Is privatization a socially responsible reform?

Is privatization a socially responsible reform?

Accepted Manuscript Is privatization a socially responsible reform? Narjess Boubakri, Omrane Guedhami, Chuck C.Y. Kwok, He Wang PII: DOI: Reference: ...

NAN Sizes 0 Downloads 63 Views

Accepted Manuscript Is privatization a socially responsible reform?

Narjess Boubakri, Omrane Guedhami, Chuck C.Y. Kwok, He Wang PII: DOI: Reference:

S0929-1199(18)30463-2 https://doi.org/10.1016/j.jcorpfin.2018.12.005 CORFIN 1431

To appear in:

Journal of Corporate Finance

Received date: Revised date: Accepted date:

4 July 2018 2 November 2018 27 December 2018

Please cite this article as: Narjess Boubakri, Omrane Guedhami, Chuck C.Y. Kwok, He Wang , Is privatization a socially responsible reform?. Corfin (2019), https://doi.org/ 10.1016/j.jcorpfin.2018.12.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Is Privatization a Socially Responsible Reform?* Narjess Boubakri, Omrane Guedhami. Chuck C.Y. Kwok, He Wang Narjess Boubakri Bank of Sharjah Chair, American University of Sharjah, UAE [email protected] Corresponding author. Phone: +971 6 515 2587.

CR

IP

T

Omrane Guedhami Moore School of Business, University of South Carolina, USA [email protected]

US

Chuck C.Y. Kwok Moore School of Business, University of South Carolina, USA [email protected]

AC

CE

PT

ED

M

AN

He Wang School of Business, Renmin University of China, China [email protected]

* We gratefully acknowledge the helpful comments from an anonymous reviewer, Douglas Cumming (editor), Najah Attig, Ruiyuan Chen, Sadok El Ghoul, Kateryna Holland, Rezaul Kabir, Walid Saffar, Xiaolan Zheng, Ying Zheng, seminar participants at West Virginia University, and participants at the Frontiers of Business Research in China International Symposium and the Development Bank of Japan Conference on “CSR, the Economy and Financial Markets”.

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

1

ACCEPTED MANUSCRIPT Is Privatization a Socially Responsible Reform?

ABSTRACT We assess the link between corporate social responsibility (CSR) and government ownership using a

T

unique sample of privatized firms (PFs) from 41 countries over the 2002 to 2014 period. We find that PFs

IP

have, on average, better CSR intensity than other publicly listed firms. Further tests show a nonlinear

CR

relation between residual state ownership and CSR intensity that depends on the trade-off between political objectives of the government and profit maximization objectives of private owners. In addition,

US

country-level institutions affect the state ownership–CSR intensity relation, and PFs benefit from higher

AN

valuation and lower equity financing costs through improved CSR.

Classification code: G32, G34

AC

CE

PT

ED

M

Key Words: Privatization; State Ownership; Corporate Social Responsibility; Institutions

2

ACCEPTED MANUSCRIPT 1. Introduction Privatization, the deliberate sale by a government of state-owned enterprises (SOEs) or assets to private economic agents (Megginson & Netter, 2001), has been the subject of renewed interest following the 2008 to 2009 financial crisis amid extensive government participation in distressed firms (Borisova,

T

Fotak, Holland, & Megginson, 2015; Megginson, 2017). Previous privatization studies provide evidence

IP

on both the determinants of privatization1 and the post-privatization performance of privatized firms

CR

(PFs).2 However, studies on the effects of privatization generally focus on such effects in relation to specific stakeholders in PFs – shareholders (domestic and foreign), creditors, and the state as a residual

US

owner – rather than take a broader social welfare perspective. This is surprising given that 1) privatization is widely perceived to be costly to society as a whole,3 2) former SOEs are often viewed as policy tools

AN

used to achieve socio-political objectives, and 3) the state, the ultimate owner of such firms, is theoretically the guardian of social welfare. In this paper, we aim to fill this void in the literature by

M

assessing the social performance of PFs as measured by their corporate social responsibility (CSR)

ED

intensity. To do so, we first compare the CSR intensity of PFs after privatization to that of comparable listed peers.4 We then examine whether government withdrawal from the firm increases or decreases PFs’

PT

post-privatization CSR intensity.

1

CE

CSR refers to a firm’s commitment to various stakeholders such as shareholders, employees,

2

AC

These include legal and political institutions, economic conditions, and capital market conditions (e.g., Bortolotti, Fantini, & Siniscalco, 2004; Boubakri, Cosset, & Guedhami, 2005a; Megginson, Nash, Netter, & Poulsen, 2004). Post-privatization performance is evaluated in terms of profitability, leverage, dividends, corporate governance, cost of debt and equity, risk-taking, and investment efficiency (e.g., Borisova & Megginson, 2011; Boubakri & Cosset, 1998; Boubakri, Cosset, & Guedhami, 2005a, b; Boubakri, Cosset, & Saffar, 2013; Chen, El Ghoul, Guedhami, & Wang, 2017; D’Souza & Megginson, 1999; Megginson, Nash, & van Randenborgh, 1994). 3

For example, privatization may lead to a decrease in employment (Chong & López-de-Silanes, 2005; Dewenter & Malatesta, 1997), a decrease in income distribution (Birdsall & Nellis, 2003), and an increase in poverty (Bayliss, 2002). These concerns have led to demonstrations against privatization reforms in Thailand, Mexico, Pakistan, Italy, and Greece. See also Megginson and Netter (2001), who raise important questions in their survey on the social impact of privatization programs. 4

Because most cross-country CSR data are available starting in 2002, we do not observe the CSR intensity of PFs before privatization, while under full state ownership, which prevents us from running an event study to assess the change in CSR from before to after privatization. However, we are confident that by comparing PFs from the date of privatization to publicly listed peers that are privately owned, we can draw strong inferences as to the relation between CSR and the ownership structure of PFs.

3

ACCEPTED MANUSCRIPT customers, the environment, and the community.5 Prior evidence shows that higher CSR intensity, which is typically based on measures of environmental and social performance, is associated with higher firm value (Ferrell, Liang, & Renneboog, 2016; El Ghoul, Guedhami, & Kim, 2017), a lower cost of capital (e.g., El Ghoul, Guedhami, Kwok, & Mishra, 2011), and better merger performance (Deng, Kang, & Low, 2013).6 However, this evidence is based on publicly traded companies; no prior CSR study to date

IP

T

examines the CSR intensity of PFs.7 By examining the CSR activities of privatized firms, we shed light

CR

on the link between the government as residual owner and social responsibility8, and we contribute to the debate on the costs and benefits of privatization.

US

We begin our analysis by contrasting the CSR intensity of PFs to that of comparable publicly listed firms. The new private owners in PFs may engage more in socially responsible activities to increase the

See European Commission (2001).

M

5

AN

firms’ competitive position because of the value-increasing role of CSR. In addition, because CSR

6

PT

ED

This literature suggests that CSR increases firm value by reducing information asymmetry between firms and investors, since high CSR firms’ disclosures are more transparent and reliable (Kim, Park, & Wier, 2012), by inducing more analyst coverage, which leads to lower analyst forecast errors (Dhaliwal, Li, Tsang, & Yang, 2011), and by decreasing agency problems, since CSR encourages managers to focus less on short-term gains in line with stockholders’ interests and more on long-term profit maximization in line with the interests of other stakeholders (e.g., employees, consumers, suppliers, and shareholders) (Freeman, 1984). 7

8

CE

There are several related studies examining CSR in the Chinese setting, which is characterized by the strong role of state ownership and regulation (e.g., Cumming, Hou, & Lee, 2016). In this paper, we exploit the privatization reform to examine the relation between state ownership, CSR intensity, and performance of privatized firms in 41 countries, which allows us to study how changes in government ownership affect CSR intensity and firm performance, and the role of country level institutions in shaping these relations.

AC

As an example of this dynamic, we examine the CSR report of Royal mail, a privatized firm in our sample. Royal Mail, is a postal service and courier company in the U.K. that was initially privatized in 2013, and fully privatized in 2015. In its 2012-2013 CSR report, in the year of initial privatization, the firm which then showcased a 30% residual state ownership, indicates the following: “We are now a listed company, majority-owned by investors, including employees (who currently hold the single biggest stake other than the Government), institutions and members of the public. Being a listed company means we now have access to capital when we need it to support our strategy. We believe we will combine the best of the public and private sectors. Our strong commitment to our corporate responsibility agenda and to our stakeholders remains unchanged”. In fact, the company established among other new initiatives related to CSR, a new Environment Governance Board (EGB) to strengthen its environmental management, and introduced a 12-month volunteering program to support and encourage employees to give back to the communities. Another example from our sample includes Safran S.A., which is a French SOE engaging in the aerospace and defense business. The government privatized 30% stake of the company in 2010, and currently (2018) holds 11%. In 2010, the company overhauled its organizational structure by starting a new Sustainable Development Department to spearhead its Health and Safety and Environmental (HSE) policy. In 2010, Safran S.A. launched an initiative to reduce greenhouse gas emissions. Since 2012, the company started to report its CSR activities in a separate section in its annual report and lay down a formal CSR strategy and governance system in 2012.

4

ACCEPTED MANUSCRIPT activities are reputation-enhancing (Porter & Kramer, 2006, 2011), we expect the residual state owners to invest more in CSR to signal the PF’s commitment to various stakeholders in an effort to reduce potential opposition to privatization (socio-political motive for CSR).9 This is all the more likely now that governments can transfer the cost of CSR activities from taxpayers to the new owners. As a consequence,

T

we predict that PFs have a higher CSR intensity than their listed peers.

IP

Alternatively, the shareholder theory of CSR first put forward by Friedman (1970) predicts that the

CR

CSR intensity of PFs should be indistinguishable from that of other listed firms. Under this alternative view, the sole social responsibility of a firm is to generate profits for its shareholders, in which case high

US

CSR investment represents a misuse of resources that should be allocated to positive NPV projects or returned to shareholders (profit maximization motive for CSR).10

AN

To test these predictions, we follow Boubakri et al. (2013) and Chen, El Ghoul, Guedhami, and Nash (2018) and merge a sample of 578 privatized firms – the largest sample of privatized firms to date – with

M

data on CSR intensity from Thomson Reuters ASSET4. The merged sample comprises 18,816 firm-year

ED

observations from 41 countries over the period 2002 to 2014. Consistent with the value-increasing and reputation-enhancing roles of CSR, we find that PFs have statistically significantly higher CSR intensity

PT

than their peers. This effect is economically significant, as privatized firms observe on average 10.3%

CE

higher CSR intensity than comparable listed firms. One concern with the above analysis is potential endogeneity of privatization decisions. For example,

AC

as Megginson and Netter (2001) note, governments may privatize the healthiest firms to make privatization “look good.” Similarly, data may be more readily available for better-performing firms in more developed countries. To address this concern, we use propensity score matching to identify comparable firms that are not privatized but otherwise have similar characteristics as the PFs in our

9

One feature of privatization programs around the world is the prevalence of partial (rather than full) divestitures, a direct implication of which is that the government retains residual ownership in PFs (Bortolotti & Faccio, 2009; Boubakri, Cosset, & Guedhami, 2005b, 2009; Megginson et al., 1994). Thus, the government can still affect the performance, value, and strategy of PFs, including CSR activities. 10

Friedman (1970) even argues that CSR activities signal agency problems whereby managers use CSR to further their own careers instead of maximizing shareholder returns.

5

ACCEPTED MANUSCRIPT sample, instrumental variable (IV) approach, and the Heckman (1979) sample selection approach. We continue to find that PFs have better CSR intensity than their peers. We next focus on the sample of PFs, which we separate into partially and fully privatized firms (i.e., those with positive versus those with zero residual state ownership). When we compare their respective

T

post-privatization CSR intensity, we find that partial privatizations exhibit higher CSR intensity than their

IP

fully privatized counterparts. Combined with our main evidence that privatization leads to higher CSR

CR

intensity, this suggests that CSR intensity has a nonlinear relation with the extent of privatization, first increasing then decreasing. To shed light on this result, we conduct quadratic regression analysis in the

US

spirit of Borisova and Megginson (2011) and find that the relation between residual state ownership and CSR intensity is indeed nonlinear: the positive effect of residual state ownership on CSR intensity is more

AN

(less) pronounced at high (low) levels of residual ownership. Taken together, these results show that at high levels of residual state ownership, governments’ political motives for CSR (including soft budget

M

constraint, cost-shifting and reputation building) prevail, whereas at lower levels of residual state

ED

ownership, economic value enhancement and profit-maximization motives for CSR dominate. We conduct additional analyses to further improve our understanding of the CSR intensity of PFs.

PT

First, we examine the role of country-level institutions on the relation between CSR and residual state

CE

ownership. Prior privatization literature shows that country-level institutions condition the performance outcomes of government ownership in PFs (e.g., Borisova, Brockman, Salas, & Zagorchev, 2012;

AC

Boubakri et al., 2013; Guedhami, Pittman, & Saffar, 2009; Chen et al., 2018). In this vein, we investigate whether the relation between CSR and residual state ownership is moderated by the quality of country-level institutions. Motivated by the privatization literature, we consider indicators of the quality of legal institutions, political institutions, and government financial constraints. These indicators provide us with the possibility to differentiate between the arguments related to the political motivations for CSR from the economic motivations for CSR. We find that the significant relation between residual state ownership and CSR intensity persists only in countries with weak legal institutions, left-wing and financially constrained governments: indeed, in these countries, political motivations are more likely to 6

ACCEPTED MANUSCRIPT dominate economic motivations for CSR. Second, we examine how the interaction between residual state ownership and CSR affects valuations and cost of equity capital of PFs. Prior research suggests that residual government ownership could be detrimental to firm performance and cost of equity capital (Ben-Nasr, Boubakri, & Cosset, 2015).

