Capital market pressures and managerial decisions:
the influence of shareholders on firms’ time horizons
Master thesis
Davide Perazzoli
Student number: 10986820
University of Amsterdam – Faculty of Economics and Business MSc Business Administration – International Management track Supervisor: Robert Kleinknecht
Statement of originality
This document is written by Davide Perazzoli who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Acknowledgements
Writing my master thesis was a process that took me much effort and dedication. There are some people that I would like to thank for supporting me and encouraging me to do my best during these last months. Especially, I would like to thank my supervisor Robert Kleinknecht for assisting me in writing my thesis with his useful insights and expertise. Then, I would like to thank my family and my friends for being supportive and motivating me even when I was away. A special thanks goes to Therese Anschütz for always being on my side during the whole master and for encouraging me in the last months.
Abstract
In the recent years, a growing number of researchers have focused their attention on the assessment of firms’ time horizons. In particular, scholars pointed out a trend for which managers are focusing on short-term results rather than aiming at the long-term profitability of their firms. As an effect, a debate on this phenomenon – which has been defined as short-termism – has arisen much interest and it has been attributing it to various causes. Among them, capital market pressures were found to encourage firms to perform in the short-term. The current study, contributing to this topic, investigated the influence of ownership concentration, institutional investors and stock liquidity on firms’ time horizons. Overcoming the difficulties of previous research in assessing managers’ temporal orientations, content analysis was used to measure time perspectives through voluntary disclosure in quarterly earnings call. The study performed an analysis on a panel of 5,493 quarterly observations of 366 firms between 2008 and 2013. The findings showed a positive impact of institutional investors on firms’ time horizons, highlighting a role of these types of shareholders in limiting short-termism by exercising more active monitoring on the management of firms.
Keywords: firms’ time horizons, short-termism, managerial myopia, ownership structure,
Table of content
1. Introduction ... 6
2. Theoretical development ... 8
2.1 Firms’ time horizons ... 8
2.2 Short-termism ... 9
2.3 Determinants of short-termism ... 10
2.4 Capital markets and ownership structure ... 14
2.4.1 Ownership concentration ... 15
2.4.2 Stock liquidity ... 17
2.4.3 Institutional investors ... 19
3. Method ... 21
3.1 Sample and time frame... 22
3.2 Variables of interest ... 22
3.2.1 Dependent variable ... 22
3.2.2 Independent and moderating variables... 24
3.2.3 Control variables ... 26 4. Results ... 27 4.1 Preliminary analysis ... 28 4.2 Regression analysis ... 30 5. Discussion... 37 5.1 Discussion of results ... 37 5.1.1 Ownership concentration ... 37 5.1.2 Stock liquidity ... 39 5.1.3 Institutional investors ... 40 5.2 Theoretical implications ... 41 5.3 Managerial implications ... 42
5.4 Limitations and suggestions for further research ... 43
6. Conclusion ... 44
References ... 45
1. Introduction
When making any strategic decisions, managers are necessarily in the need of choosing among different costs and benefits spread over time (Souder & Bromiley, 2012). This happens because investment decisions imply differences in the time needed to payback the costs (Souder & Shaver, 2010). Usually, executives rely on discounted utility models to compare different distributions of costs and returns over time, based on the net present value of investments (Loewenstein, 1988). This theory is based on the fact that investments with more immediate cash flows are preferred to more deferred results (Laverty, 1996). The choice of temporal orientation by firms, defined as the priority given to strategies with different payoffs over times (Souder & Bromiley, 2012), has arisen a wide interest in many scholars.
In particular, in the last decades scholars have been recognizing a tendency for which firms’ strategic choices are increasingly focusing on the short-term, causing the loss of competitiveness of countries that adopt this type of temporal orientation, such as the U.S. (Porter, 1992; Hayes & Abernathy, 1980). This phenomenon has been defined as short-termism, meaning that firms choose strategies that are optimizing only the short-term, but are not considering the long-term implications (Laverty, 1996). This attitude has been argued to reduce investments in tangible and intangible assets (Porter, 1992; Marginson & McAulay, 2008). In this way, managers undermine long-term performance in order to boost short-term earnings.
Several scholars have investigated short-termism trying to figure out its underlying reasons. The phenomenon was attributed to four main areas of interest: performance measurement systems, capital market pressures, managerial myopic behavior and organizational characteristics (Marginson & McAulay, 2008). In the current study, a particular focus will be put on the role of capital markets. Executives are seen as influenced
by stock markets in the choices of their temporal orientations, as the owners are pressuring the management to pursue short-term results (Jacobs, 1991; Marginson & McAulay, 2008). More specifically, the present research will investigate the role of shareholders in conditioning the time horizons of firms, contributing to the literature about capital market pressures. It will be expected that the concentration of ownership and a higher presence of institutional investors will be positively linked with the temporal orientation of managers. On the other hand, the liquidity of a firm’s stock will be estimated to have mixed effects on firms’ time horizons, having a negative effect that becomes positive when accounting for a higher ownership concentration.
The research on firms’ time horizons was constrained by the difficulty of measuring time orientation (Laverty, 1996; Souder & Bromiley, 2012). Previous research has often used R&D investments as a proxy for temporal orientation, but Laverty (1996) arose doubts on the efficacy of this measure. More recent studies used surveys (e.g. Laverty, 2004; Marginson & McAulay; 2008), accounting data (e.g. Souder & Shaver, 2010; Souder & Bromiley, 2012), and earnings management (Chen, Rhee, Veeraraghavan & Zolotoy, 2015) to identify managers’ time horizons. In the current study, temporal orientation will be measured with a content analysis method, considering the voluntary disclosure of information by managers, similarly to what already done by Brochet, Loumioti and Serafeim (2015) and DesJardine and Bansal (2014).
The following section will build on previous research to define more in depth firms’ time horizons and short-termism, leading to the theoretical predictions that will be tested. Afterwards, the methodology used to collect data and measure the different variables will be explained in detail. Subsequently, the analyses performed on the data collected will be described, presenting their results. Later, the results will be discussed and related to the previous literature on the topic, outlining the implications for practitioners and scholars, and
suggesting directions for further research. Lastly, in the conclusion the outcome of this study will be summarized.
2. Theoretical development
In this section, the current state of the research on short-termism will be assessed. Firstly, an analysis of the concept of firms’ time horizon will be provided. Secondly, the phenomenon of short-termism will be illustrated. Thirdly, various attempts to understand the determinants of the phenomenon will be described. Finally, the chapter will analyze more in depth the role of ownership structures in the determination of firms’ time horizons, and will ultimately propose the hypotheses that will be investigated in the study.
