6. Empirical Results of Model

Chapter 5 finds that the present study’s empirical discretion model fulfils every single assessment test of statistical conclusion validity, reliability, construct validity, internal validity, and external validity along a cascading hierarchy of tests, even when faced with particularly conservative thresholds from the literature.

As shown in Fig. 3.​3 in Sect.​ 3.​2.​2, nearly 30,000 firms with 300–2,000 employees are recorded in official government statistics in the manufacturing sector in China. As the sample of the present study applies to manufacturing firms with 150–5,000 employees, the number of comparable firms in the theoretical population is expected to be even larger. Moreover, with generally one or more plant managers per firm, the number of plant managers is expected to be larger yet, and is therefore here expressed as ‘tens of thousands’.

While these sample parameters can be interpreted in terms of the direction and magnitude of the impact of discretion on performance when top management at corporate headquarters in China adjust the plant manager’s managerial discretion, the scope of this interpretation is limited to the particular sample of 467 firms rather than the entire population of all comparable firms, because despite the sample’s representativeness, sampling error (i.e. unaccountable chance differences among sampling units) may prevent the sample parameters from the 467 firms to accurately estimate the population parameters for all relevant firms (see Sects.​ 4.​3.​3 and 5.​5). Section 6.2 overcomes this limitation by accounting for sampling error by means of hypothesis tests so as to generalise the empirical results from this section to the entire relevant population of the unit of analysis (Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Wooldridge 2002, pp. 298–299).

See Sect.​ 4.​3.​1 (e.g. Arnold 1982; Baron and Kenny 1986, p. 1174; Carte and Russell 2003, pp. 480–495; Chin et al. 1996, pp. 25–27, 2003, pp. 196–199; Cohen 1978; Cronbach 1987; Eggert et al. 2005, p. 103; Gentle 2007, pp. 479–491; Gujarati 2004, pp. 77, 310, 516; Harville 2008, pp. 1–10; Henseler and Fassott 2010, pp. 714–733; Iacobucci and Duhachek 2003, pp. 4–6; Irwin and McClelland 2001, p. 105; Jaccard and Turrisi 2003, pp. 1–2; Kenny and Judd 1984; Knapp 2007, pp. xxi–xxiv; Rigdon et al. 1998, p. 1; Simon 1954, pp. 477–478; Venkatraman 1989, pp. 424–428; Wilson and Henseler 2006).

Section 4.​3.​2 further notes that as all constructs are standardised to means of zero and variances of unity, the parameters in Table 6.1 measure the number of standard deviations by which a dependent variable is predicted to change when an independent variable is increased by one standard deviation. For example, for an average-sized Chinese firm (A ₂ = 0), the impact of new product introduction discretion (D ₃) on performance (P) is reduced to the direct effect (d ₃), as \( {{\left. {{{{\partial P}} \left/ {{\partial {D_3}}} \right.}} \right|}_{{{A_2}=0}}}={d_3}+{m_{3,2 }}\cdot 0={d_3}=-\mathbf{0}.\mathbf{23} \). Hence, increasing the new product introduction discretion of a plant manager in the average-sized Chinese firm in the sample by one standard deviation is predicted to reduce performance by 0.23 standard deviations (\( {d_3}=-\mathbf{0}.\mathbf{23} \)).

The exact or observed level of significance in terms of p-values (see Sect.​ 4.​3.​3) is reported in Sect. 6.2 for parameters relating to this study’s hypotheses and in Chap.​ 5 for parameters used for assessing validity.

See Sect.​ 4.​3.​3 (e.g. Ahuja and Thatcher 2005, p. 446; Boßow-Thies and Albers 2010, p. 595; Chin 1998, pp. 318–320, 2000, pp. 1–2, 2003; Chin and Dibbern 2010; Chin and Newsted 1999, pp. 309–314; Dibbern and Chin 2005, pp. 146–159; Efron and Gong 1983, pp. 40–46; Efron and Tibshirani 1993, pp. 145–147; Esposito Vinzi et al. 2010, pp. 48–60; Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Greene 2003, pp. 892–896; Gujarati 2004, pp. 65–139; Hayashi 2000, pp. 3–45; Henseler 2007, pp. 104–107; Henseler and Fassott 2010, pp. 721–730; Herrmann et al. 2006, p. 59; Hsieh et al. 2008, p. 109; Keil et al. 2000, pp. 312–315; Krafft et al. 2005, p. 83; Lohmöller 1989, p. 49; Northrop and Arsenault 2007, p. 231; Qureshi and Compeau 2009, pp. 197–208; Salvatore and Reagle 2002, pp. 87–95; Scholderer and Balderjahn 2005, pp. 88–91; Spanos 1986, pp. 213–311; Temme et al. 2006, p. 11; Tenenhaus et al. 2005, p. 176; Thas 2010, pp. 335–337; Venkatesh and Morris 2000, p. 126; Wooldridge 2002, pp. 116–299; Zhu et al. 2006, pp. 529–530).

For the plant manager of small and medium-sized enterprises in the manufacturing sector in mainland China at the time of investigation, Figs. 6.1 and 6.2 demonstrate how the new coherent model for investigating the impact of managerial discretion on performance integrates managerial discretion theory, principal-agent theory, and stewardship theory (see Chap.​ 2) in a way that takes greater granularity into account in order to work towards resolving the discretion puzzle (see Chap.​ 1). The sample is based on 467 detailed ‘double-blind’ interviews with plant managers of Chinese firms and multinationals in China for which Chap.​ 5 has exhaustively demonstrated strong validity and reliability. The parameters are again displayed to two decimal places and appear in bold face if statistically significant at the 10 % level.

While the structural relationships in the empirical discretion model are described below in terms of constructs impacting on other constructs, as explained in Box 1.1 in Sect.​ 1.​2 on the research objective’s delimitations, the present study does not aim to conclusively prove that the estimated relationships represent causal relationships with a particular direction of causality, given that demonstrating causality is a general caveat of observational cross-sectional studies (e.g. Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sánchez 2008, p. 5; Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sect.​ 5.​4.​1).

Managerial discretion theory (e.g. Abrahamson and Hambrick 1997; Berman et al. 2005; Carpenter and Golden 1997; Caza 2007, 2011; Huiyuan Chen 2006; Crossland 2007; Crossland and Hambrick 2007; Datta et al. 2003; Finkelstein and Boyd 1998; Finkelstein and Hambrick 1990; Finkelstein and Peteraf 2007; Haleblian and Finkelstein 1993; Hambrick and Abrahamson 1995; Hambrick and Finkelstein 1987; Hambrick et al. 1993; Hutzschenreuter and Kleindienst 2007; Keegan 2006; Keegan and Kabanoff 2008; Key 2002; Yougen Li and Zhao 2004; Magnan and St-Onge 1997; Quigley and Hambrick 2009; Rajagopalan and Finkelstein 1992; Thomas and Peyrefitte 1996; Zhang and Li 2008b; Zhang et al. 2006a, b).