T

If CSR signals firm commitment to stakeholders, then the negative effects of residual state ownership on

IP

firm value should decrease as PFs invest more in CSR, hence driving firm valuations up. In line with this

CR

expectation, we find that higher CSR intensity mitigates the effects of state ownership on firm value and cost of equity capital, suggesting that PFs benefit from higher valuation and lower equity financing costs

US

through improved CSR.

This paper contributes to the privatization literature by examining the social impact of privatization.

AN

To our knowledge, this is the first study to assess the CSR intensity of PFs in a cross-country setting. We predict that privatization leads to increased CSR investment in PFs as governments use CSR to reduce

M

opposition to privatization (reputation-enhancing motive) and to increase the privatized firm’s

ED

competitive advantage vis-à-vis comparable listed firms (value-increasing motive). We find support for this prediction by showing that PFs observe higher CSR intensity than their listed peers. This paper

PT

further contributes to the literature on the determinants of CSR by showing that CSR investment in PFs

CE

depends on the degree of government influence, with CSR investment being higher in PFs with high residual government ownership. Our study relates to prior research showing that political interference

AC

matters to CSR, mostly in the context of China (see, Maung, Wilson, & Tang, 2016; Cumming, Hou, & Lee, 2016).11 In particular, our findings that governments shape firms’ CSR activities are consistent with those of Lei and Zhang (2010), who document a positive association between government ownership and CSR intensity in SOEs in China. Reinforcing the role of political influence on CSR, the authors show that political connections, another form of government influence, is positively associated with CSR levels in

11

Dam and Scholtens (2012) examine the relations between different types of owners and CSR using firm-level data for more than 600 European firms from 16 countries. The authors do not find evidence that government ownership affects CSR intensity.

7

ACCEPTED MANUSCRIPT non-SOEs. In addition, our evidence that the effect of residual government ownership on CSR intensity is moderated by the quality of a country’s institutions is in line with their findings that government influence on CSR is reduced in more developed areas in China. Finally, our paper contributes to the corporate finance literature by providing evidence on the impact

T

of CSR on PFs’ financial performance. We show that higher CSR intensity in PFs is associated with

IP

higher firm value and lower cost of equity capital, suggesting that CSR helps mitigate uncertainty

CR

associated with residual state ownership (e.g., Boubakri et al., 2013; Borisova & Megginson, 2011). Our findings have implications for the debate on the costs and benefits of privatization. First, we

US

extend evidence that the privatization of SOEs is beneficial from an economic perspective [see Megginson and Netter (2001) and Megginson (2017) for a literature review] by showing that it is also

AN

beneficial from a social perspective. Specifically, PFs observe on average higher CSR intensity than other publicly listed firms. Second, we provide evidence that residual state ownership distorts firms’ decisions

M

by showing that high residual ownership is associated with a high level of CSR investment, as shown by

ED

contrasting partially privatized with fully privatized firms. The remainder of the paper is organized as follows. Section 2 reviews prior related literature and

PT

develops our hypotheses. Section 3 discusses our data and variables, and provides descriptive statistics for

CE

our sample. Section 4 presents our empirical analysis. Section 5 concludes the paper.

AC

2. Literature Review and Hypotheses Classic public finance characterizes SOEs as social enterprises that aim to cure a market failure (Atkinson & Stiglitz, 1980; Shapiro & Willig, 1990) that private firms fail to address. Under this view, welfare-maximizing SOEs may be expected to invest more in CSR than their private counterparts. In reality, however, government officials use SOEs to pursue political goals (Shleifer, 1998). Boycko, Shleifer, and Vishny (1993), for instance, show that governments often suboptimally locate SOEs in remote regions, employ too many people, or subsidize prices to help politicians win political support among voters rather than improve social welfare.

8

ACCEPTED MANUSCRIPT To reduce the inefficiency of SOEs, privatization reforms that transfer both control and cash flow rights from governments to private investors have occurred worldwide over the past three decades.12 Such ownership changes are accompanied by a change in firm focus from pursuing political objectives such as those above to profit maximization. Prior studies show that after government divestiture, firm

T

performance improves significantly (Megginson, Nash, & van Randenborgh, 1994; Boubakri & Cosset,

IP

1998; D’Souza & Megginson, 1999; Dewenter & Malatesta, 2001). We next develop our hypotheses on

2.1. CSR intensity of privatized firms versus other listed firms

CR

the consequences of privatization for firms’ CSR intensity.

US

Extant empirical evidence shows that higher CSR intensity leads to larger abnormal stock returns (e.g., Dimson, Karakas, & Li, 2015), lower idiosyncratic risk (e.g., Lee & Faff, 2009), lower cost of

AN

capital (e.g., El Ghoul et al., 2011), improved access to external finance (e.g., Cheng, Ioannou, &

M

Serafeim, 2014), and better merger performance (Deng et al., 2013). Given this value-increasing role of CSR, the new private owners in PFs may make large CSR investments to increase the firm’s competitive

ED

position vis-à-vis other publicly listed firms.

PT

Another reason PFs might invest more in CSR relates to its reputation-enhancing role. Privatization reform is fiercely opposed by those who view it as beneficial to politically connected investors and

CE

detrimental to employees, other investors, and society as a whole. This opposition, which is supported by evidence of job loss (Chong & López-de-Silanes, 2005; Dewenter & Malatesta, 1997) and increased

AC

poverty after divestiture (Birdsall & Nellis, 2003; Bayliss, 2002), sometimes leads to unrest. By increasing investment in CSR activities, the government as residual owner can mitigate concerns about social welfare following the reform. This politically motivated CSR is particularly likely as the government can shift the costs of such activities to the firm’s new private investors.

12

Besides improving efficiency, other motivations for privatization reform include: raising revenue for the state, reducing government interference in the economy, promoting wider ownership, increasing market competition, subjecting SOEs to market discipline, and developing a national capital market (Price Waterhouse, 1989a, b; Megginson & Netter, 2001).

9

ACCEPTED MANUSCRIPT Taken together, the above arguments suggest that PFs invest more in CSR activities and hence observe higher CSR intensity than their publicly listed peers. Hypothesis 1a: PFs have higher CSR intensity than other publicly listed firms. However, the shareholder theory of CSR introduced by Friedman (1970) suggests that PFs may not

T

exhibit higher CSR intensity than other publicly listed firms. According to this theory, the firm’s sole

IP

social responsibility is to generate profits for its shareholders. This theory therefore suggests that PFs

CR

should focus less on social motives for CSR than on profit-maximizing motives for CSR. In addition, because CSR is likely to increase monitoring and public scrutiny, shareholders of PFs operating in

US

environmentally sensitive industries may argue against investing in CSR activities. Taken together, these alternative arguments predict that there is no difference in CSR intensity

AN

between PFs and other publicly listed companies:

M

Hypothesis 1b: There is no difference in CSR intensity between PFs and other publicly listed firms. 2.2. CSR intensity and residual state ownership in privatized firms

ED

To disentangle the political motivations for CSR from the economic and profit-maximization

PT

motivations for CSR, in relation to residual state ownership, we focus our attention on partially (i.e., the government remains a residual shareholder) and fully (i.e., the government has transferred all control and

CE

ownership rights to private owners) privatized firms. In the same fashion as Borisova and Megginson (2011), we invoke four potential mechanisms, albeit somehow different, that can affect the expected

AC

relation between residual state ownership and CSR intensity in PFs. These mechanisms can be classified in two broad categories related to political objectives of CSR (arguments a, b, and c below), and economic value-enhancing motives of CSR (argument d below), and can be summarized as follows: a. Government Ownership (Soft Budget Constraint) Distorts Investment Decisions According to previous studies on privatization, continued involvement of the government in the firm negatively affects its performance, as a result of conflicting interests of the government versus the other firm owners. While the government pursues political/social objectives and uses the PFs to this end, the

10

ACCEPTED MANUSCRIPT other shareholders seek profits and value maximization. This conflict of interest, dubbed extreme agency problems by Shleifer and Vishny (1997), forces the firms into making suboptimal decisions from the shareholders’ perspective, while being beneficial to the government that seeks to maximize political support. The conflict between economic benefits and political benefits can distort investment decisions

T

(e.g., Jaslowitzer, Megginson, & Rapp, 2016; Chen et al., 2017), including those related to CSR activities.

IP

In partially privatized firms, a higher residual state ownership can thus lead to an over-emphasis on social

CR

goals, which the PF will be used to achieve. From the perspective of shareholders, it is expected that they will conform to this expectation given that continuing government ownership provides them with a soft

US

budget constraint (e.g., preferential rates of financing, easy access to finance) and an implicit guarantee of bailout in times of distress. Consequently, although the residual state ownership is decreasing after

AN

privatization, firms will increase CSR to cater for the government’s social mission and political

b. Cost-Shifting of CSR Investment

M

objectives.

ED

In partially privatized firms, the cost of investing in CSR can be transferred from the state budget and taxpayers to other shareholders in the firm: that is, as privatization progresses, the government has further

PT

incentives to pressure firms to invest in CSR activities, with the new owners bearing the resulting cost. By

CE

doing so, the government can reap the reputational benefits of increased CSR (i.e., political and popular support) without having to bear the full costs of CSR investment. We thus expect that even with

hypothesis.

AC

decreasing residual state ownership, we will observe an increase in CSR intensity under the cost-shifting

c. Government Reputation Building Privatization, which is by definition a transfer of ownership, is accompanied by policy uncertainty about future government actions that may affect firms’ decisions. This asymmetrical information, or policy risk as described by Perotti (1995), pertains to the future level of government interference (as a residual owner) in the firm’s decisions and environment (through regulation changes, for instance), all of

11

ACCEPTED MANUSCRIPT which affect the future performance of the firm. In this asymmetrical information setting, the government needs to signal to investors its commitment to the privatization reform. Improved CSR intensity can then be adopted as a signaling, reputation-building mechanism that will convey to the market information on the firm’s quality and the level of government commitment to market-oriented policies (low policy risk)

T

without having to keep high levels of ownership as in Perotti (1995). Based on this argument, as residual

IP

state ownership is declining, we are likely to observe an increase in CSR intensity.

CR

d. Economic Goals and Profit-Maximization

As state residual ownership decreases in privatized firms, so does the grabbing hand of the

US

government and the resulting distortions in firms’ behavior and decisions. The managerial objectives of the new owners lead them to focus on economic goals and profit maximization, and to de-emphasize any

AN

political goals that prevailed under state ownership. Reflecting this, firms now seek to rationalize their investments in the pursuit of profitability, and hence are likely to reduce their engagement in CSR below

M

what they were used to under government control. Consequently, as residual state ownership continues to

ED

decline towards zero, we expect to observe a decline in CSR intensity. To summarize, declining residual state ownership can have a positive or a negative impact on CSR

PT

intensity depending on the trade-off between economic value enhancing (argument d) and political

CE

motives (arguments a, b, and c). This is illustrated in Figure 1. Given these contrasting views, we draw and submit the following non-directional hypothesis to data:

AC

Hypothesis 2: Residual state ownership is positively or negatively related to CSR intensity of PFs. ------------------------------Figure 1 goes about here -------------------------------

2.3. The role of country-level institutions The strength of country-level institutions may affect the relation between residual state ownership and CSR intensity in PFs. Borisova et al. (2012), for instance, suggest that in countries with better investor protection, state ownership is associated with greater firm and market support in times of crisis.

12

ACCEPTED MANUSCRIPT Guedhami et al. (2009) further argue that government ownership is related to poor corporate governance incentives in countries with weak institutions, where government predation is more likely to occur (Durnev & Fauver, 2009) and investors’ rights are less likely to be protected. Consistent with this view, Chen et al. (2018) show that when country-level institutions are less effective in constraining the

T

consumption of political private benefits, firms’ decisions are more likely to be influenced by government

IP

ownership. Overall, these studies imply that the adverse effects of government ownership are less

CR

pronounced in countries with a stronger institutional environment. More related to our focus on CSR, Li and Zhang (2010) report a positive relation between government influence and CSR intensity in China,

US

with this relation being weaker in more developed areas of the country. In our setting, to the extent that higher residual state ownership distorts the firms’ objectives by over-weighting the political motives

AN

(arguments a, b, and c above) compared to the profit-maximization motives of CSR, we expect the effect of residual state ownership on CSR intensity in PFs to be moderated by the quality of country institutions.

M

More formally:

ED

Hypothesis 3: The relation between residual state ownership and the CSR intensity of PFs is

PT

moderated by the strength of country-level institutions. 2.4. Financial outcomes of CSR in PFs

CE

To shed further light on the link between government ownership and CSR, we examine how the

AC

relation between residual state ownership and CSR intensity affects firm-level financial outcomes – in particular, firm value and the cost of equity capital. With respect to firm value, prior privatization literature suggests that the government’s use of PFs to pursue socio-political objectives reduces firm performance and value (Boubakri et al., 2005a, b). The weaker corporate governance and in turn greater information asymmetry associated with government ownership have also been shown to contribute to lower financial performance in PFs (e.g., Boubakri et al., 2013; Chen et al., 2017, 2018; Guedhami et al., 2009). In the CSR literature, prior research suggests that firms with better CSR intensity observe better financial performance, as CSR helps reduce agency costs 13

ACCEPTED MANUSCRIPT and information asymmetry (Cheng et al., 2014; Ferrell et al., 2016). These arguments suggest that high CSR intensity mitigates the negative effect of residual state ownership on PFs’ valuation. Turning to the cost of equity capital, prior privatization research documents a negative effect of residual state ownership on earnings quality (Ben-Nasr et al., 2015) and transparency (Guedhami et al.,

T

2009), which imply higher costs of capital, while Ben-Nasr, Boubakri, and Cosset (2012) and Borisova

IP

and Megginson (2011) provide direct evidence that residual state ownership increases PFs’ cost of capital.