2.1 Firms’ time horizons
Many scholars in the last decades have been increasingly interested in the analysis of firms’ time horizons and their effects on managerial decisions. In particular, temporal orientation has been defined as the “importance given in strategic choices to investments with differing distribution of costs and benefits over time” (Souder & Bromiley, 2012: 551). In other words, managers have to face decisions that imply the comparison of investments with payoff in different times. Relatedly, previous research defined the problem of intertemporal choice (Loewenstein, 1988). An intertemporal choice dilemma arises when decision-makers are confronted with choices about “costs and benefits that are spread out over time” (Loewenstein & Thaler, 1989: 181).
In general, firms’ investments have diverse payoff horizons, which means they will differ in the time needed for the returns to cover the costs (Souder & Shaver, 2010). As a consequence, managers have to decide about investments that will pay back in different times
and they need to assess the value of the outcomes. Thus, executives are expected to base their intertemporal choice decisions on a discounted utility (DU) model (Loewenstein, 1988). According to this model, managers have a preference for more immediate results, meaning that a present cash flow is preferred to an equivalent one in the future (Laverty, 1996). More specifically, managers assess the value of investments basing on their net present value (NPV), calculated by devaluing future returns by a discount factor (1+r)t, with r being the cost of capital per period and t the number of time periods (Laverty, 1996). Additionally, the value of future cash flows is reduced by the presence of risk and uncertainty about future outcomes, which reduce the expected utility (EU) of investments. Therefore, the value of future cash flows from an investment is discounted because of time and uncertainty (Laverty, 1996).
2.2 Short-termism
As seen in the previous section, managers have to make decisions about investments with returns that are spread over time. Porter (1992) suggested that managers should undertake actions that ensure the long-term value of the firm. Nevertheless, Laverty (1996) highlighted a trend of firms focusing on investments aiming at short-term payoffs rather than at long-run value. As seen earlier, managers are expected to base their decisions about investment on DU models (Loewenstein, 1988). However, Graham, Harvey and Rajgopal (2005) found that more than 75% of managers choose investments with lower NPV than other alternatives because they have an earlier payoff. This short-term oriented behavior is linked to managerial myopia, which represents cognitive limitations of managers in assessing the future implications of the temporal dimension of their decisions (Miller, 2002). When myopic behavior is related to a systematic and repeated adoption of shortsighted measures it has been identified as short-termism (Laverty, 2004). Economic short-termism was defined as
representing “decisions and outcomes that pursue a course of action that is best for the short term but suboptimal over the long run” (Laverty, 1996: 826). Van Der Stede (2000) identified the term as a one-year orientation within the budgeting cycle. The choice of a short-term horizon however is not necessarily linked to short-short-termism, as it is not always a disadvantage to choose such a time horizon on a one-year basis, as in some cases financial trouble may induce to such strategies in order to ensure the survival of the firm (Marginson, McAulay, Roush & Van Zijl, 2010). Short-termism instead is related to short-term oriented actions that undermine the long-term profitability of a firm (Laverty, 1996).
From a practical point of view, short-termism has been argued to cause firms to underinvest in assets and capabilities that are necessary to ensure a long-term competitive advantage (Laverty, 2004). More specifically, companies adopting a short-term horizon invest too less in intangible assets such as R&D and employees’ skills development (Porter, 1992). Furthermore, Hayes and Abernathy (1980) claimed that firms cut short-term costs rather than investing in their long-term technological competitiveness, which means they spend less on innovation. As an effect, Porter (1992) pointed out that short-termist firms are at a competitive disadvantage towards more long-term investing companies. In his paper, he showed that the U.S., where many short-term oriented companies have been identified, are being outperformed by countries that have a longer time horizon, such as Japan or Germany.
2.3 Determinants of short-termism
After identifying the problem of short-termism, many scholars undertook research in order to understand why firms adopt this time horizon. Marginson and McAulay (2008) structured the debate on the causes of short-termism in four areas of interest that have been investigated by research: stock markets, performance measurement systems, managers’ individual behavior and preferences, and organizational characteristics. Firstly, analyzing the role of capital
markets, shareholder myopia has been defined as a tendency of shareholders to favor investments in stocks that guarantee short-term rather than long-term returns (Samuel, 2000). As a consequence, executives are motivated to follow short-term strategies at the expense of long-term performance in order to satisfy stock market expectations (Jacobs, 1991). The relationship between stock markets and short-termism will be analyzed more in depth in the subsequent section.
Secondly, considering the role of performance measurement systems, many scholars agreed that the ways in which firms evaluate and report their results are encouraging the adoption of a short-term horizon. Hayes and Abernathy (1980) pointed out that firms rely too much on short-term financial measures to assess their performance. A focus on quarterly earnings results leads firms to adopt a short-term horizon (Brochet et al., 2015; Jackson & Petraki, 2011). Furthermore, researchers discovered that considering previous performance is limiting investments on the future, as evaluations about previous cash flows are a crucial variable in determining long-term investments (Souder & Shaver, 2010). Porter (1992) further supported the detrimental role of performance measurement systems, as they evaluate projects only quantitatively, overseeing the chances of profitable investments in intangible assets such as R&D, information systems, employees training and skill development.
However, besides these studies finding support for a link between performance measurement and short-termism, Marginson et al. (2010) provided an alternative view. In their research they found that financial or non-financial measures do not shorten time horizons, but rather the way they are used. Accordingly, they distinguished two different ways to exploit these measures: a diagnostic use, which relates to the monitoring of performance compared to previously set standards, and an interactive use, encouraging dialogue about strategic choices. Their findings provided insights that refute other studies, as they showed that financial performance measurement does not have an influence on
short-termism, neither with a diagnostic nor interactive use. On the other hand, they highlighted that a diagnostic use of non-financial measures is linked with short-termism, while an interactive use encourages the choice of a long-term horizon. Their findings showed also a preference for a diagnostic use of non-financial indicators, which ultimately encourage the choice of a shorter time horizon. Houston, Lev and Tucker (2010) also found evidence that contradicts the link between performance measurement systems and short-termism. Their findings showed that firms that stopped quarterly earnings guidance did not increase the level of investment in the future, refuting the claim that a focus on quarterly results leads to shorter time horizons.
Thirdly, on the individual level, managerial myopia is seen as a further determinant of short-termism because of uncertainty (Marginson & McAulay, 2008). In general, managers face uncertainty when making decisions, which is due to a lack of information (Galbraith, 1973). Marginson and McAulay (2008) predicted that when the time horizon is extended, managers are likely to be missing even more information. Relatedly, their study found that role ambiguity, defined as the difference between information necessary to make a decision and information available, is associated with short-termism. In other words, managers sometimes do not have all the information to assess the long-term profitability of investments and choose strategies that do not consider properly the future outcomes.
Furthermore, researchers found that short-termism is linked to the way managers assess the value of future cash flows. As seen earlier, decision makers use a DU model, which has been assumed to imply a constant discount rate over time (Streich & Levy, 2007). However, Streich & Levy (2007) found that managers use a variable discount factor, with losses discounted more than gains and with different discount rates as the outcome is more or less temporally close. Furthermore, Laverty (1996) explained that when managers analyze the NPV of an investment, they calculate future cash flows with a too high discount rate.