As described in Sect.​ 2.​3.​1, managerial discretion theory was developed as a reconciliation of population ecology (e.g. Aldrich 1979; Baum 1996; Baum and Amburgey 2002; Carroll 1988; Freeman et al. 1983; Hannan and Freeman 1977, 1984; Singh and Lumsden 1990; Tushman and Romanelli 1985; Zohar and Luria 2005) and strategic choice theory (e.g. Child 1972, 1997, 2002; Child et al. 2003; Elbanna and Child 2007; Hitt and Tyler 1991; Hrebiniak and Joyce 1985; Judge and Zeithaml 1992; Marlin et al. 1994; Miles and Snow 1978; Stienstra et al. 2004).

A ‘Control Effect (c)’ of firm size on performance was expected as part of the theoretical discretion model derived from the literature (e.g. Baum 1996; Berman et al. 2005, p. B5; Chang and Wong 2003, p. 21; Datta et al. 2003, p. 101; Gammelgaard et al. 2010, p. 9; Haleblian and Finkelstein 1993, p. 853; Khanchel 2009, p. 100; Misangyi 2002, p. 62; Singh and Lumsden 1990; Xu et al. 2005, p. 9; see Fig. 4.​6 in Sect.​ 4.​1.​2).

A ‘Control Effect (c)’ of industry technology intensity on performance is likewise consistent with the literature embedded in the theoretical discretion model (e.g. Aldrich 1979; Baum 1996; Baum and Amburgey 2002; Carroll 1988; Chang and Wong 2003, p. 21; Cheng et al. 2006, p. 355; Datta et al. 2003, p. 101; Freeman et al. 1983; Gammelgaard et al. 2010, p. 9; Hannan and Freeman 1977, 1984; Khanchel 2009, p. 100; Singh and Lumsden 1990; Tushman and Romanelli 1985; Werner and Tosi 1995, p. 1678; Zohar and Luria 2005). For example, some scholars have posited that performance may be stronger in higher-technology industries than in lower-technology industries (see Sect.​ 5.​3.​3; e.g. Albors et al. 2008; Ang 2008; Bound et al. 1984; Griliches and Mairesse 1990; Hausman et al. 1984; Sarkar et al. 2006).

A ‘Direct Effect (a)’ of firm size on the level of discretion was integrated into the theoretical discretion model based on the literature (e.g. Caza 2007, pp. 30–56; Hambrick and Finkelstein 1987, pp. 378–389; Walters 1995; Zhang et al. 2006a). There are various studies in support of both positive effects of firm size on discretion (e.g. Caza 2007, p. 19; Galbraith 1967; Kaysen 1960, p. 90; Reid 1968; Williamson 1963, p. 1032) and negative effects (e.g. Aldrich 1979; Baum 1996; Finkelstein and Hambrick 1990, p. 493; Hambrick and Finkelstein 1987, pp. 379–384; Mintzberg 1978; Tushman and Romanelli 1985; Zohar and Luria 2005)—both of which imply that the effect of firm size on discretion can differ from zero.

The ‘Direct Effect (a)’ of industry technology intensity on the level of discretion is derived from studies finding that the industry in which a manager works can affect the manager’s level of perceived managerial discretion (e.g. Cheng et al. 2006; Finkelstein and Boyd 1998; Hambrick and Abrahamson 1995; Hambrick and Finkelstein 1987, pp. 378–389; Lieberson and O’Connor 1972; Luce and Raiffa 1957; Porter 1980, p. 230).

Principal-agent theory (e.g. Agrawal and Knoeber 1996; Baysinger and Butler 1985; Berger et al. 1997; Brush et al. 2000; Chang and Wong 2003; Childs and Mauer 2008; Denis et al. 1997; Eisenhardt 1989; Fama 1980; Fama and Jensen 1983a, b; He et al. 2009; Jensen 1986; Jensen and Meckling 1976; Jensen and Murphy 1990; Jensen and Ruback 1983; Laffont and Martimort 2002; Lang et al. 1995; Levinthal 1988; Ongore 2011; Shleifer and Vishny 1997; Spremann 1987; Thépot 2007; Thomsen and Pedersen 2000; Walters 1995; Wang et al. 2008; Weidenbaum and Jensen 1993; Werner and Tosi 1995, p. 1673; Xu et al. 2005; Zou 1989).

Stewardship theory (e.g. Albanese et al. 1997; Arthurs and Busenitz 2003; Corbetta and Salvato 2004; Davis et al. 1997a, b; Dicke and Ott 2002; Donaldson 1990; Donaldson and Davis 1989, 1991, 1993, 1994; Eddleston and Kellermanns 2007; Fox and Hamilton 1994; Lane et al. 1999; Liu and Cai 2004; Miller and Le Breton-Miller 2006; Mills and Keast 2009; Muth and Donaldson 1998; Salvato 2002; Tian and Lau 2001; Tosi et al. 2003; Van Slyke 2007; Vargas Sánchez 2001, 2004, 2005; Zahra 2003).

Specifically, granting a plant manager in an average-sized multinational an additional standard deviation of hiring discretion (D ₂) leads to a statistically significant reduction in performance (P) by \( {d_2}=-\mathbf{0}.\mathbf{23} \) standard deviations. Granting a plant manager in an average-sized Chinese firm an additional standard deviation of new product introduction discretion (D ₃) likewise leads to a statistically significant drop in performance (P) by \( {d_3}=-\mathbf{0}.\mathbf{23} \) standard deviations. The performance impact of discretion within the other domains is not found to be significant in either average-sized Chinese firms (\( {d_2}=-0.04 \) and \( {d_4}=-0.04 \)) or average-sized multinationals (\( {d_3}=-0.05 \) and \( {d_4}=-0.00 \)).

See Box 2.6 in Sect.​ 2.​4.​2 (e.g. Aminpour and Woetzel 2006, pp. 46–47; Hanne Chen 2004, p. 128; Crossland and Hambrick 2007, p. 14; Davis et al. 1997b, p. 42; Hexter 2006, pp. 4–6; Hexter and Woetzel 2007a, p. 4; Hofstede 1980, 1991; Hoover 2006, p. 92; McGregor 2005, pp. 273–274).

Similarly, granting a plant manager a standard deviation of capital investment discretion (D ₁) in a Chinese firm that is larger than average by two standard deviations (\( {A_2}=2 \)) is predicted to have an even stronger impact on performance of +0.64 standard deviations, because \( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=2}}^{Chinese }=\left( {+\mathbf{0}.\mathbf{40}} \right)+\left( {+\mathbf{0}.\mathbf{12}} \right)\cdot (2) \). By contrast, in an average-sized multinational, the positive performance impact of capital investment discretion (\( {d_1}=+\mathbf{0}.\mathbf{41} \), which is similar to that in an average-sized Chinese firm, \( {d_1}=+\mathbf{0}.\mathbf{40} \)) is dampened according to the ‘Moderating Effect (m)’ of \( {m_{1,2 }} = -\mathbf{0}.\mathbf{18} \) as the firm’s size increases: Granting capital investment discretion (D ₁) to a plant manager in a multinational that exceeds the average size by one standard deviation (\( {A_2}=1 \)) does not improve performance by \( {d_1}=+\mathbf{0}.\mathbf{41} \) standard deviations as in the average-sized multinationals but rather by only +0.23 standard deviations, since \( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=1}}^{Multi. }={d_1}+{m_{1,2 }}{A_2}=\left( {+\mathbf{0}.\mathbf{41}} \right)+\left( {-\mathbf{0}.\mathbf{18}} \right)\cdot (1) \). Firm size therefore amplifies the positive impact of capital investment discretion in a Chinese firm (\( {m_{1,2 }}=+\mathbf{0}.\mathbf{12} \)) but dampens the positive impact in a multinational (\( {m_{1,2 }} = -\mathbf{0}.\mathbf{18} \)). This potential dependence of the moderating effects of firm size (m) on the moderating effects of firm type (g) is discussed more fully in Sect. 6.2.3.