CR

In the CSR literature, CSR intensity is found to be positively associated with the strength of internal corporate governance (Jo & Harjoto, 2011), ease of access to financial capital (e.g., Cheng et al., 2014),

US

and abnormal returns (e.g., Dimson et al., 2015), which imply a lower cost of capital, while El Ghoul et al. (2011) provide direct evidence that CSR lowers firms’ cost of capital. This evidence together suggests that

AN

high CSR intensity should decrease the negative effect of residual state ownership on PFs’ cost of capital. Combining these two lines of argument leads to our fourth hypothesis:

M

Hypothesis 4: High CSR intensity mitigates the negative effect of residual state ownership on PFs’

ED

valuation and cost of equity capital.

3.1. Sample selection

PT

3. Sample, Variables, and Summary Statistics

CE

To construct our sample, we begin by collecting data from several sources (see the Appendix). We

AC

obtain data on privatized firms from Boubakri et al. (2013) and Chen et al. (2018).13 To the best of our knowledge, this is the largest sample of privatized firms collected to date. Next, we merge the privatization data with CSR data obtained from Thomson Reuters ASSET4, which provides information on a firm’s environmental, social, and governance (ESG) activities. This information is collected from publicly available sources (e.g., annual reports, NGO websites, and CSR reports) and is updated biweekly.

13

Following Chen et al. (2018), a privatized firm refers to an entity in which state ownership has been reduced through privatization. A privatized firm may therefore exhibit a zero (fully privatized firm) or a positive (partially privatized firm) level of residual state ownership.

14

ACCEPTED MANUSCRIPT Finally, we hand-match the resulting data set with firm-level financial data obtained from the Compustat Global database. To construct the regression variables, we drop firms with insufficient financial data. We also drop countries and industries without privatized firms. The final sample contains 18,816 observations from 41 economies over the period 2002 to 2014. 3.2. Key variables

IP

T

CSR Variables. Following Ioannou and Serafeim (2012) and El Ghoul et al., (2017), we construct our primary measure of a firm’s CSR intensity (CSR) as the average of the firm’s environmental score (EP)

CR

and social score (SP). A firm’s environmental score (EP) captures “a company’s impact on living and

US

non-living natural systems, including the air, land and water, as well as complete ecosystems,” and is based on the firm’s energy use, CO2 emissions, waste recycling, and the like.14 A firm’s social score (SP)

AN

measures “a company’s capacity to generate trust and loyalty with its workforce, customers and society, through its use of best management practices” and is based on factors such as employee turnover, injury

M

rate, training hours, percentage of female employees, and the amount donated to charitable organizations.

ED

3.2.1. Privatization and ownership variables

We use the dummy variable PRIVATIZED to indicate whether a firm is privatized. To capture state

PT

ownership in PFs, we use STATE, the level of government ownership in a PF, and PARTIALLY, a dummy variable equal to one if state ownership in a PF is greater than zero. We also employ YRS_PRIV, the

CE

number of years since privatization, in our analysis.15

AC

3.2.2. Firm-level control variables We include a number of firm-level variables to control for factors that may affect CSR intensity. In particular, we control for firm size (SIZE), the logarithm of total assets in millions of $US, return on assets (ROA), the ratio of net income before extraordinary items to total assets, R&D expenditure (RDS),

14

More detailed descriptions of these scores are available from Thomson Reuters’ website at http://thomsonreuters.com/en/about-us/corporate-responsibility-inclusion/esg-performance.html. 15 Following Megginson et al. (2004), Boubakri and Cosset (1998), and Boubakri et al. (2005a), privatization date is determined by the first time the government sells a stake in the firm. Privatization is a gradual process as the government continues to relinquish ownership over a long period of time (e.g., Boubakri et al., 2005b).

15

ACCEPTED MANUSCRIPT the ratio of research and development expenses to total sales, leverage (LEV), the ratio of total debt to total assets, cash flow riskiness (STD_ROA, the standard deviation of ROA), and stock riskiness (STD_RET, the standard deviation of daily stock returns). To mitigate the impact of outliers, we winsorize all firm-level variables at the 1% and 99% levels. In addition to these firm-level variables, we control for

T

country, industry, and year fixed effects in all of our regressions.

IP

3.3. Summary statistics

CR

Table 1 summarizes our sample composition by country, industry, and year in Panels A, B, and C, respectively. The full sample comprises 18,816 observations from 2,583 unique firms over the period

US

2002 to 2014. Of the 2,583 firms, 220 are privatized. Panel A, which presents the data by country, shows that the number of privatization observations varies considerably across countries. For instance,

AN

Kazakhstan, Mexico, and Morocco each have fewer than five observations, while France has the most at 125. Panel B, which summarizes the sample distribution by industry using Fama and French’s (1997)

M

12-industry classification, shows that there is also a large degree of variation in the number of

ED

privatization observations across industries, with the Wholesale and Retail sectors having the fewest observations (21) and Utilities having the most (287). Panel C, which summarizes the sample

------------------------------Table 1 goes about here -------------------------------

AC

CE

except in 2013 and 2014.

PT

composition by fiscal year, shows that the number of privatization observations has increased over time

Table 2 presents summary statistics for the key variables used in our analyses. Panel A provides results for the firm-level regression variables. Our main dependent variable, CSR, ranges between 97.950 and 6.200, with a mean of 56.560 and a standard deviation of 29.407, suggesting that there is considerable variation in CSR intensity across sample firms. The mean value for PRIVATIZED is 0.076, which implies that 7.6% of sample observations correspond to privatized firms. Panel B presents Pearson correlation coefficients between the firm-level variables. We find that PRIVATIZED is positively related to our proxies for CSR intensity (CSR, EP, and SP), which provides preliminary evidence that privatized firms 16

ACCEPTED MANUSCRIPT tend to have higher CSR intensity. The correlation coefficients between the key explanatory variables are low, indicating that multicollinearity is not likely to affect our regression results. Panel C reports the means of our institutional indices by country. All country-level institutions exhibit substantial variation across countries. -------------------------------

IP

T

Table 2 goes about here -------------------------------

CR

In Table 3, we conduct univariate tests of differences in means between PFs and other publicly listed firms. The average CSR is 68.181 for privatized firms, compared to 55.598 for other publicly listed firms,

US

with the difference significant at the 1% level. Similarly, privatized firms have significantly higher environmental and social scores than other firms. These results confirm the preliminary evidence in Table

AN

2 that PFs tend to have higher CSR intensity. However, this analysis does not control for other variables

M

that could affect CSR. We investigate such effects using multivariate analysis next. -------------------------------

ED

Table 3 goes about here -------------------------------

PT

4. Empirical Results

4.1. CSR intensity of PFs versus other listed firms

CE

To examine the CSR intensity of PFs, we employ the following specification:

AC

CSRit = α + β1 ⋅ PRIVATIZEDi + β2 ⋅ SIZEit + β3 ⋅ ROAit + β4 ⋅ RDSit + β5 ⋅ LEVit + β6 ⋅ STD_ROAit + β7 ⋅ STD_RETit+ Fixed effects +εit

(1)

where CSR is one of our three proxies for CSR (CSR, EP, SP), PRIVATIZED is a dummy variable indicating whether a firm is privatized, SIZE, ROA, RDS, LEV, STD_ROA, and STD_RET are firm-specific control variables, Fixed effects is a vector that includes country, industry, and year fixed effects, and i and t index firm and time, respectively. Following Petersen (2009), all of the regressions in this analysis cluster standard errors by firm and year.

17

ACCEPTED MANUSCRIPT Table 4 presents our main evidence on the relation between privatization and CSR intensity. In Column 1, the dependent variable is CSR, which is the average of a firm’s environmental and social scores. The coefficient on PRIVATIZED is positive and significant at the 1% level, suggesting that, in line with Hypothesis 1a, PFs outperform other publicly listed firms in terms of CSR intensity. This result is

T

economically significant as well: given a mean CSR of 56.560 (Table 2, Panel A), the coefficient on

IP

PRIVATIZED of 5.848 implies that on average PFs have 10.3% greater CSR intensity (from 56.560 to

CR

62.408) than their listed peers. Turning to the control variables, the results show that firm size, ROA, and R&D investment are positively associated with CSR intensity, while cash flow riskiness and stock

US

riskiness are negatively associated with CSR intensity.

In Columns 2 and 3 we use a firm’s environmental and social scores, respectively, as the dependent

AN

variable and again find a positive and significant coefficient on PRIVATIZED. These results complement the finding in Column 1 by showing that privatized firms are associated with higher environmental score

M

as well as higher social score. In terms of economic significance, privatized firms demonstrate 10.5%

61.694) than their peer firms.

ED

higher environmental score (from 57.105 to 63.123) and 10.1% higher social score (from 56.015 to

PT

In sum, consistent with Hypothesis 1a, our results show that privatized firms demonstrate better CSR

CE

intensity than other publicly listed firms, both when we consider a firm’s overall CSR intensity and when we separately consider the environmental and social dimensions of CSR intensity. These results are in line

AC

with the value-increasing and reputation-enhancing roles of CSR. ------------------------------Table 4 goes about here -------------------------------

4.2. Endogeneity of privatization If privatization is systematically related to differences in firm characteristics, the relation between CSR intensity and privatization that we document above may be due in part to differences in firm characteristics. For instance, governments may privatize the healthiest SOEs to make privatization “look good” (Megginson & Netter, 2001). In our main regression analysis in Table 4, we control for variables 18

ACCEPTED MANUSCRIPT such as size, ROA, R&D investment, leverage, cash flow riskiness, and stock riskiness to try to separate the effects of firm-specific characteristics from the effect of privatization. However, to the extent that privatization and these firm characteristics are endogenously determined, our results may be biased. It is therefore crucial that we account for such differences in our analyses.

T

In Table 3, which compares characteristics of privatized firms and other publicly listed firms, we find

IP

that privatized firms are larger and younger, and they have lower R&D investment, higher leverage, and higher ROA. To assess whether these differences influence our results, we use propensity score matching

CR

(PSM) where we match each PF with a control firm that has similar characteristics but is not privatized.16

US

We first estimate propensity scores using a Probit model in which the dependent variable is PRIVATIZED and the explanatory variables are as in Table 4. Panel A of Table 5 reports the results. In

AN

Panel A1 we compare characteristics of PFs and matched firms. The differences between privatized firms and matched firms are no longer significant, but PFs still outperform on CSR. In Panel A2 we rerun the

M

regression analysis using privatized firms and matched firms. Similar to our main evidence in Table 4, the

ED

coefficients on PRIVATIZED are positive and significant. These findings confirm our main results in

outperformance on CSR.

PT

Table 4, mitigating concerns that differences in firm characteristics could be the source of privatized firms’

CE

In Panel B of Table 5, we rerun the tests in Panel A [using propensity scores based on an alternative Probit model in which we employ firm size and industry as explanatory variables (Faccio, Masulis, &

AC

McConnell, 2006)]. Since many privatized firms are large and monopolistic in a given country, using too many explanatory variables may not give the best match for each privatized firm. The results are consistent with those in Panel A. ------------------------------Table 5 goes about here ------------------------------To further address endogeneity concerns, we employ instrumental variables (IV) analysis. 16

We perform one-to-one matching without replacement of matches. The results remain unchanged, however, when we allow replacement.

19

ACCEPTED MANUSCRIPT Specifically, we consider an endogenous binary variable model estimated using a three-stage procedure following Adams, Almeida, and Ferreira (2009).17 In the first stage, we estimate a Probit of the determinants of privatization by regressing the privatization dummy variable on the full set of control variables from our main specification and an instrument for privatization. Because privatization is

T

influenced by whether a country is ruled by a right-wing or left-wing government (Bortolotti & Pinotti,

IP

2008; Bortolotti & Faccio, 2009), we follow Beck, Clarke, Groff, Keefer, and Walsh (2001) and use

CR

RIGHT, which captures whether the government has a right-wing orientation (that is, it is dominated by a conservative, Christian democratic, or right-wing party), as an instrument. In the second stage, we regress

US

PRIVATIZED on the fitted probability from the Probit regression and all explanatory variables. In the third stage, we regress CSR on the PRIVATIZED predicted from the second stage and our control

AN

variables. The results (Column 1 of Table 6) confirm that privatized firms observe better CSR intensity. We also use the Heckman selection estimation procedure to address endogeneity. The first stage of

M

the Heckman model is a Probit model where we regress the privatization dummy on all of the control

ED

variables and the instrument. In the second stage, we regress CSR intensity on the inverse Mills’ ratio (λ) estimated from the first stage, the privatization dummy, and the control variables. The results of the

PT

Heckman model are presented in Column 2 of Table 6. We continue to find that privatized firms have

CE

better CSR intensity than other publicly listed firms. In sum, we find consistent evidence that PFs’ higher

AC

CSR intensity is robust to endogeneity concerns. ------------------------------Table 6 goes about here -------------------------------

4.3. CSR intensity of PFs and the role of residual state ownership In this section, we examine CSR intensity during the privatization process. More specifically, we focus on the sample of privatized firms and examine the changes in CSR intensity as shaped by residual

17

When the endogenous variable is a dummy, consistency of estimates from the traditional two-stage least squares model is highly sensitive to specification bias in the first-stage regression. The three-stage procedure does not suffer from this shortcoming. See Adams et al. (2009) for more details.