Accordingly, a previous experimental study found that the discount rate for making a future gain indifferent to an immediate one would be as high as 289% on average (Thaler, 1981). As an effect, managers discount excessively future returns, making future outcomes become disproportionately less preferable than immediate profits and thus they focus more on the short-term (Miles, 1993; Laverty, 1996, Jackson & Petraki, 2011).
Additionally, Laverty (1996) claimed that managers make decisions that may be optimal on a personal level but may not be for the organization. As a consequence, this approach leads them to focus on short-term results, undermining the long-term profitability of the firm. Antia, Pantzalis and Park (2010) supported this claim, finding that CEO tenure has a role in determining time horizons. More precisely, CEO with shorter expected tenure have been found to increase agency costs, as they experience a higher pressure to deliver positive results in a shorter time frame to be perceived positively in the labor market. Relatedly, Rumelt (1987) argued that managers can show opportunistic behavior, choosing investments that will perform well in the short-term and leave the firm before the long-term effects become clear, in order to show a superior managerial ability.
Moreover, a closely connected topic is managerial compensation schemes, that have been argued to be linked with short-termism. As Thanassoulis (2013) explained, managers’ salaries need to be partly related to the results of the firm. These managerial incentives can be either based on earlier performance or they can be deferred in the future. Accordingly, Jackson and Petraki (2011) suggested that when firms link executives’ pay too much to short-term results, they will be influenced to choose strategies emphasizing short-short-term results. Earlier empirical findings showed that stock-based compensation is linked to a higher chance that managers will focus on short-term results, aiming to maximize stock price (Coates, Davis & Stacey, 1995). Moreover, another study provided support for a correlation between managerial incentives and firms’ temporal orientation, finding that shorter term based
compensation is related to a shorter time horizon in the management (Gopalan, Milbourn, Song & Thakor, 2014).
Fourthly, some studies investigated the role of organizational factors that are linked to short termism. Porter (1992) argued that a high degree of decentralization has led firms to create highly autonomous business units. This in turn caused firms to manage inefficiently information flows both vertically and horizontally, creating flaws in the decision making process. Laverty (1996) further explained how this multidivisional structure increased the necessity of the headquarters to evaluate the performance of their divisions, pressuring them to provide short-term results. A different approach was taken by Marginson and McAulay (2008), who considered in particular the role of workgroup conformity. In their study they found that managers’ short-termism is linked to social pressure. In other words, an executive short-term orientation is influenced by the behavior and opinions of their team members.
2.4 Capital markets and ownership structure
In the current section, a deeper focus will be taken on the role of stock markets and ownership structure as sources of short-termism. Jacobs (1991) argued that “impatient capital” makes shares being treated as a commodity, with less shareholders interested in keeping their share longer or knowing any detail about the firm’s business. Porter (1992) further pointed out that capital markets are increasingly formed by transient owners, who seek for immediate profits holding their share for a limited time. In other words, the purpose of most shareholders is mainly to maximize the profit from selling their shares. As a consequence, managers are influenced towards the choice of strategies maximizing short-term returns to meet the interests of the ownership. (Jacobs, 1991; Marginson & McAulay, 2008). By aiming at maximizing stock value, managers undermine long-run value, ultimately leading to short-termism (Freeman, 1984). In the next sections the effects of capital markets
and ownership structure on firms’ temporal orientation will be considered, analyzing the literature on the role of ownership concentration, stock liquidity and institutional investors.
2.4.1 Ownership concentration
In this section, the role of ownership concentration as an influence on time horizon will be analyzed, considering it within an agency theory context. This assumption is based on the fact that the separation between ownership and control has implications for the business practices of firms. Agency theory identifies the relationship between owners and managers as a contract for which one or more entities (the principals) entitle another person (the agent) to make some decision on his/their behalf (Jensen & Meckling, 1976). According to this theory, the agent theoretically should always act in the best interests of the principal, but sometimes his own goals may clash with them. As a consequence, principals have to keep monitoring the work of the agent in order to make sure that he will act in his interests. This activity will involve costs, that are defined as agency costs (Jensen & Meckling, 1976).
Scholars investigated the role that ownership concentration has on innovation by firms, finding that a higher concentration is linked to more investment in R&D and innovation (Hill & Snell, 1988; Lee & O’Neill, 2003). More specifically, these studies argued that this is due to the fact that large shareholders influence management to pursue strategies in line with their interests. Lee and O’Neill (2003) further claimed that blockholders (shareholders owning at least five percent of the total ownership) are more interested in investing on innovation in order to secure the future profitability of the firm. In order to do so, these shareholders are motivated to gather more information about the firm and engage in active monitoring. On the other hand, small investors are not motivated to collect additional information about the firm and they tend to base their investment decisions on quarterly earnings, as also claimed by Jacobs (1991).
The findings by Hill and Snell (1988) and Lee and O’Neill (2003) suggest a relationship between ownership concentration and firms’ time horizons, considering that a higher focus on innovation and R&D has been often linked to investment on the long-term. The two studies, in light of what described earlier about agency theory, allow to make predictions about the relationship between concentration and short-termism. Managers, as described in the previous section, are seen as influenced towards the choice of strategies maximizing stock price at the expenses of long-term profitability when shareholders are myopic. Indeed, when the concentration of ownership is lower, and thus there are many small owners, it is likely that they are basing their investment decisions on earnings reports, as they are not willing to sustain agency costs (Lee & O’Neill, 2003; Jacobs; 1991). However, managers will be motivated anyway to take actions maximizing short-term value with detrimental effects for the long-term profitability of the firm, as because of managerial incentives (Jackson & Petraki, 2011; Coates et al., 1995; Gopalan et al., 2014), their interests will be aligned with those of small myopic shareholders.
On the other hand, the effect of agency theory seems to be different when the concentration of ownership is higher. Agency theory, combined with the findings by Lee and O’Neill (2003), suggests that blockholders are expected to take the burden of agency costs and exert control over the firm’s management. Thus, from an agency point of view, big shareholders push to get the strategy of the firm in line with their interests. As an effect, a higher concentration of ownership is likely to make managers influenced towards strategies aiming at the goals of blockholders, who are interested in the long-term profitability of the firm (Lee & O’Neill, 2003).
Basing on the argumentations above, I predict that a lower concentration of ownership will be associated with firms having a higher focus on short-term strategies, while I expect
that a higher concentration will be linked with a long-term horizon. Therefore, I formulate the following hypothesis:
Hypothesis 1: Firms with a higher (lower) concentration of ownership have a longer (shorter) time horizon.
