The under-researched family firm: New insights from unique Norwegian
This chapter reports main findings from a comprehensive study of how Norwegian family firms are governed and how they behave and perform as economic entities. Analyzing all firms from 2000 to 2015, we show that the family firm represents the most widespread way of organizing economic activity, and that family firms differ fundamentally from other firms. Our results suggest that deeper insight into the economics of family firms may make the firms better, and the public debate more informed.Keywords:: Corporate governance, corporate finance, population, family firms, majority control
We have received insightful comments to earlier versions of this chapter from Morten Bergesen, Herman Høyskel, Simone Møkster, Amir Sasson, and Marius Steen.
7.1 Motivation and overview
In this chapter we show that family firms are not like other firms. We also show that the family firm represents the most common way of organizing enterprise in the Norwegian society. Although we are the first to document these properties in the entire population of all firms, there is no reason to think this situation is peculiar to Norway, but rather reflects a global trend.
Nevertheless, the economics of the family firm remains heavily under-researched in any country. This means little is known about how family firms behave as economic entities. The objective of this chapter is to improve on this unfortunate situation for family firm stakeholders. Owners and employees making important decisions inside the family firm lack systematic insight into typical practice, atypical practice, and best practice. Politicians regulating family firms establish rules with strong impact, but may have only superficial information on what they are regulating. Business school professors training their students can only tell anecdotes about the firms most of the students will be working for or transacting with.
This setting suggests that several stakeholders would benefit from research on the behavior of the family firm in order to understand most firms in the economy better. This chapter contributes to closing this information gap by reporting the main findings from a study of the governance and finance of all Norwegian family firms during the period 2000–2015. These findings come from a project that establishes the broad, overall picture rather than the detailed specifics (Berzins, Bøhren, and Stacescu, 2018b). Accordingly, we will report descriptive statistics rather than tests of behavioral hypotheses, which is the next step in our ongoing project.
In section 7.2, we define the family firm, report why family firms are under-researched, and explain why economists expect family firms to differ from other firms. We describe our database and document the macro-economic significance of family firms in section 7.3. We report governance characteristics in section 7.4, finance characteristics in section 7.5, while we summarize and conclude in section 7.6.
7.2 Why are family firms unique and under-researched?
Unfortunately, there is no common definition of the family firm in the literature. More than 90 different definitions exist (European Commission, 2009), making it difficult to compare findings across empirical studies. The lack of a common definition manifests itself as ambiguous conclusions about the main economic relationship analyzed in the family firm literature, which is how performance depends on whether the firm is a family firm (O’Boyle, Pollack, and Rutherford, 2012; Amit and Villalonga, 2014). To illustrate, Villalonga and Amit (2006) find that this relationship between performance and family firm status is either positive, negative, or insignificant, depending on how a family firm is defined. We think an understanding of why some definitions make more sense than others requires an understanding of why family firms are special.
7.2.1 Defining the family firm
We define a family firm as a firm that is majority-owned by individuals related by blood or marriage. This definition reflects both governance and sociology, which are the two dimensions that jointly produce the uniqueness of family firms that we will outline in section 7.2.2.
The governance dimension of the family firm definition reflects that control of the firm’s decision making is the fundamental right (Tirole, 2001). Because the owners elect the board, which recruits and replaces the CEO, owners with a majority stake at the shareholder meeting (general assembly) can control every formal governance position without the other owners’ consent. Therefore, our definition requires that a group of owners holds more than half the shares. These controlling owners can single-handedly choose its participation intensity in the firm’s governance, such as whether to be on the board, be the chair, or the CEO.
The sociological dimension of our family firm definition reflects that we only consider firms where the controlling owners are individuals who constitute a particularly coherent group. We require that the group is tied together by blood or marriage up to the fourth degree of kinship. This means the family also includes great-great-grandparents, great-aunts and great-uncles, aunts and uncles, cousins, grandnieces, and grandnephews.
We prefer this definition of using majority control and sociological coherence to definitions in the literature, which mostly use either control thresholds lower than 50%, sociological thresholds looser than blood or marriage, or governance positions held rather than ownership. For instance, a family firm in Maury (2006) is one where the largest owner has at least 10% of the equity and is either a true family, all the firm’s personal owners regardless of the relationship between them, or a private firm. This definition classifies too many firms as family firms from both a control perspective and a sociological perspective.
