According to Griffin and Mahon (1997), majority of 52 literatures reviewed chose to take the samples with multiple industries, while only ‘Bromiley and Marcus (1989), Davidson, Chandy and Cross, (1987), Rockness, Schlachter and Rockness. (1986)’ three journals focused on only one industry.
 
Both samples from multiple industries and samples within one industry have their own merits and flaws. Griffin and Mahon (1997) chose to focus on only chemical industry. Although they did not work out a clear answer for the relationship question, their work gives a useful direction for further research. However, taking only one industry into account does not consider all the factors. In another words, it has a limitation, namely it reflects only one-industry characters.
 
There are also some disadvantages for samples from various industries. Carroll (1979) pointed out that ‘the issues change and they differ for different industries’. For example, issues, such as governmental regulations, public visibility, consumer-oriented nature of companies, (Arlow and Gannon, 1982) differ for multiple industries, due to the industry specialisation of social interests. (Griffin and Mahon, 1997) Particularly, accounting-based measures are not suitable to assess the financial performance within the multiple industries (Davidson and Worrell, 1990), unless using some control variables. If samples are taken from multiple industries, then the industry classification should be one of control variables, so that the flaws could be overcome.
 
Different studies select not only different industry choices, but also the number of samples. Earlier researches use small population, such as Moskowitz (1972) compared only 14 firms and Bragdon and Martin (1972) examined 17 firms; while later studies chose to use large population, such as Surroca, Tribo and Waddock (2010) selected 599 companies. The studies, which used larger population, have more possibility to lead to valuable results.
 
These literatures stem from different countries; therefore, the samples chosen might have the regional characters. For example, Robert (1992) found weak negative relationship between CER and risk for US data, while Hasseldine, Salama and Toms (2005) achieved a result of significant negative association for UK data. In order to achieve more reasonable results, Surroca, Tribo and Waddock (2010) chose to test the companies from 28 countries.
 
2.2 Different environmental performance measures
 
When the concept of corporate environmental performance came out, it was vague; therefore, it was a bit difficult to measure. (Pava and Krausz, 1996) However, corporate social responsibility (CSR) was defined as ‘policies or actions which identify a company as being concerned with society-related issues’ in the empirical studies. (Roberts, 1992)#p#分頁標題#e#
 
Ullmann (1985) pointed out that the CEP measures chosen were quite different, which leads to the inconsistent results. Meanwhile, Aupperle, Carroll and Hatfield (1985) also suggested that some indexes of CER were questionable and it was difficult to develop valid measures.
 
There are three main methods to measure CER in the previous studies, namely ‘expert evaluations of corporate policies’, analysis of relative corporate documents, (McGuire, Sundgren and Schneeweis, 1988) and using a proxy measure. Each method has its own advantages and disadvantages.
 
Now I take ‘expert evaluations of corporate policies’ for example. McGuire, Sundgren and Schneeweis (1988), Fombrun and Shanley (1990) and Stanwick and Stanwick (1998) used the Fortune’s Corporate Reputation Index, which is a popular CER measure in the research studies. And they all found a positive relationship between CER and CFP. Though McGuire, Sundgren and Schneeweis (1988) was not sure that whether Fortune data is valid or not, they suggested that Fortune Index is different from the other measures, as it could be affected by the biases of the evaluators. Moreover, Fryxell and Wang (1994) argued that Corporate Reputation Index could not represent the firm’s environmental commitment well, as it is weighted based on the financial perspective of the company. Then there is definitely a strong positive relationship between Corporate Reputation Index and financial performance; therefore, the results might be less valuable. Later Stanwick and Stanwick (1998) further checked the validity of Fortune Index, and the results show that Fortune Index is valid indicator of corporate environmental performance.
 
Salama (2005) chose Corporate Reputation Index of ‘Britain’s Most Admired Companies’ (MAC), which follows the same methodology as Fortune Index, to measure CER and found a strong positive relationship. This index reflects the CER of 239 firms in forms of ‘Community and Environmental Responsibility’, and it was published in Management Today since 1994. (Salama, 2005)
 
In the recent studies, the ratings of corporate social performance provided by the firm Kinder, Lydenberg, Domini (KLD) is widely used as the indicator of CEP. KLD rates the CEP of companies according to a range of attributes, which are related to different stakeholder concerns. (Waddock and Graves, 1997) Apart from being a multidimensional assessment, KLD has many other merits, such as independent researchers, the same set of criteria, and both internal and external information of a wide range of companies considered. Therefore, KLD is a useful measurement of CEP, and many studies chose KLD as an indicator of CEP achieved the satisfied results, such as Waddock and Graves (1997); Hull and Rothenberg (2008); Godfrey, Merrill and Hansen (2009).
 
There are also other professional ratings of CEP used in the empirical studies, such as the Social Investment Forum (Barnett and Salomon, 2006); the Ethical Investment Research Service (EIRIS) (Brammer, Brooks and Pavelin, 2006); the Council of Concerned Businessmen Index (Sturdivant and Ginter, 1977).#p#分頁標題#e#
 
Some studies used content analysis of relative corporate documents as the measures of corporate social responsibility. For example, Bowman and Haire (1975) investigated how corporate social responsibility in annual reports affected financial performance, found a positive relationship. And Parket and Eibert (1975) assessed the environmental performance by ranking the willingness of corporation to respond to a related questionnaire.
 
However, these measures have some disadvantages, as they might mix the concept of social issues and corporate actions. (Aupperle, Carroll and Hatfield, 1985) McGuire, Sundgren and Schneeweis (1988) pointed out that due to the purposes of these corporate documents, such as public relations management, the information included in the document would be less useful to assess the actual corporate actions.
 
The third method of assessing the corporate environmental performance is using a proxy measure, such as pollution control or corporate philanthropy. Using pollution control has limitations, as it is only valid for certain industry. (Bragdon and Martin, 1972) For corporate philanthropy, researchers usually use information from ‘Corporate 500 Directory of Corporate Philanthropy’. (Griffin and Mahon, 1997)
 
Some studies chose to use multiple corporate social responsibility measures, so that the limitations of any one data could be overcome by the other measures, and they could compare which one is the best according to their results. (Griffin & Mahon, 1997) Griffin and Mahon (1997) selected four measures, namely ‘the Fortune reputation survey’, ‘the KLD Index’, ‘the Toxics Release Inventory’ and ‘corporate philanthropy’.
 
2.3 Different corporate financial performance measures
 
There are four main different financial performance criteria, namely stock-market-based, accounting-based, risk measures and other firm-specific characteristics. (Pava and Krausz, 1996) In the following, I will compare these four criteria based on the pervious research results.
 
Stock-market-based measures include Earning Per Share (EPS), market price, market value to book value and so on. McGuire, Sundgren and Schneeweis (1988) reviewed the literatures using Stock-market-based measures, and the results are conflict. Examining the same samples, Moskowitz (1972) found a positive relationship; but Vance (1975) showed a negative relationship; and Alexander & Buchholz (1978), which adjusted for risk, found little association, and achieved a result that both Moskowitz (1972) and Vance (1975) were invalid.
 
Later Pava and Krausz (1996) found a positive relationship between CER and the market value to book value ratios, which ‘relates the market capitalization of firm to the accounting valuations’. On the contrast, Brammer, Brooks and Pavelin (2006) achieved the result that CEP is negatively related to stock returns. Barnett and Salomon (2006) suggested that there is a curvilinear relationship between CEP and market value change with risk adjustment.#p#分頁標題#e#
 
There are some problems for using stock-market-based measures. (McGuire, Schneeweis and Hill, 1986) McGuire, Sundgren and Schneeweis (1988) pointed out that it is not sufficient enough to only concentrate on investors’ evaluations. In another word, stock-market-based measures do not take everything into account, so that they have flaws.
 
Measures like Return on Assets (ROA), Return on Equity (ROE), Return on Capital Employed (ROCE) and total assets are all accounting-based. Each accounting-based measure has its own merits and flaws. (Pava and Krausz, 1996) ROA has an advantage that it could distinguish financial actions from both operating and investing actions. ROE, which is related to ROA, is also a useful method to assess the profitability.
 
The results of the studies using accounting-based performance measures are generally positive relationships, for example, Bowman and Haire (1975), Parket and Eibert (1975), and Waddock and Graves (1997). There is one exception, that is, Aupperle, Carroll and Hatfield (1985) found no link between CER and ROA.
 
Similarly, accounting-based measures have disadvantages. First, it only reflects the historical performance of the firms. (McGuire, Schneeweis and Hill, 1986) What’s more, Branch (1983) pointed out that managerial manipulation and the differential accounting procedures could lead to the incorrect results while using the accounting-based measures. The most important thing is that the control variables, like industry and risk adjustment, should be taken into account in order to achieve more valid results. (Aaker and Jacobson, 1987; Davidson, Worrell and Gilberton, 1986)
 
McGuire, Sundgren and Schneeweis (1988) supposed that market-based measures could overcome the disadvantages of accounting-based measures, as market measures have less risk for managerial manipulation and the accounting procedures differences. And market measures can reflect the future performance rather than the past one, as they employ the investors’ valuation. What’s more, Shane and Spicer (1983) asserted that market-based measures reflect other perspectives beyond just financial outcome of the firms.
 
Interestingly, the results of McGuire, Sundgren and Schneeweis (1988) show that accounting-based measures are ‘better predictors of corporate social responsibility than market measures’. There are two main reasons explaining the results in the journal. One is that accounting measures are more sensitive to the unsystematic social responsibility, and the other is that they are more stable than market measures. For my part, researchers should select the most suitable measure according to different situations, so that the results they achieve would be more useful.
 