The statistically significant t-value for this mediating effect, \( {d_1}\cdot {a_{1,2 }}=+\mathbf{0}.\mathbf{04} \), is estimated from bootstrapping (see Sect.​ 4.​3) as 2.06 by following the recommendations in the literature, including Chin et al. (2003, pp. 195–211), Cohen (1988, pp. 410–413), Eggert et al. (2005, pp. 105–106), Götz and Liehr-Gobbers (2004, pp. 730–731), Herrmann et al. (2006, pp. 57–61), Iacobucci and Duhachek (2003), Krafft et al. (2005, pp. 83–85), Shrout and Bolger (2002, pp. 430–432), Sobel (1982), and Venkatraman (1989, pp. 429–430).

The test results are extended to non-average-sized firms in Sect. 6.2.3 on firm size.

The discretion puzzle that the present study works towards resolving was defined in Sect.​ 1.​1 as the ostensible paradox that empirical evidence coexists for positive, neutral, and negative impacts of discretion on performance that the existing theories do not unequivocally explain, either individually or collectively (i.e. managerial discretion theory, principal-agent theory, and stewardship theory).

The empirical literature reviewed in Sect.​ 2.​2.​2 has comprised divergent results on the impact of managerial discretion on performance, ranging from positive (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Chang and Wong 2003; Gammelgaard et al. 2010; Khanchel 2009) to neutral (e.g. Caza 2011; Groves et al. 1994; Yougen Li and Zhao 2004; López-Navarro and Camisón-Zornoza 2003; Venaik 1999) and even negative (e.g. He et al. 2009; Heinecke 2011; Stano 1976; Williamson 1963; Xu et al. 2005). This contradictory evidence as to whether discretion tends to increase, not alter or decrease performance spans different levels of management (e.g. top management and middle management) as well as different geographies (e.g. United States of America, European countries, and China; see Sect.​ 2.​2.​2). Even for a given level of management in a given country, such as top management in China, empirical studies have found positive, neutral, and negative impacts of managerial discretion on performance (e.g. Chang and Wong 2003, 2004; Yougen Li and Zhao 2004; Xu et al. 2005; Zhang 1997).

See Sect.​ 2.​2 (e.g. Adams et al. 2005; Huiyuan Chen 2006; Cheng et al. 2006; Heinecke 2011; Thomas and Peyrefitte 1996; Walters 1995; Wang et al. 2008; Xu et al. 2005; Yan et al. 2010; Zhang and Li 2008b). For instance, if two different studies adopt two different measures of discretion, each of which taps into a distinct dimension of discretion, then one study might find a positive and the other a negative impact of discretion on performance, provided the distinct discretion dimensions have different performance impacts.

See Sect.​ 2.​2.​3 (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Caza 2011; Crossland and Hambrick 2007; Khanchel 2009; Yougen Li and Zhao 2004; López-Navarro and Camisón-Zornoza 2003; Ongore 2011; Tang 2008; Williamson 1963).

See Sect.​ 2.​4.​2 (e.g. Bowen et al. 2008; Chang and Wong 2003; Gammelgaard et al. 2010; Groves et al. 1994; He et al. 2009; Xiaoyang Li 2007; Stano 1976; Venaik 1999; Werner and Tosi 1995; Zhang 1997).

See Sect.​ 2.​1.​2 (e.g. Barnabas and Mekoth 2010; Carpenter and Golden 1997, p. 195; Caza 2007, pp. 26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245; Groves et al. 1994, p. 190; Hambrick and Abrahamson 1995, p. 1439; Hambrick and Finkelstein 1987, pp. 371–402; Hambrick et al. 1993, p. 409; March and Simon 1958; Perrone et al. 2003, pp. 422–423).

For example, a given manager might use additional discretion for making capital investments in a way that improves performance but use additional discretion for hiring workers in a way that reduces performance. Under this postulate, discretion could have positive, neutral, and also negative impacts on performance depending on the dimensions of discretion (e.g. capital investment discretion versus hiring discretion).

Although managerial discretion theory hints that discretion may be multidimensional (see Sect.​ 2.​1.​2), both principal-agent theory (see Sect.​ 2.​3.​2) and stewardship theory (see Sect.​ 2.​3.​3) tend to treat discretion as unidimensional (e.g. Chang and Wong 2003; Dicke and Ott 2002, p. 468; Fox and Hamilton 1994, p. 74; He et al. 2009; Spremann 1987, p. 18; Vargas Sánchez 2005, p. 19; Xu et al. 2005). Furthermore, many existing empirical studies have tended to make the simplifying assumption in their research designs that discretion is unidimensional and have thus often combined a manager’s discretion in different areas of his/her work into a single unidimensional discretion construct (e.g. Barnabas and Mekoth 2010; Bloom et al. 2008; Caza 2007, 2011; Chang and Wong 2003; Cheng et al. 2006; Gammelgaard et al. 2010; Marin and Verdier 2006).

If \( H_1^0 \) cannot be rejected, the assumption might be justifiable, but if \( H_1^0 \) can be rejected in favour of \( H_1^1 \) (i.e. multidimensionality), it is demonstrated by virtue of proof-by-counter-example that the simplifying assumption of unidimensionality is not universally tenable, because it is then shown to be untenable at least in the instance of this particular study.

This test of multidimensionality for Hypothesis 1 differs from other tests of multidimensionality in the literature that assess whether indicators tend to vary together (i.e. unidimensionality) or tend to vary separately (i.e. multidimensionality; e.g. Carmines and Zeller 1979, p. 59; Caza 2007, p. 40; Esposito Vinzi et al. 2010, pp. 50–51; Rosenberg 1965; Tenenhaus et al. 2005, p. 163). For example, Caza’s (2007, pp. 26–82) factor analysis confirms that research and development managers in Europe perceive their discretion as multidimensional, meaning that they may perceive high discretion in one area of their work but low discretion in another (i.e. not all discretion indicators vary together). This multidimensionality is empirically confirmed in the present study by both principal components analysis (PCA, which seeks to reproduce the total variance of the indicators) and common factor analysis (CFA, which seeks to reproduce only the common variance of the indicators) for both Chinese firms and multinationals: The perceived discretion of plant managers in China can vary separately across different areas of work (e.g. be high for capital investment discretion but low for hiring discretion) instead of consistently tending to be high, medium or low across all areas of work. Nevertheless, this test of multidimensionality in the literature is not suitable for testing Hypothesis 1, because it assesses whether or not discretion indicators vary together rather than whether or not they impact on performance differently, as the present study postulates. Another drawback of this test of multidimensionality by reliability estimates (Esposito Vinzi et al. 2010, p. 50; Tenenhaus et al. 2005, p. 164) and factor analysis (Bohrnstedt 1970, p. 92; Carmines and Zeller 1979, pp. 60–61; Esposito Vinzi et al. 2003, p. 5; Götz and Liehr-Gobbers 2004, p. 727; Krafft et al. 2005, pp. 73–75) when applied to the study’s discretion indicators that are formative in nature (see Sect.​ 4.​2.​2; e.g. Coltman et al. 2008) is that the test is applicable only to reflective indicators (which are expected to strongly covary) and not to formative indicators (which cannot be expected to strongly covary; see Box 5.6 in Sect.​ 5.​3; Blalock 1964; Chin 1998, p. 306; Diamantopoulos 1999, pp. 447–453; Diamantopoulos and Siguaw 2006, pp. 270–271; Esposito Vinzi et al. 2010, p. 51; Götz and Liehr-Gobbers 2004, p. 728; Herrmann et al. 2006, p. 30; Krafft 1999, p. 124; Krafft et al. 2005, p. 76; Krafft et al. 2003, p. 102; Rossiter 2002, p. 307; Sánchez 2009, p. 3).