Other definitions reflect only whether the family holds governance positions, regardless of whether the family is an owner (Anderson and Reeb, 2003; Villalonga and Amit, 2006; Bennedsen et al., 2007). Such definitions ignore ownership altogether, instead using participation in governance as the only threshold.
We think a family firm definition should reflect the family’s option to take governance positions, but not whether this option has actually been exercised. Hence, what matters is majority ownership, because it produces the option to govern. This option will presumably be exercised whenever the family finds it optimal to do so. A firm that is majority-owned by a family holding neither a board seat nor the CEO position will not be a family firm if the definition uses only governance positions. Conversely, the firm will be classified as a family firm if the family owns nothing, but does hold a board seat. In contrast, our definition classifies the first firm as a family firm regardless of the family’s participation in governance, but not the latter firm, despite the family’s participation. What matters is the right to participate, not actual participation. That right is produced by ownership.
7.2.2 Sources of uniqueness
Family firms are special because the controlling owner is a group of people who are more tightly related sociologically than are other controlling owners. These properties are captured by our definition in section 7.2.1. The important and interesting implication for the economist is that the firm’s behavior may reflect the joint maximization of family goals and business goals, which may not be identical. This situation may make characteristics of the owner unusually important for the behavior of the firm. That is, the governance of the firm may depend on the governance of the family controlling the firm (Bennedsen, Perez-Gonzalez, and Wolfenzon, 2010). Three characteristics seem particularly important.
First, demographic characteristics may matter, such as the number, age, and talent of the family members, the family’s location, the presence of the founder in the firm’s governance, the distribution of ownership within the family, as well as the size, illiquidity, and risk of the family’s wealth. For instance, larger families will have a larger pool of family members to select qualified candidates from. Families with illiquid wealth may prioritize dividends before investments. Families with undiversified wealth and income may make the firm behave more conservatively. The family may also make the firm diversify in order to reduce the risk of the overall family portfolio.
Second, family owners may be special because they have information benefits. Family members know each other particularly well after having interacted more or less intensively all their lives. Accordingly, the information asymmetry within the group of controlling owners is unusually small, making it easier to find their best representatives as officers and directors.
The family is also often close to the firm’s operations. For instance, the family holds both the chair and CEO positions in 79% of Norwegian family firms (Bøhren et al., 2018). This means the owner is unusually well informed about the firm’s future performance. The resulting low information asymmetry between the firm and the controlling owner reduces the family firm’s cost of capital (Leland and Pyle, 1977; Anderson, Mansi, and Reeb, 2003).
Third, family firms may be special due to private benefits, which accrue to the controlling family only (Jensen and Meckling, 1976). Examples of private benefits are when a firm with the family’s name has high reputation in society (social prestige), when a family-controlled newspaper influences common opinion (political impact), when the firm employs family members with lower skills than outside candidates have (nepotism), and when family members trade with the firm at below-market prices (tunneling).1These examples illustrate that private benefits may or may not be costly for the minority owners. The first two examples may not produce negative consequences for them, while the third and fourth do. Thus, private benefits increase the family’s utility of controlling the firm, while the effect on the other owners is neutral or negative.
Just like demographics and information advantages, private benefits may influence the firm’s behavior. A feeling of pride for the family firm’s name, and loyalty to the founders, may make the firm’s survival particularly important to the controlling family. This concern for survival may make the family firm more long-termist and more patient than other firms in its investment, financing, and employment decisions (Sraer and Thesmar, 2007). Concerns for survival may also make the family firm adopt less aggressive growth strategies and choose industries and products with less risk (Almeida and Wolfenson, 2006).
7.2.3 Reasons for low attention
The existing research on corporate governance (Hermalin and Weisbach, 2017) and corporate finance (Eckbo, 2007) is heavily biased toward firms that are public (listed on a stock exchange and widely held) rather than private (not listed and closely held). Because almost every family firm is private, the lack of research on private firms carries over to family firms.
There are at least three reasons why economists have paid limited attention to private firms in general and to family firms in particular. First, public firms may look more attractive because the quality of the firm’s behavior may be measured by the observable market value and not just by the book (accounting) value, which is normally the only option in private firms. Thus, performance is harder to measure when the firm is private.