Measures of risk, including market beta, interest coverage and current ratio, are also a kind of valid financial performance measure. (Pava and Krausz, 1996) Orlitzky and Benjamin’s (2001) meta-survey of 1968-1985 found an overall negative correlation of 0.0965 between CER and risk. What’s more, Roberts (1992), who examined the US data, found weak negative relationship, and Hasseldine, Salama and Toms, (2005), who tested the UK data, found a significant negative association. (Salama, Anderson and Toms, 2011) Salama, Anderson and Toms (2011) achieved the result that there would be only a 0.02 reduction in firm’s risk when a 1.0 increase in CER.#p#分頁標題#e#
 
Pava and Krausz (1996) defined these measures, such as size, capital investment intensity, as the fourth financial performance criteria, namely other firm-specific characteristics. The results show that there are few associations when using these financial performance measures.
 
Some studies employ more than one kind of criteria to assess financial performance, as they would like to achieve more representative results while testing which measure would be better indicator of financial performance, such as McGuire, Sundgren and Schneeweis (1988).
 
To sum up, it is important to select the most suitable measures for financial performance in the research, as different measures lead to different results. There should not be only one best choice, as these measures have their own advantages and disadvantages; therefore, researchers should choose the one according to their own research characteristics.
 
2.4 Different control variables
 
Control variables, which have been mentioned above, are important factors in the research of the relation between CEP and CFP.
 
Early studies did not take control variables into account when testing the relationship. However, later studies have paid more and more attention to control variables. Some studies only control for the basic variables, such as size, risk, and industry. (eg., Waddock and Graves,1997; Stanwick and Stanwick, 1998) The most recent studies started the research of the moderators of the relationship between CEP and CFP, and they achieved some valuable results, which give a direction for future research in this field. Russo and Fouts (1997) took industry growth as the moderator, and only indicated that there is a positive relationship between industry growth and CFP. Hull and Rothenberg (2008) found that corporate social performance have more impact on the financial performance of firms with low innovation and low level of differentiation in the industry. Both Mcwilliams and Siegel (2000) and Surroca, Tribo and Waddock (2010) suggested that there is no direct link between CEP and CFP when considering the moderators. The former study chose R&D investment as the moderator, while the latter one selected intangible resources of the firm, including ‘innovation resources’, ‘human resources’, ‘reputation’ and ‘culture’. (Surroca, Tribo and Waddock, 2010)
 
Control variables significantly affect the results of the research of the link between CEP and CFP. Therefore, it is important to select the correct control variables.
 
2.5 Applied different theories
 
Theory background is also an important part of the research. There are several theories related to the research in this field, such as Stakeholder Theory (eg. Roberts, 1992; Clarkson, 1995), Legitimacy Theory (eg. Guthrie & Parker, 1989) Political Economy Theory (eg. Cooper & Sherer, 1984), Resource-Based View Theory (eg. Toms, 2002; Russo and Fouts, 1997), Slack Resources Theory and Good Management Theory. (Waddock and Graves, 1997)#p#分頁標題#e#
 
Stakeholder theory is popular in the empirical studies to explain the relationship between CER and CFP. Roberts (1992) asserted stakeholder theory was a reasonable approach to explain the issues between environmental performance and financial performance.
 
There are some examples for other theories, for instance, Russo and Fouts (1997), which is based on the resource-based perspective, suggested a positive correlation between CER and profitability. Moreover, Waddock and Graves (1997) explained the positive relationship between CER and CFP in two directions of causality by applying slack resources theory and good management theory.
 
2.6 The direction of causality
 
The direction of causality is an important issue in the research of relationship between CEP and CFP. Orlitzky, Schmidt.and Rynes, (2003) proved that the relationship between CEP and CFP should be ‘bidirectional and simultaneous’. This means there are two directions of causality, namely prior CFP could affect the CER while CER is a predictor of CFP.
 
McGuire, Sundgren and Schneeweis (1988) argued that the social responsibility might be also linked to the prior firm financial performance. The results show that it is worthy to examine how financial performance affects the CER, as prior financial performance seems to be a good predictor of corporate social performance. Later Stanwick and Stanwick (1998) found that better financial performance has a positive effect on CSP by using the samples in the Fortune 500 listing from year 1987 to 1992.
 
It is obviously that a majority of studies chose to research the second direction of causality, namely CER affects CFP. For example, Salama (2005) selected CER as an independent variable and CFP as a dependent variable, finally achieved the result that there is a positive relationship between two variables.
 
2.7 Different analysis methods
 
The inconsistent results, not only derive from various variables measures, but also due to different analysis methods.
 
The most popular analysis methods used in the studies should be descriptive statistics, correlation analysis and regression analysis. Stanwick and Stanwick (1998) applied these analysis methods and achieved a positive relationship between CSP and CFP. However, Pava and Krausz (1996) only used descriptive statistics.
 
Salama (2005) examined the relationship between CEP and CFP using median regression. And there are some assumptions for applying the median regression. (Kahane, 2001) The results show that there is a stronger positive relationship when employing the median regression.
 
2.8 Possible research errors
 
Apart from the reasons mentioned above, there might be some errors in the researches, which could lead to the inconsistent findings. The possible errors include ‘stakeholder mismatching (Wood and Jones, 1995), the general neglect of contingency factors (Ullmann, 1985), measurement errors (Waddock and Graves, 1997), and failing to see important differences between theory and operational context (Margolis and Walsh, 2001)’. (as cited in Orlitzky, Schmidt.and Rynes, 2003)#p#分頁標題#e#
 
These empirical studies give a direction for further research in this field. I conclude the most representative literatures in this research area in Table 1, which is showed below. It is obvious that a good and reasonable study in the CER-CFP link should carefully consider several issues, including sample selection, CER and CFP measurements, control variables and analysis method. Therefore, in the following sections, I will indicate how I deal with these issues in order to achieve more reasonable and valuable results.
 
Table 1: The most representative literatures in this research area
 
StudySamplesCSR MeasuresCFP MeasuresControl Variables or/and ModeratorsCasual DirectionResultsLimitations
 
Moskowitz (1972)14 firmsN.AStock price change 01/1972-07/1972N.A.CEP->CFPPositiveNo adjustment for risk; CFP measure is questionable; the time period is too short; no test for significance.
 
Bragdon & Martin (1972)17 firms in the paper and pulp industryPollution performance IndexAverage Return on Equity; Average ROC; Earnings per share during five-year periodN.A.CEP->CFPPositiveNo adjustment for risk; CSR/CFP measures are not suitable; sample is not large enough; no test for significance.
 
Bowman & Haire (1975)82 food processing firms Corporate annual reportsFive year return on equityN.A.CEP->CFPU shaped performance curveNo adjustment for risk; CSR/CFP measures are not suitable; sample is not large enough; no test for significance.
 
Vance (1975)14 firms of Moskowitz’s list
 
45(50) firms from surveysBusiness and Society ReviewStock price increases over timeN.A.CEP->CFPNegativeNo adjustment for risk; samples are questionable; time period is too short; regression line does not fit the data; performance criterion is in adequate.
 
Sturdivant & Ginter (1977)28 firms of Moskowitz’s listBusiness and Society Review10 year earnings per share growthN.A.CEP->CFPPositiveNo adjustment for risk; used t-test with very small sample; inconsistent industry categories.
 
Alexander & Buchholz (1978)14 firmsBusiness and Society ReviewStock price increases over 2 years and 5 yearsN.A.CEP->CFPNo relationshipQuestionable samples and inadequate performance measures.
 
Aupperle et al. (1985)241 CEOs listed in Forbes 1981 Annual Directory responded the survey CEO’s orientationsOne year and five years return on assetsN.A.CEP->CFPNo relationshipCEO’s orientations might not represent actual organisation actions.
 
MuGuire et al. (1988)131 firms rated by FortuneThe Fortune SurveyBoth accounting-based and market based measuresN.A.CEP->CFP CFP->CEPPositive;
 
Prior CFP worked better than subsequent ones.Questionable CSR measures.
 
Fombrun & Shanley (1990)The 292 firms included in Fortune's 1985 studyThe Fortune Corporate Reputation Index.Accounting-based and market-based measuresSize, advertising intensity, diversification posture. CFP->CEPComplexNeed better specifying the dimensionality.#p#分頁標題#e#
 
Pava & Krausz (1996)106 firmsCouncil on Economic PrioritiesAccounting-based, market-based measures, measures of risk, and other firm-specific characteristicsN.A.CEP->CFPComplex and nuancedNeed better understanding of CSR.
 
Griffin & Mahon (1997)Firms in the chemical industryFortune; KLD; TRI and corporate philanthropyFive common accounting-based measuresN.A.CEP->CFPInconsistent Only one industry examined.
 
Russo & Fouts (1997)243 firmsThe FRDC ratingsData from COMPUSTATIndustry concentration; firm growth rate; firm size; capital intensity; R&D intensity; advertising intensity. CEP->CFPFour of five are positiveN.A
 
Waddock & Graves (1997)469 firmsKLDReturn on assets, Return on equity, Return on salesSize, Risk and IndustryCEP->CFP;
 
CFP->CEPPositiveOnly test 1-year lag;
 
Not control for quality of management.
 