With the latter approach, the four formative discretion indicators (measuring discretion in the four areas of the plant manager’s work) have differently signed indicator weights for the single discretion construct (D)—e.g. hiring discretion has a positive weight and capital investment discretion has a negative weight. As the single discretion construct’s (D) single direct effect on performance is estimated as negative (\( d<0 \)), the model still reveals discretion’s multidimensionality in terms of hiring discretion having a negative total effect on performance in multinationals (i.e. positive weight times negative direct effect d) and capital investment discretion having a positive total effect on performance (i.e. negative weight times negative direct effect; Fornell et al. 1990, p. 1252; Herrmann et al. 2006, p. 60). However, with this latter approach, the single discretion construct (D) is conceptually inconsistent with the definition of managerial discretion as the ‘latitude of managerial action’ (Hambrick and Finkelstein 1987, pp. 371–378; see Sect.​ 2.​1.​1): It is increasing in the latitude of hiring workers but decreasing in the latitude of making capital investments. This incoherence is not resolved by inverting the signs of all discretion weights (i.e. ‘reverse coding’; Chin 2000, p. 1), which is permissible given the indeterminacy of the signs of the latent variables (see Box 5.2 in Sect.​ 5.​1.​1; Chin 2000, p. 1; Esposito Vinzi et al. 2010, p. 52; Temme and Kreis 2005, pp. 203–206). The inverted single discretion construct has a positive single direct effect on performance (d > 0) and e.g. hiring discretion has a negative weight in multinationals while capital investment discretion has a positive weight.

For the average-sized firm (\( {A_2}=0 \)), the direct effects of discretion on performance (\( {d_k} \)) measure the total impact of discretion on performance (\( {{{\partial P}} \left/ {{\partial {D_k}}} \right.}={d_k}+{m_{k,2 }}{A_2} \)) (see Sect.​ 4.​3.​2 on comparative statics; e.g. Aiken and West 1991, pp. 35–129; Dowling 2000, p. 284; Finney et al. 1984; Henseler and Fassott 2010, pp. 727–729; Hirschey 2009, p. 99; Ringle et al. 2005).

See Sect.​ 4.​3.​3 (e.g. Ahuja and Thatcher 2005, p. 446; Boßow-Thies and Albers 2010, p. 595; Chin 1998, pp. 318–320, 2000, pp. 1–2, 2003; Chin and Dibbern 2010; Chin and Newsted 1999, pp. 309–314; Dibbern and Chin 2005, pp. 146–159; Efron and Gong 1983, pp. 40–46; Efron and Tibshirani 1993, pp. 145–147; Esposito Vinzi et al. 2010, pp. 48–60; Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Greene 2003, pp. 892–896; Gujarati 2004, pp. 65–139; Hayashi 2000, pp. 3–45; Henseler 2007, pp. 104–107; Henseler and Fassott 2010, pp. 721–730; Herrmann et al. 2006, p. 59; Hsieh et al. 2008, p. 109; Keil et al. 2000, pp. 312–315; Krafft et al. 2005, p. 83; Lohmöller 1989, p. 49; Northrop and Arsenault 2007, p. 231; Qureshi and Compeau 2009, pp. 197–208; Salvatore and Reagle 2002, pp. 87–95; Scholderer and Balderjahn 2005, pp. 88–91; Spanos 1986, pp. 213–311; Temme et al. 2006, p. 11; Tenenhaus et al. 2005, p. 176; Thas 2010, pp. 335–337; Venkatesh and Morris 2000, p. 126; Wooldridge 2002, pp. 116–299; Zhu et al. 2006, pp. 529–530).

See Sect.​ 5.​3 on construct validity (e.g. Ahuja and Thatcher 2005, p. 446; Albers 2010, p. 411; Albright and Malloy 2000, p. 349; Arnold 1982; Arteaga et al. 2010, p. 164; Bagozzi and Yi 1988, p. 82; Balderjahn 1986, p. 236; Barroso et al. 2010, p. 437; Baumgartner and Homburg 1996; Bido, 2007; Blalock 1964; Bohrnstedt 1970, p. 92; Bollen and Lennox 1991, p. 308; Boßow-Thies and Albers 2010, p. 596; Bromley 2002, p. 35; Campbell and Fiske 1959, p. 81; Carmines and Zeller 1979, p. 53; Carte and Russell 2003, pp. 493–494; Caza 2007, p. 40; Chin 1998, p. 318, 2000, pp. 1–2, 2010, p. 670; Chin et al. 2003, p. 194; Churchill 1979, 1987; Coltman et al. 2008; Cronbach and Meehl 1955; Diamantopoulos 1999, pp. 447–453; Diamantopoulos and Siguaw 2006, p. 271; Diamantopoulos and Winklhofer 2001, p. 272; Donsbach and Traugott 2008, p. 364; Duarte and Raposo 2010, p. 463; Eggert and Fassott 2003, pp. 5–9; Esposito Vinzi et al. 2010, pp. 50–51; Esposito Vinzi et al. 2003, p. 5; Fassott and Eggert 2005, p. 32; Fornell and Cha 1994, pp. 71–73; Fornell and Larcker 1981, pp. 45–46; Fornell et al. 1990, p. 1252; Fritz 1995, p. 136; Garson 2002, pp. 195–196; Geisser 1975, pp. 320–328; Gliner and Morgan 2000, pp. 321–322; Götz and Liehr-Gobbers 2004, p. 727; Hahn 2002, p. 104; Helm 2005, pp. 249–252; Henseler and Fassott 2010, pp. 719–721; Herrmann et al. 2006, pp. 24–30; Hinkel 2001, p. 291; Homburg and Baumgartner 1995b, p. 1093; Homburg and Dobratz 1998, p. 450; Homburg and Giering 1996, p. 12; Hsieh et al. 2008, p. 109; Hu and Olshfski 2007, p. 207; Hulland 1999, pp. 198–199; Jarvis et al. 2003, p. 202; Jöreskog and Wold 1982, p. 270; Keil et al. 2000, pp. 312–315; Krafft 1999, p. 124; Krafft et al. 2005, pp. 73–75; Krafft et al. 2003, p. 102; Lohmöller 1989, p. 36; Mosier 1947; Nunnally 1978, p. 111; Ping, 2005, p. 1; Qureshi and Compeau 2009, pp. 197–199; Reinartz et al. 2004, p. 298; Rigdon et al. 1998, p. 1; Ringle et al. 2005; Rodgers and Pavlou 2003, p. 25; Rossiter 2002, p. 315; Ruiz et al. 2010, pp. 546–548; Sambamurthy and Chin 1994, pp. 231–232; Sánchez 2009, p. 3; Schepers et al. 2005, p. 504; Seltin and Keeves 1994, p. 4356; Stone 1974; Tenenhaus et al. 2005, pp. 163–174; van Oppen et al. 2005, p. 19; Venkatesh and Morris 2000, p. 126; Venkatraman 1989, p. 426; Wold 1982, p. 10; Zhu et al. 2006, pp. 529–530).