2.4.2 Stock liquidity
Stock liquidity has been defined as the possibility to trade large quantities of an asset quickly and at a low cost without causing a change in its price (Pástor & Stambaugh, 2003). Porter (1992) argued that the U.S. stock market is characterized by a fluid capital, with external capital moved rapidly from one company to another. Bhide (1993) suggested that the benefits of stock liquidity, such as an enhanced possibility to diversify investors’ portfolio, can be overcome by the disadvantages, as “stock liquidity discourages internal monitoring by reducing the costs of ‘exit’ of unhappy stockholders” (Bhide, 1993: 31). Further research linked liquidity to short-termism, finding that a higher stock liquidity is limiting innovation, due to an increased risk of hostile takeovers and a bigger presence of non-dedicated institutional investors (Fang, Tian & Tice, 2014). Furthermore, Holden and Subrahmanyam (1996) found that a high extent of information-less trading, which is linked to a higher stock liquidity, increases the tendency to collect information about the short-term. This, as a result, strengthens short-termism, as stock prices become more informative about the short- rather than the long-term.
Oppositely, another stream of research supported a positive relationship between stock liquidity and time horizons. Edmans, Fang and Zur (2013) found that a higher stock liquidity fosters the likelihood of institutional investors to acquire a block (at least 5 percent of shares). Additionally, Bharath, Jayaraman and Nagar (2013) investigated the role of stock
liquidity as a measure of exit threat, as when it is easier to sell shares, an investor’s threat of leaving the ownership is stronger. Exit threat represents a governance mechanism for blockholders, who therefore can influence the choices of firms’ management (Edmans et al. 2013). Large shareholders are also more interested in gathering information about the firm they are investing and in creating a long-term profitable relationship (Lee & O’Neill, 2003). As a consequence, when stock liquidity is higher, managers feel more pressure to satisfy the interests of shareholders, focusing on the maximization of long-term firm value rather than stock price. Chen, Rhee, Veeraraghavan and Zolotoy (2015) followed this line of thought, finding out that when a firms’ stock is more liquid, blockholders can have a stronger influence on strategic choices, making managers less likely to influence earnings management in order to pursue short-term results.
Overall, the previous literature on stock liquidity provided different points of view on its link with firms’ time horizons. The current study will aim to reconcile these contrasting perspectives. Building on the previous argumentations, when the presence of blockholders is not taken into account, stock liquidity seems to have a negative effect on firms’ temporal orientation, encouraging uninformed trading (Holden & Subrahmanyam, 1996) and the presence of non-dedicated investors (Fang, Tian & Tice, 2014). Therefore, I propose the following hypothesis:
Hypothesis 2a: Firms with a higher (lower) stock liquidity have a shorter (longer) time horizon.
On the other hand, what seems to change the role of stock liquidity is the presence of blockholders that can influence the governance of the firm through exit threat (Bharath et al., 2013; Edmans et al., 2013; Chen, Rhee, Veeraraghavan & Zolotoy, 2015). Therefore, these
findings point in the direction of an effect of ownership concentration on the relationship between liquidity and temporal orientation. While normally stock liquidity would encourage firms to choose a shorter time horizon, a higher concentration of ownership seems to counteract this effect, through the governance by exit threat exerted by big shareholders. Thus, I expect ownership concentration to make the relationship between stock liquidity and firms’ time horizons become positive. Building on these argumentations, I suggest the following hypothesis:
Hypothesis 2b: When ownership concentration is higher, firms with a higher (lower) stock liquidity have a longer (shorter) time horizon.
2.4.3 Institutional investors
Although not many scholars investigated the effect of institutional investors on time horizons, extensive research aimed to assess the effects of institutional investors on R&D expenditure and innovation. Porter (1992) defined three groups of institutional investors: (1) “transient” owners, who hold small shares of different firms and trade them often; (2) “dedicated” owners, who possess shares of a few firms and with long-term focus; (3) “quasi-indexers”, who pursue diversification strategies and have low portfolio turnover. In his dissertation, he claimed that an increased presence of transient institutional investors is increasing the focus on short-term results. A few empirical studies reinforced the negative relationship between institutional ownership and R&D and innovation. This negative link was supported in a research limited to the computer industry (Graves, 1988) and in another one not linked to a specific industry (Samuel, 2000). Furthermore, Chen, Lin & Yang (2015) in a study limited to Taiwan found that domestic institutional ownership is linked to a higher likelihood for managers to cut R&D spending in order to boost short-term earnings. However, they also
identified a higher foreign institutional ownership seems to be positively linked to R&D expenditure.
Even though a limited number of studies suggested a negative link between institutional investors and R&D, a bigger stream of research highlighted the opposite. Accordingly, several studies recognized institutional ownership to increase the focus of firms’ management on R&D and innovation (Aghion, Van Reenen & Zingales, 2013; Baysinger, Kosnik & Turk, 1991; Brossard, Lavigne & Sakinç, 2013; Bushee, 1998; Hansen & Hill, 1991; Kochhar & David, 1996; Pozen, 2015). More specifically, the higher focus on innovation and R&D has been linked with a more active monitoring by institutional investors, (Aghion et al., 2013). These investors are increasingly able to influence firms’ management and thus they can engage in order to encourage a higher focus on long-term value creation (Pozen, 2015). Additionally to considering these findings, it can be noticed that the previously mentioned studies have been restricted to a single industry (Graves, 1988) or country (Chen, Lin & Yang, 2015), arising doubts about their validity. As an effect, the more extensive support for the positive relationship between institutional investors and R&D seems to prevail.
Although research associated R&D expenditure to a longer time horizon (Porter, 1992), Jackson and Petraki (2011) casted doubts on the efficacy of R&D expenditure as a proxy for measuring short-termism, arguing that R&D is not the only driver of long-term orientation and in some industries R&D is unlikely to be reduced. Laverty (1996) also argued that R&D is not a suitable measure of long-term investment, as evidence has shown that R&D is not creating economic value (Erickson and Jacobson, 1992). Thus, in the present study these limitations will be overcome and it will be investigated whether institutional investors have also a positive influence on time horizons of firms. Therefore, I propose the following hypothesis:
Hypothesis 3a: Firms with a higher (lower) percentage of ownership by institutional investors have a longer (shorter) time horizon.
Additionally, in the current analysis, ownership concentration will be linked to the relationship predicted between institutional investors and temporal orientation of firms. As mentioned earlier, a higher presence of institutional investors is linked to a more active monitoring of firms (Aghion et al., 2013) and to a bigger influence on firms’ management (Pozen, 2015). Considering that large shareholders are also interested in the long-term results of a firm and in informing themselves (Lee & O’Neill, 2003), it can be expected that the presence of blockholders when institutional ownership is higher should foster even more the pursuit of long-term goals. Therefore, when ownership is more concentrated, the positive relationship between institutional ownership and firms’ time horizons should be stronger. Thus, I suggest the following hypothesis:
Hypothesis 3b: A higher concentration of ownership strengthens the positive relationship between the percentage of institutional investors and firms’ time horizons.
