Sage Journals: Discover world-class research

Abstract

We examine the association between strategic deviation and investment inefficiency. We conceptualize strategic deviation as the extent to which the pattern of a firm’s resource allocation deviates from its industry peers. We posit that firms pursuing deviant strategies are prone to increased information asymmetry and hence, are able to engage in self-serving behaviour as manifested in inefficient investments. Our results suggest that deviant firms have sub-optimal investments. A battery of robustness tests validates our findings. We further provide evidence to suggest that weaker monitoring, high product market competition and a low-quality information environment moderate the relation between strategic deviation and investment inefficiency.

JEL Classification:

M41, G41

Keywords

Information asymmetry investments product market competition strategic deviation

1. Introduction

In this study, we examine whether strategic deviation, conceptualized as resource allocation that deviates from industry peers, is associated with investment inefficiency. Managers can avoid competition by pursuing strategic choices that are different from their peers (Porter, 1996), thereby making it difficult for shareholders to evaluate the managerial performance of such firms (Carpenter, 2000). Accordingly, prior research finds that firms that deviate from common industry strategies hold cash for opportunistic reasons (Dong et al., 2021), have less synchronous stock returns (Ye et al., 2021) and exhibit extreme performance compared to their industry peers (Tang et al., 2011). We extend the literature on the consequences of strategic deviation by examining its association with firms’ investment decisions.

Prior research in this space have primarily captured strategy based on strategy typologies, such as Miles and Snow’s (1978) prospector, analyzer and defender strategies. For example, Navissi et al. (2017) find that prospector (defender)-type firms are more likely to over (under)-invest. They argue that prospector-type firms enjoy greater managerial discretion and less stringent monitoring, thereby enabling them to over-invest for self-serving behaviour, including, but not limited to, reducing their career-related risks. Defender-type firms, on the other hand, are subject to a relatively higher level of managerial monitoring and less managerial discretion that results in under-investments.

We differentiate our study form Navissi et al. (2017) by examining the consequences of strategic deviation, which is conceptually different from strategy typology. While strategy typology focuses on resource allocation within firms, strategic deviation focuses on an inter-firm strategy perspective and captures the differences in the focal firm’s resource allocation decisions when compared to industry peers (Finkelstein and Hambrick, 1990; Geletkanycz and Hambrick, 1997). While strategy typologies, such as prospector and defender strategies, are entrenched in a firm’s competitive environment (Porter, 1980), strategic deviation captures a firm’s competitive positioning relative to its peers. Furthermore, researchers arbitrarily choose cut-off scores to categorize firms into prospector, defender and analyzer groups (Bentley et al., 2013; Dong et al., 2021). In contrast, strategic deviation focuses on the resource allocation of a firm relative to the industry which reflects how managers pursue strategies in comparison to the commonly adopted industry practices. We complement the strategy-based explanation for corporate investment policies by providing evidence that the adoption of a deviant strategy increases investment inefficiencies.

Business strategy determines the resource allocations for investments (Miles and Snow, 1978). A strategy that is in line with the industry norms, that is, conforming strategy, is likely to provide benchmarks for evaluation (Porter, 1996). Deviant strategy, that is, a strategy that is different from industry norms, on the other hand, fails to provide appropriate benchmarks for comparisons (Carpenter, 2000). The lack of benchmarks creates information risk and uncertainties for capital market participants. Furthermore, managers of firms pursuing deviant strategies are required to invest in new technology and other projects to attract new customers, segments and markets: investments that reduce the benefit of economies of scale and create higher risks. While some researchers argue that being different has first-entry benefits (Porter, 1980, 1986), we take the view that deviant firms face high risk and uncertainty in cash flow generation and escalate information asymmetry and an opaque information environment (Carpenter, 2000; Deephouse, 1999; Litov et al., 2012). Taken together, the lack of benchmark comparison, information asymmetry and escalated needs for capital expenditures among deviant firms incentivize managers to pursue sub-optimal investments. We empirically investigate this proposition.

We then explore the settings under which the expected relation between strategic deviation and investment inefficiency might vary cross-sectionally. First, we examine whether internal and external governance mechanisms moderate this association. Effective governance mechanisms exert effective monitoring (Fama and Jensen, 1983). Shleifer and Vishny (1997) argue that independent directors are effective in monitoring managerial actions. In addition, monitoring by institutional investors reduces over-investments (Ferreira and Matos, 2008). Strategic deviation creates information asymmetry, and thus constrains strong monitoring compared to other industry peers. Therefore, investment inefficiency for firms with high strategic deviation will be exacerbated by poor monitoring.

Then, we test whether product market competition moderates the association between strategic deviation and investment inefficiency. A stream of research supports the disciplinary view of product market competition (Babar and Habib, 2021). However, intense competition increases information risk because the managers operating in highly competitive industries are reluctant to release proprietary information: an action that hinders obtaining product profitability information ex ante by peers (Stoughton et al., 2017), thus enabling managers of deviant firms to engage in sub-optimal investments.

We then investigate the moderating role of information asymmetry on the association between strategic deviation and investment inefficiency. Ye et al. (2021) suggest that deviant firms are different from industry peers and, therefore, have high information processing costs and increased information asymmetry. Prior research argues that increased information asymmetries arising from information opaqueness increase investment inefficiency (Chen et al., 2017; Lin et al., 2021). We, therefore, expect that the positive association between strategic deviation and investment inefficiency will be more pronounced in a high-information-asymmetry setting. Finally, we consider the role of financial statement comparability. More comparable financial statements increase the ability of capital market participants to benchmark managerial and firm performance and, as a result, reduces information acquisition and processing costs (Choi et al., 2019; De Franco et al., 2011; Imhof et al., 2022; Kim et al., 2016). As argued before, since strategic deviation enables managers to avoid strict monitoring, the incentives for producing more comparable financial statements are likely lower for such firms. We, therefore, argue that the positive association between strategic deviation and investment inefficiency will be more pronounced when financial statement comparability is low.

A majority of research on investment efficiency has modelled in(efficient) investment in two-stages with residuals obtained in the first stage being used as the dependent variable in second stage as proxy for investment efficiency (Biddle et al., 2009). However, several studies have questioned the appropriateness of such a research design highlighting biased coefficients and t-statistics, which are not in the expected direction (Chen et al., 2018, 2022; Christodoulou et al., 2018; Jackson, 2022). Jackson (2022) suggests a one-stage regression procedure with a set of indicator variables for industries, years, the independent variable from the first-step regression and interaction terms between each of industries, years and the independent variable from the first-step regression. We follow this procedure in testing our hypothesis. Using a US sample of 56,133 firm-year observations from 1987 to 2020, we document a positive and significant relationship between strategic deviation and investment inefficiency.¹ Fixed effect estimates, two-stage least square (2SLS) estimates, the entropy balancing test and two-step system GMM (generalized method of moments) to allay endogeneity concerns validate our original findings. Furthermore, weaker monitoring, increased information asymmetry and low financial statement comparability exacerbate the positive association between strategic deviation and investment inefficiency. In an additional test, we find that investments by the deviant firms are discounted by the capital market.

Our study is motivated based on the calls for investigations into the repercussions of adopting a conforming or a deviant strategy (Deephouse, 1996), as it is important to understand the consequences of deviating from industry strategy norms (Chen and Hambrick, 1995). While there are some studies that examine the consequences of strategy typology, only a few have looked into the strategy deviation aspect (Dong et al., 2021; Provaty et al., 2022; Tang et al., 2011; Ye et al., 2021), and We also differ from strategy differentiation, which identifies resource allocation between segments (Dong et al., 2021), and instead focus on inter-firm differences of resource allocation across functions, such as production, marketing, innovation and finance (Finkelstein and Hambrick, 1990). Thus, we capture a firm’s strategy positioning in a competitive market. Furthermore, we respond to calls for further research on examining the first-order determinants of investment in(efficiency) (Biddle et al., 2009) by providing evidence from a strategic deviation perspective. We contribute to the strategic management literature by answering the call (Deephouse, 1999) for the implications, investment inefficiency in our case, of a firm being different from its peers. We further extend corporate investment literature by identifying that being different from industry peers increases investment inefficiency. Thereby, we extend strategy and investments literature, such as Navissi et al. (2017).

The paper is organized as follows. This introduction is followed by literature and hypotheses development in section 2. Section 3 presents the methodology, while section 4 discusses the empirical results and robustness tests. Section 5 concludes the paper.

2. Literature and hypotheses development

According to Modigliani and Miller (1958), firms are likely to pursue optimal investment strategies in a perfect market, but agency frictions and financial constraints make markets imperfect and hence, affect optimal investment levels (Jensen, 1986; Myers, 1977; Shleifer and Vishny, 1989). Under-investments occur when managers pass on positive NPVs to protect their career concerns, while over-investments occur when they invest in negative NPV projects for empire-building incentives (Biddle et al., 2009). These agency problems manifest through managerial empire building, career motives, herding behaviour and managerial myopia (Bebchuk and Stole, 1993; Holmström, 1999; Jensen, 1986; Malmendier and Tate, 2005). Moral hazard–based agency frictions escalate under-investments, while adverse selection-induced agency frictions increase over-investments. Shleifer and Vishny (1997) suggest that weak corporate governance aggravates agency frictions and, consequently, accentuates investment inefficiencies. Agency theory, therefore, posits that managers subject to weak monitoring engage in over-investments with the intent of empire building (Jensen, 1986).