See Sect.​ 5.​4.​2 on no multicollinearity (e.g. Backhaus et al. 2006; Diamantopoulos and Siguaw 2006, p. 270; Diamantopoulos and Winklhofer 2001, p. 272; Eckey et al. 2004, p. 92; Fornell and Bookstein 1982a; Greene 2003, pp. 57–58; Gujarati 2004, pp. 342–363; Hair et al. 1998, p. 208; Helm 2005, pp. 248–249; Kleinbaum et al. 1998, p. 214; Krafft et al. 2005, pp. 72–80; Kutner et al. 2004; Motulsky 2003, p. 106; Temme et al. 2006, p. 18; Wooldridge 2002, p. 95).

See Sect.​ 5.​5 on external validity (e.g. Abraham et al. 2007; Bureau van Dijk 2005, p. 2, 2006/2007, p. 2; Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Groves 1990, p. 233; Groves et al. 2009, pp. 54–56; Groves and Lyberg 2001, p. 195; Guojia tongji ju [National Bureau of Statistics] 2003, 2007, 14–1, 14–2, 14–18; McCarty 2003, p. 397; ISIC Rev.3.1; National Bureau of Statistics 2002; Northrop and Arsenault 2007, pp. 235–236; Oliinik 2008; Poncet et al. 2008, p. 8; Ringle et al. 2005; Schofield 2006, pp. 28–29; Schwester 2007, pp. 272–273; Stuart 1984; Temme et al. 2006, pp. 7–8; The American Association for Public Opinion Research 2008, pp. 34–35; United Nations 2007, p. 63; Whyte 2000, p. 62; Wooldridge 2002, pp. 298–299).

As discussed in Sect. 6.1, these direct effects estimate the total impact of discretion on performance in both average-sized Chinese firms (\( {d_1}=+\mathbf{0}.\mathbf{40} \), \( {d_2}=-0.04 \), \( {d_3}=-\mathbf{0}.\mathbf{23} \), \( {d_4}=-0.04 \)) and average-sized multinationals (\( {d_1}=+\mathbf{0}.\mathbf{41} \), \( {d_2}=-\mathbf{0}.\mathbf{23} \), \( {d_3}=-0.05 \), \( {d_4}=-0.00 \)). The finding that certain types of discretion (i.e. capital investment discretion D ₁) have a positive impact on performance while other discretion types (i.e. new product introduction discretion D ₃ in Chinese firms; hiring discretion D ₂ in multinationals) have a negative impact on performance supports the alternative hypothesis (\( H_1^1 \)) that managerial discretion is multidimensional in its impact on performance. Yet it is difficult to infer whether e.g. hiring discretion (\( {d_2}=-0.04 \)) and new product introduction discretion (\( {d_3}=-\mathbf{0}.\mathbf{23} \)) differ significantly, as both of these discretion types have a negative direct effect on performance (P) in the sample of 367 Chinese firms.

For example, although hiring discretion (D ₂) and new product introduction discretion (D ₃) both have a negative direct effect on performance (P) in the sample of 367 Chinese firms (\( {d_2}=-0.04 \) and \( {d_3}=-\mathbf{0}.\mathbf{23} \)), the difference between these direct effects, \( \left( {{d_2}-{d_3}} \right)=\left( {-0.04} \right)-\left( {-\mathbf{0}.\mathbf{23}} \right)=+\mathbf{0}.\mathbf{19}\), is shown to have a p-value on a two-tailed test of only 2.3 % in Table 6.2. This constitutes significant evidence at the 2.3 % level that the direct effects of hiring discretion on performance and new product introduction discretion on performance differ in the population of relevant Chinese firms.

For instance, the direct effect of capital investment discretion on performance exceeds that of hiring discretion by \( +\mathbf{0}.\mathbf{44} \) in the sample of Chinese firms and \( +\mathbf{0}.\mathbf{64} \) in the sample of multinationals, both of which are significant at the 0.0 % level (to one decimal place).

It should be noted that as explained in Box 1.1 in Sect.​ 1.​2 on the delimitations of the research objective, the present study’s approach is subject to the caveat of observational cross-sectional studies in terms of demonstrating causality (e.g. Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sánchez 2008, p. 5; Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sects.​ 5.​4 and 7.​6).

As is explained in Sect.​ 7.​1, erroneously treating discretion as unidimensional when it is in fact multidimensional might produce misleading estimates of the impact of discretion on performance that could potentially be a cause of the contradictory empirical evidence that gives rise to the discretion puzzle. When failing to account for this granularity, discretion’s performance impact could arbitrarily be found to be positive, neutral or even negative simply depending on (inter alia) how discretion is measured. For example, a study gauging overall discretion by capital investment discretion-related indicators might find that discretion has a positive impact on performance, while a study focusing on hiring discretion-related indicators might find a negative impact, and a study mixing the two dimensions might find a neutral impact on performance.

The test results are extended to non-average-sized firms in Sect. 6.2.3 on firm size.

See Sect.​ 2.​2 (e.g. Adams et al. 2005; Huiyuan Chen 2006; Cheng et al. 2006; Groves et al. 1994; Heinecke 2011; López-Navarro and Camisón-Zornoza 2003; Thomas and Peyrefitte 1996; Wang et al. 2008; Werner and Tosi 1995; Williamson 1963; Xu et al. 2005; Yan et al. 2010; Zhang and Li 2008b; Zhang and Xie 2008; Zhang 1997).

See Table 2.​9 in Sect.​ 2.​3.​2.​2 (e.g. Caza 2007, 2011; Chang and Wong 2003; Cheng et al. 2006; Davis et al. 1997b; Eisenhardt 1989; Fama and Jensen 1983b; Finkelstein and Peteraf 2007; Jensen and Meckling 1976; Xiaoyang Li, 2007; Wang et al. 2008; Zhang and Li 2008b).

See Sect.​ 2.​3.​3.​2 (e.g. Argyris 1973a, b; Brown 1969; Caza 2007; Davis et al. 1997b, pp. 27–43; French and Raven 1959; Gibson et al. 1991; Hofstede 1980, 1991; Katz and Kahn 1978; Khanchel 2009, p. 98; Lane et al. 1999, p. 1079; Lawler 1986, 1992; Maslow 1970; McGregor 1960; Mills and Keast 2009; Simon 1957a, b; Triandis 1990, 1995; Triandis et al. 1993; Turner 1981; Van Slyke 2007, p. 164; Vargas Sánchez 2001, 2004, 2005; Walton 1980, 1985).