Stanwick & Stanwick (1998)Firms listed in the top 500 companies of pollution emissions in the United States Environment Protection Agency’s Toxic Release Inventory Report The Fortune Corporate Reputation Index.Profitability (yearly profits divided by the annual sales level)Size CEP->CFP;
 
CFP->CEPPositiveThe use of pollution emissions is questionable; based on one time unique extraordinary circumstances; only selected the large organizations.
 
Mcwilliams & Siegel (2000)524 firmsKLDROASize, risk, industry, R&D Intensity, Advertising intensityCEP->CFPNeutral N.A.
 
Salama (2005)201 firms listed in the Management Today 2000 surveyThe corporate reputation index of Britain’s MACData from DataStream Size, beta, industry and R&D intensity. CEP->CFPPositiveN.A.
 
Barnett & Salomon (2006)61 SRI fundsSocially responsibility investingRisk-adjusted performance (RAP)Age, size, risk, bond funds or stock funds, yearly dummy variables. CEP->CFPCurvilinear relationshipN.A.
 
Hull & Rothenberg (2008)Firms listed KLD, around 500 firms.KLDROASize, risk, industry, innovation; industry differentiationCEP->CFPPositiveSelected CSP measure has some limitations.
 
Godfrey et al. (2009)178 negative actions against firmsKLDShareholder valueFirm size, market-to-bookCEP->CFPPositiveN.A.
 
Surroca et al. (2010)599 companies from 28 countriesKLDTobin’s qInnovation, human capital, reputation, culture; tangible resources; size, risk, year, industry. CEP->CFP;
 
CFP->CEPNo direct relationshipSustainalytics database is not free from criticism; other variables might influence the CEP-CFP link.
 
Salama (2011)Firms listed in the Management Today 1994-2006The corporate reputation index of Britain’s MAC 1994-2006Firm’s beta risk (Beta)Size, dividend payout, liquidity, capital gearing, asset growth, firm profitability, industry. CEP->CFPNegativeLimited amount of data yet available.#p#分頁標題#e#
 
3. Theoretical Framework
 
As mentioned in the literature review section, there are several theories related to the research of relation between CER and CFP. This study applies ‘slack resources theory’ and ‘good management theory’ (Waddock and Graves, 1997) to explain the relationship between CER and CFP.
 
3.1 Slack Resources Theory
 
Waddock and Graves (1997) suggested that firms with better financial performance would have slack resources, which could be invested on the activities positively related to social environmental performance. As a result, their environmental reputation would increase. This is Slack Resources Theory. For example, when IBM had a good financial performance in the earlier days, it had slack resources to invest on philanthropy programs, (Waddock and Graves, 1997) which are obviously related to environmental reputation. Ullmann (1985) model’s third dimension pointed out that financial performance did influence the financial capability to conduct social and environmental related activities. (as cited in Roberts, 1992) In fact, financial performance presents not only financial resources, but also other relative resources. (Waddock and Graves, 1997) There are several empirical studies support the slack resources theory, such as Mcguire, Sundgren and Schneeweis (1988) and Waddock and Graves (1997).
 
3.2 Good Management Theory
 
Good management theory can explain the other direction of causality for CER-CFP relationship. (Waddock and Graves, 1997) It is known as that good management practice, including paying attention to social and environmental issues, leads to overall performance improvement. As a result, corporate financial performance will be higher.
 
Waddock and Graves (1997) further discussed good management theory in the aspects of employee relations, community relations and positive customer perceptions. They argued that keep good relationship with employees could enhance productivity and efficiency, which lead to cost reduction and sales increase. And good community relations might make government to provide some useful benefits that would reduce the cost. Moreover, positive customer perceptions for quality of products, could give a signal to stakeholders that the corporations care about them. As a result, sales increase while stakeholder management costs reduce. Cost reduction and sales increase can lead to better corporate financial performance.
 
Empirical studies, such as Mcguire, Sundgren and Schneeweis (1988) and Waddock and Graves (1997), proved the good management theory.
 
3.3 Hypothesis
 
According to slack resources theory and good management theory, there are two directions of causality for testing the relation between CER and CFP. One is prior corporate financial performance predicts subsequent environmental reputation; another is prior environmental reputation affects the subsequent financial performance. This study examines two directions of causality, while finding out the most reasonable financial performance measures. Thus, I hypothesise that CER is both a predictor and consequence of CFP. There are three hypotheses in this study.#p#分頁標題#e#
 
Hypothesis 1: There is a positive relationship between prior corporate financial performance and subsequent corporate environmental reputation.
 
Hypothesis 2: There is a positive relationship between prior corporate environmental reputation and subsequent corporate financial performance.
 
Hypothesis 3: Accounting-based measures are better indicators for financial performance when examining the CER-CFP link than stock market-based measures.
 
4. Methodology
 
In this section, I will present the methodology of examining the relation between corporate environmental reputation and financial performance based on UK evidence. As the data related to this research question is numerical data, the quantitative research method is selected. And it is impossible to collect the primary data required in this research within the limited time and resources, thus secondary data is used. In the following, details for how research is conducted will be presented.
 
4.1 Sample
 
I select a total of 163 UK companies, excluding the firms with missing financial and control variables, which are listed in the Management Today survey of Britain’s Most Admired Companies (MAC) in year 2006, to examine the relation between environmental reputation and financial performance. These British firms are from 38 different sectors, including Oil & Gas Producers; Travel & Leisure; Food Producers; Software & Computer Services; Banks and so on.
 
4.2 Corporate environmental reputation measure
 
As discussed in the literature review section, it is essential to select the suitable corporate environmental performance measure. This study selects the corporate reputation index published in the Management Today survey of Britain’s Most Admired Companies (MAC) to assess corporate environmental reputation.
 
This reputation index contains the ratings of ‘Community and Environmental Responsibility’ of British companies from 1992 to 2006 (excepting year 1993). This survey selected a total of 239 British companies with the largest market capitalisation. As a result, it includes ‘all the FTSE100 British companies and, on average, 90% of the top 200 companies by market capitalisation’. (Salama, Anderson and Toms, 2011) And there are 38 sectors in total, including Oil & Gas Producers; Travel & Leisure; Food Producers; Software & Computer Services; Financial Service, covered in this survey.
 
Each year senior specialist business analysts and senior managers of 260 British companies mark the corporate environmental performance of the companies. Analysts and executives only make an appraisal on the companies in the sectors they conversant with, but excluding the one they work for. To ranking the reputation, these professionals are asked to get a score for nine issues, including ‘quality of management, financial soundness, quality of service/products, quality of marketing, ability to attract & retain top talent, long-term investment value, capacity to innovation, use of corporate assets, and community and environmental responsibility’ (Salama, Anderson and Toms, 2011), which all have impacts on major stakeholders. The marks are from 0 for poor to 10 for excellent. The final figures representing the CER is the average of the marks given by analysts and executives. (Salama, Anderson and Toms, 2011) Therefore, the CER ratings of this survey are range from 0 to 10. This study selects the data of year 2006. There are 163 British companies in total after excluding those firms with missing variables.#p#分頁標題#e#
 
There are several reasons why this corporate reputation index is chosen. First, this research is based on the UK evidence, so the survey of British Most Admired Companies fits the sample requirements of the research question well. Several representative researches based on US evidence selected the Council on Economic Priorities (CEP) ratings (Freedman and Jaggi, 1982) to assess the corporate social responsibility. However, there is no equivalent rating for British corporations. For this point, MAC rating is the most suitable measure for this research.
 
Second, MAC ratings provide the required data for this research. CEP ratings mentioned above only consider one factor (pollution), rather than wider aspects of environmental performance; therefore, it is only available for certain industries. (Toms, 2002) On the contrast, MAC ratings are ranked by the average scores of nine aspects of environmental management, and it contains the sufficient sample size within 38 sectors. As mentioned above, this research examines the corporations from various industries. Thus, MAC ratings suit the needs of this research.
 
Third, the methodology of MAC survey is more reasonable, which leads to more valuable ratings. The data of MAC ratings come from the answers of professional executives and analysts. They only mark the firms within the sector they are familiar with. All these contribute to the availability and representativeness of the MAC ratings.
 
What’s more, Toms (2000) asserted that MAC ratings could represent the actual managerial strategies, as ‘MAC scores were related to disclosure levels’.
 
However, there are still several weaknesses of MAC index. This survey only presents the figures, but cannot tell the true factors which contribute to the environmental performance change. And Fryxell and Wang (1994) pointed out that MAC survey was slightly different from the other variables, such as ‘management quality, financial variables’, in the MAC data. It fails to present ‘a high degree of co-linearity’. (Toms, 2002) Due to the methodology of MAC survey, the results may be influenced by the preconception and bias of respondents. In another word, the MAC ratings may be not objective enough.
 
Notwithstanding these weaknesses, MAC rating is still a qualified and reasonable measure for environmental reputation in this research, since its advantages can overcome disadvantages.
 
4.3 Corporate financial performance measures
 
As discussed in the literature review section, there are four main kinds of measures to assess the corporate financial performance, which could lead to different CER-CFP relationship outcomes. Accounting-based and stock market-based measures are the most popular ones to assess the financial performance in the empirical studies. This study selects both accounting-based and market-based measures, and compares their CER-CFP relationship results to examine which one would be better.#p#分頁標題#e#
 
4.3.1 Return on Capital Employed (ROCE)
 
For accounting-based measure, this study chooses Return on Capital Employed (ROCE) to assess financial performance. ROCE is the operating income to capital employed ration. It takes sources of financing into accounting, while being similar to Return on Assets (ROA). Aaker and Jacobson (1987) suggested that risk; industry characteristics should be considered as control variables, when accounting-based measures were selected, due to the characteristics of accounting-based measures, namely the possible bias from ‘managerial manipulation’ and accounting procedures differences. (Branch, 1983) The issues of control variables will be detailed in the following. The yearly ROCE is obtained for sample companies from DataStream (Datatype: Worldscope 08376).
 