Second, regulation puts stronger requirements on the external communication of public firms than of private firms. Public firms must regularly publish standardized, audited accounting statements that data vendors make easily accessible to investors, analysts, and researchers worldwide. Reliable accounting data for private firms are harder to obtain in most countries. Correspondingly, governance data for public firms are easily accessible, but not for private firms. Moreover, even if governance data had been available, family firms cannot be identified and analyzed reliably without knowing the family relationships between owners, directors, and CEOs. This task requires census data for larger samples, which are not publicly available.
Finally, economists tend to consider the family firm an anachronism (Bennedsen, Perez-Gonzalez, and Wolfenzon, 2010). The reason may be a lack of recognition of the fact that family firms continue to play a strong economic role around the world (Franks, Mayer, and Rossi, 2009; Mehrotra et al., 2013). This prevalence of the family firm, despite lower frictions in capital and labor markets, may jointly refute the idea that the family firm is a viable organizational form only in underdeveloped markets (Khanna and Yafeh, 2007). Moreover, most governance researchers may simply have overlooked the family firm because the ruling paradigm concerns the widely held firm, and the resulting separation between weak owners and strong managers (Berle and Means, 1932; Roe, 1994; Hermalin and Weisbach, 2017).
Unfortunately, existing findings from public firms may not apply to family firms. We will establish that family firms face different environments and have different governance. For instance, family firms are almost always private and cannot finance themselves in active equity markets. This means their shares can be traded only at high transaction costs, and that their minority shareholders enjoy less regulatory protection than in public firms. Our data show that family firms have unusually concentrated ownership, small boards, and many owners in CEO and chair positions. Theory suggests that such characteristics matter for the firm’s investments, capital structure, dividend policy, growth, and risk management. This behavior may in turn influence performance, such as the return on capital invested. The literature has just started addressing the relationship between performance and the sources of family firm uniqueness. The only exception is the small subsample of public family firms, which may differ strongly from their private counterparts.
7.3 Data, groups, and macro-economic significance
Our database captures the entire population of Norwegian firms with limited liability (AS and ASA) during the period 2000–2015. Many private firms are organized in corporate groups with parents and subsidiaries. We report a firm separately if it has no majority parent, reporting one observation for a set of firms organized as a corporate group by consolidating the firms’ activities.2We find that groups controlled by families are up to eight levels deep and may contain almost 100 subsidiaries.
Experian (www.experian.no) provides the accounting, ownership, and board data, the census data on family relationships are from the National Registry (folkeregisteret) (www.skatteetaten.no/en/person/national-registry), and the Centre for Corporate Governance Research (www.bi.edu/ccgr) organizes these two data sets as one integrated database. Tax return data on shareholder wealth and income are from Statistics Norway (www.ssb.no/en) and the Norwegian Tax Administration (Skattedirektoratet) (www.skatteetaten.no). The average sample contains about 86,000 family firms and nonfamily firms per year. The detailed construction of the sample is described in chapter 4 in Berzins, Bøhren, and Stacescu (2018b).
Figure 7.1 shows how family firms contribute to economic activity in the Norwegian economy in 2015. According to panel A, 70% of all firms are family firms, and they account for 39% of aggregate employment, 27% of aggregate sales, and 17% of all assets. Panel B shows that family firms are more common in certain industries. The fraction of family firms varies between 79% and 37%. For instance, 69% of the firms are family-controlled in retail and wholesale, which also has the largest number of family firms in the economy (22,780 firms; not shown in Figure 7.1). Only 53% are family firms in mining and oil, which also has few family firms (432 firms; not shown in Figure 7.1).
These numbers document that the family firm is by far the most common way of organizing an economic enterprise. Because the share of firms exceeds the share of employment, sales, and assets, family firms tend to be smaller and more labor intensive than nonfamily firms. Family firms also seem to gravitate toward industries that use more labor and less assets. These findings may suggest that, compared to nonfamily owners, family owners are more constrained by limited funds, more reluctant to grow even if they could, more often choose industries with lower capital requirements, and more often specialize in managing labor. We address these questions using corporate finance data in section 7.5.