Box 2.6 in Sect.​ 2.​4.​2 discusses the rationale for why these influences on managers (derived from principal-agent theory and stewardship theory) are expected to differ by firm type and firm size. In simple terms, as to firm type, plant managers in firms of different types might be subject to different influences specified by principal-agent theory and stewardship theory that potentially affect the manager’s natural predisposition to performance maximisation and compensation control mechanisms—and therefore moderate the impact of discretion on performance. E.g. the organisation’s situational characteristics (i.e. management philosophy and cultural differences) that stewardship theorists have argued affect the manager’s natural predisposition to performance maximisation may differ by firm type, that is between domestic Chinese firms and foreign multinationals (e.g. Aminpour and Woetzel 2006, pp. 46–47; Hanne Chen 2004, p. 128; Crossland and Hambrick 2007, p. 14; Hexter 2006, pp. 4–6; Hexter and Woetzel 2007a, p. 4; Hoover 2006, p. 92; McGregor 2005, pp. 273–274). Similarly, with different types of firms attracting different people, the psychological characteristics and expectations that stewardship theorists have argued influence how managers use discretion might differ between Chinese firms and multinationals, as could compensation control mechanisms from principal-agent theory (see Box 2.6 in Sect.​ 2.​4.​2).

As explained in Sect.​ 4.​1.​2, three-way interaction stems from synthesising the moderating effects derived from principal-agent theory and stewardship theory in Chap.​ 2 into the study’s new discretion model.

See Sect.​ 4.​3.​3 for a discussion of bootstrapping (e.g. Chin 1998, pp. 318–320; Efron and Gong 1983, pp. 40–46; Efron and Tibshirani 1993, pp. 145–147; Esposito Vinzi et al. 2010, pp. 59–60; Krafft et al. 2005, p. 83; Tenenhaus et al. 2005, p. 176; Thas 2010, pp. 335–337).

This test has been widely used in the literature (e.g. Ahuja and Thatcher 2005, p. 446; Hsieh et al. 2008, p. 109; Keil et al. 2000, pp. 312–315; Qureshi and Compeau 2009, pp. 197–199; Venkatesh and Morris 2000, p. 126; Zhu et al. 2006, pp. 529–530). With the test’s assumptions fulfilled (see Sect.​ 5.​3; e.g. Bido 2007; Carte and Russell 2003, pp. 493–494; Chin 2000, pp. 1–2; Dibbern and Chin 2005, p. 146; Henseler and Fassott 2010, p. 721; Hsieh et al. 2008, p. 109; Qureshi and Compeau 2009, p. 199), the testing procedure’s validity for verifying Hypothesis 2 is secured.

Specifically, plant managers may use hiring discretion in a more performance-reducing way (as principal-agent theory predicts; see Sect.​ 2.​3.​2) in average-sized multinationals than in Chinese firms. This indicates that the agency problem is relatively more severe for hiring workers in average-sized multinationals than in Chinese firms. The moderating effect \( {g_2}=+\mathbf{0}.\mathbf{19} \) is thus positive, measuring by how many standard deviations the impact of hiring discretion on performance improves when moving from an average-sized multinational (\( {d_2}=-\mathbf{0}.\mathbf{23} \)) to an average-sized Chinese firm (\( {d_2}=-0.04 \)), all other things being held equal.

As explained in Sect.​ 5.​1.​2, the finding that the coefficient of determination for Chinese firms and multinationals combined (\( {R^2}=0.24 \)) is evidently lower than that of either Chinese firms (\( {R^2}=0.28 \)) or multinationals (\( {R^2}=0.32 \)) when calibrated individually manifests the moderating effects of firm type (i.e. Chinese firms versus multinationals) on all structural relationships in the empirical discretion model at once by reasoning related to the Chow test (Chow 1960, pp. 595–604; Henseler and Fassott 2010, pp. 730–732).

E.g. for firms exceeding average firm size by two standard deviations (\( {A_2}=2 \)), the model predicts that capital investment discretion’s impact on performance significantly falls to only +0.05 in large multinationals (\( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=2}}^{Multi. }={d_1}+{m_{1,2 }}\cdot (2)=\left( {+\mathbf{0}.\mathbf{41}} \right)+\left( {-\mathbf{0}.\mathbf{18}} \right)\cdot (2) \)) and by contrast significantly rises to +0.64 in large Chinese firms, whereby granting a plant manager one more standard deviation of capital investment discretion boosts performance by 0.64 standard deviations (\( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=2}}^{Chinese }={d_1}+{m_{1,2 }}\cdot (2)=\left( {+\mathbf{0}.\mathbf{40}} \right)+\left( {+\mathbf{0}.\mathbf{12}} \right)\cdot (2) \)).

It should be noted that as explained in Box 1.1 in Sect.​ 1.​2 on the delimitations of the research objective, the present study’s approach is subject to the caveat of observational cross-sectional studies in terms of demonstrating causality (e.g. Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sánchez 2008, p. 5; Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sects.​ 5.​4 and 7.​6).

While firm type in terms of domestic Chinese firms versus foreign multinationals has thus been shown to have a significant moderating effect in the present study, differentiating firm type among state-owned enterprises in China between employee-owned stock cooperatives, limited liability companies, and limited liability stock companies has been found to be insignificant in the literature (Xu et al. 2005).

See Sect.​ 2.​2 (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Bowen et al. 2008; Caza 2007, 2011; Chang and Wong 2003; Huiyuan Chen 2006; Cheng et al. 2006; Gammelgaard et al. 2010; He et al. 2009; Khanchel 2009; Yougen Li and Zhao 2004; Stano 1976; Werner and Tosi 1995; Williamson 1963; Zhang and Li 2008b; Zhang 1997).

Box 2.6 in Sect.​ 2.​4.​2 has described that firm size (e.g. 150 versus 5,000 employees; see Sect.​ 4.​2.​4) might affect monitoring control mechanisms, which aim to reduce the principal’s asymmetry of information regarding the manager’s actions and thereby aim to improve the impact of discretion on performance in principal-agent theory (see Sect.​ 2.​3.​2; e.g. Berger et al. 1997, p. 1411; Burkart et al. 1997, p. 705; Caza 2007, 2011; Chang and Wong 2003, p. 6; Eisenhardt 1989, p. 60; Fama and Jensen 1983b; Levinthal 1988, p. 153; Spremann 1987, pp. 10–11; Verhoest 2003, pp. 2–5; Wang et al. 2008; Werner and Tosi 1995, p. 1673). E.g. Caza (2007, pp. 12–20) argues that a larger firm with multiple units (here plants) that are similar to that led by the (plant) manager under consideration provides the principal with more informational benchmarks to assess the (plant) manager’s performance and thus potentially constrain his/her misconduct.

A plant manager in a smaller firm (e.g. 150 employees) might, for example, respond differently to additional discretion than a plant manager in a larger firm (e.g. 5,000 employees), leading to a different total impact of discretion (D) on performance (P) depending on firm size (equal to a direct effect and a moderating effect).