4.3.2 Earnings Per Share (EPS)
 
For stock market-based measure, this study chooses Earnings Per Share (EPS). EPS is a useful proxy for the profitability of a company. It is calculated as (net income - dividends on Preferred Stock)/ Average Outstanding Shares. As mentioned in the literature review section, market-based measures can overcome the disadvantages of accounting-based measures while having its own flaws. Similarly, EPS can be obtained for sample companies from DataStream (Datatype: EPS).
 
4.4 Control variables
 
Waddock and Graves (1997) suggested some factors affect both environmental performance and financial performance, such as size, risk and industry. Therefore, size, risk and industry are selected to be control variables in this study.
 
4.4.1 Size
 
There is some evidence that firm size plays a role in the corporate environmental performance. (eg. Trotman and Bradley, 1981; Stanwick and Stanwick, 1998) For example, larger companies have higher scores in the Fortune’s Corporate Reputation Index (Fombrun and Shanley, 1990), which indicates larger firms should have a higher level of corporate environmental performance. It is because larger firms receive a higher level of attention from external constituents, and then they need to act more openly to meet the needs of stakeholders, (Burke et al., 1986) which leads to a higher level of environmental performance finally.
 
Therefore, firm size should be taken into account as a control variable when testing the relation between environmental reputation and financial performance. According to Adams and Hardwick (1998), size is measured by the logarithm of firm total assets in this study. Total assets of selected firms are obtained for sample companies from DataStream (Datatype: Worldscope 02999), and then firm size is calculated in the Microsoft Office- Excel.
 
4.4.2 Risk
 
Systematic risk, namely market risk, ‘is defined as the covariance between returns on a risky asset (eg. a corporation's common stock) and market portfolio, divided by the variance of the market portfolio (Copeland & Weston, 1983).’ (as cited in Roberts, 1992)#p#分頁標題#e#
 
Roberts (1992) pointed out that the firms with lower market risk have a higher level of environmental performance. There are two main reasons to explain this result. First, the firms with lower riskiness, which indicates more stock return, would have more money to invest in the activities that have a positive impact on environmental reputation. Thus, the environmental performance could be improved by these activities. Waddock and Graves (1997) listed three kinds of activities, which could be affected by the level of systematic risk. They are these activities, which have the potential to (1) ‘elicit saving’, (2) ‘incur future or present costs’, or (3) ‘build or destroy markets’. Second, the high level of environmental performance will benefit the firm in different aspects, such as more investment, enhancing employee morale and productivity. (Moskowitz, 1972) This leads to riskiness reduction finally.
 
Due to these factors, risk should be considered as a control variable in this study. According to Waddock and Graves (1997), the long-term debt to total assets ratio is used as a proxy for riskiness. Similarly, long-term debt and total assets can be obtained for sample companies from DataStream (Datatypes: Worldscope 03255 and 02999), and the ratios can be calculated in the Microsoft Office- Excel afterwards.
 
4.4.3 Industry
 
As mentioned above, the research is based on the corporations from various industries. Griffin and Mahon (1997) asserted that different industries have different industries characteristics, such as ‘intensity of competition, consumer visibility, regulatory risk’ (Roberts, 1992), ‘levels of R&D’ (Waddock and Graves, 1997) or consumer-oriented nature of companies (Arlow and Gannon, 1982). All these differences contribute to different levels of environmental performance. Roberts (1992) asserted that higher profile industries should have a higher level of environmental performance. Therefore, industry classification should be taken into account as a control variable.
 
There are a total of 38 sectors in this study. And I classify these industries into two categories, namely high-profile and low-profile industries. High-profile industries are those who could have a significant impact on the natural environment. (Salama, 2005) Industry effects are measured by dichotomous classification (1/0) by reference to Roberts (1992). The sectors, including Oil Equipment & Services; Mining; Aerospace & Defense; Oil & Gas Producers; Travel & Leisure (Airline sectors); Gas, Water & Multiutilities; Automobiles & Parts; Alternative Energy; Construction & Materials; Industrial Transportation; Industrial Engineering are high-profile industries. The variable ‘industry’ of the companies from these industries is measured as 1. The other industries are low profile, so the industry variable is measured as 0.
 
4.5 Model specification#p#分頁標題#e#
 
Two main models are employed to test the hypotheses listed above in this study.
 
Model 1: Testing the link between prior corporate financial performance and subsequent environmental reputation (Testing Hypothesis 1)
 
CER (i) = a CFP (i-1) + b SIZE (i-1) + c RISK (i-1) + d IND + e
 
Model 2: Testing the link between prior corporate environmental reputation and subsequent financial performance (Testing Hypothesis 2)
 
CFP (i+1) = A CER (i) + B SIZE (i+1) + C RISK (i+1) + D IND + E
 
where
 
CER (i) = a proxy for corporate environmental reputation for year i
 
CFP = yearly ROCE or EPS is used as a proxy for corporate financial performance
 
SIZE = a proxy for the firm size (log of total assets)
 
RISK = a proxy for the riskiness of the firm (long-term debt/ total assets ratio)
 
IND = industry classification of the firm (1 for high-profile industry and 0 for low-profile industry)
 
i = selected year ( year 2006)
 
For Model 1, testing Hypothesis 1, prior CFP is used as independent variable; subsequent CER is used as dependent variable, while controlling for size, risk and industry.
 
For Model 2, testing Hypothesis 2, prior CER is used as independent variable; subsequent CFP is used as dependent variable, while controlling for size, risk and industry.
 
The research methods, including Descriptive Statistics, Correlation Analysis and Ordinary Least Squares (OLS) regression, are employed to analyse the data. The software SPSS 17.0 helps me to do the analysis. There are a total of four models tested in the SPSS, namely model 1a (ROCE as independent variable), model 1b (EPS as independent variable), model 2a (ROCE as dependent variable), model 2b (EPS as dependent variable). Then each Hypothesis will be tested. The results will discuss in the next section.
 
5. Analysis of Findings
 
In this section, the details of numerical results will be presented, and implications for management and further research will be discussed according to the results achieved. Moreover, limitations of this study will also be listed in this section.
 
5.1 Data Analysis
 
Table 2 shows descriptive statistics for main variables, including CER, ROCE, and EPS, used in this study. Then it is easy to get an overview of these variables. For example, there are a total of 163 companies in the sample. But only 161 companies have the ROCE data (2007), while 158 companies’ EPS (2005) being included. The highest CER score in the study is 8.0, while the lowest is 3.5.
 
Table 2: Descriptive Statistics
 
NMinimumMaximumMeanStd. Deviation
 
CER(06)1633.58.05.629.8578
 
ROCE(07)161-22.90112.6417.420516.04936#p#分頁標題#e#
 
EPS(07)162.0390.846.86954.6904
 
ROCE(05)163-32.6135.513.60515.9925
 
EPS(05)158.00345.2636.538442.56569
 
Industry16301.24.428
 
5.2 Results Analysis
 
5.2.1 Results related to Hypothesis 1
 
The results of correlations analysis showed in Table 3 can be used to test Hypothesis 1. As we can see from Table 3, the correlation coefficients for CER (06) and financial data (05) are 0.151 (for ROCE) and 0.136 (for EPS) with significant at p=0.002 (<0.01) and 0.006 (<0.01) respectively. According to Field (2009), these results mean that subsequent CER is significant positively related to both prior ROCE and prior EPS. In another word, better corporate financial performance leads to higher level of subsequent corporate environmental reputation, which supports Hypothesis 1. Moreover, it is obvious that size (05) is also significant positively related to CER (06) as expected. (The correlation coefficient is 0.145, with p=0.003 <0.05) And there is a negative relationship between CER (06) and risk (05), although it is less strong, still significant. (The correlation coefficient is -0.71, with p=0.09 <0.1) All these results are consistent with the empirical studies I reviewed. For example, Fombrun and Shanley (1990) asserted a positive relation between CER and size; Roberts (1992) suggested that firm with lower risk should have higher level of CER.
 
To better test Hypothesis 1, Ordinary Least Squares (OLS) regression is employed, and related results are showed in Table 4 and full details of results are in the Appendix 1a and Appendix1b. There are two models, namely Model 1a and Model 1b, which using 1-year lag the financial performance (2005 data) and environmental reputation (2006 data). Model 1a uses CER (2006) as the dependent variable, ROCE (2005) as independent variable and controlling for Size (2005), Risk (2005) and Industry; and Model 1b changes to use EPS (2005) as independent variable while keep other variables the same as Model 1a.
 
Table 3: Correlations with CER (2006) and financial variables (2005) and control variables (2005)
 
Correlations
 
CER(06)ROCE(05)EPS(05)size(05)risk(05)Industry
 
CER(06)1.000.151**.136**.145**-.071.089
 
Sig. (1-tailed)..002.006.003.090.084
 
N163163158163163163
 
ROCE (05).151**1.000.100*-.120*-.193**.066
 
Sig. (1-tailed).002..031.012.000.151
 
N163163158163163163
 
EPS(05).136**.100*1.000.171**.024-.059
 
Sig. (1-tailed).006.031..001.328.184
 
N158158158158158158
 
size(05).145**-.120*.171**1.000.119*-.019
 
Sig. (1-tailed).003.012.001..012.382
 
N163163158163163163
 
risk(05)-.071-.193**.024.119*1.000-.113*
 
Sig. (1-tailed).090.000.328.012..040#p#分頁標題#e#
 
N163163158163163163
 
Industry.089.066-.059-.019-.113*1.000
 
Sig. (1-tailed).084.151.184.382.040.
 