Strategies determine resource allocation in an entity. Therefore, it is important to understand the type of strategies that firms need to follow to pursue effective resource allocation for efficient investments. Corporate strategy can be defined as a pattern that echoes a series of decisions, which determines product markets, technology deployment, organizational structures and business models (Mintzberg, 1978). Miles and Snow’s (1978) strategy typology suggests that firms adopt strategies to remain competitive in the market. According to Miles and Snow (1978), three types of business strategies can exist: prospectors, defenders and analyzers. These strategies vary depending on product markets, processes, organizational structures and technology. Specifically, prospector strategy focuses on innovation and market leadership, while the defender strategy focuses on competition based on price, service or quality. The analyzer strategy falls in between the prospector and defender strategy continuum.

Bentley et al. (2013) find that prospector firms are more likely to have financial irregularities, despite high audit efforts, due to the inherent business risk. Prospector firms have high risk and uncertainty relative to defender firms, and, as a result, increase the incremental information acquisition costs for the analysts (Bentley-Goode et al., 2019). However, prospector firm managers have incentives to reduce information asymmetry to gain access to financial markets for pursuing their innovative strategies. In line with this argument, Bentley-Goode et al. (2019) find that prospector firms make more management earnings forecasts to attract more analyst coverage compared to defender firm managers: actions that result in lower information asymmetry. As the prospector-type firms seek new innovations, they are more likely to engage in over-investments, while defender-type firms suffer from under-investments (Navissi et al., 2017): sub-optimal investments that is associated with poor future performance.

Strategy deviation differs from strategy typology in the sense that strategy deviation is based on firms’ preference to conform to industry peers. Based on the institutional theory, managers tend to adopt practices that conform to their peers, which drives inter-organizational homogeneity (DiMaggio and Powell, 1983). As a result, when a strategy deviates from the industry norms, it induces costs to investors. While a deviant strategy can assist firms to explore new markets, build unique customer and supplier relationships and achieve competitive advantage (Porter, 1980, 1986), the non-conformity eliminates the benchmarks for comparison (Carpenter, 2000). This creates higher information processing costs for investors to evaluate firm and managerial performance (Litov et al., 2012). As a result, strategic deviation increases agency costs and hinders firms with high strategic deviation to access external resources at a cheaper cost (Deephouse, 1999), creates information asymmetry and increases cash holdings (Dong et al., 2021).² Strategic deviation, therefore, allows managers to pursue self-serving behaviour at the expense of shareholders’ interest (Jensen and Meckling, 1976; Jensen, 1986). Furthermore, the lack of benchmarks emanating from high strategic deviation reduces competition and escalates agency problems (Shleifer and Vishny, 1997). Therefore, managers of firms with high strategic deviation are likely to over-invest to capture customers and retain suppliers who otherwise may stick with peers who follow conforming strategies. Similarly, deviant firm managers may increase capital expenditures and research and development (R&D) expenditures to seek new customers and technology: investments that are risky and may adversely affect future performance Based on the preceding arguments, we develop the following directional hypothesis:

H1: There is a positive association between strategic deviation and investment inefficiency.

It is important to understand what factors exacerbate the strategic deviation and investment inefficiency association. We argue that corporate governance, product market competition and the quality of the information environment moderate the association between strategic deviation and investment inefficiency.

2.1. Corporate governance

Independent monitoring is important to alleviate managerial opportunistic behaviour (Fama and Jensen, 1983). Weakly monitored CEOs may engage in empire building by investing in unprofitable projects, leading to over-investments (Jensen, 1986). Similarly, CEOs are likely to subsidize poorly performing divisions to gain personal benefits (Jensen and Meckling, 1976), which also results in investment inefficiencies. At the same time, excessive monitoring distorts efficient decision-making. Therefore, it is important to implement effective monitoring mechanisms to encourage efficient investments. Independent board members act in the best interest of shareholders by effectively monitoring managerial investment behaviour, among other activities (Fama and Jensen, 1983). Rajkovic (2020) finds that board independence increases investment efficiency, but board independence compromised by CEO and director social ties decreases investment efficiency (Kang et al., 2021). Institutional ownership, an external governance mechanism, can also play a role in determining investment efficiencies through stronger monitoring (Biddle et al., 2009). Ferreira and Matos (2008) find lower capital expenditure in firms with high institutional ownership, which suggests that institutional ownership reduces over-investments. Therefore, we argue that the positive association in H1 above is likely to be stronger for deviant firms lacking effective governance mechanisms.

H2: The positive association between strategic deviation and investment inefficiency is stronger for poorly governed firms.

2.2. Product market competition

Two competing arguments exist on the association between product market competition and investment efficiency (Babar and Habib, 2021). Product market competition acts as an external governance mechanism, and thus disciplines managerial behaviour. In line with this disciplinary argument, a stream of studies find that product market competition increases capital expenditure and R&D but curbs over-investments (Jiang et al., 2015; Laksmana and Yang, 2015) and increases cash flow–enhancing investments (Abdoh and Varela, 2017). A competing view based on the signal precision perspective posits a negative association between product market competition and investment efficiency. Firms operating in a highly competitive market are reluctant to obtain precise signals about their rivals’ actions due to the marginal effect of a single firm’s signal. As information is costly, managers tend to gather information only when it generates higher profits, ex ante. The impact of one firm’s signal is marginal when there are a lot of firms in the industry. Therefore, there is less incentive for information gathering in a competitive market. Investment efficiency is weaker in such a setting. Stoughton et al. (2017) find support for this prediction. Furthermore, prior research provides evidence on the risk-increasing effect of competition (Irvine and Pontiff, 2008). Given the inconclusive evidence on the association between product market competition and corporate investment efficiency (Babar and Habib, 2021), we develop the following hypothesis:

H3: Product market competition moderates the association between strategic deviation and investment inefficiency.

2.3. Information environment

We examine how the information environment proxied by analyst following and financial statement comparability moderates the association between strategic deviation and investment inefficiency. Low analyst following reduces the quality of the information environment because of less transparency and high information processing costs for investors. High analyst following, on the other hand, reduces mispricing and information asymmetry (Brennan and Subrahmanyam, 1995; Chung et al., 1995; Roulstone, 2003). As strategic deviation augments information asymmetry, capital market participants tend to discount firms with unique strategies (Litov et al., 2012). Litov et al. (2012) argue that unique strategies create substantial cost for analysts to understand a firm’s resource allocation behaviour, and hence, such firms are followed by fewer analysts. Weaker monitoring courtesy of a low analyst following for firms with deviant strategies provides incentives for managers of such firms to engage in sub-optimal investments. We, therefore, hypothesize the following:

H4a: The positive association between strategic deviation and investment inefficiency is stronger for firms with high information asymmetry.

Financial statement comparability (Comparability) is the extent of similarity in accounting choices between two or more entities under similar economic conditions (Financial Accounting Standards Board (FASB, 2010). Comparability enables investors to properly identify similarities and differences in firm performance within an industry (Francis et al., 2014). Higher comparability helps outsiders to identify peer firms’ risk of under- and over-investments (i.e. investment inefficiencies) due to the availability of benchmark information. As a result, high comparability reduces agency frictions of empire building and moral hazard behaviour. Al-Hadi et al. (2021) find that comparability increases investment efficiency. On the other hand, low comparability increases information asymmetry and reduces the quality of the information environment (De Franco et al., 2011). Firms that follow deviant strategies have extremely poor performance (Tang et al., 2011), and thus provide incentives for managers to produce less comparable financial statements to conceal this poor performance. As sub-optimal investments by firms following deviant strategies may be associated with poor future performance, managers of such firms are likely to produce less comparable financial statements to conceal such sub-optimal behaviour. We, therefore, hypothesize as follows:

H4b: The positive association between strategic deviation and investment inefficiency is stronger for firms with low financial statement comparability.

3. Methodology

3.1. Sample

The sample covers data ranging from 1987 to 2020. We begin with 1987, as the data on cash flow from operations are available on the Compustat dataset from 1987. We obtain governance data from Thomson Reuters and analyst following data from I/B/E/S. Since we require five years of data for calculating strategic deviation and also require five years of lagged values for cash flows to calculate the control variables (e.g. cash flow volatility), our actual regression spans the period 1992 to 2020. We winsorize all the continuous variables at the top and bottom one percentiles to reduce the impact of outliers. We also exclude financial firms (SIC codes between 6000 and 6999). We further exclude observations with missing values for the measurement of key dependent, independent and control variables. Our final sample consists of 56,133 firm-year observations. The number of observations in any given regression varies depending on the model-specific data requirements. Panel A in Table 1 presents the sample selection procedure, while Panel B provides details of the industry distribution of the sample based on the two-digit SIC industry classification. About 33% observations come from machinery, electrical, computer equipment, followed by business services (16%) and chemical, petroleum and rubber and allied products (12%).