3. Method
The current section will explain the methodology adopted to retrieve and structure the data necessary for the subsequent analysis. The first part will describe how the sample was created and justify the choice of the time frame. The second part then will explain in detail how the data about the dependent, independent and control variables were collected and how the final values for each variable were calculated. More specifically, for the dependent variable, a
deeper focus will be taken in illustrating the method, explaining the choice of content analysis for assessing firms’ time horizon.
3.1 Sample and time frame
The sample was built from the intersection of data about (1) quarterly earnings call transcripts, collected from SeekingAlpha, (2) shareholder ownership of firms, retrieved from Thomson Reuters Eikon, (3) bid and ask historical stock values from the CRSP database, and (4) firms’ fundamentals from the Compustat capital IQ database. Data were collected on a quarterly basis over a six years period from 2008 to 2013. The choice of a six years period was motivated by the possibility of avoiding biases due to periodical changes in the business cycle (DesJardine & Bansal, 2014). Additionally, a multi-year time frame allowed to capture enough variations in the time horizons of managers using content analysis (Yadav, Prabhu and Chandy, 2007). The sample was restricted to U.S. firms, in order to avoid the possibility of differences in corporate governance across different countries (Aguilera & Jackson, 2003). During the data collection about ownership, 46 firms presented a total percentage of ownership covered going over 100 percent, and therefore were cut out of the sample. After this step, the final panel is formed by 366 firms, for a total of 5493 firm-quarter observations. The panel is unbalanced due to the unavailability of information regarding some variables in some periods of time.
3.2 Variables of interest 3.2.1 Dependent variable
The dependent variable in the current study is firms’ time horizon. Although the measurement of temporal orientation and short termism have arisen an interest in various scholars, research on intertemporal choice has been limited by the difficulty implied by
measuring it (Souder & Bromiley, 2012). Laverty (1996) casted doubts on the usage of R&D as a measurement of time horizons which has been traditionally used by researchers, and identified as crucial to find an alternative measure to advance the debate on the topic. Subsequently, many studies tried to develop different methods to quantify firms’ temporal orientation. Among them, some chose to use surveys to measure the time perceptions of managers (e.g. Laverty, 2004; Marginson & McAulay; 2008). Other studies based their calculations on accounting data, identifying the temporal orientation of investments (e.g. Souder & Shaver, 2010; Souder & Bromiley, 2012). Another study based its measurement on earnings management, building on the fact that managers use financial reporting and investment decisions to follow short-term objectives (Chen, Rhee, Veeraraghavan & Zolotoy, 2015).
The present study instead will use a content analysis method to measure time horizon of firms. Content analysis has been defined as “any methodological measurement applied to text (or other symbolic materials) for social science purposes” (Shapiro & Markoff, 1997: 14). Duriau, Reger and Pfarrer (2007) investigated the content analysis literature, pointing out several advantages of this approach. Among these advantages, they pointed out the possibility to understand people’s cognitive schemas (Huff, 1980; Gephart, 1993; Woodrum, 1984), to ensure flexibility and scalability and to be non-intrusive (Woodrum, 1984). Additionally, in a business context, this method enables to use quantitative methods to analyze in depth the meaning of organizational documents (Duriau et al., 2007), and to implement longitudinal research designs thanks to the availability of corporate information over long periods of time, such as annual reports (Jauch, Osborn, & Martin, 1980; Kabanoff, 1996; Weber, 1990).
More specifically, in the short-termism literature, a few researchers have already chosen content analysis as a method. Brochet et al. (2015), DesJardine and Bansal (2014), and DesJardine (2015) used full-text earnings conference calls transcripts from several firms
to perform text analysis. Brochet et al. (2015) justified the choice of this method, finding evidence that voluntary corporate disclosures are informative about the temporal focus of firms. In each of the studies mentioned above, the researchers created a list of words indicating either a short-term or a long-term orientation. DesJardine and Bansal (2014), and DesJardine (2015) explained in detail how they calculated time horizons. They made a ratio of the number of words linked to a long-term orientation to the total number of words showing either a short-term and long-term orientation.
In the current study, basing on a similar approach, quarterly earnings call transcripts were collected to perform a content analysis. An earnings call consists of a speech by one or more executives, typically followed by a question and answer session with analysts. The transcripts were analyzed in order to find the ratio explained above (DesJardine & Bansal, 2014; DesJardine, 2015), measuring the temporal orientation of firms, basing on the words used by their managers. The list of words relating to the long- and short-term was the same used by DesJardine (2015), and can be found in the appendix. However, in the current study, differently from what done previously, the counting of words included all terms referring to time horizons. Instead, DesJardine and Bansal (2014) and DesJardine (2014), excluded some words from the counting when they were followed by certain words, when repeated within two words and when anticipated by negation words. In the current study all the words referring to time horizons were counted regardless of possible word combinations. The aim was to avoid excluding from the count any word which could be relevant in assessing firms’ temporal orientations.
3.2.2 Independent and moderating variables
The first independent variable, ownership concentration, which was also tested as a moderator on the other two independent variables, was calculated basing on data collected
from Thomson Reuters Eikon. The output from the database consisted of quarterly information on the share of ownership held by each shareholder. To limit the size of the data to a processable amount, the collection was limited to the first 100 shareholders, which allowed to include most relevant owners, as necessary for the subsequent analysis. Once the ownership breakdowns about each firm were collected, ownership concentration was assessed on a quarterly basis. In particular, ownership concentration was determined basing on the measurement used by Lee and O’Neill (2003), calculating the total percentage owned by shareholders holding at least a three percent stake into each firms.
For the second independent variable, stock liquidity, data about bid and ask price historical stock values were collected from the CRSP database. Previously, Chordia, Roll and Subrahmanyam (2001) calculated stock liquidity using intra-day transaction data to identify the percentage bid-ask spread. Accordingly, a higher spread between bid and ask implies a lower liquidity (Fang et al., 2014), therefore it represents an illiquidity measure. In the current study, due to the difficulty to access data about bid and ask values on single transactions, the data were collected on a daily basis relatively to the 2008-2013 period. Subsequently, the percentage bid-ask spread was calculated as the ratio of the difference between ask and bid price to the ask price. Lastly, an average of the percentage spread of each day was determined for every time period, enabling to identify a mean illiquidity value for each quarter. Additionally, as suggested by Fang, Tian & Tice (2014), the natural logarithm of illiquidity was calculated, in order to avoid a non-normal distribution. Lastly, the scale of values was inverted, enabling to measure stock liquidity instead of illiquidity.