N163163158163163163
 
**. Correlation is significant at the 0.01 level (1-tailed).
 
*. Correlation is significant at the 0.05 level (1-tailed).
 
For Model 1a, B (for ROCE) is 0.008 with significant= 0.069 (<0.1), which means there is less strong positive relationship between prior ROCE and subsequent CER, but it is still significant. For Model 1b, B (for EPS) is 0.003 with significant= 0.115 (>0.1), which means there is positive relationship between prior EPS and subsequent CER, but it is not significant. The results of two models are slightly different, and I will further discuss the differences in the next part. All these contribute to the result that better accounting-based measurement can lead to better CER, while there is a weak relationship between market-based measurement and CER. In the case of using accounting-based measurements, Hypothesis 1 is supported. In other words, accounting measured financial performance has a positive impact on CER.
 
It can be seen from the results related to variable size, B for Model 1a is 0.262 with Sig. = 0.001 while B for Model 1b is 0.209 with Sig. = 0.011 (<0.05). All these further prove the positive relationship between size and CER.
 
5.2.2 Results related to Hypothesis 2
 
Similarly, results of correlations analysis showed in Table 5 can be used to test Hypothesis 2. As we can see from Table 5, the correlation coefficients for financial data (07) and CER (06) are 0.144 (for ROCE) and 0.134 (for EPS) with significant at p=0.003 (<0.01) and 0.006 (<0.01) respectively. According to Field (2009), these results mean that subsequent financial performance (both ROCE and EPS) is significant positively related to prior CER. In another word, higher level of CER leads to better financial performance, which supports Hypothesis 2. However, the data shows that there is no significant relationship between prior CER and subsequent risk. (The correlation coefficient is -0.048, with p=0.184 >0.1) This means prior CER has little effect on the subsequent risk.
 
Table 4: OLS Regression Analysis- Model 1a, 1b
 
BSig.
 
Dependent Variable: Corporate Environmental Reputation
 
Model 1a
 
Independent Variable: ROCE0.0080.069
 
Control Variables:
 
Size0.2620.001
 
Risk-0.3280.420
 
Industry0.1900.220
 
R Suqare0.088
 
Adj. R Suqare0.065
 
F3.8020.006
 
Model 1b
 
Independent Variable: EPS0.0030.115
 
Control Variables:
 
Size0.2090.011
 
Risk-0.4700.245
 
Industry0.2100.181#p#分頁標題#e#
 
R Suqare0.087
 
Adj. R Suqare0.063
 
F3.6510.007
 
Table 5: Correlations with CER (2006) and financial variables (2007) and control variables (2007)
 
Correlations
 
ROCE(07)EPS(07)CER(06)size(07)risk(07)Industry
 
ROCE(07)1.000.081.144**-.160**-.091*.183**
 
Sig. (1-tailed)..064.003.001.043.002
 
N161160161161161161
 
EPS(07).0811.000.134**.193**.055-.027
 
Sig. (1-tailed).064..006.000.149.340
 
N160162162162162162
 
CER(06).144**.134**1.000.162**-.048.089
 
Sig. (1-tailed).003.006..001.184.084
 
N161162163163163163
 
size(07)-.160**.193**.162**1.000.058-.009
 
Sig. (1-tailed).001.000.001..138.444
 
N161162163163163163
 
Risk (07)-.091*.055-.048.0581.000-.089
 
Sig. (1-tailed).043.149.184.138..083
 
N161162163163163163
 
Industry.183**-.027.089-.009-.0891.000
 
Sig. (1-tailed).002.340.084.444.083.
 
N161162163163163163
 
**. Correlation is significant at the 0.01 level (1-tailed).
 
*. Correlation is significant at the 0.05 level (1-tailed).
 
Ordinary Least Squares (OLS) regression is also employed to further examine Hypothesis 2. Similarly, related results are showed in Table 6 and full details of results are in the Appendix 2a and Appendix 2b. There are also two models, namely Model 2a and Model 2b, which using 1-year lag environmental reputation (2006 data) and financial performance (2007 data). Model 2a uses ROCE (2007) as dependent variable, CER (2006) as the independent variable, and controlling for Size (2007), Risk (2007) and Industry; and Model 2b changes to use EPS (2007) as dependent variable while keep other variables the same as Model 2a.
 
When ROCE (2007) is used as the dependent variable, B for CER (2006) is 2.793 with significant= 0.058 (<0.1). This means prior CER has a positive effect on subsequent ROCE. Although the relation is not strong enough as 0.5<P<0.1, ( significant with 6.353 is (2006) CER for B Because CER. and performance financial between relationship the indicate to evidence little there performance, assess EPS using of case In significant. still it>0.1), which means the relation between two variables is not significant enough. (Field, 2009) Thus, results are inconsistent when employing ROCE and EPS respectively. Then I conclude that Hypothesis 2 is supported when using accounting-based measures, but for employing market-based measures, the evidence is not significant enough to support Hypothesis 2. Thus, prior CER has a positive impact on accounting measured financial performance.
 
5.2.3 Results related to Hypothesis 3
 #p#分頁標題#e#
As we can see from the discussions above, results of using accounting-based measures and stock market-based measures are inconsistent. Therefore, I will further discuss the related results and make a conclusion.
 
Table 6: OLS Regression Analysis- Model 2a, 2b
 
BSig.
 
Model 2aDependent Variable: ROCE
 
Independent Variable: Corporate Environmental Reputation2.7930.058
 
Control Variables:
 
Size-4.2110.006
 
Risk-17.8410.018
 
Industry4.2250.145
 
R Suqare0.107
 
Adj. R Suqare0.084
 
F4.6830.001
 
Model 2bDependent Variable: EPS
 
Independent Variable: Corporate Environmental Reputation6.3530.214
 
Control Variables:
 
Size14.5590.006
 
Risk17.5240.497
 
Industry3.4630.728
 
R Suqare0.071
 
Adj. R Suqare0.048
 
F3.0120.020
 
When examining the link between prior financial performance and subsequent CER, results are consistent with Hypothesis 3. First, the results of correlation analysis for CER (06) and financial data (05) support Hypothesis 3. The correlation coefficient for ROCE (0.151) is higher than the one for EPS (0.136), with more significant (0.002<0.006). These results suggest that ROCE as accounting-based measure is a better predictor for subsequent CER than EPS as stock market-based measure. When using regression analysis, Model for ROCE achieved a significant result, while Model for EPS failed to get a significant result. This means only accounting-based measured financial performance has a positive effect on CER. Moreover, R Square values indicate that accounting-based performance can produce a better explanatory model than stock market-based performance. (R Square= 0.088 and 0.087 respectively.) This evidence further supports Hypothesis 3.
 
The situation is similar in the case of testing the relationship between prior CER and subsequent financial performance. For correlation analysis, results for ROCE are more supportive than those for EPS. Because r for ROCE (0.144) is larger than r for EPS (0.134), while p for ROCE (0.003) is smaller than p for EPS (0.006). For regression analysis, there is a significant positive relationship between prior CER and subsequent ROCE. However, when using EPS to measure financial performance, it achieved a positive relationship, but it is not significant. R Square value for accounting-based performance (0.107) is larger than the one for market-based performance (0.071). This further proves that accounting-based measures have better explanatory value than market-based measures, when testing the relationship between prior CER and subsequent financial performance.
 
To sum up, accounting-based measures can produce a better explanatory model than stock market-based measures, when examining the CER-CFP link. This is consistent with Hypothesis 3.#p#分頁標題#e#
 
5.3 Implications for Management
 
Results reported here show that there is a positive relationship between corporate environmental reputation and financial performance, when using CER as the independent or dependent variable. This means they are consistent with slack resources theory and good management theory.
 
And it supports the concept of ‘a virtuous circle’ raised by Waddock and Graves (1997). ‘A virtuous circle’ concept means that there is a simultaneous and interactive relationship between CER and CFP. (Waddock and Graves, 1997) Moreover, the results also indicate other variables that have an impact on CER, CFP or CER-CFP relationship, such as size, risk and industry. In the following, implications for management will be presented.
 
Results related to Hypothesis 1, namely better financial performance leads to higher level of CER, support slack resources theory. As mentioned earlier, firms with higher level of financial performance have more slack resources (both financial and other ones) to invest on the activities that could enhance level of environmental performance, such as improving community and employ relations. As a result, the level of corporate environmental reputation will increase. The CER measure I selected, namely MAC survey, takes several aspects of CER into account to achieve the ratings. Therefore, even if investing on the programs that related to only one aspect of CER, the total CER rate will increase. All these contribute to explain why there is a positive relationship between prior financial performance and subsequent environmental reputation. Then managers should raise the awareness that it is essential to improve corporate financial performance. Because it is not just financial figures, it could influence other important performances of the corporation, such as environmental reputation.
 