Or equivalently, Hypothesis 3 refers to whether influences on managers in terms of firm size affect the way that managers use their discretion to impact on performance, i.e. the impact of discretion on performance.

The reason is that this moderating effect precisely measures the difference in the impact of discretion on performance when firm size is altered by one standard deviation. This can be seen by differentiating the total impact of e.g. new product introduction discretion (D ₃) on performance (P), derived in Sect.​ 4.​3.​2 as \( {{{\partial P}} \left/ {{\partial {D_3}}} \right.}={d_3}+{m_{3,2 }}{A_2} \), by firm size (\( {A_2} \)). It yields the moderating effect of firm size as the parameter \( {m_{3,2 }} \) (e.g. Aiken and West 1991, pp. 35–129; Carte and Russell 2003, pp. 480–495; Chin et al. 2003, pp. 196–200; Dowling 2000, p. 284; Finney et al. 1984; Henseler and Fassott 2010, pp. 727–729; Hirschey 2009, p. 99; Irwin and McClelland 2001, p. 105; Ringle et al. 2005). In addition to the ‘Moderating Effect (m)’ of firm size derived from principal-agent theory, Fig. 6.5 retains the general theoretical discretion model’s ‘Direct Effect (a)’ of firm size on the level of discretion (e.g. Caza 2007, pp. 30–56; Hambrick and Finkelstein 1987, pp. 378–389; Walters 1995; Zhang et al. 2006a) and the ‘Control Effect (c)’ of firm size on the level of performance (e.g. Baum 1996; Berman et al. 2005, p. B5; Chang and Wong 2003, p. 21; Datta et al. 2003, p. 101; Gammelgaard et al. 2010, p. 9; Haleblian and Finkelstein 1993, p. 853; Khanchel 2009, p. 100; Misangyi 2002, p. 62; Singh and Lumsden 1990; Xu et al. 2005, p. 9; see Fig. 4.​6 in Sect.​ 4.​1.​2).

As explained in Box 2.2 in Sect.​ 2.​3.​2.​2, the manager’s natural predisposition to performance maximisation (linked to firm type in Sect. 6.2.2) and monitoring control mechanisms (linked to firm size above) may both have moderating effects on the impact of discretion on performance that might depend on each other (as depicted in the general theoretical discretion model in Fig. 4.​4 in Sect.​ 4.​1.​2). E.g. monitoring control mechanisms might either mitigate or aggravate discretion’s impact on performance depending on the manager’s natural predisposition to performance maximisation. Therefore, Sect.​ 4.​1.​3 has incorporated this potential three-way interaction (e.g. Henseler and Fassott 2010, p. 722) between the manager’s natural predisposition to performance maximisation, monitoring control mechanisms, and discretion in its impact on performance into the study’s specific theoretical discretion model by allowing for three-way interaction between firm type, firm size, and discretion in its impact on performance.

For the average-sized Chinese firm and multinational (\( A_2^{Chinese }=A_2^{Multi. }={A_2}=0 \)), the moderating effect of firm type on the total impact of discretion on performance in Equation (6.4) reduces to the firm type’s moderating effect on discretion’s direct effect on performance (e.g. \( {g_3} \)), which was tested for significance in Sect. 6.2.2.

See Sect.​ 4.​3.​3 (e.g. Ahuja and Thatcher 2005, p. 446; Boßow-Thies and Albers 2010, p. 595; Chin 1998, pp. 318–320, 2000, pp. 1–2, 2003; Chin and Dibbern 2010; Chin and Newsted 1999, pp. 309–314; Dibbern and Chin 2005, pp. 146–159; Efron and Gong 1983, pp. 40–46; Efron and Tibshirani 1993, pp. 145–147; Esposito Vinzi et al. 2010, pp. 48–60; Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Greene 2003, pp. 892–896; Gujarati 2004, pp. 65–139; Hayashi 2000, pp. 3–45; Henseler 2007, pp. 104–107; Henseler and Fassott 2010, pp. 721–730; Herrmann et al. 2006, p. 59; Hsieh et al. 2008, p. 109; Keil et al. 2000, pp. 312–315; Krafft et al. 2005, p. 83; Lohmöller 1989, p. 49; Northrop and Arsenault 2007, p. 231; Qureshi and Compeau 2009, pp. 197–208; Salvatore and Reagle 2002, pp. 87–95; Scholderer and Balderjahn 2005, pp. 88–91; Spanos 1986, pp. 213–311; Temme et al. 2006, p. 11; Tenenhaus et al. 2005, p. 176; Thas 2010, pp. 335–337; Venkatesh and Morris 2000, p. 126; Wooldridge 2002, pp. 116–299; Zhu et al. 2006, pp. 529–530).

This means that there is an estimated 29.4 % chance (p-value) that the moderating effect of \( m_{4,2}^{Multi. }=+0.13 \) in the sample of 100 multinationals diverges from zero by as much as is observed due to sampling error when the true population parameter is zero.

For instance, plant managers may use capital investment discretion in a more performance-enhancing way (as stewardship theory predicts; see Sect.​ 2.​3.​3) in larger Chinese firms compared to smaller Chinese firms.

It should be noted that as explained in Box 1.1 in Sect.​ 1.​2 on the delimitations of the research objective, the present study’s approach is subject to the caveat of observational cross-sectional studies in terms of demonstrating causality (e.g. Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sánchez 2008, p. 5; Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sects.​ 5.​4 and 7.​6).

See Sect.​ 2.​2 (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Bowen et al. 2008; Caza 2007, 2011; Chang and Wong 2003; Huiyuan Chen 2006; Cheng et al. 2006; Gammelgaard et al. 2010; He et al. 2009; Khanchel 2009; Yougen Li and Zhao 2004; Stano 1976; Werner and Tosi 1995; Williamson 1963; Zhang and Li 2008b; Zhang 1997).

Specifically, Table 6.5 combines the individual ‘Direct Effect (d)’ of each dimension of discretion (D ₁, D ₂, D ₃, and D ₄) in each firm type (Chinese firms and multinationals) with the relevant ‘Moderating Effect (m)’ of firm size (\( {A_2} \)) according to the equation on the left of Table 6.4. This produces the total impact of discretion on performance as a combination of direct and moderating effects (e.g. \( {{\left. {{{{\partial P}} \left/ {{\partial {D_3}}} \right.}} \right|}^{Chinese }}=d_3^{Chinese }+m_{3,2}^{Chinese }A_2^{Chinese } \) for D ₃) as a function of firm size for each dimension of discretion and for each firm type. A total of eight such comparative statics (due to four discretion dimensions and two firm types) result from combining the direct and moderating effects in Table 6.4. All eight equations enter Table 6.5, which presents the total empirical impact of discretion on performance by dimension of discretion, firm type, and firm size.