Table 1.

Sample selection and distribution.

Panel A: sample selection.
Initial sample for the period 1992–2020			238,691
Less: observations in the financial institutions (#60 – #69)			(67,490)
Less: observations lost for strategy deviation calculation			(40,081)
Less: observations lost for investment efficiency and control variable calculation			(74,987)
Final sample			56,133
Panel B: sample distribution by industry.
2-Digit SIC	Industry	Observations	%
01-14	Agriculture & mining	2,276	4.05
15-17	Building construction	444	0.79
20-21	Food & kindred products	1,553	2.77
22-23	Textile mill products & apparels	1,111	1.98
24-27	Lumber, Furniture, Paper, and Printing	2,088	3.72
28-30	Chemical, Petroleum, and Rubber & Allied Products	6,769	12.06
31-34	Metal	2,547	4.54
35-39	Machinery, electrical, computer equipment	18,416	32.81
40-49	Railroad, Communications and Other Transportation	2,541	4.53
50-51	Wholesale goods, building materials	2,783	4.96
53-59	Store merchandise, auto dealers, home furniture stores	3,372	6.01
70-79	Business services	9,085	16.18
80-99	Others	3,148	5.60
	Total	56,133	100.00

Note: This table presents the sample selection procedure (Panel A) and sample distribution by industry (Panel B). SIC = Standard Industrial Classification.

3.2. Variable measurements

3.2.1. Independent variable: Strategic deviation (STRAT_DEV)

We follow previous research (Dong et al., 2021; Ye et al., 2021) and measure strategic deviation $(S T R A T_D E V$ ) using six indicators, namely, (a) Advertising intensity, measured as advertising expenditure (XAD) over sales (REVT); (b) R&D intensity, measured as R&D expenditure³ scaled by sales; (c) Plant and equipment newness, measured as net property, plant and equipment (PPENT) scaled by gross property, plant and equipment (PPEGT); (d) Non-production overhead, measured using selling, general and administrative (SG&A) expenses scaled by sales; (e) Inventory level, measured using inventories (INVT) scaled by sales; and (f) Financial leverage, measured as debt (DLTT) over equity (CEQ). For each fiscal year, we standardize each of the measures by industry (based on two-digit SIC codes) and calculate the absolute difference between a firm’s score and its industry average. $S T R A T_D E V$ is the sum of the six standardized difference scores. A higher value of $S T R A T_D E V$ implies a larger deviation of a firm’s strategy from industry norms.

3.2.2. Dependent variable: Investment ( $I N V) .$

Following prior research (Biddle et al., 2009; Rajkovic, 2020), we first use the following regression.

$I N V_{i t} = α_{0} + β_{1} Δ S a l e s_{i t - 1} + ε_{i, t},$ (1)

where $I N V$ is total investments, calculated as the sum of R&D expenditure, capital expenditure (CAPEX) and acquisition expenditure (AQC), less cash receipts from sale of property, plant and equipment (SPPE), and scaled by the lagged total assets (AT). $Δ S a l e s_{i t - 1}$ is the sales change from the last year (for the remainder of the paper we denote this as SALESG). An overwhelming body of empirical research uses two-step regression to estimate and then examine the key determinant(s) of investment in(efficiency). More specifically, the residuals from Equation (1) above is considered by researchers to represent in(efficient) investment and is used as the dependent variable in the second stage regression. Due to the concerns regarding the use of two-step regression models (see Chen et al., 2018, 2022; Jackson, 2022), we use a one-step regression model which includes the first step control (i.e. SALESG), years and interaction terms between each of industries, years and the independent variable from the first-step regression to alleviate the biased coefficient and other econometrics concerns (Chen et al., 2022; Jackson, 2022).

3.3. Empirical model

We test the association between $S T R A T_D E V$ and $I N V$ using the following OLS regression model:

$\begin{array}{l} I N V_{i, t} = β_{0} + β_{1} S T R A T_D E V_{i, t} + β_{2} M T B_{i, t} + β_{3} S I Z E_{i, t} + β_{4} C F L_V O L_{i, t} \\ + β_{5} S A L E S_V O L_{i, t} + β_{6} I N V_V O L_{i, t} + β_{7} T A N G_{i, t} + β_{8} C F O S A L E S_{i, t} \\ + β_{9} S L A C K_{i, t} + β_{10} L O S S_{i, t} + β_{11} A G E_{i, t} + β_{12} Z S C O R E_{i, t} + β_{13} D I V_{i, t} \\ + β_{14} O P C Y C L E_{i, t} + β_{15} | D A C C_{i, t} | + β_{16} S A L E S G_{i, t} + \sum_{j} λ_{j} I N D U S T R Y_{j, t} \\ + \sum_{k} λ_{k} Y E A R_{k, t} + \sum_{j} λ_{j} I N D U S T R Y_{j, t} * S A L E S G_{j, t} + \sum_{k} λ_{k} Y E A R_{k, t} * S A L E S G_{i, t} + ε_{i, t} \end{array}$ (2)

where $I N V$ is the investment and $S T R A T_D E V$ is the strategic deviation measure as mentioned above. β₁₆ is the first-stage variable which is also interacted with INDUSTRY and YEAR variables. A positive and significant coefficient on β₁ would support our H1. Detailed variable definitions of the control variables are in Appendix 1.

4. Empirical results

4.1. Descriptive statistics

We present descriptive statistics of the variables in Table 2. Results show that the mean and median INV is 0.160 and 0.103 $. S T R A T_D E V$ has a mean and a median of 2.237 and 1.817. The mean (median) of $M T B$ is 3.384 (2.015), $C F L_V O L$ is 0.579 (0.015), $S A L E S_V O L$ is 0.211 (0.147) and $I N V_V O L$ is 0.151 (0.058). Of the sample observations, 36% have reported a loss and about 28% have paid dividends. The average absolute discretionary accruals ( $| D A C C |$ ) are 10% of lagged total assets.

Table 2.

Descriptive statistics.

Variables	Obs	Mean	SD	0.25	0.50	0.75
INV	56,133	0.160	0.194	0.045	0.103	0.202
STRAT_DEV	56,133	2.237	1.725	1.259	1.817	2.594
MTB	56,133	3.384	5.261	1.193	2.015	3.570
SIZE	56,133	5.210	2.214	3.622	5.150	6.732
CFL_VOL	56,133	0.579	4.218	0.002	0.015	0.108
SALES_VOL	56,133	0.211	0.206	0.081	0.147	0.264
INV_VOL	56,133	0.151	1.623	0.026	0.058	0.129
TANG	56,133	0.240	0.214	0.079	0.170	0.336
CFOSALES	56,133	0.005	0.214	–0.003	0.060	0.117
SLACK	56,133	4.206	13.451	0.125	0.623	2.757
LOSS	56,133	0.357	0.479	0.000	0.000	1.000
AGE	56,133	2.913	0.534	2.565	3.045	3.367
ZSCORE	56,133	1.143	1.679	0.637	1.345	1.999
DIV	56,133	0.275	0.447	0.000	0.000	1.000
OPCYCLE	56,133	4.719	0.793	4.322	4.804	5.208
\|DACC\|	56,133	0.103	0.105	0.031	0.070	0.137
SALESG	56,133	0.230	0.788	–0.033	0.078	0.247
BIND	18,686	0.745	0.124	0.667	0.750	0.833
INSTOWN	15,168	0.367	0.482	0.000	0.000	1.000
COMPETITION	53,204	6.817	3.668	4.160	6.015	8.646
ANALYST	27,766	2.125	0.667	1.609	2.079	2.639
FCOMP	33,313	–3.133	1.686	–3.820	–2.800	–2.070

Note: This table shows summary statistics of the variables used in the regression models.

Table 3 shows the correlation coefficients between the variables used in the empirical models. $I N V$ and $S T R A T_D E V$ are positively significantly correlated (0.004). The correlation between $S T R A T_D E V$ and all of the independent variables is significant at the 1% level, except INV_VOL. As per the unreported variance inflation factors (VIFs), the maximum $v i f$ is 1.85 with a mean of 1.32, suggesting that multicollinearity is not a concern.

Table 3.