Results related to Hypothesis 2, namely there is a positive relationship between prior environmental reputation and subsequent financial performance, support good management theory. The results are consistent with the findings of Waddock and Graves (1997). As mentioned above, CER ratings I used are the average scores of nine issues related to environmental performance, including quality of management, quality of service/products, quality of marketing, ability to attract & retain top talent and so on. Then high levels of CER generally represent those companies have better management strategies. For instance, the firms with higher CER might have higher quality products as well as service, and then more customers would be attracted, which leads to sales increase. On the other hand, better relationship with employees, which is also an aspect of CER rating, could attract and retain top talent. This is an advantage of human resources, which can increase productivity, efficiency and reduce cost. What’s more, firms with high CER might have a reputation that attracting more investments. All these contribute to financial performance increase.#p#分頁標題#e#
 
Some studies suggested that programs for increase CER also have a cost, which could be a competitive disadvantage. However, it is obvious that the benefits described above will overcome the small amount of CER cost. Thus, CER programs should be a competitive advantage rather than disadvantage. (Waddock and Graves, 1997)
 
Early studies (eg. Carroll, 1979) defined CSP as ‘a discretionary activity on the part of management’. However, the results reported in this study show that CSP is not a discretionary activity; it is tightly related to management. Therefore, it is essential for senior managers to pay more attention to environmental reputation related activities. And it is necessary to take environmental performance into account, when doing business plan or making a strategical decision. For example, managers should conduct some activities to improve key stakeholder relationships. Annual Celebration is a good way to improve the relationship with key stakeholders by inviting employees, investors and other stakeholders. And more activities that can issue signals to public about the reputation of the firm, such as Philanthropic Contributions, Volunteering, should be planned. Moreover, advertising is also necessary to send the information about what and how the firm is doing to the public. All these can improve the image and reputation of the firm; as a result, more investors, customers and top talent will be attracted, which lead to better financial performance.
 
5.4 Limitations
 
5.4.1 Data collection
 
There are a total of 163 companies examined in this study. For testing Hypothesis 1, a total of 161 companies tested as the sample after removing the missing data; for testing Hypothesis 2, only 158 companies tested as the sample because of missing data. The number of companies tested is enough to achieve a significant result. However, more companies included, more valuable results will be. Due to limit time and resources, there are some missing data, which might have a negative impact on the results. This is one limitation of this study. Moreover, this study only considers the large companies by market capitalisation due to CER measure selection.
 
5.4.2 Control variables: R&D, Advertising Intensity
 
There are three control variables in this study, namely size, risk and industry. However, it would be better to control for more variables in this kind of research. Moreover, some variables should be taken in account as moderators for the CER-CFP relationship.
 
R&D intensity could be considered as a control variable or moderator, as it has impact on CER-CFP link. Mcwilliams and Siegel (2000) suggested that the model, which did not control for investment in R&D, would be misspecified. Griliches (1979) asserted that investment in R&D could lead to better financial performance. (as cited in Mcwilliams and Siegel, 2000) It is because investment in R&D will lead to innovation, which in turn improves the technical issues and as a result productivity enhanced. (Mcwilliams and Siegel, 2000) And there are many empirical studies support this result. For instance, both Lichtenberg and Siegel (1991) and Hall (1999) found a positive relationship between R&D and financial performance. Mcwilliams and Siegel (2000) also discussed the relation between CER and R&D. First, both process innovation and product innovation, which represent the CER improvement, might have an impact on consumers’ decisions. Then companies would like to improve environmental reputation though innovation to attract more customers. Therefore, R&D intensity does have an impact on the CER-CFP relationship. Mcwilliams and Siegel (2000) argued that there is no direct relationship between CER and CFP, when using R&D intensity as a moderator. And Hull and Rothenberg (2008) also suggested that innovation should be considered as a moderator in those models, which testing the CER-CFP relationship. However, it is impossible for me to collect the relative data for R&D intensity within the limit time and energy, the models in this study ignored the control variable R&D intensity. This means results I achieved in this study might be not reasonable enough.#p#分頁標題#e#
 
Advertising intensity is a proxy for level of product differentiation and barriers to entry at the industry level. (Mcwilliams and Siegel, 2000) Corporate social performance can lead to improve the level of differentiation of the firm. But firms in the industries with high levels of differentiation respond less to the changes in differentiation made by CSP than those within low-differentiation industries. Because firms within low-differentiation industries are not familiar with differentiation, so once a firm improves the differentiation by CSP itself, the others may not know how to imitate even if they had the ability to do so. According to Hull and Rothenberg (2008), high differentiation leads to better financial performance. Therefore, CSP has more positive impact on financial performance in the low-differentiation industries than it does in the high-differentiation industries Then it is reasonable to consider advertising intensity as a control variable or moderator in the model. Similarly, it is too difficult to collect data for advertising intensity within limit time, so I did not include this control variable.
 
Empirical studies suggested several other variables, which should be considered as control variables or moderator in the model of testing CER-CFP relationship. For instance, Roberts (1992) controlled for ‘age’ of the firm. He pointed out that there was a positive relationship between firm’s age and social responsibility.
 
And Surroca, Tribo and Waddock (2010) examined how a firm’s intangible resources, including Innovation Resources, Human resources, Reputation and Culture, mediated the CER-CFP relationship. And it suggested there was no direct relationship between CER and CFP, when considering intangible resources. Moreover, Russo and Fouts (1997) took ‘industry growth’ into account, while Toms (2002) suggested ‘governance structures’ as moderators.
 
To sum up, it is not enough to consider only three control variables (size, risk and industry), which is a limitation of this study.
 
5.4.3 Measurements for financial performance and CER
 
As discussed in the Methodology section, there are some limitations for both financial performance and CER measures I selected.
 
I examined the CER-CFP relationship by using accounting-based (ROCE) and market-based (EPS) measures to assess CFP. Both of them have their own advantages and disadvantages, which have been mentioned in the literature review and methodology sections. The mixed results of these two measures show that accounting-based measure is better than market-based measure. Maybe, further research could try to find out a new kind of measure, which is most reasonable and objective, to assess financial performance.
 
For MAC ratings, the most important flaw is that the survey might be not objective enough, as senior managers and analysts marked those companies. They are still human beings, and then there might be bias or preconception. Overall, MAC survey is still a representative and reasonable measure.#p#分頁標題#e#
 
5.4.4 Analysis method: median regression analysis
 
In this study, OLS regression is employed to investigate the relationship between CER and CFP. However, Salama (2005) found out that median regression should be better than OLS regression to test CER-CFP link. There are many outliers of financial performance indicators influencing the analysis. Salama (2005) suggested that median regression could reflect majority of the outliers, thus it had better statistical properties than OLS regression. Though examining 201 firms, median regression shows a better result than OLS. In this respect, OLS regression has its own limitation, namely leading to a partial and misleading influence on the CER-CFP relationship. (Salama, 2005)
 
5.5 Implications for further research
 
The relation between corporate environmental reputation and corporate financial performance is a big and meaningful research question. The relative results could give useful directions for senior managers to better develop the corporation. And further research should investigate this CER-CFP relationship more deeply and objectively by improving current limitations.
 
First, wider ranges of firms from more industries should be taken as the sample. It should be more meaningful to examine both large and small firms in different countries rather than large firms in only one country. Then the results could have a worldwide meaning while being more representative. Meanwhile, recent data should be considered, although it might be a bit difficult to assess to most recent data.
 
Second, researchers should pay attention to control variables and moderators in the further research. More possible variables should be taken into account, and then examining their influences on the CER-CFP relationship. The actual causality of CER-CFP relationship might be further investigated.
 
Third, further study should pay attention to different aspects of environmental performance rather than only one score representing the CER. There are many aspects of environmental performance, and they should have various impacts on the economic performance. Thus, it is necessary to examine the effect of each aspect of CEP. Then managers would know which kinds of activities related to CER should pay more attention in order to enhance profitability.
 
Fourth, results related to Hypothesis 3 indicate that accounting-based measures are better proxy for financial performance than market-based measures, when investigating the CER-CFP relationship. According to Mcguire, Sundgren and Schneeweis (1988), accounting-measured performance is more sensitive to firm-specific social responsibility than market-measured performance, which presents systematic market trends. Davidson, Worrell and Gilberton (1986) pointed out that environmental performance could affect different aspects of financial performance in various ways. Thus, further research could investigate the details about how CER affects each aspects of financial performance. Then more valuable results could be achieved to give a direction for managers to do business plan and make strategical decisions.#p#分頁標題#e#
 
To sum up, it is important for further researchers to select the most suitable and reasonable financial performance measures, when testing CER-CFP relationship. Additionally, a new kind of financial performance measure might be found out in future.
 
Fifth, more studies need to test which regression analysis should be best to examine the CER-CFP link. It is important to select the most suitable regression analysis method in order to achieve the valuable results.
 
At last, further research could put effort to investigate the extent to the impacts of CER or CFP. For example, it is meaningful to find out that 1% CER increase will lead to how much percent CFP rise. The outcome will get a more detailed direction for corporate management.
 
6. Conclusion
 
This study has attempted to achieve an answer for the research question raised in the beginning, namely whether there is a relationship between corporate environmental reputation and financial performance, and in what direction of causality. A total of 163 British companies are investigated in a reasonable methodology, for example, suitable CER (MAC ratings) and CFP measures (ROCE and EPS); control variables including size, risk and industry; correlation analysis and OLS regression analysis. And the findings achieved successfully answered the research question. There is definitely a positive relationship between corporate environmental reputation and accounting-measured financial performance, in both two directions of causality. One is that better financial performance lead to higher level of CER, which is consistent with slack resources theory; another is that higher level of CER lead to better financial performance, which supports good management theory. The financial performance is measured by accounting method; as there is a weak link between stock market-measured financial performance and CER.
 