7.4 Corporate governance
Corporate governance concerns the relationship between owners, directors, and officers (Shleifer and Vishny, 1997). We count the family as one owner by adding each family member’s equity stake in the firm into the family’s stake. To capture differences across subgroups of family firms, we show the results separately for all family firms, for sole entrepreneurships versus classic firms, and for large versus medium and small firms.
We define sole entrepreneurships as firms less than ten years old and controlled by one person.3According to this definition, only one family member can own shares in an entrepreneurship. The idea is to isolate the case where one individual in the family rather than several start the firm. Sole entrepreneurships may have less family firm flavor than other family firms, which we call classic. We define a large firm as having at least 10 employees and sales above NOK 10 million at 2015 purchasing power on average over the sample period. Groups are reported as one observation, and we carefully measure governance and finance characteristics at the appropriate level in the group.
We analyze sample characteristics of the ownership structure in section 7.4.1, the board in section 7.4.2, and the CEO in section 7.4.3. While we consider all family firms as one sample in sections 7.4.1–7.4.3, we compare subsamples of family firms in section 7.4.4.
7.4.1 The ownership structure
Our database includes all firms in the Norwegian economy. Therefore, we can measure ultimate ownership, which is the owner’s direct equity stake plus the indirect stake owned through corporate intermediaries. Accounting for indirect ownership is particularly important in our sample period, as the number of holding companies more than tripled in 2005, when increased dividend taxation for persons discouraged the use of direct ownership (Berzins, Bøhren, and Stacescu, 2018a). Similarly, ship-owners reorganized their corporate groups in order to adapt to the new tax regime for ship-owning firms in 2007.
Panel A of Table 7.1 shows properties of the ownership structure in 2015. The most common concentration measure is the largest equity stake, which exceeds 50% in all family firms by definition. Counting the family as one owner, the average largest owner holds 50% in nonfamily firms and 95% in family firms. Thus, while the average largest owner in both firm types has simple majority (1/2), the largest owner in family firms has supermajority (2/3) by a wide margin. To illustrate, while the largest average owner can single-handedly elect the board in both firm types, that owner can also amend the charter in family firms. In fact, the controlling family can even buy out the minority owners, which requires a 90% stake (Bøhren and Krosvik, 2013).
|All firms||Non-family firms||All||Classic||Sole entre-preneurships||Large||Medium & small|
|Panel A: Ownership|
|Number of owners||3.05||5.74||1.87||2.42||1.28||3.22||1.69|
|Second largest owner||15%||24%||11%||18%||3%||16%||10%|
|Percent single-owner firms||63%||14%||81%||75%||88%||65%||83%|
|Panel B: Board|
|Family is on the board||99%||99%||98%||98%||99%|
|Family's share of seats||91%||91%||90%||86%||91%|
|Family has board majority||85%||87%||84%||77%||87%|
|Family has all seats||81%||80%||83%||67%||83%|
|Family has chair||80%||89%||70%||80%||81%|
|Minority owner has board seat||8%||9%||6%||11%||7%|
|Proportion female directors||21%||19%||22%||22%||22%||17%||22%|
|Proportion female directors in 2000||16%||12%||17%||19%||15%||13%||18%|
|Age female directors||49.9||49.2||50.3||51.7||46.1||51.6||50.1|
|Age male directors||52.6||51.6||53.1||55.3||48.3||54.5||52.9|
|Standard deviation director age||8.2||7.2||9.3||9.8||7.8||10.1||9.0|
|Standard deviation director gender||33%||29%||35%||33%||37%||29%||35%|
|Panel C: CEO|
|Family has CEO||83%||82%||85%||73%||85%|
|Family has CEO and chair||65%||74%||61%||58%||67%|
|CEO has majority||64%||0%||81%||66%||98%||64%||82%|
|Proportion female CEOs||19%||18%||19%||18%||20%||10%||21%|
|Proportion female CEOs in 2000||14%||12%||14%||14%||14%||7%||15%|
|Age female CEOs||46.9||45.7||47.3||49.9||44.1||50.1||47.2|
|Age male CEOs||50.0||48.6||50.6||53.9||45.9||52.1||50.4|
|Number of firms||115,259||35,077||80,182||32,281||32,422||9,560||70,622|
This table shows ownership, board, and CEO characteristics of all Norwegian firms with limited liability in 2015. All figures are mean values. A family firm is majority-owned by individuals related by blood or marriage up to the fourth degree of kinship, while nonfamily firms are all other firms. A sole entrepreneurship is less than ten years old and is controlled by one person, while classic firms are the remaining family firms. A large firm has average sales above NOK 10 million and at least 10 employees, while the remaining family firms are medium and small. Single-owner firms have just one owner, which may be a family with several individual owners. Ownership is measured as the sum of the shareholder’s direct and indirect equity holdings in the firm. “Inside owners” is the equity fraction held by the firm’s officers and directors as a group. Business groups represent one observation each.