Correlation analysis.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. INV	1.000
2. STRAT_DEV	0.004	1.000
3. MTB	0.182	0.159	1.000
4. SIZE	0.088	–0.145	0.191	1.000
5. CFL_VOL	0.085	0.079	0.104	–0.114	1.000
6. SALES_VOL	–0.021	0.042	0.012	–0.245	0.095	1.000
7. INV_VOL	0.054	–0.001	0.024	–0.006	0.096	0.041	1.000
8. TANG	0.011	0.020	–0.083	–0.002	–0.041	–0.146	0.002	1.000
9. CFOSALES	–0.187	–0.195	–0.174	0.294	–0.187	–0.063	–0.037	0.175	1.000
10. SLACK	0.100	0.047	0.077	–0.023	0.141	0.015	0.016	–0.281	–0.247	1.000
11. LOSS	0.106	0.126	0.098	–0.317	0.115	0.075	0.033	–0.060	–0.657	0.134	1.000
12. AGE	–0.100	–0.071	–0.043	0.260	–0.085	–0.151	–0.030	0.017	0.204	–0.067	–0.223	1.000
13. ZSCORE	–0.289	–0.137	–0.218	0.143	–0.180	0.134	–0.045	0.024	0.675	–0.214	–0.535	0.185	1.000
14. DIV	–0.146	–0.038	–0.041	0.339	–0.081	–0.167	–0.027	0.131	0.260	–0.117	–0.311	0.291	0.264	1.000
15. OPCYCLE	0.013	0.050	–0.021	–0.067	0.025	–0.140	–0.000	–0.301	–0.059	–0.015	0.023	0.044	–0.099	–0.036	1.000
16. \|DACC\|	0.185	0.060	0.133	–0.164	0.100	0.111	0.025	–0.117	–0.281	0.109	0.206	–0.122	–0.232	–0.165	0.042	1.000

Note: This table presents the correlation coefficient between variables used in the base models. Bold-faced correlations are significant at p < 0.01. Variables are defined in Appendix 1.

4.2. Regression results

Table 4 presents the baseline regression results whereby we regress $I N V$ on $S T R A T_D E V$ and a set of control variables with standard errors clustered at the firm level. As elaborated in Section 3.2.2, we use one-step regression model. In Column (1), we present pooled OLS regression results where $I N V$ is regressed on $S T R A T_D E V$ , a set of control variables, the first stage control variable (SALESG), industry and year indicators, SLAESG*industry effects and SALESG*year effects. The coefficient on $S T R A T_D E V$ is positive and significant (β = 0.003, p < 0.01), suggesting a positive relationship between strategic deviation and investment inefficiency, thereby supporting H1. The reported coefficient implies that a one-standard-deviation increase in $S T R A T_D E V$ increases $I N V$ by 3.23% relative to mean INV ((0.003*1.725 (S.D. of STRAT_DEV)/0.16 (mean INV)). In Column 2, we present robust regression results. The coefficient on $S T R A T_D E V$ continues to be positive and significant (β = 0.001, p < 0.01). Among the control variables, firm size (SIZE) is positively and significantly associated with INV. Sales volatility (SALES_VOL) is positively and significantly associated with INV in the pooled model but insignificant in the robust model. Other controls, such as, SLACK, ZSCORE and OPCYCLE, are negatively and significantly associated with INV. Discretionary accruals (|DACC|) is positively related to INV. These are consistent with prior research (Biddle et al., 2009; Rajkovic, 2020).

Table 4.

Baseline regression: Strategic deviation and investment inefficiency.

	(1)	(2)
	Pooled	Robust
STRAT_DEV	0.003***(0.001)	0.001**(0.000)
MTB	0.000***(0.000)	0.000***(0.000)
SIZE	0.016***(0.001)	0.009***(0.000)
CFL_VOL	0.000(0.000)	–0.000(0.000)
SALES_VOL	0.032***(0.006)	0.001(0.001)
INV_VOL	0.003***(0.001)	0.001***(0.000)
TANG	0.077***(0.008)	0.089***(0.003)
CFOSALES	–0.028***(0.010)	–0.000***(0.000)
SLACK	–0.000**(0.000)	–0.000***(0.000)
LOSS	–0.016***(0.002)	–0.004***(0.001)
AGE	–0.014***(0.002)	–0.008***(0.001)
ZSCORE	–0.020***(0.001)	–0.006***(0.000)
DIV	–0.023***(0.002)	–0.011***(0.001)
OPCYCLE	–0.007***(0.002)	–0.002***(0.001)
\|DACC\|	0.162***(0.011)	0.045***(0.003)
First step control
SALESG	0.060**(0.029)	0.112***(0.024)
Industry effects	Yes	Yes
Year effects	Yes	Yes
SALESG*Industry effects	Yes	Yes
SALESG*Year effects	Yes	Yes
Constant	0.097***(0.025)	0.1094***(0.024)
Observations	56,133	56,133
Adjusted R²	0.290	0.570

Note: This table shows the regression estimates of the association between strategic deviation (STRAT_DEV) and investment inefficiency (INV) for the pooled model (column 1) and for the robust model (column 2). Both models include the first step control variable, SALESG, and the interactions between SALESG and Fyear and SALESG and Industry (Chen et al., 2022; Jackson, 2022). Robust standard errors clustered at firm-level are in parentheses. **, *** denote a two-tailed p-value of less than 0.10, 0.05 and 0.01, respectively. Variables are defined in Appendix 1.

4.3. Addressing endogeneity

Although the preceding analyses control for a variety of firm characteristics that might explain the association between $S T R A T_D E V$ and investment inefficiency, endogeneity could be a concern that may bias the findings. Endogeneity could arise from omitted variable concern, reverse causation problem (i.e. firms that are less investment efficient could deviate more from their peers), and design choices. We use several endogeneity tests and report the results in Table 5. First, we estimate fixed effect models. As shown in Panel A, our original results hold when we use fixed effects estimates (β = 0.004, p < 0.01). Next, we use the entropy balancing method to address any design choices. McMullin and Schonberger (2020) document that entropy balancing noticeably improves covariate balance when compared with propensity-score approaches. In addition, entropy balancing improves the balance quality without removing any observations from either treatment or control group (Hainmueller, 2012). To conduct entropy balancing estimates, we divide our sample into two groups based on the mean STRAT_DEV. We consider firm-year observations with above (below) mean STRAT_DEV as the treated (control) group. In Panel B, we first report means, variances and skewness of all covariates before and after balancing, and the results suggest a reasonable covariate balance. Using the entropy balanced sample, we then run the baseline regression and find the coefficient on $S T R A T_D E V$ positive and significant (β = 0.004, p < 0.01).

Table 5.

Endogeneity tests.

Panel A: Fixed effect model.
	INV
STRAT_DEV	0.004***(0.001)
Other control variables	Yes
First step control
SALESG	0.078***(0.001)
Constant	–0.117***(0.015)
Industry effects	No
Year Effects	Yes
SALESG*Year effects	Yes
Firm effect	Yes
Observations	56,133
Adjusted R²	0.08

Panel B: Entropy balancing models.
Covariate matching table.
	Before balancing						After balancing
	Treat			Control			Control
	Mean	Variance	Skewness	Mean	Variance	Skewness	Mean	Variance	Skewness
MTB	3.534	39.270	5.826	3.235	16.910	6.185	3.534	39.270	5.826
SIZE	4.921	5.048	0.213	5.588	4.811	0.135	4.921	5.048	0.213
CFL_VOL	0.676	19.830	13.810	0.452	14.320	17.700	0.676	19.830	13.810
SALES_VOL	0.227	0.050	2.270	0.196	0.036	2.502	0.227	0.050	2.270
INV_VOL	0.144	2.614	86.320	0.173	4.871	75.040	0.144	2.614	86.320
TANG	0.258	0.050	1.074	0.231	0.043	1.487	0.258	0.050	1.074
CFOSALES	–0.012	0.048	–1.688	0.025	0.042	–1.850	–0.012	0.048	–1.688
SLACK	4.527	235.600	6.585	3.538	108.900	8.303	4.527	235.600	6.585
LOSS	0.381	0.236	0.493	0.334	0.223	0.703	0.381	0.236	0.493
AGE	2.900	0.285	–0.647	2.933	0.280	–0.744	2.900	0.285	–0.647
ZSCORE	1.133	3.501	–2.193	1.143	2.092	–2.252	1.133	3.501	–2.193
DIV	0.284	0.203	0.958	0.282	0.202	0.971	0.284	0.203	0.958
OPCYCLE	4.658	0.771	–0.563	4.754	0.499	–0.799	4.658	0.771	–0.563
\|DACC\|	0.105	0.011	1.993	0.100	0.011	2.070	0.105	0.011	1.993
SALESG	0.160	0.404	5.630	0.179	0.319	6.217	0.160	0.404	5.630

Entropy balancing regression
Variable	INV
STRAT_DEV	0.004***(0.0007)
Controls	Yes
First step control
SALESG	0.074**(0.033)
Constant	0.085***(0.0198)
Industry effects	Yes
Year Effects	Yes
SALESG*Industry effects	Yes
SALESG*Year effects	Yes
Observations	56,133
Adjusted R²	0.301