There are some limitations due to limit time and resources. Although CER and CFP measures as well as analysis methods are valid, they still have some flaws. And more control variables should be considered. Nevertheless, the findings of this study can give a direction for both corporate management and further research in this area. For management, it is worthy and necessary to invest on programs, which enhance the environmental reputation, because corporate financial performance will increase as a result. Further research should be pay more attention on possible control variables or moderators; other valid CER and CFP measures; and it is important to find out the extent to how CER affects CFP and how CFP affects CER. All these improvements contribute to achieve more valuable results, which have an impact on corporate management.
 
7. Bibliography
 
Aaker, D. and Jacobson, R. (1987). The Role of Risk In Explaining Differences In Profitability. Academy of Management Journal, 30, 277-296.
 
Adams, M. and Hardwick, P. (1998). An analysis of corporate donations: United Kingdom evidence. Journal of Management Studies, 35(5), 641–654.#p#分頁標題#e#
 
Alexander, J. and Buchholz, A. (1978). Corporate Social Responsibility and Stock Market Performance. Academy of Management Journal, 21, 479-486.
 
Arlow, P. and Gannon, M. (1982). Social responsiveness, corporate structure, and economic performance. Academy of Management Review, 7, 235-241.
 
Aupperle, K., Carroll, A. B.and Hatfield, J. (1985). An empirical examination of the relationship between corporate social responsibility and profitability. Academy of Management Journal, 28, 446-463.
 
Barnett, M. and Salomon, R. (2006). Beyond Dichotomy: The curvilinear relationship between social responsibility and financial performance. Strategic Management Journal, 27, 1101-1122.
 
Bowman, E.and Haire, M. (1975). A strategic posture towards CSR. California Management Review, 18(2), 49-58.
 
Bradgon, J. H. and Martin, J. (1972). Is pollution profitable? Risk Management, 19(4), 9-18.
 
Brammer, S., Brooks, C. and Pavelin, S. (2006). Corporate social performance and stock returns: UK evidence from disaggregate measures. Financial Management, 97-116.
 
Branch, B. (1983). Misleading accounting: The danger and the potential. Working paper, University of Massachusetts, Amherst
 
Bromiley and Marcus. (1989).The deterrent to dubious corporate behaviour: Profitability, probability, and safety recalls. Strategic Management Journal, 10, 233-250.
 
Burke, L., J. M. Logsdon, W. Mitchell, M. Reiner and D. Vogel (1986). Corporate community involvement in the San Francisco Bay Area. California Management Review, XXVIII(3), 122–141.
 
Carroll, A. B. (1979). A three-dimensional conceptual model of corporate social responsibility. Academy of Management Review, 4, 497-505.
 
Clarkson, M. B. E. (1995). A stakeholder framework for analyzing and evaluating corporate social performance. Academy of Management Review, 20(1), 92-116.
 
Cooper, D. and Sherer, M. (1984). The value of corporate accounting reports: arguments for a political economy of accounting. Accounting, Organizations and Society, 9 (¾), 207-232.
 
Copeland, T.E. and Weston, J.F. (1983). Solutions Manual for Financial Theory and Corporate Policy. 2nd ed. Addison-Wesley Publishing Company.
 
Davidson, W. N., Worrell, D. L. and Gilberton, D. (1986). The appropriateness of using accounting data in studies relating corporate social responsibility to firm financial performance. Academy of Management annual meeting, Chicago.
 
Davidson, W.N., Chandy, P. R. and Cross, M. (1987). Large Losses, Risk Management and Stock Returns in the Airline Industry. Journal of Risk and Insurance, 55, 162-172.
 
Davidson, W. N.and Worrell, D. L. (1990). A comparison and test of the use of accounting and stock market data in relating corporate social responsibility and financial performance. Akron Business and Economic Review, 21(3),7-19.#p#分頁標題#e#
 
Field, A. (2009). Discovering Statistics Using SPSS, 3rd edition. London: Sage.
 
Fombrun, C.and Shanley, M. (1990). What's in a name? Reputation building and corporate strategy. Academy of Management Journal, 33, 233-258.
 
Freedman, M. and Jaggi, B. (1982). The SEC's pollution disclosure requirements- Are they meaningful, California Management Review, 24(2), 60-67.
 
Fryxell, C. and Wang, J. (1994). The Fortune corporate ``reputation'' index: Reputation for what?' Journal of Management, 20(1), 1-14.
 
Godfrey, P., Merrill, C. and Hansen, J. (2009). The relationship between corporate social reasonability and shareholder value: An empirical test of the risk management hypothesis. Strategic Management Journal, 30, 425-445.
 
Griffin, J.J. and Mahon, J.F., (1997). The corporate social performance and corporate financial performance debate: twenty-five years of incomparable research. Business and Society, 36 (1), 5–31.
 
Griliches, Z. (1979 ). Issues in assessing the contribution of R&D to productivity growth. Bell Journal of Economics, 10(1), 92–116.
 
Hall, B. H. (1999). Innovation and market value. National Bureau of Economic Research Working Paper #6984.
 
Hasseldine, J. Salama, A. and Toms, S. (2005) Quantity versus Quality: The Impact of Environmental Disclosures on the reputations of UK plcs, British Accounting Review, 37(2), 231-248.
 
Hull, C. and Rothenberg, S. (2008). Firm performance: The interactions of corporate social performance with innovation and industry differentiation. Strategic Management Journal, 29, 781-789.
 
Ingram, R.W. and Frazier, K.B. (1983) Narrative disclosures in annual reports. Journal of Business Research, 11, 49-60.
 
Kahane, L.H. (2001). Regression Basics. London: Sage Publications.
 
Lichtenberg, F. and D. Siegel (1991). The impact of R&D investment on productivity: New evidence using linked R&D-LRD data, Economic Inquiry, 29, 203–228.
 
Margolis, J. D. and J. P. Walsh. (2001). People and Profits? The Search for a Link between a Company’s Social and Financial Performance. Mahwah, NJ: Lawrence Erlbaum Associates.
 
Moskowitz, M. (1972). Choosing socially responsible stocks. Business and Society, 1, 71-75.
 
McGuire, J., Schneeweis, T.and Branch, B. (1986). A comparison of alternative measures of corporate performance. Unpublished manuscript, University of Massachusetts, Amherst.
 
McGuire, J., A. Sundgren and T. Schneeweiss. (1988). Corporate Social Responsibility and Firm Financial Performance, Academy of Management, 31, 854–872.
 
Mcwilliams, A. and Siegel, D. (2000). Corporate social responsibility and financial performance: Correlation or Misspecification? Strategic Management Journal, 21, 603-609.#p#分頁標題#e#
 
Orlitzky, M.and Benjamin, J. D. (2001). Corporate social performance and firm risk: A meta- analytic review. Business & Society, 40(4), 369-396.
 
Orlitzky, M., Schmidt, F.L.and Rynes, S.L. (2003). Corporate social and financial performance: a meta-analysis. Organization Studies, 24 (3), 403–441.
 
Parket, R.and Eibert, H. (1975). Social responsibility: The underlying factors. Business Horizons,18, 5-10.
 
Pava, M.L.and Krausz, J. (1996). The paradox of social cost. Journal of Business Ethics, 15, 321–357.
 
Prahalad, C. K. and G. Hamel. (1994). Strategy as a field of study: Why search for a new paradigm? Strategic Management Journal, Summer Special Issue (15), 5–16.
 
Roberts, R.W. (1992). Determinants of corporate social responsibility disclosure: an application of stakeholder theory. Accounting, Organizations and Society, 17 (6), 595–612.
 
Rockness, J., Schlachter, P.and Rockness, H. (1986). Hazardous waste disposal, corporate disclosure, and financial performance in the chemical industry. Advances in public interest accounting, 1, 167-191.
 
Russo & Fouts. (1997). A Resource-Based Perspective on Corporate Environmental Performance and Profitability. The Academy of Management Journal, 40(3), 534-559.
 
Salama. (2005).A note on the impact of environmental performance on financial performance, Structural Change and Economic Dynamics, 16, 413–421
 
Salama, A., Anderson, K. and Toms, S. (2011). Does Community and Environmental Responsibility Affect Firm Risk? Evidence from UK Panel Data 1994-2006. Business Ethics: A European Review (forthcoming) 20(2).
 
Shane, P. and Spicer, B. H. (1983). Market response to environment information produced outside the firm. Account. Rev, 58(3), 521-536.
 
Stanwick & Stanwick. (1998). The Relationship Between Corporate Social Performance, and Organizational Size, Financial Performance, and Environmental Performance: An Empirical Examination. Journal of Business Ethics, 17, 195–204.
 
Sturdivant, F. D.and Ginter, J. L. (1977). Corporate social responsiveness. California Management Review, 19(3), 30-39.
 
Surroca, J., Tribo, J. and Waddock, S. (2010). Corporate responsibility and financial performance: The role of intangible resources. Strategic Management Journal, 31, 463-490.
 
Toms, J. (2002). Firm resources, quality signals and the determinants of corporate environmental reputation: Some UK evidence. British Accounting Review, 34, 257-272.
 
Trotman, K.T. and Bradley, G.W. (1981). Associations between social responsibility disclosure and characteristics of companies. Accounting, Organizations and Society, 6 (4), 355-362.
 