Although both firm types in our sample have few owners on average, there are still considerably fewer owners in family firms than in nonfamily firms (1.87 vs. 5.74, respectively). The average stake of the second largest owner is smaller, being 11% in family firms and 24% in nonfamily firms. Moreover, while 81% of the family firms have no other owners than the family, such single-owner firms constitute only 14% of the nonfamily firms. Finally, officers and directors, who are what we call the firm’s insiders, own 67% in nonfamily firms and 94% in family firms.
The figures in panel A reflect two properties of the family firm’s ownership structure with important implications for governance (Edmans and Holderness, 2017): Ownership concentration is very high, and insider ownership is also very high. These two properties speak directly to what the literature calls the first and the second agency problem, respectively (Villalonga and Amit, 2006). The first agency problem concerns conflicts of interest between the firm’s owners and insiders (Shleifer and Vishny, 1997). This alignment problem is less serious when insiders have incentives to act as the owners would have done themselves (Demsetz and Lehn, 1985). That automatically happens when the insiders are in fact large owners, as in our case. Therefore, the first agency problem is mostly negligible.
The second agency problem is potentially serious when ownership is concentrated (La Porta et al., 2000). This situation may tempt the controlling family to make decisions that benefit themselves at the other owners’ expense. However, this problem is smaller the more the controlling stake exceeds the minimum level of 50%. For instance, a family owning 51% pays only 51% of the loss they cause the firm that underprices goods to the family. The remaining 49% must be paid by the minority. In contrast, a family owning 99% pays 99% of the loss, while the minority pays only 1%. Panel A shows that the average family is rather close to the latter scenario of holding a very high control stake. Moreover, there is no second agency problem whatsoever in 82% of the family firms, because they have no minority owners.
We conclude that, compared to other firms, the ownership structure of family firms makes them less exposed to conflicts of interest both between owners, officers, and directors, and also less exposed to conflicts between large and small owners.
7.4.2 The board
Panel B of Table 7.1 shows board characteristics.4Because our data cannot tell how nonfamily directors are related to the owners, the table has no entries for seat identity in nonfamily firms. The controlling family is on the board in 99% of the family firms, holds at least half the seats in 85%, every seat in 81%, and the chair in 80%. The directors own more equity in family firms than in nonfamily firms (83% vs. 66%). The tendency is even stronger for the chair (66% vs. 26%).
This situation means that, just like the shareholder meeting in panel A, the board meeting is totally dominated by the controlling family. In fact, the family’s average fractions of share capital and of board seats are quite close, being 95% and 91%, respectively. Thus, once more, we find that the first agency problem is minuscule in family firms. However, the composition of the board does not mitigate the potential second agency problem. Although minority owners hold a higher fraction of seats than of shares (9% vs. 5%), their formal power is mostly nil because the board always decides by simple majority.
Family firms tend to have smaller boards than nonfamily firms (1.64 vs. 2.98 seats) and directors with longer tenure (2% vs. 7% likelihood of at least one director being replaced in a given year).5Norwegian boards are very small by international standards. For instance, the average size of 6.2 including employee directors in public firms is among the smallest boards in the world (Bøhren and Strøm, 2010). Female directors are more common in family firms (22% vs. 19% of the seats), and more common than fifteen years earlier in both firm types. The average female director is about two and a half years younger than males in both firm types. Finally, diversity as measured by the standard deviation of director age and of gender both suggest that family firms have the more heterogeneous board.
Taken together, these figures reflect that the family is very much present on the family firm’s board. This active involvement reduces the separation between ownership and control, which is positive from a corporate governance perspective. The potential problem is conflicts with minority owners, who seldom have formal power in the boardroom.