20

ACCEPTED MANUSCRIPT state ownership and by the number of years since privatization. In Panel A of Table 7, we compare the CSR intensity of partially and fully privatized firms using univariate analysis. Consistent with Hypothesis 2, we find that partially privatized firms outperform fully privatized firms on overall CSR intensity and social score. The mean difference in CSR intensity (social

T

score) is statistically significant at the 5% (1%) level.

IP

In Panel B of Table 7, we examine the effect of partial privatization using regression analysis. More

CR

specifically, restricting our attention to the sample of privatized firms, we regress our CSR intensity measures on a dummy for partial privatization (PARTIALLY) and the control variables in Column 1 of

US

Panel B. The coefficient on PARTIALLY is positive and significant, suggesting that partially privatized firms exhibit higher CSR intensity than fully privatized firms. Columns 4 and 5 show similar effects when

AN

we consider EP and SP measures, respectively. Overall, these results are consistent with the prediction in Hypothesis 2 that state influence over partially privatized firms is positively related to CSR intensity

M

when political objectives prevail over economic objectives.

ED

Turning to the control variables, we find that SIZE and YRS_PRIV load positively in all regressions, suggesting that CSR intensity increases with firm size and number of years elapsed since privatization:

PT

Larger firms are more likely to have the resources required to conduct CSR activities, consistent with

CE

evidence in Attig et al. (2016) and El Ghoul et al. (2016). One potential interpretation of the positive coefficient on YRS_PRIV is that with privatization being gradually implemented, governments continue to

AC

control privatized firms several years following privatization. For example, in our sample, the average residual state ownership is 39.66% (34.36%) 6 to 10 years (and more than 11 years) after privatization.18 At first glance, these results may seem at odds with the evidence of higher CSR intensity of privatized firms where governments relinquished ownership compared to publicly listed firms, as reported

18

In untabulated results, we further analyze the time effects by replacing YRS_PRIV with dummy variables indicating 1 to 5 years since privatization (1-5 YRS_PRIV), 6 to 10 years since privatization (6-10 YRS_PRIV), and more than 11 years since privatization (>11 YRS_PRIV). We find that the coefficient on 1-5 YRS_PRIV is statistically insignificant, while those on 6-10 YRS_PRIV and >11 YRS_PRIV dummies load positively and significantly. We thank the reviewer for suggesting this analysis.

21

ACCEPTED MANUSCRIPT in Table 3. These findings indicate a possible nonlinear relation between residual state ownership and CSR intensity: In partially privatized firms, the government’s political motives for CSR prevail over shareholders’ profit-maximizing objectives, whereas in fully privatized firms, the new private owners’ profit-maximizing objectives drive CSR investment decisions. -------------------------------

IP

T

Table 7 goes about here -------------------------------

CR

To explore the nonlinear relation between state ownership and CSR intensity, we use a quadratic model that includes a continuous measure of residual state ownership (STATE) and its square (STATE_SQ).

US

Specifically, we regress CSR intensity on STATE and STATE_SQ. The results are presented in Column 1 of Table 8. We find that STATE loads with a positive coefficient that is statistically significant at the 5%

AN

level. More importantly, we find that the quadratic term STATE_SQ has a negative coefficient that is statistically significant at the 10% level. CSR intensity is shown to be highest at 45.5% government

M

ownership. These results lend further support to the idea that the relation between residual state ownership

ED

and CSR intensity depends on the trade-off between the new private owners’ profit-maximizing objectives and the government’s political objectives. More specifically, consistent with Figure 1, declining residual

PT

state ownership is positively related to CSR intensity (arguments a, b, and c in Section 2.2) up to the

CE

threshold of 45.5% government ownership. Any subsequent decline in residual state ownership then leads

AC

to a decline in CSR intensity (argument d in Section 2.2).

------------------------------Table 8 goes about here -------------------------------

4.4. The role of country-level institutions The privatization literature suggests that the institutional environment conditions privatization design and the performance outcomes of PFs (e.g., Borisova et al., 2012; Boubakri et al., 2013; Guedhami et al., 2009; Chen et al., 2017, 2018). Accordingly, we next investigate whether the nonlinear relation between residual state ownership and CSR intensity is influenced by country-level institutions as predicted in 22

ACCEPTED MANUSCRIPT Hypothesis 3. To conduct our analysis, we consider three indicators: (i) the quality of legal institutions, (ii) political institutions, and (iii) government financial constraints. Specifically, we run split sample analysis using the following proxies for the quality of legal institutions: (1) LEGAL ORIGIN, which identifies the legal origin of a country’s corporate law, (2) ANTI-SELF-DEALING, which measures the degree to which

T

minority shareholders are protected against expropriation by controlling owners, and (3) JUDICIAL

IP

INDEPENDENCE, which measures the efficiency of the judiciary using the country risk ratings of

CR

Business International Corp. We also use (4) POLITICAL ORIENTATION to capture political institutions, and (5) GOVERNMENT DEFICITS to proxy for government financial constraints. POLITICAL

US

ORIENTATION is measured with party orientation with respect to economic policy. Right-oriented parties are those defined as conservative, Christian democratic, or right-wing; Left-oriented parties are those

AN

defined as communist, socialist, social democratic, or left-wing. In the center are parties defined as centrist or when party position can best be described as centrist. GOVERNMENT DEFICITS is measured

M

as the general government total expenditure minus revenue, as a percentage of GDP. Importantly for our

ED

purposes, these indicators allow us to disentangle the arguments related to the political motivations for CSR (arguments a, b, and c) from the economic and profit-maximization motivations for CSR (argument

PT

d).

CE

More specifically, legal institutions variables (LEGAL ORIGIN, ANTI-SELF-DEALING, and JUDICIAL INDEPENDENCE) condition the level of government participation and influence in the

AC

economy, and hence, the extent of distortion between economic (profit maximization) and political objectives associated with state ownership (Boycko et al., 1996; Nash, 2017). For instance, prior privatization literature suggests that performance gains are lower in privatized firms operating in countries with weak institutions (e.g., Boubakri, Cosset, & Guedhami, 2008), where residual state ownership remains high and changes towards private ownership are slow (i.e., more time to full relinquishment of control) (see, for example, Boubakri et al., 2005b; Guedhami and Pittman, 2006; Bortolotti and Faccio, 2009; Boubakri et al., 2018). These firms, it is argued, are used by governments as tools to fulfill political objectives, including maximizing votes and political support, and ensuring 23

ACCEPTED MANUSCRIPT political tenure. In this context, CSR constitutes a valuable tool to help achieve such objectives. Accordingly, we expect that government political motives for CSR prevail over economic objectives in countries with weak institutions (i.e., with civil law legal origin, weak investor protection, and low judicial independence).

T

Political orientation (POLITICAL ORIENTATION) captures the reputation-building mechanism

IP

related to CSR. The literature on the political economy of privatization suggests that partial privatizations

CR

with government gradual control relinquishment are used to signal governments’ commitment to market-oriented policies (Perotti, 1995). As explained in the hypothesis section, by relinquishing their

US

control rights, governments signal that privatization is credible and implies no policy risk (i.e., risk of interference in the operations of privatized firms through regulation changes for instance). By retaining

AN

residual ownership, governments thus signal their willingness to share in any remaining policy risk and their commitment to the privatization process. In the same vein, related studies argue that right-wing

M

governments, which are typically more committed to privatization and economic reforms and are less

ED

fiscally constrained, favor less state control in the economy, and hence divest control more quickly (Biais and Perotti, 2002; Boubakri et al., 2011). As a result, right-wing governments do not need to signal any

PT

commitment. Applied to our setting, this means that they do not need to use CSR as such a signaling

CE

mechanism, unlike left-wing-oriented governments. Therefore, we expect that political objectives related to reputation building (argument c—government reputation building) will prevail more in left-wing

AC

governments than in right wing governments. Finally, government deficit (GOVERNMENT DEFICITS) reflects its financial constraints and hence, its incentives to shift costs of CSR activities to the firm’s new private investors. Therefore, we expect political objectives related to cost shifting (argument b) to prevail more in countries with higher government deficit: governments with high deficit have more incentives to pressure new owners of PFs to bear the costs of CSR investments. Table 8 reports the split sample results. Columns 2 and 3 of Table 8 split the sample based on LEGAL ORIGIN (civil law and common law, respectively), while Columns 4 and 5 (6 and 7) split the sample 24

ACCEPTED MANUSCRIPT based on the median of ANTI-SELF-DEALING (JUDICIAL INDEPENDENCE). The results indicate that the inverted U-shaped relation between CSR intensity and residual state ownership as reflected by STATE and STATE_SQ concentrates in countries with weak legal protection of minority investors. These findings suggest that, consistent with Hypothesis 3, country-level legal institutions do indeed moderate the relation

T

between residual state ownership and CSR intensity in PFs. In particular, strong country-level legal

IP

institutions constrain governments from distorting the objectives of privatized firms.

CR

We interpret these findings as suggesting that strong legal institutions decrease the political motivations for CSR by governments, and increase the importance of economic objectives. This means

US

that in countries with better legal environments, CSR increases with residual state ownership, reflecting economic objectives of profit maximization, rather than political objectives (arguments a, b, c: distortions,

AN

cost shifting, and reputation building) and a lack of trade-off between both. Indeed, in common law countries, governments are less associated with the distortion of economic objectives, exhibit less

M

presence in the economy, and are more market-oriented. In addition, the time to fully privatize (relinquish

ED

control) is much shorter in common law countries compared to civil law countries, suggesting that economic objectives prevail sooner in the former. In contrast, in weak legal environments (such as civil

PT

law countries), the relation is nonlinear reflecting a trade-off between political and economic objectives:

CE

in fact, we determine that CSR is highest at an inflection point of 30.8% government ownership in civil law countries, suggesting that beyond this inflection point (i.e., when state residual ownership is higher

AC

than 30.8%), political objectives are more important than economic objectives in determining privatization decisions, and CSR investment in turn, becomes guided by the distorted objectives associated to high residual government ownership. Below this inflection point (i.e., when residual government ownership is lower than 30.8%), the CSR investment decision is rather guided by the economic objectives (profit maximization) of the new private owners. This evidence is consistent with findings in Liang and Renneboog (2017) that CSR ratings are higher in civil law countries, and together reflect the dominance of governments in the economy of countries with weak institutions (e.g., civil law)

25

ACCEPTED MANUSCRIPT as documented in Bortolotti and Faccio (2009) and Boubakri et al. (2005b).19 We obtain similar inflection points when we examine ANTI-SELF-DEALING and JUDICIAL INDEPENDENCE at 30.1% and 35.5%, respectively. Again, these inflection points reflect the stage at which the government relinquishes control of the firm to the benefit of the new private shareholders, and suggest that the relinquishment of

T

government control leads to a readjustment of the PFs objectives from political to economic motivations.

IP

Columns 8 and 9 of Table 8 split the sample based on the political ideology of the government (i.e.,

CR

right versus left orientation). We find that the inverted U-shaped relation between CSR intensity and residual state ownership holds in countries with left-wing governments. The inflection point of 43.3% in

US

these countries indicates that beyond this point (i.e., when state residual ownership is higher than 43.3%), the CSR decision is guided by political objectives reflecting the need for left-wing governments to signal

AN

their reputation of market commitment through CSR. Below this inflection point (i.e., when state residual ownership is lower than 43.3%), governments relinquish control and the need for signaling is reduced,

M

such that the new private owners adjust the CSR levels to meet the economic objectives of the firm

ED

(argument c).

Our final set of results in Columns 10 and 11 shows the role of government financial constraints in

PT

influencing the state ownership-CSR intensity relationship. The results suggest that the inverted U-shaped

CE

relation between CSR intensity and residual state ownership holds only in countries with high government deficit, consistent with our prediction based on the cost shifting argument (argument b). The

AC

coefficient estimates on STATE and STATE_SQ suggest an inflection point at 34.8%. When state residual ownership is higher than 34.8%, governments under higher financial constraints are more likely to pressure the new owners to bear the cost of CSR investments. This pressure is increasing in the residual state ownership. Below the 34.8% inflection point, governments gradually lose their influence over firms’ decisions and the new private owners, who are then able to adjust the CSR intensity levels to meet the profit maximization objectives of the firm.

19

We do not report inflection points for strong institutions countries because the relation is not nonlinear.