Panel C: 2SLS.
Variables	1st stage regression	2nd stage regression	1st stage regression	2nd stage regression
	DV = SRAT_DEV	DV = INV	DV = SRAT_DEV	DV = INV
	Geographical Proximity index		Tightness/Looseness index
PROXIMITY	0.001***(0.000)
LOOSENESS			0.057***(0.013)
SRAT_DEV_PRED		0.015*(0.009)		0.202***(0.055)
MTB	0.000***(0.000)	0.000(0.000)	0.001***(0.000)	–0.000***(0.000)
SIZE	–0.048***(0.004)	0.020***(0.001)	–0.061***(0.004)	0.031***(0.003)
CFL_VOL	0.002(0.002)	0.001***(0.000)	0.011**(0.002)	–0.001**(0.001)
SALES_VOL	0.179***(0.037)	0.0324***(0.005)	0.202***(0.034)	–0.006(0.014)
INV_VOL	0.000(0.003)	0.003***(0.000)	–0.009**(0.004)	0.005***(0.001)
TANG	–0.631***(0.052)	0.040***(0.008)	–0.571***(0.045)	0.169***(0.034)
CFOSALES	–0.838***(0.052)	–0.020*(0.011)	–0.976***(0.056)	0.156***(0.055)
SLACK	0.008***(0.001)	–0.000(0.000)	0.006***(0.001)	–0.001***(0.000)
LOSS	–0.018(0.019)	–0.021***(0.003)	–0.027(0.018)	–0.020***(0.004)
AGE	–0.024*(0.015)	–0.022***(0.002)	–0.010(0.013)	–0.017***(0.003)
ZSCORE	–0.141***(0.006)	–0.024***(0.001)	–0.129***(0.006)	0.002(0.007)
DIV	–0.109***(0.018)	–0.044***(0.003)	–0.089***(0.017)	–0.027***(0.007)
OPCYCLE	0.295***(0.011)	–0.013***(0.003)	0.319***(0.010)	–0.072***(0.018)
\|DACC\|	0.485***(0.066)	0.207***(0.010)	0.444***(0.064)	0.122***(0.029)
Industry effects	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes
Constant		0.025(0.196)		–0.341**(0.142)
Observations	40,951	40,951	55,172	55,172
Adj R²		0.176	–	0.310
Kleibergen-Paap rk LM statistic: p-value	0.000		0.000
Weak identification test:
Cragg-Donald Wald F-stat.	203.44		19.14
Stock-Yogo critical value	16.38		16.38

Panel D: Two-step system generalized method of moments (GMM).
Variables	(1)
STRAT_DEV	0.012***(0.003)
MTB	0.0010(0.001)
SIZE	0.0446***(0.005)
CFL_VOL	0.0013(0.001)
SALES_VOL	0.0358**(0.017)
INV_VOL	0.0048***(0.001)
TANG	0.1193***(0.039)
CFOSALES	0.1192**(0.059)
SLACK	–0.0011**(0.000)
LOSS	–0.0832***(0.012)
AGE	–0.0395(0.028)
ZSCORE	–0.0400***(0.006)
DIV	–0.0265**(0.012)
OPCYCLE	–0.0594***(0.014)
\|DACC\|	0.2827***(0.080)
First stage control
SALESG	0.0724(0.137)
AR(1) (p value)	0.000
AR(2) (p value)	0.445
Industry effects	Yes
Year effects	Yes
SALESG*Year effects	Yes
Observations	56,133

Note: This table shows the endogeneity tests. Panel A presents the fixed effect estimates. Panel B shows the results from the entropy balanced test. Panel C shows the 2SLS models. Panel D presents two-step system GMM results. Robust standard errors clustered at firm-level are in parentheses. *, **, *** denote a two-tailed p-value of less than 0.10, 0.05 and 0.01, respectively. Variables are defined in Appendix 1. GMM = generalized method of moments.

Then we perform 2SLS estimates⁴ and present the results in Panel C of Table 5. First, we use geographical proximity (PROXIMITY) to local largest strategic-deviant firm as an instrumental variable. Prior research suggests that decision making strategies are similar among companies in the local community (Pool et al., 2015). Therefore, it is more likely that the focal firm pursues a deviant strategy if that firm is geographically closer to the largest strategic-deviant firm in the city. The largest deviant firm in the city is known as the local-largest deviant firm. Following prior research (Dong et al., 2021; Pool et al., 2015), we first find the latitudes and longitudes for all the observations.⁵ We then determine the latitudes and longitudes of the largest strategic-deviant firm at the city level. Then we use the Vincenty’s formula⁶ to calculate the geographical proximity between the focal firm and the largest strategic-deviant firm based on longitudes and latitudes of the focal firm and the local largest strategic-deviant firm. Column 1 reveals that the instrument PROXIMITY is positively and significantly associated with STRAT_DEV (β = 0.001, p < 0.01), supporting our assertion that geographical proximity to the local largest strategic-deviant firm increases focal firm’s strategic deviation. As shown in the second stage, the STRAT_DEV_PRED is positively and significantly associated with INV (β = 0.015, p < 0.10). The LM statistic reveals that the excluded instruments are ‘relevant’. The Cragg–Donald F-statistic of 203.44 is higher than the Stock and Yogo (2005) critical value of 16.38, implying that our instrument does not suffer from weak identification.

Second, following Harrington and Gelfand (2014), we use the US state-level tightness-looseness index (LOOSENESS) as the instrumental variable. Harrington and Gelfand (2014) argue that the extent of enforcement of laws and regulations among the states are different and hence, the tolerance level of deviation from law implementation differs between the states. Provaty et al. (2022) argue that the state-level looseness (i.e. if the state’s tolerance for lack of law enforcement is high) has a positive association with strategy deviation. We, therefore, use LOOSENESS index as our second instrumental variable. We take the view that when the LOOSENESS score is high, there is more tolerance towards strategic deviation, hence, we expect a positive association between LOOSENESS and STRAT_DEV. As shown in Column 3, we indeed find a positive association between LOOSENESS and STRAT_DEV (β = 0.057, p < 0.01). In Column 4, we find the coefficient on STRAT_DEV_PRED positive and significant (β = 0.202, p < 0.01). The Cragg–Donald F-statistic, however, is much lower than the first instrument.

Finally, we use the two-step system GMM (generalized method of moments) approach adopted by Arellano and Bover (1995) and Blundell and Bond (1998) to validate our results documented in Table 4. This should also alleviate any concerns with reverse causality concern. The coefficient on STRAT_DEV is positive and significant (coefficient 0.012, p < 0.01). Given that errors in levels are serially uncorrelated, we expect significant first-order serial correlation, but insignificant second-order correlation in the first-differenced residuals. Test results reported at the bottom of Panel D confirm the desirable statistically significant AR(1), and statistically insignificant AR(2).

Based on these endogeneity tests above, we conclude that our results are not affected by endogeneity concerns.

4.4. Strategic deviation and investment inefficiency: Cross-sectional tests

Table 6 presents the cross-sectional test results for the association between strategic deviation and investment inefficiency. As discussed in Section 2, we consider corporate governance (board independence and institutional ownership), product market competition, and the quality of the information environment proxied by analyst following and financial statement comparability as the three cross-sectional settings that are likely to moderate the relation between strategic deviation and investment inefficiency. Regression results are based on a sub-sampling procedure which allows the coefficients on all the control variables to vary between the groups.

Table 6.

Strategic deviation and investment efficiency: Cross-sectional tests.