7.4.3 The CEO
Panel C shows that the family’s dominance in shareholder meetings and boardrooms carries over to the CEO position. For instance, the family holds the position of CEO in 83% of the firms, and both the chair and CEO positions in 65%. Again, the first agency problem is practically nonexistent, while there is no obvious mitigation of the second.
The proportion of female CEOs is close to 20% in both firm types, increasing by about five percentage points over the sample period. Both male and female CEOs are on average about two years older in family firms than in nonfamily firms.
7.4.4 Types of family firms
We next compare governance characteristics across different types of family firms. The figures in the four rightmost columns of Table 7.1 show that the sole entrepreneurship and the large family firm differ more from the average family firm than do classic or middle-sized and small family firms. Because sole entrepreneurships and large family firms are also the ones that differ the most from each other, we compare these two family firm types in the following.
Regarding ownership, the ownership concentration and the insider holdings are both highest in sole entrepreneurships, and lowest in large firms. As for the board, family domination is strongest—and size smallest—in sole entrepreneurship, while the opposite is true for large firms. Both types use external chairs more often than other family firms do. The CEO in sole entrepreneurship is more often a controlling owner, while large firms are at the opposite end. Sole entrepreneurships use female CEOs more often than do large firms. They also have the youngest directors and CEOs regardless of gender, while large firms have the oldest.
Summarizing section 7.4, we find that the conflict potential between owners, directors, and officers is very small in family firms due to their ownership structure, board composition, and the CEO’s background.6This situation differs widely from the widely held public firm, where large owners are mostly much smaller, owners are much less present on the board, and the CEO is very rarely a large owner. Conflicts between the family and minority owners are also mitigated by the ownership structure, but not by the background of officers and directors. This conclusion holds across different types of family firms, although sole entrepreneurships and large firms have governance that resembles the widely held public firm the least and the most, respectively. The very strong family dominance in board and CEO positions may create settings where the beneficial effects of family control are offset by the negative effect of recruiting officers and directors from a pool of talent that is too limited.
7.5 Corporate finance
In this section, we briefly describe micro-economic characteristics of the family firm in 2015 as summarized in Table 7.2. Like in section 7.4, we first compare all family firms as a group to nonfamily firms in sections 7.5.1–7.5.3, comparing different types of family firms to each other in section 7.5.4.
|All firms||Nonfamily firms||All||Sole entrepre-neurships||Classic||Large||Medium & small|
|Panel A: Size|
|Sales (million NOK)||53.1||128.0||20.3||9.2||29.6||132.5||5.1|
|Assets (million NOK)||124.1||339.6||29.8||5.3||38.1||205.0||6.1|
|Sales (million NOK)||3.3||5.6||2.7||1.8||4.0||32.6||2.2|
|Assets (million NOK)||2.2||3.4||1.9||0.9||3.4||23.5||1.4|
|Panel B: Factor intensity|
|Assets per employee (million NOK)||4.53||7.84||3.08||1.38||4.17||3.94||2.97|
|Sales per employee (million NOK)||1.98||2.89||1.58||1.26||1.83||2.58||1.45|
|Sales to assets||2.78||3.24||2.58||3.20||2.09||1.56||2.72|
|Panel C: Financing|
|Debt to assets||0.75||0.79||0.73||0.80||0.67||0.64||0.74|
|Short- to long-term debt||4.32||4.94||4.04||4.01||3.93||4.06||4.03|
|Cash to assets||0.31||0.30||0.32||0.36||0.27||0.15||0.34|
|Payout ratio for payers||80.0%||82.4%||78.9%||79.9%||77.7%||64.3%||83.0%|
|Proportion of payers||16.8%||16.0%||17.1%||13.0%||20.3%||31.0%||15.2%|
|Panel D: Growth, risk, and profitability|
|Return on assets (ROA)||3.6%||0.6%||5.0%||4.3%||5.7%||9.4%||4.2%|
|Number of firms||115,259||35,077||80,182||32,281||32,422||9,560||70,622|
This table shows behavioral characteristics of all Norwegian firms with limited liability in 2015. A family firm is majority-owned by individuals related by blood or marriage up to the fourth degree of kinship, while nonfamily firms are all other firms. A sole entrepreneurship is less than ten years old and is controlled by one person, while classic firms are the remaining family firms. A large firm has average sales above NOK 10 million and at least 10 employees, while the remaining family firms are medium and small. “Payout ratio” is dividends divided by earnings, while “Risk” is the standard deviation of sales divided by average sales in 2013–2015. Every ratio in panel C except “Fraction payers” is winsorized at the top 2.5% of the distribution. In panel D, the first two ratios are winsorized at the top 5%, while ROA is winsorized at the bottom and top 2.5% of the distribution. Business groups represent one observation each.