26

ACCEPTED MANUSCRIPT Taken together, our findings show that the trade-off between political and economic objectives affects the residual state ownership-CSR intensity relation in countries with weak legal institutions, left-wing and financially constrained governments. In such countries, high residual state ownership distorts firms’ objectives, and allows governments to transfer the cost of CSR investment to private new owners, and to use CSR investment decisions as a signal of commitment for reputation building

IP

T

purposes.20

CR

4.5. Financial outcomes of CSR in PFs

In this section, we empirically examine the effect of post-privatization ownership structure and CSR

US

intensity on PFs’ financial performance – firm value and cost of equity. Specifically, we examine the impact of CSR on firm value and on cost of equity using the following specifications:

AN

Qit = α + λ1 ⋅ STATEit + λ2 ⋅ PCSRit + λ3 ⋅ STATEit × PCSRit + λ 4 ⋅ SIZEit + λ 5 ⋅ ROAit + λ 6 ⋅ RDSit + λ 7 ⋅ LEVit+ λ 8 ⋅ STD_ROAit + λ 9 ⋅ STD_RETit+ λ 10 ⋅ CASHit+ λ 11 ⋅ CAPXit (2.1)

M

+ λ 12 ⋅YRS_PRIVit+ λ13⋅ANTI-SELF-DEALINGi + Fixed effects +εit

ED

COEit = α + γ1 ⋅ STATEit + γ 2 ⋅ PCSRit + γ 3 ⋅ STATEit × PCSRit + γ 4 ⋅ BETAit + γ 5 ⋅ LOG_MVEit + γ 6 ⋅Qit + γ 7 ⋅ LEVit+ γ 8 ⋅ LTGit + γ 9 ⋅ DISPit+ γ 10 ⋅ YRS_PRIVit+ γ 11 ⋅ INFLit

PT

+ γ 12 ⋅ ANTI-SELF-DEALINGi + Fixed effects +εit,

(2.2)

CE

where Q is Tobin’s Q and COE is the firm’s cost of equity capital, STATE is our continuous measure of residual state ownership in PFs, PCSR is predicted values of CSR intensity or its components (PEP and

AC

PSP), and Fixed effects is a vector that contains year and industry fixed effects. We also include a vector of the usual firm-level control variables shown to influence firm value or the cost of equity capital. In this analysis, we cluster standard errors by firm and year. In Panel A of Table 9, we first examine the impact of CSR and state ownership on Tobin’s Q. Following Boubakri, Guedhami, and Cosset (2005b) and Guedhami, Pittman, and Saffar (2014), we

20

In unreported results, we examine whether the relation between state residual ownership and CSR intensity varies according to the industry characteristics, such as industry concentration (measured with Herfindahl–Hirschman Index) and regulated industries. However, the results do not show different effects across these characteristics.

27

ACCEPTED MANUSCRIPT implement a two-stage estimation procedure. In the first stage, we regress CSR intensity measures (CSR, EP, and SP) for each firm-year observation using the model from Column 1 of Table 4. The first-stage fitted values for CSR intensity (PCSR, PEP, and PSP) are then used in the second-stage regressions explaining Tobin’s Q and the firm’s cost of equity capital as shown in equations 2.1 and 2.2. More

T

specifically, in the second stage, we regress Q on STATE, a PF’s predicted CSR intensity from the first

IP

stage regression, and the interaction STATE * PCSR, as well as the controls and fixed effects. We find

CR

that the coefficients on the interactions between residual state ownership and predicted CSR intensity (PCSR, PEP, and PSP) are positive and significant. These results suggest that, in line with Hypothesis 4,

US

state ownership is valued more highly in PFs with higher CSR intensity.

Recent research suggests that higher CSR intensity is associated with lower cost of equity capital

AN

(Dhaliwal et al., 2011; El Ghoul et al., 2011). Accordingly, in Panel B of Table 9 we extend the above analysis by looking at the effect of CSR and state ownership on PFs’ cost of equity capital (COE).

M

Following prior studies (e.g., Hail & Leuz, 2006; El Ghoul et al., 2011), we construct COE as the average

ED

of four implied cost of equity capital measures: KGLS of Gebhardt, Lee, and Swaminathan (2001), KCT of Claus and Thomas (2001), KOJN of Ohlson and Juettner-Nauroth (2005), and KMPEG of Easton (2004).21

PT

The results reported in Panel B of Table 9 show that the coefficient on the interactions between state

CE

ownership and predicted CSR variables (PCSR, PEP, PSP) are negative and significant, suggesting that PFs with higher CSR intensity benefit from lower equity financing costs.

AC

Taken together, the results in this section support prior evidence that residual state ownership is associated with lower financial performance, but also show that under high CSR intensity, the negative impact of residual state ownership on PFs’ financial performance is reduced. ------------------------------Table 9 goes about here -------------------------------

21

For further details about the assumptions and the estimation of the implied cost of equity models, see Hail and Leuz (2006) and El Ghoul et al. (2011).

28

ACCEPTED MANUSCRIPT 5. Conclusion Prior privatization literature has largely focused on the effects of residual state ownership in relation to main stakeholders, such as shareholders and creditors (e.g., Borisova & Megginson, 2011; Ben-Nasr et al., 2015), and has overlooked the broader societal effects (Megginson & Netter, 2001). This is surprising

T

given the common perception that the state is the guardian of social welfare, and that privatization could

IP

undermine the socio-economic development (or such a role). Against this backdrop, in this paper we

CR

investigate the CSR intensity of privatized firms, with a particular focus on the CSR intensity of PFs compared to publicly listed firms, and the link between residual state ownership and CSR intensity within

US

PFs.

Using a large sample of 18,816 firm-year observations from 41 countries over the period 2002 to

AN

2014, we find that CSR intensity is significantly higher for PFs in comparison to other publicly listed

M

firms. This finding, which is robust to addressing endogeneity concerns, is consistent with PFs investing more in CSR to increase their competitiveness (and in turn profitability) or to enhance their reputation

ED

(and in turn signal the privatized firm’s commitment to stakeholders). Our next analysis compares partially and fully privatized firms and shows that partially privatized firms have significantly higher CSR

PT

intensity, which suggests that when the government’s political objectives dominate new owners’

CE

profit-maximization objectives, the government is likely to use CSR investment to pursue the reputation benefits of CSR, particularly when it can transfer the costs of CSR investment to the firm’s new owners.

AC

However, as PFs become more independent of government control, shareholders’ profit-maximization objectives start to prevail, which results in CSR investment being adjusted downward. In line with Borisova and Megginson (2011), who examine the cost of debt implication of residual state ownership in PFs, we find a nonlinear relation between the extent of residual state ownership and CSR intensity in PFs. We further find that this relation concentrates in countries with a weak institutional environment, which suggests that strong institutions constrain governments from distorting PFs’ CSR decisions and leads private owners to freely pursue economic objectives of profit maximization. Finally, we investigate the

29

ACCEPTED MANUSCRIPT financial outcomes of PFs’ CSR activities. We find that higher CSR intensity helps align the objectives of new private owners and the government, leading to higher firm value and lower cost of capital in PFs. This paper contributes to the privatization literature by showing that PFs have higher CSR intensity than other publicly listed firms. To our knowledge, this is the first study to assess the CSR intensity of

T

PFs. Our evidence shows indeed that as long as the firms are under government control, their objectives

IP

will be distorted and geared towards achieving political goals. In other words, they remain policy tools in

CR

the hands of the government. In addition, concerns that privatization may be socially and environmentally costly are not supported by our results, which show instead that even after full relinquishment of

US

government ownership, CSR intensity in PFs remains higher than that of matching publicly listed firms. This paper also contributes to the literature on the determinants of CSR by linking CSR to a

AN

macroeconomic policy that is politically driven, namely, privatization, and by showing that CSR varies with the degree of government influence in privatized firms. Finally, our paper contributes to the

M

corporate finance literature by providing evidence on the impact of CSR on firm valuation and the cost of

ED

equity in PFs. We find that higher CSR intensity helps mitigate the adverse effect of residual state ownership on PFs’ financial performance.

PT

While we provide strong evidence for the effect of privatization on CSR, there are limits to this study

CE

that arise from the nature of the data that we use. In short, the present study provides cross-country evidence on the relation between state ownership and CSR intensity in PFs. With more data on state

AC

ownership, future research could examine how changes in state ownership affect CSR intensity. In addition, given the prevalence of state acquisitions around the world, it would be interesting to examine how these acquisitions affect CSR intensity.

30

ACCEPTED MANUSCRIPT Appendix

Variable Definitions and Data Sources Variable CSR EP

SP

PCSR PEP PSP PRIVATIZED YRS_PRIV STATE PARTIALLY SIZE ROA RDS

Definition Panel A: CSR Variables The overall CSR intensity is equal to the average of environmental score and social score.

Source

T P

The environmental score measures a company’s impact on living and non-living natural systems, including the air, land, and water, as well as complete ecosystems. It reflects how well a company uses best management practices to avoid environmental risks and capitalize on environmental opportunities in order to generate long-term shareholder value. The social score measures a company’s capacity to generate trust and loyalty with its workforce, customers, and society through its use of best management practices. It reflects the company’s reputation and the health of its license to operate, which are key factors in determining its ability to generate long-term shareholder value. Predicted CSR intensity using regressions in Table 4. Predicted environmental score using regressions in Table 4. Predicted social score using regressions in Table 4. Panel B: Privatization and Ownership Variables A dummy variable that equals 1 for privatized firms, 0 otherwise.

I R

C S

U N

A

D E

M

Years since privatization. Percentage of shares held by a government. A dummy variable that equals 1 if state ownership is not zero in a PF, 0 if state ownership is zero. Panel C: Firm-level Variables The natural logarithm of total assets in millions of $US.

T P

E C

LOG_MVE

The ratio of net income before extraordinary items to total assets. Ratio of research and development expenses to total sales. Missing research and development expenses are set to zero. The ratio of total debt to total assets. Cash flow riskiness, the standard deviation of ROA. Stock riskiness, the standard deviation of daily stock returns. Cash holdings, cash and marketable securities to total assets. The ratio of capital expenditures to total assets. Market beta obtained from regressions of a firm’s monthly excess stock returns on the corresponding CRSP value-weighted index excess returns using at least 24 months and up to 60 months ending in June of each year. Excess returns are monthly returns minus the 1-month Treasury bill rate. The natural logarithm of market value of equity in millions of $US.

LTG

Average long-term growth forecast reported in June of each year.

LEV STD_ROA STD_RET CASH CAPX BETA

C A

Authors’ calculation based on ASSET4 data ASSET4

As above

Authors’ calculation Authors’ calculation Authors’ calculation Firms’ annual reports and offering prospectuses As above As above As above Authors’ calculation based on Compustat data As above As above As above As above As above As above As above Authors’ calculation based on CRSP data Authors’ calculation based on Compustat data I/B/E/S

31

ACCEPTED MANUSCRIPT DISP

Q COE

LEGAL ORIGIN ANTI-SELF-DEALING JUDICIAL INDEPENDENCE POLITICAL ORIENTATION

GOVERNMENT DEFICITS

Dispersion of analyst forecasts defined as the coefficient of variation of one-year-ahead analyst forecasts of earnings per share in June of each year. Panel D: Financial Outcome Measures Market value of equity, minus book value of equity, plus book value of assets, all scaled by book value of assets. Implied cost of equity premium= (KGLS+ KCT+KOJN+KMPEG)/4. KGLS follows the Gebhardt et al. model (2001), KCT follows the Claus and Thomas model (2001), KOJN follows the Ohlson and Juettner-Nauroth model (2005), and KMPEG follows the Easton model (2004). Panel E: Country-level Institutions A dummy variable that equals 1 if the legal origin of the company law is common law, 0 if civil law. Average of ex-ante and ex-post private control of self-dealing. The average of three variables: (1) tenure of supreme court judges; and (2) tenure of the administrative court judges; and (3) case law. The index ranges from 0 to 1 with higher scores indicating more judicial independence. Party orientation with respect to economic policy. Right (1): for parties that are defined as conservative, Christian democratic, or right-wing. Left (3): for parties that are defined as communist, socialist, social democratic, or left-wing. Center (2): for parties that are defined as centrist or when party position can best be described as centrist. 0: for all those cases that do not fit into the above-mentioned categories or no information. The general government total expenditure minus revenue, as percentage of GDP.

T P

I R

C S

U N

A

D E

M

RIGHT

A dummy variable that equals 1 if party orientation is right: for parties that are defined as conservative, Christian democratic, or right-wing.

INFL

Realized inflation rate over the next year.

T P

E C

Authors’ calculation based on I/B/E/S data As above Authors’ calculation based on I/B/E/S and Compustat data La Porta et al. (1998) Djankov et al. (2008) La Porta et al. (2004) The World Bank Database of Political Institutions

Authors’ calculation based on IMF's World Economic Outlook Database. Authors’ calculation based on The World Bank Database of Political Institutions World Development Indicators

C A

32

ACCEPTED MANUSCRIPT

REFERENCES Adams, R., Almeida, H., & Ferreira, D. 2009. Understanding the relationship between founder–CEOs and firm performance. Journal of Empirical Finance, 16: 136-150. Atkinson, A., & Stiglitz, J. 1980. Lectures on public economics. Maidenhead, UK: McGraw-Hill.

IP

T

Attig, N., Boubakri, N., El Ghoul, S., & Guedhami, O. 2016. Firm internationalization and corporate social responsibility. Journal of Business Ethics, 134: 171-197.

CR

Bayliss, K. 2002. Privatization and poverty: The distributional impact of utility privatization. Annals of Public

US

and Cooperative Economics, 73: 603-625.

Beck, T., Clarke, G., Groff, A., Keefer, P., & Walsh, P. 2001. New tools in comparative political economy:

AN

The database of political institutions. World Bank Economic Review, 15: 165-176. Ben-Nasr, H., Boubakri, N., & Cosset, J.-C. 2012. The political determinants of the cost of equity: Evidence

M

from newly privatized firms. Journal of Accounting Research, 50: 605-646.

ED

Ben-Nasr, H., Boubakri, N., & Cosset, J.-C. 2015. Earnings quality in privatized firms: The role of state and foreign owners. Journal of Accounting and Public Policy, 34: 392-416.