	1	2	3	4	5	6	7	8	9	10
	BIND		INSTOWN		COMPETITION		ANALYST		FCOMP
	Low	High	Low	High	Low	High	Low	High	Low	High
STRAT_DEV	0.001(0.001)	0.003(0.002)	0.004*(0.002)	–0.000(0.002)	0.001(0.001)	0.004***(0.001)	0.003*(0.002)	0.001(0.001)	0.004***(0.001)	–0.001(0.001)
MTB	0.000***(0.000)	0.000*(0.000)	0.000(0.000)	0.000*(0.000)	0.000***(0.000)	0.000(0.000)	–0.000**(0.000)	0.000***(0.000)	0.000(0.000)	0.000(0.000)
SIZE	0.017***(0.001)	0.010***(0.002)	0.010***(0.002)	0.008***(0.002)	0.012***(0.001)	0.014***(0.001)	0.010***(0.002)	0.001(0.002)	0.017***(0.001)	0.011***(0.001)
CFL_VOL	0.001(0.001)	0.001(0.001)	0.003*(0.001)	0.005*(0.003)	0.001(0.001)	–0.000(0.000)	0.002*(0.001)	–0.001(0.001)	0.001(0.001)	0.003***(0.001)
SALES_VOL	0.045***(0.013)	0.017(0.018)	0.033**(0.015)	0.040**(0.019)	0.044***(0.007)	0.027***(0.008)	0.037***(0.011)	–0.006(0.013)	0.047***(0.011)	0.002(0.010)
INV_VOL	0.004***(0.001)	–0.005***(0.001)	0.002(0.001)	–0.026(0.016)	0.003***(0.000)	0.002(0.002)	0.002***(0.000)	–0.030***(0.010)	–0.001(0.001)	0.003***(0.001)
TANG	0.018(0.016)	0.033*(0.018)	0.011(0.023)	0.022(0.022)	0.092***(0.011)	0.076***(0.013)	0.040**(0.016)	0.062***(0.016)	0.035**(0.015)	0.060***(0.014)
CFOSALES	–0.024(0.018)	0.001(0.023)	–0.028(0.025)	–0.019(0.023)	0.023*(0.013)	–0.037***(0.013)	0.013(0.017)	–0.050**(0.020)	–0.018(0.016)	–0.085***(0.017)
SLACK	–0.000*(0.000)	–0.000(0.000)	–0.000*(0.000)	–0.001***(0.000)	–0.000**(0.000)	–0.000***(0.000)	–0.001**(0.000)	0.000(0.000)	–0.000(0.000)	–0.001***(0.000)
LOSS	–0.008(0.006)	0.007(0.007)	0.000(0.007)	0.001(0.007)	–0.015***(0.003)	–0.014**(0.004)	–0.019***(0.005)	0.006(0.007)	–0.008*(0.004)	–0.013***(0.005)
AGE	–0.023***(0.006)	–0.029***(0.007)	–0.015**(0.008)	–0.027***(0.007)	–0.009***(0.003)	–0.014***(0.003)	–0.012***(0.004)	–0.006(0.004)	–0.014***(0.004)	–0.014***(0.004)
ZSCORE	–0.026***(0.003)	–0.028***(0.003)	–0.022***(0.003)	–0.030***(0.004)	–0.022***(0.002)	–0.018***(0.002)	–0.038***(0.003)	–0.024***(0.004)	–0.024***(0.002)	–0.035***(0.003)
DIV	–0.011**(0.005)	–0.029***(0.005)	–0.013**(0.006)	–0.003(0.006)	–0.015***(0.003)	–0.023***(0.004)	–0.013***(0.004)	–0.041***(0.005)	–0.021***(0.004)	–0.032***(0.004)
OPCYCLE	–0.013***(0.004)	–0.010*(0.005)	–0.023***(0.007)	–0.005(0.005)	0.001(0.003)	–0.010***(0.003)	–0.012***(0.004)	–0.003(0.004)	–0.015***(0.004)	–0.006*(0.004)
\|DACC\|	0.171***(0.027)	0.164***(0.032)	0.120***(0.037)	0.195***(0.033)	0.115***(0.015)	0.190***(0.017)	0.203***(0.021)	0.249***(0.027)	0.148***(0.020)	0.196***(0.019)
First step control
SALESG	0.408***(0.068)	0.108(0.091)	–0.185**(0.087)	0.686***(0.049)	0.048(0.038)	0.096*(0.050)	0.109(0.087)	0.272(0.181)	–0.029(0.058)	0.191**(0.096)
Industry effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
SALESG*Industry effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
SALESG*Year effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.127***(0.032)	0.085(0.060)	0.164**(0.071)	0.175***(0.046)	0.072**(0.032)	0.064*(0.033)	0.221***(0.045)	0.218***(0.057)	0.066*(0.039)	0.332***(0.045)
Observations	10,460	8,226	7,078	8,090	24,594	28,610	15,477	12,289	17,195	16,118
Adjusted R²	0.272	0.260	0.316	0.333	0.301	0.262	0.323	0.268	0.321	0.259

Note: This table presents the cross-sectional test results. Columns 1 and 2 shows the association between STRAT_DEV and INV for Low and High board independence (BIND) sub-samples. Columns 3 and 4 shows the association between STRAT_DEV and INV for Low and High institutional ownership (INSTOWN) sub-samples. Columns 5 and 6 show regression estimates for the association between STRAT_DEV and INV for the Low and High competition (COMPETITION) sub-sample, when the competition is measured using product market fluidity. Columns 7 and 8 show the association between STRAT_DEV and INV for the Low and High sub-samples of number of analysists following (ANALYST). Columns 9 and 10 show the regression estimates of the association between STRAT_DEV and INV for the Low and High sub-samples of financial statement comparability (FCOMP). Robust standard errors clustered at firm-level are in parentheses *, **, *** denote a two-tailed p-value of less than 0.10, 0.05 and 0.01, respectively. Variables are defined in Appendix 1.

Columns 1 and 2 present regression estimates of the association between STRAT_DEV and $I N V$ for the board independence ( $B I N D$ ) context. The low- (high-) BIND sub-sample consists of firm-year observations that have board independence (i.e. proportion of independent directors on board) lower (higher) than the industry-year median, respectively. We do not find empirical evidence to support our assertion for board independence. Columns 3 and 4 present regression results for the low- and high-institutional ownership ( $I N S T O W N$ ) sub-sample. The low- (high-) $I N S T O W N$ sub-sample includes observations where the $I N S T O W N$ is lower (higher) than the industry-year median. As shown in Column 3, we find a positive and marginally significant (β = 0.004, p < 0.10) association between $S T R A T_D E V$ and $I N V$ for the low- $I N S T O W N$ sub-sample, suggesting that strategic deviation increases investment inefficiencies in the presence of weaker external monitoring (i.e. low $I N S T O W N$ ). The coefficient on $S T R A T_D E V$ for the high- $I N S T O W N$ sub-sample is insignificant. The results of board independence and institutional ownership, therefore, provides partial support for H2.

Then, we examine the moderating effect of product market competition on the association between $S T R A T_D E V$ and $I N V .$ We proxy product market competition using the fluidity measure (FLUID) developed by Hoberg et al. (2014). The data are sourced from the Hoberg-Phillips Data Library.⁷ The low- (high-) product-market-competition sub-sample includes observations that have a FLUID value lower (higher) than the industry-year median FLUID. The coefficient on $S T R A T_D E V$ is positive and significant for the high-product-market-competition sub-sample, (β = 0.004, p < 0.01) (Column 6), while this is insignificant for the low-product-market-competition sub-sample (Column 5). This finding suggests that the escalated risks and agency problems stemming from high competition are associated with sub-optimal investments by deviant firms and supports signal precision view of H3.

Finally, to ascertain the moderating effects of the information environment on the association between $S T R A T_D E V$ and $I N V$ , first we divide the sample into low and high analyst following. The low- (high-) analyst-following sub-sample includes observations with analyst following ( $L N A N A L Y S T$ ) lower (higher) than the industry median for each year. Results show that the coefficient on $S T R A T_D E V$ is positive and significant for the low-analyst-following sub-sample (β = 0.003, p < 0.10) (Column 7). However, the coefficient is insignificant for the high-analyst-following sub-sample (Column 8). This supports the view that strategic deviation increases information opaqueness and, hence, managerial opportunism and consequently sub-optimal investments. The result, therefore, supports H4a. We also investigate whether the association between $S T R A T_D E V$ and $I N V$ is stronger when the financial statement comparability $(F C O M P)$ is lower. We follow De Franco et al. (2011) and measure financial statement comparability using time-series regression.⁸ We use $I N D_C O M P$ , the median financial statement comparability for all firms j in the same industry as firm $i$ during year $t,$ as our measure of financial statement comparability. The low- (high-) $F C O M P$ sub-sample includes observations that have an $I N D_C O M P$ value lower (higher) than the industry-year median. As shown in Column 9, the coefficient on $S T R A T_D E V$ is positive and significant for the low-FCOMP sub-sample only (β = 0.004, p < 0.01), thereby supporting H4b.

4.5. Robustness tests

In our first robustness test, we develop an alternative investment inefficiency measure $(A L T_I N V_I N E F F)$ following Eisdorfer et al. (2013). First, we determine the actual investments for each firm-year by scaling the gross CAPEX by the book value of total assets at the beginning of the year. We then determine expected investment where the expected investment is calculated as the median investment in the four-digit SIC industries. Then we estimate the abnormal investments by subtracting actual investments from expected investments. We use the absolute value of abnormal investment and denote this as $A L T_I N V_I N E F F .$ Table 7 shows the coefficient on $A L T_I N V_I N E F F$ positive and significant (β = 0.313, p < 0.01) (Column 1).

Table 7.

Strategic deviation and investment efficiency: robustness tests.