For the third and last independent variable, institutional investors, data were collected as well from the Thomson Reuters Eikon database. More specifically, in this dataset each shareholder was classified within a category, allowing to understand if an investor could be identified as institutional or not. Brossard et al. (2013) recognized seven categories of
institutional investors to be most common in their study: investment advisors, hedge funds, private equity, venture capital funds, pension funds, banks and trusts, and research firms. In the current study, three additional categories of institutional investors have been included: endowment funds, insurance companies and sovereign wealth funds. To calculate the total percentage of institutional investors within each firm, the percentage owned by each shareholder belonging to one of the ten categories described above were summed.
3.2.3 Control variables
To assess whether firms’ time horizons could be affected by some other firm level factors, three control variables were identified. The data collected from Compustat capital IQ allowed to control on a quarterly basis for firm size, return on assets and financial leverage. Firstly, firm size, basing on what done by DesJardine and Bansal (2014), was calculated as the natural logarithm of one plus the book value of total assets. This variable was included in order to understand if a bigger firm may be more likely to focus on the long-term than a smaller one. Secondly, return on assets (ROA) was calculated as the ratio of operating income before depreciation to the value of total assets (DesJardine & Bansal, 2014), and it was included to check whether firms might focus more on the short-term because of poor financial performance (Souder & Shaver, 2010). Thirdly, financial leverage was calculated as the ratio between the sum of long-term debt and debt in current liabilities to the value of total assets, and it was added to control if financial health might affect the time horizons of firms (DesJardine & Bansal, 2014).
Moreover, considering the dependent variable has been measured as the number of words used by management referring to the long-term, it is possible that the time horizon of firms might have been influenced by the structure of the conference calls. More specifically, the time horizon of analysts or other non-management individuals that intervened in the
conference calls was included as a control variable. Equally to the temporal orientation of firms, it was calculated considering the ratio of long-term words to the total of long- and short-term words used by analysts intervening in the Q&A. This variable allowed to control for the fact that questions during the conference calls may have induced the management to discuss about the long- rather than the short-term or vice versa, modifying the calculation of firms’ time horizon. Following the same line of reasoning, the number of analysts has also been included as a control, as a different number of people asking for questions may have increased or decreased the likelihood of covering more long- or short-term issues. Lastly, the total number of words of the conference calls, calculated as the total words of managers plus other intervening people, were included as a control variable to assess whether a longer or shorter conference call may have affected the temporal orientation calculation.
The collection of data about the variables ownership concentration and institutional investors pointed out some companies for whichthe totality of ownership is not covered, with a change in the total percentage of ownership known across the periods. This might have been due to the fact that over time the first 100 shareholders could have been changed, and as they were sorted by the most recent period, in the previous years some owners could have been left out. Therefore, after analyzing the data further, it will be considered whether the known percentage of ownership will be included as a control variable.
4. Results
In this section it will be explained how the different variables were analyzed in order to test the hypotheses previously stated. As a first step, the descriptive analyses performed on the data will be shown, examining the distributions of the variables and the correlations among
them. Subsequently, the regression analyses that were performed will be described, evaluating the relationships between the variables in light of the hypotheses.
4.1 Preliminary analysis
As a first step, descriptive statistics including skewness and kurtosis were run in order to identify the presence of non-normally distributed variables. The analysis pointed out that firms’ time horizon, return on assets, time horizon of analysts and the total number of words in the conference calls presented non-normal distributions (as it can be seen in table 1). More specifically, firms’ time horizon had a positive kurtosis, which means that some values made the distribution more pointy than a normal one. The other three non-normally distributed variables show both a high positive kurtosis and a high positive skewness, which implied their distributions were more pointy and had a longer tail on the right. To reduce extreme values that might have made the distribution non-normal, the four variables mentioned above were winsorized at a 1% and 99% level. Accordingly, the values below the 1st percentile and above the 99th percentile were replaced respectively by the values of the 1st and 99th percentiles. After this procedure, the skewness and kurtosis of the four variables were reduced, making their distributions more normal (as shown in table 2).
Table 1 – Descriptive statistics
N Min. Max. Mean SD Skewness Kurtosis
Statistic Statistic Statistic Statistic Statistic Statistic SE Statistic SE
Firms’ time horizon 5493 .0000 .5714 .1506 .0632 .8412 .0330 1.4962 .0661
Ownership concentration 5493 .0000 .7851 .3158 .1219 .5056 .0330 .0639 .0661 Institutional investors 5493 .1504 .9917 .5845 .1572 .1540 .0330 -.5391 .0661 Stock liquidity 5493 1.8466 4.7125 3.5988 .4949 -.6299 .0330 .1909 .0661 Known ownership 5493 .1930 .9990 .6198 .1584 .1545 .0330 -.6696 .0661 Firm size 5493 4.6733 13.6494 8.6379 1.5479 .3202 .0330 .1232 .0661 Financial leverage 5493 .0000 1.1306 .2396 .1695 .8241 .0330 1.0022 .0661 Return on assets 5493 -.4347 .8682 .0771 .0760 1.5540 .0330 8.4163 .0661
Analysts’ time horizon 5493 .0000 1.0000 .1336 .1059 1.6922 .0330 6.8322 .0661
Number of analysts 5493 0 28 8.1116 3.971 .3407 .0330 .1916 .0661
Table 2 – Descriptive statistics after winsorizing variables
N Min. Max. Mean SD Skewness Kurtosis
Statistic Statistic Statistic Statistic Statistic Statistic SE Statistic SE
Firms’ time horizon 5493 .0369 .3393 .1503 .0614 .6555 .0330 .3189 .0661
Ownership concentration 5493 .0000 .7851 .3158 .1219 .5056 .0330 .0639 .0661 Institutional investors 5493 .1504 .9917 .5845 .1572 .1540 .0330 -.5391 .0661 Stock liquidity 5493 1.8466 4.7125 3.5988 .4949 -.6299 .0330 .1909 .0661 Known ownership 5493 .1930 .9990 .6198 .1584 .1545 .0330 -.6696 .0661 Firm size 5493 4.6733 13.6494 8.6379 1.5479 .3202 .0330 .1232 .0661 Financial leverage 5493 .0000 1.1306 .2396 .1695 .8241 .0330 1.0022 .0661 Return on assets 5493 -.0358 .3168 .0769 .0691 1.1418 .0330 1.2987 .0661
Analysts’ time horizon 5493 .0000 .4800 .1324 .0998 1.0112 .0330 1.1348 .0661
Number of analysts 5493 0 28 8.1116 3.9707 .3407 .0330 .1916 .0661
Total words 5493 2946 14653 7939.8726 2306.0670 .2347 .0330 .1007 .0661
Table 2 provides summary statistics for all the variables. On average, 15 percent of words relating to firms’ time horizon referred to a long temporal orientation. Ownership concentration, based on the total percentage of shareholders owning at least a three percent of each firm, was on average 31.6 percent. Institutional investors represented on average the 58.5 percent of the ownership of firms. The bid-ask spread on which the liquidity measure is based had a mean of 2.7 percent. Considering the control variables, on average the 62 percent of ownership was known, the average firm had total assets of 5.6 billion $ (firm size), a return on assets of 7.7 percent and a financial leverage of 24 percent. The average percentage of long time horizon words used by analysts was 13 percent, with an average of 8 analysts and approximately a total of 7940 words in each conference call.