PT

Biais, B., & Perotti, E. 2002. Machiavellian privatization. American Economic Review, 92: 240-258. Birdsall, N., & Nellis, J. 2003. Winners and losers: Assessing the distributional impact of privatization. World

CE

Development, 31: 1617-1633.

AC

Borisova, G., Brockman, P., Salas, J. M., & Zagorchev, A. 2012. Government ownership and corporate governance: Evidence from the EU. Journal of Banking & Finance, 36: 2917-2934. Borisova, G., Fotak, V., Holland, K., & Megginson, W. L. 2015. Government ownership and the cost of debt: Evidence from government investments in publicly traded firms. Journal of Financial Economics, 118: 168-191. Borisova, G., & Megginson, W. L. 2011. Does government ownership affect the cost of debt? Evidence from privatization. Review of Financial Studies, 24: 2693-2737.

33

ACCEPTED MANUSCRIPT Bortolotti, B., & Faccio, M. 2009. Government control of privatized firms. Review of Financial Studies, 22: 2907-2939. Bortolotti, B., Fantini, M., & Siniscalco, D. 2004. Privatisation around the world: Evidence from panel data. Journal of Public Economics, 88: 305-332.

T

Bortolotti, B., & Pinotti, P. 2008. Delayed privatization. Public Choice, 136: 331-351.

IP

Boubakri, N., & Cosset, J. C. 1998. The financial and operating performance of newly privatized firms:

CR

Evidence from developing countries. Journal of Finance, 53: 1081-1110.

Boubakri, N., Cosset, J.-C., & Guedhami, O. 2005a. Liberalization, corporate governance and the performance

US

of privatized firms in developing countries. Journal of Corporate Finance, 11: 767-790. Boubakri, N., Cosset, J.-C., & Guedhami, O. 2005b. Postprivatization corporate governance: The role of

AN

ownership structure and investor protection. Journal of Financial Economics, 76: 369-399. Boubakri, N., Cosset, J.-C., & Guedhami, O. 2008. Privatization in developing countries: Performance and

M

ownership effects. Development Policy Review, 26: 275-308.

ED

Boubakri, N., Cosset, J.-C., & Guedhami, O. 2009. From state to private ownership: Issues from strategic industries. Journal of Banking & Finance, 33: 367-379.

PT

Boubakri, N., Cosset, J.-C., Guedhami, O., & Saffar, W. 2011. The political economy of residual state

244-258.

CE

ownership in privatized firms: evidence from emerging markets. Journal of Corporate Finance,17:

AC

Boubakri, N., Cosset, J.-C., & Saffar, W., 2013. The role of state and foreign owners in corporate risk-taking: Evidence from privatization. Journal of Financial Economics, 108: 641-658. Boubakri, N., El Ghoul, S., Guedhami, O. and Megginson, W.L. 2018. The market value of government ownership. Journal of Corporate Finance, 50: 44-65. Boycko, M., Shleifer, A., & Vishny, R. W. 1993. Privatizing Russia. Brookings Papers on Economic Activity, 2: 139-192. Boycko, M., Shleifer, A., & Vishny, R. W. 1996. A theory of privatisation. The Economic Journal, 106: 309-319. 34

ACCEPTED MANUSCRIPT Chen, R., El Ghoul, S., Guedhami, O., & Wang, H. 2017. Do state and foreign ownership affect investment efficiency? Evidence from privatizations. Journal of Corporate Finance, 42: 408-421. Chen, R., El Ghoul, S., Guedhami, O., & Nash, R. 2018. State ownership and corporate cash holdings. Journal of Financial and Quantitative Analysis, 53: 2293-2334.

T

Cheng, B., Ioannou, I., & Serafeim, G. 2014. Corporate social responsibility and access to finance. Strategic

IP

Management Journal, 35: 1-23

CR

Chong, A., & López-de-Silanes, F. 2005. The truth about privatization in Latin America. In A. Chong & F. López-de-Silanes (Eds), Privatization in Latin America: Myths and reality: 1-66. World Bank/Stanford

US

University Press.

Claus, J., & Thomas, J., 2001. Equity premia as low as three percent? Evidence from analysts’ earnings

AN

forecasts for domestic and international stock markets. Journal of Finance, 56: 1629-1666. Cumming, D., Hou, W., & Lee, E. 2016. Business Ethics and Finance in Greater China: Synthesis and Future

M

Directions in Sustainability, CSR, and Fraud. Journal of Business Ethics, 138: 601-626.

ED

Dam, L., & Scholtens, B. 2012. Does ownership type matter for corporate social responsibility?. Corporate Governance: An International Review, 20: 233-252.

PT

Deng, X., Kang, J.K., & Low, B. S. 2013. Corporate social responsibility and stakeholder value maximization:

CE

Evidence from mergers. Journal of Financial Economics, 110: 87-109. Dewenter, K. L., & Malatesta, P. H. 1997. Public offerings of state-owned and privately-owned enterprises:

AC

An international comparison. Journal of Finance, 52: 1659-1679. Dewenter, K. L., & Malatesta, P. H. 2001. State-owned and privately owned firms: An empirical analysis of profitability, leverage, and labor intensity. American Economic Review, 91: 320-334. Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. 2011. Voluntary nonfinancial disclosure and the cost of equity capital: The initiation of corporate social responsibility reporting. Accounting Review, 86: 59-100. Dimson, E., Karakas, O., & Li, X. 2015. Active ownership. Review of Financial Studies, 28: 3225-3268. Djankov, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 2008. The law and economics of self-dealing. Journal of Financial Economics, 88: 430-465. 35

ACCEPTED MANUSCRIPT D’Souza, J., & Megginson, W. L. 1999. The financial and operating performance of privatized firms during the 1990s. Journal of Finance, 54: 1397-1438. Durnev, A., & Fauver, L. 2009. Stealing from thieves: Firm governance and performance when states are predatory. Unpublished working paper, McGill University. Available at SSRN:

T

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=970969

IP

Easton, P. D. 2004. PE ratios, PEG ratios, and estimating the implied expected rate of return on equity capital.

CR

Accounting Review, 79: 73-95.

El Ghoul, S., Guedhami, O. and Kim, Y. 2017. Country-level institutions, firm value, and the role of corporate

US

social responsibility initiatives. Journal of International Business Studies, 48: 360-385. El Ghoul, S., Guedhami, O., Kwok, C. C., & Mishra, D. R. 2011. Does corporate social responsibility affect

AN

the cost of capital? Journal of Banking & Finance, 35: 2388-2406.

El Ghoul, S., Guedhami, O., Wang, H., & Kwok, C. 2016. Family control and corporate social responsibility.

M

Journal of Banking and Finance, 73: 131-146.

ED

European Commission. 2001. Green paper: Promoting a European framework for corporate social responsibility. Brussels: Office for Official Publications of the European Communities.

CE

Finance, 61: 2597-2635.

PT

Faccio, M., Masulis, R. W., & McConnell, J. 2006. Political connections and corporate bailouts. Journal of

Fama, E., & French, K. 1997. Industry costs of equity. Journal of Financial Economics, 43: 153-193.

AC

Ferrell, A., Liang, H., & Renneboog, L. 2016. Socially responsible firms. Journal of Financial Economics, 122(3): 585-606.

Freeman, R. E. 1984. Strategic management: A stakeholder approach. Boston: Pitman. Friedman, M. 1970. The social responsibility of business is to increase its profits. New York Times, Sept. 13: 33 & 122-126. Gebhardt, W. R., Lee, C., & Swaminathan, B. 2001. Toward an implied cost of capital. Journal of Accounting Research, 39: 135-176.

36

ACCEPTED MANUSCRIPT Guedhami, O., and Pittman, J. 2006. Ownership concentration in privatized firms: The role of disclosure standards, auditor choice, and auditing infrastructure. Journal of Accounting Research, 44: 889-929. Guedhami, O., Pittman, J. A., & Saffar, W. 2009. Auditor choice in privatized firms: Empirical evidence on the role of state and foreign owners. Journal of Accounting and Economics, 48: 151-171.

T

Guedhami, O., Pittman, J. A., & Saffar, W. 2014. Auditor choice in politically connected firms. Journal of

IP

Accounting Research, 52: 107-162.

CR

Hail, L., & Leuz, C. 2006. International differences in the cost of equity capital: Do legal institutions and securities regulation matter? Journal of Accounting Research, 44: 485-531.

US

Heckman, J. J. 1979. Sample selection bias as a specification error. Econometrica, 47: 153-161. Ioannou, I., & Serafeim, G. 2012. What drives corporate social performance? The role of nation-level

AN

institutions. Journal of International Business Studies, 43: 834-864.

Jaslowitzer, P., Megginson, W. L., & Rapp, M. 2016. Disentangling the effects of state ownership on

ED

https://ssrn.com/abstract=2735698.

M

investment – Evidence from Europe. Working paper. Available at SSRN:

Jo, H., & Harjoto, M. A. 2011. Corporate governance and firm value: The impact of corporate social

PT

responsibility. Journal of Business Ethics, 103: 351-383.

CE

Kim, Y., Park, M. S., & Wier, B. 2012. Is earnings quality associated with corporate social responsibility? The Accounting Review, 87: 761-796.

AC

La Porta, R., Lopez-de-Silanes, F., Pop-Eleches, C., & Shleifer, A. 2004. Judicial checks and balances. Journal of Political Economy, 112: 445-470. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. 1998. Law and finance. Journal of Political Economy, 106: 1113-1155. Lee, D. D., & Faff, R. W. 2009. Corporate sustainability performance and idiosyncratic risk: A global perspective. Financial Review, 44: 213-237. Li, W., & Zhang, R. 2010. Corporate social responsibility, ownership structure, and political interference: evidence from China. Journal of Business Ethics, 96: 631-645. 37

ACCEPTED MANUSCRIPT Liang, H., & Renneboog, L. 2017. On the foundations of corporate social responsibility. The Journal of Finance, 72: 853-910. Maung, M., Wilson, C., & Tang, X. 2016. Political connections and industrial pollution: Evidence based on state ownership and environmental levies in China. Journal of Business Ethics,138: 649–659.

T

Megginson, W. L. 2017. Privatization, state capitalism, and state ownership of business in the 21st century.

IP

Foundations and Trends in Finance, 11(1-2): 1-153.

CR

Megginson, W. L., Nash, R. C., Netter, J. M., & Poulsen, A. B. 2004. The choice of private versus public capital markets: Evidence from privatizations. Journal of Finance, 59: 2835-2870.

US

Megginson, W. L., Nash, R. C., & van Randenborgh, M. 1994. The financial and operating performance of newly privatized firms: An international empirical analysis. Journal of Finance, 49: 403-452.

Journal of Economic Literature, 39: 321-389.

AN

Megginson, W. L., & Netter, J. M., 2001. From state to market: A survey of empirical studies on privatization.

M

Nash, R. 2017. Contracting issues at the intersection of the public and private sectors: New data and new

ED

insights. Journal of Corporate Finance, 42: 357-366. Ohlson, J. A., & Juettner-Nauroth, B. E. 2005. Expected EPS and EPS growth as determinants of value.

PT

Review of Accounting Studies, 10: 349-365.

CE

Perotti, E. 1995. Credible privatization. American Economic Review, 85: 847-859. Petersen, M. 2009. Estimating standard errors in finance panel data sets: Comparing approaches. Review of

AC

Financial Studies, 22(1): 435-480. Porter, M. E., & Kramer, M. R. 2006. The link between competitive advantage and corporate social responsibility. Harvard Business Review, 84(12): 78-92. Porter, M. E., & Kramer, M. R. 2011. Creating shared value. Harvard Business Review, 89(1/2): 62-77. Price Waterhouse. 1989a. Privatisation: Learning the lessons from the UK experience. London: Price Waterhouse. Price Waterhouse. 1989b. Privatisation: The facts. Salisbury, UK: Addison Design Company.

38

ACCEPTED MANUSCRIPT Shapiro, C., & Willig, R. 1990. Economic rationales for the scope of privatization. In E. N. Suleiman & J. Waterbury (Eds), The political economy of public sector reform and privatization. Boulder, CO: Westview Press. Shleifer, A. 1998. State versus private ownership. NBER Working Paper No. 6665.

AC

CE

PT

ED

M

AN

US

CR

IP

T

Shleifer, A., & Vishny, R. W. 1997. A survey of corporate governance. Journal of Finance, 52: 737-783.

39

100%

80%

AN

US

CR

IP

T

CSR

ACCEPTED MANUSCRIPT

60%

40%

20%

0%

M

Residual State Ownership

ED

Figure 1.