	(1)	(2)
	ALT_INV_INEFF	INV
STRAT_DEV	0.313***(0.044)
ALT_STRAT_DEV		0.002**(0.001)
MTB	0.000(0.000)	0.000***(0.000)
SIZE	0.205***(0.039)	0.015***(0.001)
CFL_VOL	0.074***(0.027)	0.000(0.000)
SALES_VOL	2.493***(0.306)	0.030***(0.005)
INV_VOL	–0.0155(0.016)	0.003***(0.001)
TANG	13.235***(0.754)	0.077***(0.008)
CFOSALES	–1.063(0.943)	–0.029***(0.009)
SLACK	0.004*(0.002)	–0.002*(0.000)
LOSS	–0.813***(0.181)	–0.016***(0.002)
AGE	–0.673***(0.176)	–0.014***(0.002)
ZSCORE	0.258***(0.082)	–0.021***(0.001)
DIV	–1.351***(0.108)	–0.023***(0.002)
OPCYCLE	0.715***(0.240)	–0.006***(0.002)
\|DACC\|	5.887***(0.710)	0.160***(0.011)
First step control
SALESG	–	0.057**(0.029)
Industry effects	Yes	Yes
Year effects	Yes	Yes
SALESG*Industry effects	–	Yes
SALESG*Year effects	–	Yes
Constant	–2.737(2.242)	0.099***(0.025)
Observations	64,510	59,423
Adjusted R²	0.069	0.278

Note: This table shows the pooled regression estimates of the association between strategic deviation and investment using alternative measures. Column 1 shows the association between STRAT_DEV and ALT_INV_INEFF. Column 2 shows the association between ALT_STRAT_DEV and INV. Robust standard errors clustered at firm-level are in parentheses. *, **, *** denote a two-tailed p-value of less than 0.10, 0.05 and 0.01, respectively. Variables are defined in Appendix 1.

Second, we use an alternative measure of strategic deviation ( $A L T_S T R A T_D E V)$ and re-estimate our baseline model and report the result in Column 2 in Table 7. We follow prior research (Dong et al., 2021; Finkelstein and Hambrick, 1990) and construct $A L T_S T R A T_D E V,$ excluding R&D intensity and advertising intensity from the $S T R A T_D E V$ measure. We regress $I N V$ on $A L T_S T R A T_D E V$ , a set of control variables, the first stage control variable (SALESG), industry and year indicators, SLAESG*industry effects and SALESG*year effects. We find support for the assertion that $S T R A T_D E V$ and $I N V$ are positively associated using this alternative proxy of strategic deviation (β = 0.002, p < 0.05).

4.6. Strategic deviation, investment and firm value

Finally, we examine whether the investment inefficiencies that arise due to strategic deviation have negative consequences for firm value. We use the following regression to test this proposition:

$\begin{array}{l} V A L U E_{i, t + 1} = β_{0} + β_{1} S T R A T_D E V_{i, t} + β_{2} I N V_{i, t} + β_{3} S T R A T_D E V * I N V_{i, t} \\ + \sum_{i} λ_{i} C O N T R O L S_{i, t} + \sum_{j} λ_{j} I N D U S T R Y_{j, t} + \sum_{k} λ_{k} Y E A R_{k, t} + ε_{i, t} \end{array}$ (3)

Where VALUE is firm value proxied by Peters and Taylor (2017) investment-Q (QTOT)⁹ and TOBINQ measures and other variables are defined as before. We expect the coefficient β₃ to be negative and significant if the capital market discounts the sub-optimal investments undertaken by the strategically deviant firms. We present the regression results in Table 8. Columns 1 and 2 report the result for QTOT and TOBINQ measures respectively. We find the coefficient on the interactive variable STRAT_DEV*INV negative and significant for both QTOT (coefficient -0.043, p < 0.01) and TOBINQ (β = -0.345, p < 0.01) suggesting that the capital market discounts the investments undertaken by the deviant firms.

Table 8.

Strategic deviation, investment and firm value.

	(1)	(2)
	QTOT_t ₊ ₁	TOBINQ_t ₊ ₁
STRAT_DEV_t	0.396***(0.050)	3.392***(0.258)
INV_t	0.055***(0.006)	0.188***(0.028)
STRAT_DEV*INV _t	–0.043***(0.014)	–0.345***(0.072)
MTB_t	0.000*(0.000)	0.001***(0.000)
SIZE_t	0.232***(0.006)	0.872***(0.029)
CFL_VOL_t	0.020***(0.002)	0.096***(0.011)
SALES_VOL_t	0.235***(0.043)	1.093***(0.209)
INV_VOL_t	–0.002(0.006)	–0.013(0.025)
TANG_t	–0.431***(0.066)	–8.865***(0.298)
CFOSALES_t	0.197***(0.076)	0.877**(0.350)
SLACK_t	0.004***(0.001)	0.064***(0.006)
LOSS_t	–0.124***(0.015)	–0.339***(0.076)
AGE_t	0.049**(0.023)	–0.177*(0.102)
ZSCORE_t	–0.022**(0.009)	–0.458***(0.053)
DIV_t	–0.256***(0.027)	–1.207***(0.117)
OPCYCLE_t	–0.009(0.015)	–0.367***(0.078)
\|DACC\|_t	0.682***(0.061)	2.263***(0.291)
Industry effects	Yes	Yes
Year effects	Yes	Yes
Constant	0.187(0.312)	4.540***(0.824)
Observations	50,470	50,263
Adjusted R₂	0.244	0.387

Note: This table shows the association between investment and firm value when strategic deviance exists. Column 1 shows results when investment-Q model is used, following Peters and Taylor (2017). Column 2 shows the results when firm value is measured using TOBINQ. Robust standard errors clustered at firm-level are in parentheses. *, **, *** denote a two-tailed p-value of less than 0.10, 0.05 and 0.01, respectively. Variables are defined in Appendix 1.

5. Conclusion

We explore the association between strategic deviation and investment inefficiency. It is important to understand the repercussions of deviating from the industry-level strategy. At the same time, it is of vital importance to know the direct antecedents of investment inefficiencies. We argue that firms deviating from industry norms are prone to increased information asymmetry and hence, are able to engage in self-serving behaviour as manifested in inefficient investments. Using a sample of U.S. data, we find support for our prediction. Then we examine the moderating role of weak monitoring, high competition and a low-quality information environment. Our findings suggest that the positive association between strategic deviation and investment inefficiency is pronounced for firms exposed to weak monitoring and with a low-quality information environment and for firms operating in a highly competitive environment. Our results remain robust to possible endogeneity concerns. By examining the association between strategic deviation and investment inefficiency, we contribute to both the strategic management and the investment literature. Although there are a number of studies that examine the determinants of investment inefficiencies, there is no evidence on the association between deviant strategies and investment inefficiencies. This study fills this gap in the literature.

Business strategy is a choice that managers have to make, and, hence, it is important to understand the managerial incentives behind choosing a particular business strategy. As elaborated above, deviating from industry peers results in negative consequences, such as high information asymmetry, high risk and uncertainty. On the other hand, conforming to industry peers bring about homogeneity and is preferred by capital market participants due to low information processing costs and low uncertainty (DiMaggio and Powell, 1983). Litov et al. (2012) argue that pursuing unique strategies brings about economic rents which are associated with firm value in the long term. In contrast, Navissi et al. (2017) find that sub-optimal investments resulting from different strategic choices adversely affect future performance. Similarly, Dong et al. (2021) find that the capital market discounts the value of cash holdings of firms pursuing non-conforming strategies. Given these mixed implications, managers face a paradox in selecting a strategy (Litov et al., 2012). Despite the negative implications, managers might pursue deviant strategies when the costs of being different is lower than the benefits of conforming to industry standards. In other words, such firms may confront information asymmetry and thereby negative capital market consequences in the short term, but may enjoy first-mover benefits in the long term. Alternatively, the impact of strategic deviance on firm-level outcomes may be stronger in firms with peculiar characteristics. For example, Tang et al. (2011) find that firms with dominant CEOs pursue deviant strategies and encounter extreme performance. In this study, we focus on whether strategically deviant firms are associated with sub-optimal investments as a first step to understanding the implications of strategic choice.

Footnotes

We acknowledge the Associate Editor,Professor Tom Smith,Deputy Editor (Finance),Associate Professor Chelsea Liu and the anonymous reviewer for their valuable comments.

Final transcript accepted on 3 January 2023 by Tom Smith (AE Finance)

Funding

The author(s) received no financial support for the research,authorship,and/or publication of this article.

ORCID iDs

Dinithi Ranasinghe

Ahsan Habib

References

Abdoh

Varela

(2017) Product market competition, idiosyncratic and systematic volatility. Journal of Corporate Finance 43: 500–513.

Al-Hadi

Habib

Taylor

, et al. (2021) Financial statement comparability and corporate investment efficiency. Meditari Accountancy Research 29: 1283–1313.

Arellano

Bover

(1995) Another look at the instrumental variable estimation of error-components models. Journal of Econometrics 68: 29–51.

Babar

Habib

(2021) Product market competition in accounting, finance, and corporate governance: A review of the literature. International Review of Financial Analysis 73: 101607.

Bebchuk

Stole

(1993) Do short-term objectives lead to under-or overinvestment in long-term projects? The Journal of Finance 48: 719–729.

Bentley

Omer

Sharp

(2013) Business strategy, financial reporting irregularities, and audit effort. Contemporary Accounting Research 30: 780–817.

Bentley-Goode

Omer

Twedt

(2019) Does business strategy impact a firm’s information environment? Journal of Accounting, Auditing & Finance 34: 563–587.

Biddle

Hilary

Verdi

(2009) How does financial reporting quality relate to investment efficiency? Journal of Accounting and Economics 48: 112–131.