Table 3 shows the correlations between each of the variables. Firms’ time horizon shows a tendency towards a positive relationship with the control variables firm size (r = .225, p < .01) and analysts’ time horizon (r = .282, p < .01). Moreover, the dependent variable shows significant correlations with almost all the other variables, but the low Pearson correlations cast doubts on the strength of these relationships. What is also remarkable to notice is the positive and significant correlation of ownership concentration (r = .802, p < .01) and institutional investors (r = .905, p < .01) with known ownership. This may imply that the
effects of these variables on firms’ time horizons could be affected by changes in known ownership, but for now this variable will not be included in the analysis.
Table 3 – Correlation matrix
Variable 1 2 3 4 5 6 7 8 9 10 11
1. Firms’ time horizon 1
2. Ownership concentration -.100** 1 3. Stock liquidity .097** .028* 1 4. Institutional investors .014 .646** .303** 1 5. Known ownership -.040** .802** .288** .905** 1 6. Firm size .225** -.333** .207** -.157** -.228** 1 7. Financial leverage .123** .035* .025 .045** .019 .185** 1 8. Return on assets -.062** -.036** .070** .010 .014 -.222** -.104** 1
9. Analysts’ time horizon .282** -.021 .114** .048** .025 .187** .099** -.010 1
10. Number of analysts -.005 -.112** .048** -.014 -.044** .270** .052** .111** -.012 1
11. Total words .085** -.185** -.006 -.047** -.106** .291** .122** .096** .054** .532** 1
5,493 firm-quarter observations
* Correlation is significant at the .05 level (2-tailed). ** Correlation is significant at the .01 level (2-tailed).
4.2 Regression analysis
As mentioned earlier, the sample was formed by a panel of 5493 firm-quarter observations over a six years period. Therefore, to test for the effects on firms’ time horizon predicted in the hypotheses, a series of fixed-effects panel regressions were run. A fixed-effect regression model is a method that allows to control for variables that were not or could not be measured, using each variable as its own control (Allison, 2009). More specifically, there are two necessary conditions to use fixed-effects models. First, the dependent variable must have been assessed for every individual on at least two occasions, and must have been measured with the same scale. Secondly, the independent variables must change in value across the different occasions for a relevant quota of a sample (Allison, 2009). In the current analysis, fixed-effects were controlled within firms. Controlling for the effects within each firm allowed to exclude that the effects of the independent variables on the dependent could be due to differences among firms that are not measurable. The choice of this method meets the requirements described above, as firms’ time horizons are measured in more periods for each
firm, and the values of the independent variables change in most periods. Moreover, quarterly dummies for each quarter were added to control for effects within periods. Therefore, the following models were estimated:
Firms’ time horizon = α + γX + ε + Firm FE + Quarter controls (1)
Firms’ time horizon = α + β × Ownership concentration + γX + ε + Firm FE + Quarter controls
(2)
Firms’ time horizon = α + β × Stock liquidity + γX + ε + Firm FE + Quarter controls
(3)
Firms’ time horizon = α + β × Ownership concentration + β × Stock liquidity + β × Stock liquidity × Ownership concentration + γX + ε + Firm FE + Quarter controls
(4)
Firms’ time horizon = α + β × Institutional investors + γX + ε + Firm FE + Quarter controls
(5)
Firms’ time horizon = α + β × Ownership concentration + β × Institutional investors + β × Institutional investors × Ownership concentration
+ γX + ε + Firm FE + Quarter controls
(6)
Firms’ time horizon = α + β × Ownership concentration + β × Institutional investors + β × Stock liquidity + γX + ε + Firm FE + Quarter controls
(7)
where X is the vector of control variables, including firm size, financial leverage, return on assets, analysts’ time horizon, number of analysts and total words, and ε is the error term.
The first model tested whether the control variables have an effect on firms’ time horizon. No significant effect of firm characteristics that may have affected firms’ temporal orientation were found (βfirm size = .003, p > .05; βfinancial leverage = .013, p > .05; βreturnon assets =
horizons were found to have significant effects (βanalysts’ time horizon = .089, p < .001; βnumber of analysts = -.002, p < .001; βtotalwords = .000, p < .01). Model 2 then tested the effect predicted in
hypothesis 1, for which a positive relationship was expected between the independent variable ownership concentration and firms’ time horizon. The regression showed a significant relationship between them, but in a negative direction (βownership concentration = -.023,
p < .05). As an effect, hypothesis 1 was not confirmed. In model 3, hypothesis 2a was tested, predicting that a higher stock liquidity would be negatively linked to firms’ temporal orientation. The regression model also did not show a significant relationship between the two variables (βstock liquidity = -.001, p > .05). Therefore, hypothesis 2a did not find any support.
Hypothesis 2b was predicting that ownership concentration would moderate the relationship between stock liquidity and firms’ time horizon, making it positive instead of negative. The moderation was tested in model 4, and the regression analysis did not present a significant effect of the interaction term (βstock liquidity × ownership concentration = .001, p > .05), which means
hypothesis 2b was not supported.
Model 5 allowed to assess if institutional investors would have the positive effect on firms’ temporal orientation which was predicted in hypothesis 3a. The regression showed a highly significant positive relationship between the two variables (βinstitutional investors = -.006, p
> .05). As a consequence, hypothesis 3a was also not supported. Hypothesis 3b was tested in model 6, with the expectation of a positive moderating effect of ownership concentration on the relationship between institutional investors and firms’ time horizon. The regression did not show a significant effect of the interaction term on the dependent variable (βinstitutional investors × ownership concentration = -.033, p > .05), thus hypothesis 3b did not found support. Lastly,
model 7 tested the effects of all the independent variables together on the dependent variable, showing again a negative significant effect of ownership concentration on firms’ temporal orientation (βownership concentration = -.037, p < .05), but no significant effects of stock liquidity
(βstock liquidity = -.001, p > .05) and institutional investors (βinstitutional investors = .021, p > .05). A
summary of all the regression models that were run can be found below in table 4.