CSR intensity and the Extent of Privatization This figure describes the trade-off between the economic value enhancing motives of CSR (argument d)

PT

and political motives of CSR (distortion, cost-shifting, and reputation building—arguments a, b, and c,

AC

CE

respectively)

40

ACCEPTED MANUSCRIPT Table 1

Sample Composition

AC

CE

PT

ED

M

AN

US

CR

IP

T

This table presents the sample distribution by country, industry (Fama–French 12 industry groups), and year for the 18,816 observations representing 2,583 unique firms from 41 countries over the period 2002 to 2014. Privatized Obs Privatized Obs Firms Firms Full Sample 18,816 1,439 2,583 220 Panel A: By Country Australia 1,835 37 359 4 Austria 190 64 20 7 Belgium 267 12 26 2 Brazil 380 30 80 6 Chile 111 17 21 4 China 309 81 62 18 Colombia 43 13 9 3 Czech Republic 21 9 3 2 Denmark 191 5 19 1 Egypt 54 7 11 1 Finland 282 51 26 6 France 955 125 96 18 Germany 770 87 86 9 Greece 149 53 18 6 Hungary 19 7 3 2 India 391 42 86 11 Indonesia 159 50 34 12 Ireland 153 13 16 1 Israel 82 15 16 4 Italy 467 93 52 11 Japan 3,915 30 407 4 Kazakhstan 6 4 1 1 Korea 540 32 112 6 Malaysia 227 22 48 5 Mexico 146 3 29 1 Morocco 17 4 3 1 Netherlands 436 10 54 1 New Zealand 97 41 15 5 Norway 208 29 21 3 Philippines 115 16 26 3 Poland 143 41 29 10 Portugal 128 53 12 6 Russia 200 26 33 4 Singapore 353 63 41 7 South Africa 448 12 118 2 Spain 484 70 55 8 Sweden 519 25 50 3 Switzerland 589 13 64 1 Thailand 135 12 30 3 Turkey 152 26 27 5 UK 3,130 96 365 13 Panel B: By Industry Business Equipment 1,388 33 194 4 Chemicals 889 29 116 5 Consumer Durables 708 25 83 3

41

ACCEPTED MANUSCRIPT 6 20 28 26 33 41 4 50

42 47 67 88 90 100 134 145 152 157 157 136 124

398 418 889 1,172 1,189 1,261 1,410 1,639 1,986 2,076 2,165 2,218 1,995

42 47 67 88 90 100 134 145 152 157 157 136 124

T

182 151 527 319 135 108 216 552

AC

CE

PT

ED

M

AN

US

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

50 145 90 189 254 287 21 316

IP

1,351 939 3,762 2,672 927 805 1,613 3,762 Panel C: By Year 398 418 889 1,172 1,189 1,261 1,410 1,639 1,986 2,076 2,165 2,218 1,995

CR

Consumer Nondurables Energy Finance Manufacturing Telephone and Television Utilities Wholesale and Retail Others

42

ACCEPTED MANUSCRIPT Table 2

Descriptive Statistics and Correlation Matrix This table reports descriptive statistics for the main variables used in our analyses. Panels A and B present descriptive statistics and Pearson correlation coefficients, respectively, for the firm-level regression variables. Panel C reports the means of country-level institutional variables over the sample period. The full sample comprises 18,816 observations representing 2,583 unique firms from 41 countries over the period 2002 to 2014. Definitions and data sources for all variables are provided in the Appendix. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. Panel A: Descriptive Statistics Mean SD Min P25 P50 P75 Max CSR 56.560 29.407 6.200 27.962 61.422 85.085 97.950 EP 57.105 31.345 8.610 23.370 63.905 88.760 97.310 SP 56.015 31.230 3.480 25.340 60.940 86.320 98.960 PRIVATIZED 0.076 0.266 0.000 0.000 0.000 0.000 1.000 SIZE 8.864 1.809 1.687 7.696 8.710 9.918 15.146 ROA 0.041 0.069 -0.250 0.010 0.035 0.069 0.279 RDS 0.012 0.028 0.000 0.000 0.000 0.008 0.175 LEV 0.176 0.147 0.000 0.051 0.151 0.261 0.626 STD_ROA 0.040 0.104 0.001 0.009 0.020 0.041 3.313 STD_RET 2.205 1.030 0.804 1.509 1.949 2.594 6.445 Panel B: Correlation Matrix (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) CSR (1) 1.000 EP (2) 0.940*** 1.000 SP (3) 0.940*** 0.767*** 1.000 PRIVATIZED (4) 0.114*** 0.087*** 0.127*** 1.000 SIZE (5) 0.406*** 0.385*** 0.379*** 0.106*** 1.000 ROA (6) 0.009 -0.020** 0.037*** 0.033*** -0.171*** 1.000 RDS (7) 0.117*** 0.146*** 0.074*** -0.032*** -0.087*** -0.011 1.000 LEV (8) 0.090*** 0.080*** 0.089*** 0.095*** 0.098*** -0.152*** -0.151*** 1.000 STD_ROA (9) -0.132*** -0.128*** -0.121*** -0.027*** -0.256*** -0.041*** 0.044*** -0.059*** 1.000 STD_RET (10) -0.116*** -0.097*** -0.121*** -0.066*** -0.148*** -0.291*** 0.054*** -0.015* 0.184*** 1.000

T P

I R

C S

U N

A

D E

M

T P

E C

C A

43

ACCEPTED MANUSCRIPT Table 2–Continued

T

IP

CR

US

AN

M

ED

PT

CE

AC

Australia Austria Belgium Brazil Chile China Colombia Czech Republic Denmark Egypt Finland France Germany Greece Hungary India Indonesia Ireland Israel Italy Japan Kazakhstan Korea Malaysia Mexico Morocco Netherlands New Zealand Norway Philippines Poland Portugal Russia Singapore South Africa Spain Sweden Switzerland Thailand Turkey UK Average

Panel C: Means of Country-level Institutions across Years LEGAL JUDICIAL POLITICAL GOVERNMENT ANTI-SELF-DEALING ORIGIN INDEPENDENCE ORIENTATION DEFICITS 1 0.790 1.000 2.397 -1.089 0 0.209 0.670 2.358 -2.106 0 0.540 0.670 1.532 -2.438 0 0.291 0.670 3.000 -4.356 0 0.625 0.670 2.385 -1.148 0 0.778 0.000 3.000 -0.370 0 0.576 0.330 0.000 -2.284 0 0.340 . 0.000 -3.568 0 0.466 1.000 1.555 -2.062 0 0.491 0.670 0.000 -9.867 0 0.460 1.000 1.798 0.662 0 0.382 0.330 1.450 -3.862 0 0.279 1.000 1.418 -1.019 0 0.225 0.670 1.771 -8.510 0 0.204 . 1.737 . 1 0.549 1.000 3.000 -8.285 0 0.683 1.000 0.000 -1.493 1 0.787 1.000 1.725 -5.453 1 0.714 1.000 1.000 -4.163 0 0.385 0.670 1.328 -3.367 0 0.483 0.670 0.530 -5.706 0 0.480 . 1.667 2.657 0 0.461 0.670 1.037 1.237 1 0.948 1.000 0.000 -3.660 0 0.178 0.330 1.315 -4.503 0 0.569 . 0.000 -3.346 0 0.209 0.670 1.220 -0.866 1 0.950 1.000 1.784 -0.521 0 0.435 1.000 2.317 -4.101 0 0.237 1.000 2.000 -0.510 0 0.300 . 1.734 -4.490 0 0.486 0.670 2.188 -5.223 0 0.476 . 1.690 0.152 1 1.000 1.000 0.000 6.369 1 0.814 1.000 3.000 -3.090 0 0.370 0.670 2.157 -3.927 0 0.340 1.000 1.667 0.602 0 0.267 0.670 . 0.258 1 0.849 0.670 0.000 -0.457 0 0.426 1.000 0.000 -2.939 1 0.927 1.000 2.324 -5.150 0.364 0.576 0.801 1.589 -2.580

44

ACCEPTED MANUSCRIPT Table 3

Univariate Tests This table presents univariate test results for differences between privatized firms and other publicly listed firms. The full sample comprises 18,816 observations representing 2,583 unique firms from 41 countries over the period 2002 to 2014. Definitions and data sources for all variables are provided in the Appendix. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. All Firms Privatized Firms Other Firms t-statistics of Mean Differences N Mean SD N Mean SD N Mean SD CSR 18,816 56.560 29.407 1,439 68.181 25.330 17,377 55.598 29.516 15.700*** EP 18,816 57.105 31.345 1,439 66.571 26.548 17,377 56.321 31.583 11.965*** SP 18,816 56.015 31.230 1,439 69.791 27.332 17,377 54.875 31.261 17.554*** SIZE 18,816 8.864 1.809 1,439 9.528 1.444 17,377 8.809 1.826 14.559*** ROA 18,816 0.041 0.069 1,439 0.049 0.058 17,377 0.040 0.070 4.543*** RDS 18,816 0.012 0.028 1,439 0.008 0.026 17,377 0.012 0.028 -4.366*** LEV 18,816 0.176 0.147 1,439 0.224 0.146 17,377 0.172 0.147 13.032*** STD_ROA 18,816 0.040 0.104 1,439 0.031 0.056 17,377 0.041 0.107 -3.667*** STD_RET 18,816 2.205 1.030 1,439 1.968 0.891 17,377 2.224 1.038 -9.098***

T P

I R

C S

U N

A

D E

M

T P

E C

C A

45

ACCEPTED MANUSCRIPT Table 4

Privatization and CSR: Main Evidence

AC

CE

PT

ED

M

AN

US

CR

IP

T

This table reports results for regressions of CSR intensity on privatization. The full sample comprises 18,816 observations representing 2,583 unique firms from 41 countries over the period 2002 to 2014. The dependent variables are a firm’s overall CSR intensity (CSR), environmental score (EP), and social score (SP). Definitions and data sources for the variables are provided in the Appendix. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. CSR EP SP (1) (2) (3) PRIVATIZED 5.848*** 6.018*** 5.679*** (4.356) (4.276) (3.748) SIZE 9.458*** 9.227*** 9.689*** (33.020) (31.181) (30.071) ROA 13.659*** 9.783* 17.534*** (2.782) (1.784) (3.364) RDS 56.688*** 57.386** 55.990*** (2.718) (2.533) (2.608) LEV 4.497 7.345** 1.649 (1.627) (2.382) (0.588) STD_ROA -4.448** -2.714 -6.183*** (-1.992) (-1.088) (-2.684) STD_RET -1.053*** -0.989** -1.116*** (-2.837) (-2.487) (-2.759) Constant -7.109** -14.867*** 0.649 (-2.430) (-4.693) (0.197) Country FE Yes Yes Yes Industry FE Yes Yes Yes Year FE Yes Yes Yes Observations 18,816 18,816 18,816 Adjusted R2 0.442 0.418 0.409

46

ACCEPTED MANUSCRIPT Table 5

Privatization and CSR: Propensity Score Matching

AC

CE

PT

ED

M

AN

US

CR

IP

T

This table reports univariate tests and regression results by propensity score matching samples. Propensity scores are estimated using Probit regressions of treatment status on all control variables (Panel A) or on firm size and industry fixed effects (Panel B). The dependent variables in Panels A2 and B2 are a firm’s overall CSR intensity (CSR), environmental score (EP), and social score (SP). The matched sample contains 2,878 observations. Definitions and data sources for all variables are provided in the Appendix. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. Panel A: PSM By All Control Variables Panel A1: Univariate Tests Privatized Firms Matched Firms t-statistics of Mean Differences N Mean SD N Mean SD CSR 1,439 68.181 25.330 1,439 62.615 29.159 5.467*** EP 1,439 66.571 26.548 1,439 61.598 31.174 4.607*** SP 1,439 69.791 27.332 1,439 63.632 29.931 5.764*** SIZE 1,439 9.528 1.444 1,439 9.513 1.572 0.252 ROA 1,439 0.049 0.058 1,439 0.047 0.057 0.967 RDS 1,439 0.008 0.026 1,439 0.010 0.027 -1.471 LEV 1,439 0.224 0.146 1,439 0.229 0.157 -0.836 STD_ROA 1,439 0.031 0.056 1,439 0.032 0.096 -0.653 STD_RET 1,439 1.968 0.891 1,439 2.016 0.865 -1.466 Panel A2: Regression Results CSR EP SP (1) (2) (3) PRIVATIZED 4.819*** 4.354*** 5.284*** (3.732) (3.250) (3.552) SIZE 8.635*** 9.105*** 8.166*** (13.611) (13.504) (11.688) ROA 30.956*** 27.127*** 34.784*** (3.572) (3.160) (3.235) RDS -50.155** -48.869* -51.442* (-1.988) (-1.780) (-1.809) LEV -1.819 -0.907 -2.732 (-0.358) (-0.169) (-0.490) STD_ROA -8.127 -3.459 -12.794** (-1.380) (-0.475) (-2.545) STD_RET -2.549*** -1.691** -3.407*** (-3.283) (-2.089) (-4.024) Constant -6.327 -16.083* 3.430 (-0.827) (-1.935) (0.438) Country FE Yes Yes Yes Industry FE Yes Yes Yes Year FE Yes Yes Yes Observations 2,878 2,878 2,878 Adjusted R2 0.478 0.479 0.410

47

ACCEPTED MANUSCRIPT

IP

T

t-statistics of Mean Differences 6.241*** 4.641*** 7.042*** 2.276** 4.384*** 0.218 0.676 -1.370 -2.865*** SP (3) 4.451*** (2.765) 9.135*** (15.772) 18.686* (1.841) 13.677 (0.408) -7.298 (-1.312) -8.035 (-1.348) -3.284*** (-4.071) -0.174 (-0.024) Yes Yes Yes 2,878 0.393