Blundell

Bond

(1998) Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics 87: 115–143.

10.

Brennan

Subrahmanyam

(1995) Investment analysis and price formation in securities markets. Journal of Financial Economics 38: 361–381.

11.

Carpenter

(2000) The price of change: The role of CEO compensation in strategic variation and deviation from industry strategy norms. Journal of Management 26: 1179–1198.

12.

Chen

M-J

Hambrick

(1995) Speed, stealth, and selective attack: How small firms differ from large firms in competitive behavior. Academy of Management Journal 38: 453–482.

13.

Chen

Xie

Zhang

(2017) How does analysts’ forecast quality relate to corporate investment efficiency? Journal of Corporate Finance 43: 217–240.

14.

Chen

Hribar

Melessa

(2018) Incorrect inferences when using residuals as dependent variables. Journal of Accounting Research 56: 751–796.

15.

Chen

Hribar

Melessa

(2022) On the use of residuals as dependent variables. Journal of Financial Reporting 7: 69–83.

16.

Choi

Myers

, et al. (2019) Financial statement comparability and the informativeness of stock prices about future earnings. Contemporary Accounting Research 36: 389–417.

17.

Christodoulou

Vasnev

(2018) Inference-in-residuals as an estimation method for earnings management. Abacus 54: 154–180.

18.

Chung

McInish

Wood

, et al. (1995) Production of information, information asymmetry, and the bid-ask spread: Empirical evidence from analysts’ forecasts. Journal of Banking & Finance 19: 1025–1046.

19.

Deephouse

(1996) Does isomorphism legitimate? Academy of Management Journal 39: 1024–1039.

20.

Deephouse

(1999) To be different, or to be the same? It’s a question (and theory) of strategic balance. Strategic Management Journal 20: 147–166.

21.

De Franco

Kothari

Verdi

(2011) The benefits of financial statement comparability. Journal of Accounting Research 49: 895–931.

22.

DiMaggio

Powell

(1983) The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review 48: 147–160.

23.

Dong

Chan

Cui

, et al. (2021) Strategic deviance and cash holdings. Journal of Business Finance & Accounting 48: 742–782.

24.

Eisdorfer

Giaccotto

White

(2013) Capital structure, executive compensation, and investment efficiency. Journal of Banking & Finance 37: 549–562.

25.

Fama

Jensen

(1983) Separation of ownership and control. The Journal of Law and Economics 26: 301–325.

26.

Ferreira

Matos

(2008) The colors of investors’ money: The role of institutional investors around the world. Journal of Financial Economics 88: 499–533.

27.

Financial Accounting Standards Board (FASB) (2010) Statement of financial accounting concepts No. 8. Conceptual framework for financial reporting. Norwalk, CT: FASB.

28.

Finkelstein

Hambrick

(1990) Top-management-team tenure and organizational outcomes: The moderating role of managerial discretion. Administrative Science Quarterly 35: 484–503.

29.

Francis

Pinnuck

Watanabe

(2014) Auditor style and financial statement comparability. The Accounting Review 89: 605–633.

30.

Geletkanycz

Hambrick

(1997) The external ties of top executives: Implications for strategic choice and performance. Administrative Science Quarterly 42: 654–681.

31.

Hainmueller

(2012) Entropy balancing for causal effects: A multivariate reweighting method to produce balanced samples in observational studies. Political Analysis 20: 25–46.

32.

Harrington

Gelfand

(2014.) Tightness–looseness across the 50 united states. Proceedings of the National Academy of Sciences 111: 7990–7995.

33.

Hoberg

Phillips

Prabhala

(2014) Product market threats, payouts, and financial flexibility. The Journal of Finance 69: 293–324.

34.

Holmström

(1999) Managerial incentive problems: A dynamic perspective. The Review of Economic Studies 66: 169–182.

35.

Imhof

Seavey

Watanabe

(2022) Competition, proprietary costs of financial reporting, and financial statement comparability. Journal of Accounting, Auditing & Finance 37: 114–142.

36.

Irvine

Pontiff

(2008) Idiosyncratic return volatility, cash flows, and product market competition. The Review of Financial Studies 22: 1149–1177.

37.

Jackson

(2022) Residuals from two-step research designs. Accounting & Finance 62: 4345–4358.

38.

Jensen

(1986) Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review 76: 323–329.

39.

Jensen

Meckling

(1976) The theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3(4): 305–360.

40.

Jiang

Kim

Nofsinger

, et al. (2015) Product market competition and corporate investment: Evidence from China. Journal of Corporate Finance 35: 196–210.

41.

Kang

Lee

H-Y

Park

H-Y

, et al. (2021) Social ties, managerial overconfidence, and investment efficiency. Finance Research Letters 46: 102300.

42.

Kim

J-B

, et al. (2016) Financial statement comparability and expected crash risk. Journal of Accounting and Economics 61: 294–312.

43.

Laksmana

Yang

Y-w

(2015) Product market competition and corporate investment decisions. Review of Accounting and Finance 14: 128–148.

44.

Lin

Y-E

Y-W

Cheng

, et al. (2021) Corporate social responsibility and investment efficiency: Does business strategy matter? International Review of Financial Analysis 73: 101585.

45.

Litov

Moreton

Zenger

(2012) Corporate strategy, analyst coverage, and the uniqueness paradox. Management Science 58: 1797–1815.

46.

McMullin

Schonberger

(2020) Entropy-balanced accruals. Review of Accounting Studies 25: 84–119.

47.

Malmendier

Tate

(2005) CEO overconfidence and corporate investment. The Journal of Finance 60: 2661–2700.

48.

Miles

Snow

(1978) Organizational Strategy, Structure, and Process. New York: McGraw Hill.

49.

Mintzberg

(1978) Patterns in strategy formation. Management Science 24: 934–948.

50.

Modigliani

Miller

(1958) The cost of capital, corporation finance and the theory of investment. The American Economic Review 48: 261–297.

51.

Myers

(1977) Determinants of corporate borrowing. Journal of Financial Economics 5: 147–175.

52.

Navissi

Sridharan

Khedmati

, et al. (2017) Business strategy, over-(under-) investment, and managerial compensation. Journal of Management Accounting Research 29: 63–86.

53.

Peters

Taylor

(2017) Intangible capital and the investment-q relation. Journal of Financial Economics 123: 251–272.

54.

Pool

Stoffman

Yonker

(2015) The people in your neighborhood: Social interactions and mutual fund portfolios. The Journal of Finance 70: 2679–2732.

55.

Porter

(1980) Competitive Strategy: Techniques for Analyzing Industries and Competitors. New York: The Free Press.

56.

Porter

(1986) Changing patterns of international competition. California Management Review 28: 9–40.

57.

Porter

(1996) What is strategy? Harvard Business Review, November–December, 37–55.

58.

Provaty

Amin

Hasan

(2022) Strategic deviation and debt maturity structure. Finance Research Letters 50: 103317.

59.

Rajkovic

(2020) Lead independent directors and investment efficiency. Journal of Corporate Finance 64: 101690.

60.

Roulstone

(2003) The relation between insider-trading restrictions and executive compensation. Journal of Accounting Research 41: 525–551.

61.

Shleifer

Vishny

(1989) Management entrenchment: The case of manager-specific investments. Journal of Financial Economics 25: 123–139.

62.

Shleifer

Vishny

(1997) A Survey of Corporate Governance. The Journal of Finance 52(2): 737–783.

63.

Stock

Yogo

(2005) Asymptotic distributions of instrumental variables statistics with many instruments. In: Andrews

DWK

Stock

Rothenberg

(eds) Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg, pp. 109–120. Cambridge: Cambridge University Press.

64.

Stoughton

Wong

(2017) Investment efficiency and product market competition. Journal of Financial and Quantitative Analysis 52: 2611–2642.

65.

Tang

Crossan

Rowe

(2011) Dominant CEO, deviant strategy, and extreme performance: The moderating role of a powerful board. Journal of Management Studies 48: 1479–1503.

66.

Juan

Zhang

(2021) Strategic deviation and stock return synchronicity. Journal of Accounting, Auditing & Finance 36: 172–194.

Strategic deviation and investment inefficiency

Abstract

JEL Classification:

Keywords

1. Introduction

2. Literature and hypotheses development

2.1. Corporate governance

2.2. Product market competition

2.3. Information environment

3. Methodology

3.1. Sample

3.2. Variable measurements

3.2.1. Independent variable: Strategic deviation (STRAT_DEV)

3.2.2. Dependent variable: Investment ( I N V ) .

3.3. Empirical model

4. Empirical results

4.1. Descriptive statistics

4.2. Regression results

4.3. Addressing endogeneity

4.4. Strategic deviation and investment inefficiency: Cross-sectional tests

4.5. Robustness tests

4.6. Strategic deviation, investment and firm value

5. Conclusion

Footnotes

Funding

ORCID iDs

References

3.2.2. Dependent variable: Investment ( $I N V) .$