Table 4 – Fixed-effects regression analysis
Variable Model 1 Baseline Model 2 H1 Model 3 H2a Model 4 H2b Model 5 H3a Model 6 H3b Model 7 Ownership concentration -.023* -.027 -.017 -.037* (.011) (.050) (.031) (.015) Stock liquidity -.001 -.001 -.001 (.004) (.006) (.004) Institutional investors -.006 .031 .021 (.011) (.021) (.016)
Stock liquidity * Ownership
conc. .001
(.014) Inst. investors * Ownership
conc. -.033 (.044) Firm size .003 .002 .003 .002 .003 .002 .002 (.004) (.004) (.004) (.004) (.004) (.004) (.004) Financial leverage .013 .015 .013 .015 .013 .015 .015 (.011) (.012) (.012) (.012) (.012) (.012) (.012) Return on assets -.020 -.021 -.020 -.021 -.020 -.022 -.022 (.016) (.016) (.016) (.016) (.016) (.016) (.016)
Analysts’ time horizon .089*** .089*** .089** .088*** .089*** .088*** .088***
(.007) (.007) (.007) (.007) (.007) (.007) (.007) Number of analysts -.002*** -.002*** -.002*** -.002*** -.002*** -.002*** -.002*** (.000) (.000) (.000) (.000) (.000) (.000) (.000) Total words .000** .000** .000** .000** .000** .000** .000** (.000) (.000) (.000) (.000) (.000) (.000) (.000) Constant .107*** .121*** .109** .125** .111*** .109** .117** (.030) (.031) (.033) (.037) (.031) (.032) (.035)
Quarter controls Included Included Included Included Included Included Included
Firm fixed-effects Included Included Included Included Included Included Included
R-squared .074 .074 .074 .074 .074 .075 .075
Firm-quarter observations 5,493 5,493 5,493 5,493 5,493 5,493 5,493
Number of firms 366 366 366 366 366 366 366
Standard errors in parentheses.
Coefficients for quarter controls not shown in the regression summary. *** p<0.001, ** p<0.01, * p<0.05
The regression models that were run did not provide support for any of the effects expected in the hypotheses. However, observing the negative regression coefficients of the independent variables, it is possible that increasing values of these variables over time could have made the relationships negative, considering that the regression included controls for effects over
time. As it can be seen in the appendix, graphical representations of the variables show an increase over time of ownership concentration, institutional investors and stock liquidity, while time horizons do not follow a clear trend. Considering that ownership concentration and institutional investors were found to be linked to the known percentage of ownership, it is likely that an increase in known ownership over time might have affected these variables. Indeed, a graphical representation of known ownership (which can be found in the appendix), shows a clear increase of this variable across the time periods. Basing on these considerations, another series of panel fixed-effect regressions was run, adding to the previous models the known percentage of ownership as a control variable.
At first, model 1 allowed to test whether the control variables have an effect on firms’ time horizon. The regression analysis did not highlight any significant difference in firm characteristics that may have affected temporal orientation (βfirm size = .002, p > .05; βfinancial leverage = .014, p > .05; βreturnon assets = -.020, p > .05). On the other hand, the measure of firms’
time horizon was affected by the time horizon and number of analysts in the Q&A section, and by the length of the conference call (βanalysts’ time horizon = .089, p < .001; βnumber of analysts =
-.002, p < .001; βtotal words = .000, p < .01). Moreover, in this model known ownership was
negatively but not significantly related to firms’ time horizons (βknown ownership = -.020, p >
.05). Model 2 tested the positive relationship expected between ownership concentration and firms’ temporal orientation, finding that the effect is not significant (βownership concentration =
-.021, p > .05). Thus, hypothesis 1 was not supported. In model 3 it was tested whether stock liquidity has the negative effect on firms’ time horizon expected in hypothesis 2a. The regression analysis did not point out a significant effect (βstock liquidity = -.001, p > .05),
consequently hypothesis 2a is not confirmed. The moderation of ownership concentration on the relationship between stock liquidity and firms’ temporal orientation was tested in model 4. The regression analysis did not present the significant positive effect of the interaction
term that was expected (βstock liquidity × ownership concentration = .001, p > .05), therefore hypothesis
2a is not supported.
Model 5 then tested for the positive effect of institutional investors on firms’ time horizons expected in hypothesis 3a. The regression analysis showed a highly significant effect of this independent variable on the dependent variable (βinstitutional investors = .135, p <
.001) and also a significant negative effect of known ownership (βknown ownership = -.147, p <
.001), which suggests that indeed institutional investors could have been affected by the increasing percentage of ownership known over time. As an effect, hypothesis 3a was strongly supported. Hypothesis 3b, predicting a positive moderating effect of ownership concentration on the relationship between institutional investors and firms’ temporal orientation was tested in model 6. The regression did not show a significant effect of the interaction term on the dependent variable (βinstitutional investors × ownership concentration = -.044, p >
.05), providing support only for the direct effect of institutional investors (βinstitutional investors =
.148, p < .001) and known ownership (βknown ownership = -.136, p < .01) as in hypothesis 3a.
Lastly, model 7 tested the effects on firms’ temporal orientation of all the variables taken together, supporting again only a significant effect of institutional investors (βinstitutional investors
= .132, p < .001) and known ownership (βknown ownership = -.134, p < .01). Table 5 shows a
Table 5 – Fixed-effects regression analysis controlling for known ownership
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Baseline H1 H2a H2b H3a H3b
Ownership concentration -.021 -.024 .014 -.014 (.017) (.053) (.033) (.017) Stock liquidity -.001 -.001 -.001 (.004) (.006) (.004) Institutional investors .135*** .148*** .132*** (.037) (.040) (.037)
Stock liquidity * Ownership
conc. .001
(.014) Inst. Investors * Ownership
conc. -.044 (.044) Firm size .002 .002 .002 .002 .002 .002 .002 (.004) (.004) (.004) (.004) (.004) (.004) (.004) Financial leverage .014 .015 .014 .015 .016 .016 .016 (.011) (.012) (.012) (.012) (.012) (.012) (.012) Return on assets -.020 -.021 -.020 -.021 -.021 -.022 -.022 (.016) (.016) (.016) (.016) (.016) (.016) (.016)
Analysts’ time horizon .088*** .088*** .088*** .088*** .089*** .089*** .089***
(.007) (.007) (.007) (.007) (.007) (.007) (.007) Number of analysts -.002*** -.002*** -.002*** -.002*** -.002*** -.002*** -.002*** (.000) (.000) (.000) (.000) (.000) (.000) (.000) Total words .000** .000** .000** .000** .000** .000** .000** (.000) (.000) (.000) (.000) (.000) (.000) (.000) Known ownership -.020 -.004 -.019 -.003 -.147*** -.136** -.134** (.011) (.017) (.011) (.017) (.037) (.041) (.041) Constant .122*** .122*** .124** .126** .126*** .119*** .130*** (.031) (.031) (.035) (.038) (.031) (.032) (.035)
Quarter controls Included Included Included Included Included Included Included
Firm fixed-effects Included Included Included Included Included Included Included
R2 .074 .074 .074 .074 .077 .077 .077
Firm-quarter observations 5,493 5,493 5,493 5,493 5,493 5,493 5,493
Number of firms 366 366 366 366 366 366 366
Standard errors in parentheses.
Coefficients for quarter controls not shown in the regression summary. *** p < .001, ** p < .01, * p < .05