7.5.1 Firm size and factor intensity
We measure size in panel A by sales, assets, and employees, reporting both means and medians. The numbers show at the firm level what panel A of Figure 7.1 shows in the aggregate, i.e., that family firms tend to be smaller than nonfamily firms. The average family firm is about one-tenth the average nonfamily firm according to sales or assets, and about one third according to employment.
However, mean values for nonfamily firms in particular are heavily influenced by some very large firms. This effect is evident when we compare the medians, which is the observation at the center of the distribution (half the firms are above the median and half the firms are below): All three size measures suggest that the typical firm is much smaller than what the means suggest. Moreover, the typical family firm is about half the size of the typical nonfamily firm. Specifically, the median family firm employs three people and sells for NOK 2.7 million, while the median nonfamily firm employs six people and sells for NOK 5.6 million The corresponding figures for assets are NOK 1.9 million and NOK 3.4 million, respectively.
The skewness toward low size can be further illustrated by considering the shape of the distribution in more detail. For instance, half the family firms have sales between NOK 1 million and NOK 8 million, while one-tenth has sales above NOK 28 million.
The factor intensities in panel B confirm the impression that family firms are less capital intensive than nonfamily firms are. For instance, the average family firm has assets of NOK 3.08 million per employee, while the average nonfamily firm has NOK 7.84 million.
Panel C shows that the family firm is on average somewhat less debt-financed than nonfamily firms, as the debt to asset ratios are 0.73 and 0.79, respectively. Moreover, family firms tend to have less debt with short maturity, the ratio of short-term to long-term debt being 4.04 and 4.94, respectively. As the average ratio of cash to assets is practically equal in the two firm types (about 0.30), family firms tend to have the higher working capital.
The standard measure of payout from the firm to its owners is the payout ratio, which we calculate as dividends divided by earnings. The figures suggest that the dividend policy is very similar in the two firm types. About 17% of the firms pay dividends, about 19% of the earnings are paid out if we also include the firms that do not pay, while 80% of the earnings are paid by the subsample of firms that do pay.7The dividend tax increase in 2005 had a very strong effect on dividend policy in both family firms and nonfamily firms. The average payout ratio dropped from about 80% to about 15%, and the fraction of dividend payers dropped from about 50% to 15%. These numbers have increased somewhat during the last five years.
7.5.3 Growth, risk, and profitability
The average annual growth rate in sales and assets is shown in panel D. Family firms grow less than nonfamily firms do as measured by sales (5.7% vs. 6.5%), while they grow more as measured by assets (5.4% vs. 3.6%).
We measure risk by the coefficient of variation for sales, which is the standard deviation of sales divided by average sales. Using the past three years as the measurement period, we find that family firms and nonfamily firms have very similar risk on average, the coefficient of variation being 0.27 in family firms and 0.25 in nonfamily firms, respectively.
Finally, we measure profitability by return on assets (ROA), which is operating earnings after tax but before interest plus interest divided by total assets. Family firms outperform nonfamily firms by a large margin, the average ROA being 5.0% and 0.6%, respectively. If we compare the medians rather than the means, and use the entire sample period 2000–2015 rather than just 2015, the ROA is 1.8 percentage points higher in family firms than in nonfamily firms.
This ROA difference, which may be called the family firm premium, resembles the level reported for public family firms internationally (Amit and Villalonga, 2014). A comprehensive analysis is required to understand this premium. This analysis will start from the fact that the premium in Table 7.2 ignores any firm characteristic other than firm type (family vs. nonfamily) The premium also ignores any variation in ROA across firms within a firm type, using only the average. There may be at least four explanations of the premium that are not mutually exclusive.