Vance, S. (1975). Are socially responsible firms good investment risks? Management Review, 64, 18-24.#p#分頁標題#e#
 
Ullmann, A. (1985). Data in search of a theory: A critical examination of the relationships among social performance, social disclosure, and economic performance. Academy of Management Review, 10, 540-577.
 
Waddock, S.A.and Graves, S.B. (1997), The corporate social performance-financial performance link. Strategic Management Journal, 18(4), 303-319.
 
Wood, D.J. (1991). Corporate social performance revisited. Academy of Management Review, 14(4), 691-718.
 
Wood, D. J.and Jones, R. E. (1995). Stakeholder mismatching: A theoretical problem in empirical research on corporate social performance. International Journal of Organizational Analysis, 3, 229-267.
 
8. Appendices
 
Appendix 1a
 
Model Summary
 
ModelRR SquareAdjusted R SquareStd. Error of the Estimate
 
1.296a.088.065.8296
 
a. Predictors: (Constant), Industry, ROCE(05), size(05), risk(05)
 
Model Summary
 
ModelChange Statistics
 
R Square ChangeF Changedf1df2Sig. F Change
 
1.0883.8024158.006
 
ANOVAb
 
ModelSum of SquaresdfMean SquareFSig.
 
1Regression10.46842.6173.802.006a
 
Residual108.739158.688
 
Total119.207162
 
a. Predictors: (Constant), Industry, ROCE(05), size(05), risk(05)
 
b. Dependent Variable: CER(06)
 
Coefficientsa
 
ModelUnstandardized CoefficientsStandardized CoefficientstSig.
 
BStd. ErrorBeta
 
1(Constant)3.866.540 7.163.000
 
ROCE(05).008.004.1421.833.069
 
size(05).262.080.2503.254.001
 
risk(05)-.328.406-.063-.808.420
 
Industry.190.154.0951.230.220
 
a. Dependent Variable: CER(06)
 
Coefficientsa
 
ModelCorrelationsCollinearity Statistics
 
Zero-orderPartialPartToleranceVIF
 
1(Constant)
 
ROCE(05).119.144.139.9561.046
 
size(05).224.251.247.9791.021
 
risk(05)-.092-.064-.061.9491.054
 
Industry.096.097.093.9731.028
 
a. Dependent Variable: CER(06)
 
Collinearity Diagnosticsa
 
ModelDimensionEigenvalueCondition IndexVariance Proportions
 
(Constant)ROCE(05)size(05)
 
113.4001.000.00.03.00
 
2.7622.112.00.00.00
 
3.5942.392.00.68.00
 
4.2353.803.01.25.02
 
5.00821.106.99.04.98
 
a. Dependent Variable: CER(06)
 
Collinearity Diagnosticsa
 
ModelDimensionVariance Proportions
 
risk(05)Industry#p#分頁標題#e#
 
11.02.02
 
2.08.80
 
3.15.04
 
4.74.13
 
5.01.01
 
a. Dependent Variable: CER(06)
 
Appendix 1b
 
Model Summary
 
ModelRR SquareAdjusted R SquareStd. Error of the Estimate
 
1.295a.087.063.8234
 
a. Predictors: (Constant), Industry, EPS(05), risk(05), size(05)
 
Model Summary
 
ModelChange Statistics
 
R Square ChangeF Changedf1df2Sig. F Change
 
1.0873.6514153.007
 
ANOVAb
 
ModelSum of SquaresdfMean SquareFSig.
 
1Regression9.90142.4753.651.007a
 
Residual103.742153.678
 
Total113.644157
 
a. Predictors: (Constant), Industry, EPS(05), risk(05), size(05)
 
b. Dependent Variable: CER(06)
 
Coefficientsa
 
ModelUnstandardized CoefficientsStandardized CoefficientstSig.
 
BStd. ErrorBeta
 
1(Constant)4.264.524 8.138.000
 
EPS(05).003.002.1251.586.115
 
size(05).209.081.2032.573.011
 
risk(05)-.470.403-.091-1.167.245
 
Industry.210.157.1051.343.181
 
a. Dependent Variable: CER(06)
 
Coefficientsa
 
ModelCorrelationsCollinearity Statistics
 
Zero-orderPartialPartToleranceVIF
 
1(Constant)
 
EPS(05).167.127.123.9551.047
 
size(05).222.204.199.9561.046
 
risk(05)-.105-.094-.090.9741.027
 
Industry.108.108.104.9751.026
 
a. Dependent Variable: CER(06)
 
Collinearity Diagnosticsa
 
ModelDimensionEigenvalueCondition IndexVariance Proportions
 
(Constant)EPS(05)size(05)
 
113.4341.000.00.03.00
 
2.7762.103.00.03.00
 
3.5352.534.00.72.00
 
4.2473.732.01.20.01
 
5.00820.853.98.02.99
 
a. Dependent Variable: CER(06)
 
Collinearity Diagnosticsa
 
ModelDimensionVariance Proportions
 
risk(05)Industry
 
11.02.02
 
2.05.82
 
3.21.00
 
4.71.15
 
5.00.01
 
a. Dependent Variable: CER(06)
 
Appendix 2a
 
Model Summary
 
ModelRR SquareAdjusted R SquareStd. Error of the Estimate
 
1.327a.107.08415.35792
 
a. Predictors: (Constant), Industry, size(07), risk(07), CER(06)
 
Model Summary
 
ModelChange Statistics
 
R Square ChangeF Changedf1df2Sig. F Change#p#分頁標題#e#
 
1.1074.6834156.001
 
ANOVAb
 
ModelSum of SquaresdfMean SquareFSig.
 
1Regression4418.08141104.5204.683.001a
 
Residual36795.030156235.866
 
Total41213.111160
 
a. Predictors: (Constant), Industry, size(07), risk(07), CER(06)
 
b. Dependent Variable: ROCE(07)
 
Coefficientsa
 
ModelUnstandardized CoefficientsStandardized CoefficientstSig.
 
BStd. ErrorBeta
 
1(Constant)31.52211.394 2.767.006
 
CER(06)2.7931.463.1501.909.058
 
size(07)-4.2111.503-.219-2.801.006
 
risk(07)-17.8417.429-.183-2.402.018
 
Industry4.2252.882.1121.466.145
 
a. Dependent Variable: ROCE(07)
 
Coefficientsa
 
ModelCorrelationsCollinearity Statistics
 
Zero-orderPartialPartToleranceVIF
 
1(Constant)
 
CER(06).115.151.144.9301.075
 
size(07)-.175-.219-.212.9341.071
 
risk(07)-.189-.189-.182.9841.017
 
Industry.152.117.111.9781.022
 
a. Dependent Variable: ROCE(07)
 
Collinearity Diagnosticsa
 
ModelDimensionEigenvalueCondition IndexVariance Proportions
 
(Constant)CER(06)size(07)
 
113.8881.000.00.00.00
 
2.7502.277.00.00.00
 
3.3393.385.00.01.00
 
4.01516.040.03.87.33
 
5.00722.951.96.12.67
 
a. Dependent Variable: ROCE(07)
 
Collinearity Diagnosticsa
 
ModelDimensionVariance Proportions
 
risk(07)Industry
 
11.02.02
 
2.07.86
 
3.89.11
 
4.00.01
 
5.02.00
 
a. Dependent Variable: ROCE(07)
 
Appendix 2b
 
Model Summary
 
ModelRR SquareAdjusted R SquareStd. Error of the Estimate
 
1.267a.071.04853.3727
 
a. Predictors: (Constant), Industry, size(07), risk(07), CER(06)
 
Model Summary
 
ModelChange Statistics
 
R Square ChangeF Changedf1df2Sig. F Change
 
1.0713.0124157.020
 
ANOVAb
 
ModelSum of SquaresdfMean SquareFSig.
 
1Regression34320.99348580.2483.012.020a
 
Residual447236.5931572848.641
 
Total481557.586161
 
a. Predictors: (Constant), Industry, size(07), risk(07), CER(06)
 
b. Dependent Variable: EPS(07)
 
Coefficientsa
 
ModelUnstandardized CoefficientsStandardized CoefficientstSig.#p#分頁標題#e#
 
BStd. ErrorBeta
 
1(Constant)-87.90739.638 -2.218.028
 
CER(06)6.3535.089.1001.248.214
 
size(07)14.5595.217.2222.791.006
 
risk(07)17.52425.735.053.681.497
 
Industry3.4639.938.027.348.728
 
a. Dependent Variable: EPS(07)
 
Coefficientsa
 
ModelCorrelationsCollinearity Statistics
 
Zero-orderPartialPartToleranceVIF
 
1(Constant)
 
CER(06).153.099.096.9291.077
 
size(07).242.217.215.9341.071
 
risk(07).029.054.052.9801.021
 
Industry.023.028.027.9741.027
 
a. Dependent Variable: EPS(07)
 
Collinearity Diagnosticsa
 
ModelDimensionEigenvalueCondition IndexVariance Proportions
 
(Constant)CER(06)size(07)
 
113.8911.000.00.00.00
 
2.7502.278.00.00.00
 
3.3363.403.00.01.00
 
4.01516.124.03.87.34
 
5.00723.029.97.13.66
 
a. Dependent Variable: EPS(07)
 
Collinearity Diagnosticsa
 
ModelDimensionVariance Proportions
 
risk(07)Industry
 
11.02.02
 
2.07.84
 
3.88.13
 
4.00.01
 
5.03.00
 
a. Dependent Variable: EPS(07)