E.g. for firms exceeding average firm size by two standard deviations (\( {A_2}=2 \)), the model predicts that capital investment discretion’s impact on performance significantly falls to only +0.05 in large multinationals (\( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=2}}^{Multi. }={d_1}+{m_{1,2 }}\cdot (2)=\left( {+\mathbf{0}.\mathbf{41}} \right)+\left( {-\mathbf{0}.\mathbf{18}} \right)\cdot (2) \)) and by contrast significantly rises to +0.64 in large Chinese firms—whereby granting a plant manager one more standard deviation of capital investment discretion improves performance by 0.64 standard deviations (\( \left. {{{{\partial P}} \left/ {{\partial {D_1}}} \right.}} \right|_{{{A_2}=2}}^{Chinese }=\left( {+\mathbf{0}.\mathbf{40}} \right)+\left( {+\mathbf{0}.\mathbf{12}} \right)\cdot (2) \)).

Figure 6.6 also facilitates the testing of Hypothesis 4 in Sect. 6.2.4 as well as the creation of recommendations to top management in Sect.​ 7.​5.​2.

While the fictitious scatter plots in Chap.​ 2 depict performance (P) on the vertical axis as a function of managerial discretion (D) on the horizontal axis with the slope of the line measuring the impact of discretion on performance (\( {{{\partial P}} \left/ {{\partial D}} \right.} \)), the calibrated scatter plot in Fig. 6.6 shows the impact of discretion on performance (\( {{{\partial P}} \left/ {{\partial D}} \right.} \)) on the vertical axis as a function of firm size (\( {A_2} \)) on the horizontal axis with the slope of the line measuring the moderating effect of firm size (\( {m_{k,2 }} \)) on discretion’s impact on performance.

The axes are measured in standard deviations, which can differ between Chinese firms and multinationals. The horizontal axis with firm size measures by how many standard deviations a firm exceeds average size. To avoid extrapolation, the domains range from the minimum to the maximum values of firm size in the sample.

These findings represent tendencies for the study’s representative sample of 467 firms. As described in Sect.​ 7.​5.​2, due to potential sampling error, one should conclude that these impacts of discretion on performance are positive or negative in the population for a given firm size only if the lines in Fig. 6.6 lie sufficiently far above or below the horizontal axis in Fig. 6.6 (i.e. only if they are statistically significant).

As discussed in Sects.​ 1.​1 and 2.​3, existing combinations of managerial discretion theory, principal-agent theory, and stewardship theory also fail to fully explain the impact of discretion on performance (e.g. Caza 2007, 2011; Davis et al. 1997b, pp. 27–43; Finkelstein and Peteraf 2007, pp. 237–243; Lane et al. 1999, p. 1079; Vargas Sánchez 2001, 2004, 2005).

As explained in Sect.​ 2.​3.​1, managerial discretion theory does not clearly specify the nature of the impact of discretion on performance and therefore makes no predictions on e.g. positive versus negative impacts (e.g. Adams et al. 2005; Crossland and Hambrick 2007; Finkelstein and Hambrick 1990; Hambrick and Finkelstein 1987; Misangyi 2002; Quigley and Hambrick 2009; Tang 2008; Zhao et al. 2010). Although principal-agent theory can be consistent with the null hypothesis (\( H_4^0 \)) that discretion does not alter performance in exceptional cases—i.e. when control mechanisms are so effective that they unhinge the theory’s assumptions (see Box 2.1 and Box 2.2 in Sect.​ 2.​3.​2.​2; e.g. Agrawal and Knoeber 1996, p. 377; Chang and Wong 2003, p. 7)—the main prediction of principal-agent theory is the alternative hypothesis (\( H_4^1 \)) that discretion decreases performance (see Sect.​ 2.​3.​2; e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38; Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010).

Universally rejected refers to rejecting the null hypothesis after accounting for the greater granularity, e.g. rejecting the null hypothesis when differentiating discretion’s impact by multiple dimensions of discretion.

See Sect.​ 4.​3.​3 (e.g. Ahuja and Thatcher 2005, p. 446; Boßow-Thies and Albers 2010, p. 595; Chin 1998, pp. 318–320, 2000, pp. 1–2, 2003; Chin and Dibbern 2010; Chin and Newsted 1999, pp. 309–314; Dibbern and Chin 2005, pp. 146–159; Efron and Gong 1983, pp. 40–46; Efron and Tibshirani 1993, pp. 145–147; Esposito Vinzi et al. 2010, pp. 48–60; Fogiel 2000, pp. 158–160; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148; Greene 2003, pp. 892–896; Gujarati 2004, pp. 65–139; Hayashi 2000, pp. 3–45; Henseler 2007, pp. 104–107; Henseler and Fassott 2010, pp. 721–730; Herrmann et al. 2006, p. 59; Hsieh et al. 2008, p. 109; Keil et al. 2000, pp. 312–315; Krafft et al. 2005, p. 83; Lohmöller 1989, p. 49; Northrop and Arsenault 2007, p. 231; Qureshi and Compeau 2009, pp. 197–208; Salvatore and Reagle 2002, pp. 87–95; Scholderer and Balderjahn 2005, pp. 88–91; Spanos 1986, pp. 213–311; Temme et al. 2006, p. 11; Tenenhaus et al. 2005, p. 176; Thas 2010, pp. 335–337; Venkatesh and Morris 2000, p. 126; Wooldridge 2002, pp. 116–299; Zhu et al. 2006, pp. 529–530).

There is a nearly zero percent chance (see the p-values in Table 6.6, i.e. 0.0000000001 % for Chinese firms and 0.02 % for multinationals) that capital investment discretion’s impact on performance as estimated in the sample (+0.40 for average-sized Chinese firms and +0.41 for average-sized multinationals) diverges from zero by as much as is observed due to sampling error when the true population parameter is zero.

As the empirical results are not consistent with any single existing theory for the present study’s particular instance of plant managers in China, this serves as a proof-by-counter-example for the postulate that neither existing theory can fully explain the empirical impact of discretion on performance after applying the greater granularity, since neither theory can fully explain the impact of discretion on performance at least in the particular instance of the present study and therefore not universally in all studies.

Section 2.​3 has derived these influences (i.e. moderators) from the theories’ assumptions, as these determine the theories’ predicted impact of discretion on performance. For example, the principal’s and manager’s psychological characteristics, perceptions of the organisation’s situational characteristics, and expectations are all postulated to influence whether the manager’s natural predisposition to performance maximisation is as assumed in principal-agent theory (i.e. discretion is abused to pursue self-serving interests at the expense of performance) or as assumed in stewardship theory (i.e. discretion is used to diligently improve performance). The influences on managerial action derived from principal-agent theory include the manager’s natural predisposition to performance maximisation, compensation control mechanisms, the nature of managerial activities, monitoring control mechanisms, and the manager’s ability, knowledge, and information.

Potential reasons why agency versus stewardship behaviour could depend on firm type were discussed in Box 2.6 in Sect.​ 2.​4.​2. E.g. the organisation’s situational characteristics (i.e. management philosophy and cultural differences), which stewardship theorists have argued to affect the manager’s natural predisposition to agency versus stewardship behaviour, may differ by firm type, i.e. differ between domestic Chinese firms and foreign multinationals (e.g. Aminpour and Woetzel 2006, pp. 46–47; Hanne Chen 2004, p. 128; Crossland and Hambrick 2007, p. 14; Hexter 2006, pp. 4–6; Hexter and Woetzel 2007a, p. 4; Hoover 2006, p. 92; McGregor 2005, pp. 273–274).