AC

CE

PT

ED

M

AN

US

CR

Panel B: PSM By Size and Industry Panel B1: Univariate Tests Privatized Firms Matched Firms N Mean SD N Mean SD CSR 1,439 68.181 25.330 1,439 61.933 28.290 EP 1,439 66.571 26.548 1,439 61.679 29.896 SP 1,439 69.791 27.332 1,439 62.188 30.502 SIZE 1,439 9.528 1.444 1,439 9.401 1.535 ROA 1,439 0.049 0.058 1,439 0.039 0.062 RDS 1,439 0.008 0.026 1,439 0.008 0.020 LEV 1,439 0.224 0.146 1,439 0.220 0.155 STD_ROA 1,439 0.031 0.056 1,439 0.035 0.105 STD_RET 1,439 1.968 0.891 1,439 2.066 0.951 Panel B2: Regression Results CSR EP (1) (2) PRIVATIZED 5.196*** 5.940*** (3.408) (3.485) SIZE 9.638*** 10.140*** (19.504) (20.184) ROA 18.658** 18.630** (2.124) (2.177) RDS 13.775 13.872 (0.418) (0.371) LEV -4.548 -1.798 (-0.866) (-0.328) STD_ROA -7.074 -6.112 (-1.379) (-1.297) STD_RET -2.425*** -1.566** (-3.593) (-2.150) Constant -11.462* -22.750*** (-1.835) (-3.505) Country FE Yes Yes Industry FE Yes Yes Year FE Yes Yes Observations 2,878 2,878 Adjusted R2 0.450 0.455

48

ACCEPTED MANUSCRIPT Table 6

Privatization and CSR: Endogeneity

AC

CE

PT

ED

M

AN

US

CR

IP

T

This table reports results of IV and Heckman selection analyses that address the endogeneity of privatization decisions. The full sample comprises 18,816 observations representing 2,583 unique firms from 41 countries over the period 2002 to 2014. The instrumental variable is RIGHT. Definitions and data sources for all variables are provided in the Appendix. All regressions control industry and year fixed effects. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. (1) (2) IV Heckman PRIVATIZED 35.509*** 25.235*** (5.965) (3.097) SIZE 8.008*** 8.192*** (48.744) (25.120) ROA 20.251*** 21.507*** (6.584) (3.351) RDS 73.912*** 75.335*** (8.860) (3.588) LEV 14.234*** 14.615*** (10.423) (4.567) STD_ROA 0.889 0.840 (0.464) (0.250) STD_RET -1.058*** -1.210** (-4.156) (-2.144) λ -11.132*** (-2.660) Constant -17.279*** -17.841*** (-11.661) (-5.482) Industry FE Yes Yes Year FE Yes Yes Observations 18,816 18,816 Adjusted R2 0.243 0.315

49

ACCEPTED MANUSCRIPT Table 7

The Role of Residual State Ownership: Fully versus Partially Privatized Firms This table compares CSR intensity between fully privatized firms and partially privatized firms. The sample of privatized firms comprises 1,439 observations representing 220 unique privatized firms from 41 countries over the period 2002 to 2014. Panel A presents the univariate tests of differences between fully and partially privatized firms. Panel B presents results for regressions of CSR intensity on partial privatization. The dependent variables in Panel B are a firm’s overall CSR intensity (CSR), environmental score (EP), and social score (SP). Definitions and data sources for all variables are provided in the Appendix. All firm-level control variables are winsorized at the 1st and 99th percentiles. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% level, respectively. Panel A: Univariate Tests All Privatized Firms Partially Privatized Firms Fully Privatized Firms t-statistics of Mean Differences N Mean SD N Mean SD N Mean SD CSR 1,439 68.181 25.330 801 69.532 24.982 638 66.484 25.678 2.271** EP 1,439 66.571 26.548 801 67.503 26.374 638 65.399 26.739 1.494 SP 1,439 69.791 27.332 801 71.561 27.011 638 67.569 27.589 2.758*** SIZE 1,439 9.528 1.444 801 9.745 1.385 638 9.254 1.472 6.494*** ROA 1,439 0.049 0.058 801 0.051 0.062 638 0.046 0.053 1.481 RDS 1,439 0.008 0.026 801 0.007 0.020 638 0.010 0.031 -2.378** LEV 1,439 0.224 0.146 801 0.204 0.132 638 0.250 0.157 -6.006*** STD_ROA 1,439 0.031 0.056 801 0.027 0.029 638 0.034 0.077 -2.376** STD_RET 1,439 1.968 0.891 801 2.042 0.889 638 1.874 0.885 3.562*** YRS_PRIV 1,439 12.187 6.650 801 11.195 6.334 638 13.433 6.831 -6.430***

T P

I R

C S

U N

A

D E

M

T P

E C

C A

50

ACCEPTED MANUSCRIPT Table 7–Continued

SIZE ROA RDS LEV

CR

STD_ROA

T

PARTIALLY

EP (2) 3.254* (1.679) 7.423*** (6.989) 8.238 (0.605) -21.959 (-0.493) -3.188 (-0.426) -2.547 (-0.261) -0.520 (-0.514) 0.495*** (2.810) -9.566 (-0.862) Yes Yes Yes 1,439 0.556

IP

Panel B: Regression Results CSR (1) 4.019** (1.972) 6.928*** (7.020) 17.169 (1.042) -32.997 (-0.946) -0.935 (-0.114) -8.153 (-0.766) -1.277 (-1.406) 0.483*** (2.650) 1.109 (0.111) Yes Yes Yes 1,439 0.549

STD_RET

US

YRS_PRIV

AN

Constant

AC

CE

PT

ED

M

Country FE Industry FE Year FE Observations Adjusted R2

SP (3) 4.785* (1.902) 6.433*** (5.418) 26.100 (1.249) -44.035 (-1.156) 1.318 (0.132) -13.758 (-1.042) -2.033* (-1.827) 0.471** (2.175) 11.783 (1.002) Yes Yes Yes 1,439 0.474

51

ACCEPTED MANUSCRIPT Table 8

State Ownership, Institutional Environment, and CSR This table reports regressions of CSR intensity on state ownership and square of state ownership in split samples by whether firms are located in a strong (above country median) or weak (below country median) institutional environment. Definitions and data sources for the variables are provided in the Appendix. All firm-level control variables are winsorized at the 1st and 99th percentiles. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. The superscript asterisks ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. JUDICIAL POLITICAL GOVERNMENT LEGAL ORIGIN ANTI-SELF-DEALING INDEPENDENCE ORIENTATION DEFICITS All Civil Common Low High Low High Left/Center Right Low High (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) STATE 0.273** 0.308** 0.543* 0.301* 0.256 0.284** 0.009 0.434** 0.189 0.199 0.418** (2.156) (2.206) (1.774) (1.846) (1.071) (2.012) (0.040) (2.332) (1.232) (1.162) (2.214) STATE_SQUARE -0.003* -0.005** -0.006 -0.005** -0.002 -0.004** 0.000 -0.005* -0.003* -0.002 -0.006** (-1.788) (-2.434) (-1.373) (-2.262) (-0.550) (-2.117) (0.106) (-1.675) (-1.709) (-0.871) (-2.074) SIZE 6.976*** 6.331*** 6.891** 6.421*** 6.102*** 5.496*** 9.017*** 8.097*** 9.521*** 5.510*** 7.633*** (7.015) (5.889) (2.268) (5.724) (2.966) (5.780) (5.272) (6.265) (6.747) (4.011) (5.004) ROA 16.726 7.914 35.292 -9.179 48.779** -32.788 23.439 -9.595 -16.996 5.305 29.098 (1.017) (0.410) (1.260) (-0.550) (2.283) (-1.487) (1.188) (-0.483) (-0.669) (0.263) (1.402) RDS -36.285 -84.751*** -21.928 -76.185** 104.898 -53.389** 39.687 24.737 -43.755 -1.397 -155.102** (-1.017) (-3.013) (-0.173) (-2.497) (0.959) (-2.049) (0.459) (0.463) (-1.120) (-0.031) (-2.276) LEV -2.334 -0.696 -4.450 -4.219 -3.685 -5.422 15.141 3.142 -0.842 -5.255 3.416 (-0.283) (-0.086) (-0.314) (-0.434) (-0.350) (-0.623) (1.191) (0.245) (-0.064) (-0.468) (0.314) STD_ROA -8.656 8.463 2.566 -35.556 3.183 35.138 7.799 -1.547 4.974 -6.912 -18.602 (-0.820) (0.273) (0.210) (-1.293) (0.259) (1.162) (0.615) (-0.140) (0.151) (-0.587) (-0.462) STD_RET -1.210 0.426 -4.395** 0.867 -2.554 0.518 -2.840 -0.777 -0.811 -0.214 -2.079* (-1.359) (0.549) (-2.352) (0.887) (-1.615) (0.719) (-1.605) (-0.561) (-0.580) (-0.189) (-1.715) YRS_PRIV 0.458** 0.194 0.735* 0.132 0.813*** 0.008 0.515* 0.938*** -0.147 0.763*** 0.174 (2.489) (0.919) (1.944) (0.545) (2.820) (0.036) (1.932) (3.166) (-0.607) (2.967) (0.664) Constant 0.122 -49.569*** -14.834 14.875 -1.419 -43.288*** -18.501 -23.124* -22.744 -0.231 -2.622 (0.012) (-4.903) (-0.475) (1.596) (-0.071) (-4.345) (-0.942) (-1.815) (-1.567) (-0.018) (-0.185) Country FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Industry FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 1,439 1,086 353 815 624 718 630 644 536 716 723 Adjusted R-squared 0.550 0.611 0.408 0.515 0.539 0.710 0.420 0.590 0.494 0.511 0.585 Inflection Point 45.5 30.8 30.1 35.5 43.4 34.8

T P

I R

C S

U N

A

D E

M

T P

E C

C A

52

ACCEPTED MANUSCRIPT Table 9

Financial Outcomes of CSR in PFs

AC

CE

PT

ED

M

AN

US

CR

IP

T

This table reports regression results of firm performance on ownership, predicted CSR (PCSR), and the interaction between ownership and predicted CSR in PFs. The dependent variable is Tobin’s Q (Q) in Panel A, and cost of capital (COE) in Panel B. PCSR, PEP and PSP are predicted CSR, environmental, social score, respectively. STATE*PCSR (STATE*PEP, STATE*PSP) is divided by 1000. Definitions and data sources for the variables are provided in the Appendix. All models include year and industry fixed effects. Robust t-statistics adjusted for clustering by firm and year are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Panel A: Tobin’s Q (1) (2) (3) STATE -0.007** -0.006* -0.007** (-2.035) (-1.914) (-2.072) PCSR 0.004* (1.660) STATE*PCSR 0.089** (2.132) PEP 0.003 (1.250) STATE*PEP 0.082* (1.953) PCSR 0.005* (1.870) STATE*PSP 0.092** (2.188) SIZE -0.166*** -0.156*** -0.170*** (-5.909) (-5.663) (-6.060) ROA 5.217*** 5.277*** 5.141*** (6.665) (6.669) (6.617) RDS 1.600 1.731* 1.562 (1.530) (1.695) (1.449) LEV 0.173 0.199 0.169 (0.942) (1.102) (0.918) STD_ROA -0.061 -0.011 -0.090 (-0.205) (-0.035) (-0.309) STD_RET -0.055 -0.057 -0.055 (-1.348) (-1.366) (-1.356) CASH 1.586*** 1.585*** 1.569*** (4.738) (4.680) (4.702) CAPX 0.898 0.909 0.863 (1.453) (1.467) (1.405) YRS_PRIV -0.010*** -0.010*** -0.010*** (-2.722) (-2.640) (-2.726) ANTI-SELF-DEALING 0.039 0.027 0.039 (0.363) (0.252) (0.364) Constant 2.328*** 2.304*** 2.324*** (8.786) (8.600) (8.851) Industry FE Yes Yes Yes Year FE Yes Yes Yes Observations 1,359 1,359 1,359 Adjusted R2 0.549 0.544 0.552

53

ACCEPTED MANUSCRIPT Table 9–Continued

PCSR STATE*PCSR PEP STATE*PEP

CR

PCSR

Q LEV

M

LTG

ED

DISP YRS_PRIV

AC

Industry FE Year FE Observations Adjusted R2

CE

ANTI-SELF-DEALING

PT

INFL

Constant

AN

LOG_MVE

1.437*** (3.319) -0.241 (-0.802) -0.750* (-1.721) -1.011 (-0.778) 0.042*** (3.280) 4.213*** (3.841) -0.004 (-0.125) 8.625 (0.870) 0.751 (0.951) 11.197*** (4.609) Yes Yes 879 0.341

1.433*** (3.324) -0.201 (-0.675) -0.774* (-1.768) -0.886 (-0.668) 0.042*** (3.282) 4.234*** (3.845) -0.002 (-0.065) 8.593 (0.864) 0.663 (0.829) 11.162*** (4.582) Yes Yes 879 0.341

US

STATE*PSP BETA

(3) 0.052** (2.295)

T

STATE

IP

Panel B: Cost of Equity Capital (1) (2) 0.053** 0.051** (2.431) (2.529) 0.026 (1.255) -0.579** (-2.004) 0.022 (1.116) -0.579** (-2.043)

0.028 (1.370) -0.556* (-1.904) 1.440*** (3.321) -0.269 (-0.908) -0.741* (-1.702) -1.086 (-0.844) 0.042*** (3.277) 4.195*** (3.834) -0.006 (-0.167) 8.292 (0.841) 0.821 (1.052) 11.253*** (4.659) Yes Yes 879 0.341

54

ACCEPTED MANUSCRIPT Highlights 

Study examines CSR activities of privatized firms and the role of residual government ownership. Privatized firms show better CSR intensity than other publicly listed firms.



Nonlinear relation between residual state ownership and CSR intensity.



Country-level institutions affect the state ownership–CSR performance relation.



Privatized firms benefit from higher valuation and lower cost of equity through improved

CR

IP

T



AC

CE

PT

ED

M

AN

US

CSR.

55