First, family firms and nonfamily firms may have different accounting practices. For instance, family firms may be more conservative in valuing their physical assets, and may less often account for intangible assets. Second, because Figure 7.1 shows that family firms are unevenly distributed across industries, the premium may be due to differences in industry returns. Third, the premium may be due to corporate finance characteristics discussed in this section, such as the firm’s size, growth, financing, and risk. Fourth, the explanation may be corporate governance and the unique properties of family ownership we discussed in section 7.4 and quantified in Table 7.1. Our ongoing project addresses these various explanations in order to understand better why family firms behave and perform differently from nonfamily firms (Berzins, Bøhren, and Stacescu, 2018).
7.5.4 Different family firms
Like Table 7.1 on corporate governance, Table 7.2 on corporate finance shows that the atypical firms among the family firms are the sole entrepreneurships and the large firms. Sole entrepreneurships have unusually high labor intensity, high debt, and low dividends, while large firms are at the opposite end. Large firms take less risk, while both firm types have the highest growth and profitability among the family firms.
Summarizing section 7.5, the frequency distribution of size for the population of all firms is heavily skewed toward low size. The median family firm is half the size of the median nonfamily firm, and family firms are more labor intensive than nonfamily firms are. The average family firm has lower financial leverage and higher working capital, while the dividend policy is very similar in the two firm types. Family firms have higher average asset growth, but lower sales growth. Average risk is practically independent of firm type, while the average family firm outperforms the average nonfamily firm. As for corporate governance, the sole entrepreneurship and the large firm are the outliers among the family firms in corporate finance as well. One common property is high average growth and profitability.
7.6 Summary and conclusion
Surprisingly little is known internationally about the family firm as an economic entity. This situation is worrying when considering a main finding we report in this chapter. Using novel data over sixteen years from the population of Norwegian firms and their owners, we find that the family firm is the dominating organizational form in the economy, and that family firms behave and perform differently from other firms. Family firms represent 70% of all firms, employ four-tenths the labor, generate one-quarter of the sales, and own about one-sixth of the assets.
We define a family firm as being majority-owned by individuals related by blood or marriage. Family firms are special because the controlling owner is a group of people who, sociologically, are unusually tightly related. This means the governance of the family firm may depend on the governance of the family as reflected in its demographics, information advantages, and private benefits. The family’s incentives to ensure firm survival may make the family firm more long-termist, risk averse, and patient in its decision making than other firms.
We find that the potential conflict of interest between owners, directors, and officers is very small in most family firms due to their ownership structure, board composition, and CEO background. Conflicts between family and nonfamily owners are also mitigated by the ownership structure, but not by the background of officers and directors. This is true across different types of family firms, although we find that sole entrepreneurships and large family firms have the governance that resembles the widely held public firm the least and the most, respectively. The very strong family dominance in shareholder meetings, boardrooms, and CEO positions may create settings where the beneficial effects of family involvement are offset by the negative effect of recruiting governance resources from too limited a talent pool.
The typical family firm is half the size of the typical nonfamily firm, and both firm types consist of much more small firms than large. Family firms are more labor intensive, have lower leverage and higher working capital, while the dividend policy is very similar. Family firms have higher asset growth, but lower sales growth, while the risk is practically identical. Family firms outperform nonfamily firms on average as measured by returns to assets. The sole entrepreneurship and the large firm are atypical among the family firms regarding both behavior and performance.
These findings reflect a first, rough attempt at answering important questions using new and complicated data. As far as we are aware, no existing study has used population data to explore such a wide range of economic characteristics for family firms. This setting implies that our findings should be used with caution. We think a fruitful way forward is to use our results to ask deeper questions and design more careful analyses. A top priority is to explore where the performance premium for family firms comes from. One explanation is that the premium is simply due to different accounting practices in family firms than in nonfamily firms. A second explanation is our finding that family firms are overrepresented in certain industries. The premium may also be due to the fact that we have only considered averages and ignored the differences in behavior and performance across individual firms. Finally, the premium may be due to our finding that the family firm’s governance has unique properties rooted in the ownership structure and the controlling family. Our ongoing project addresses these explanations in order to understand better why family firms are different.
|1||These examples illustrate that private benefits may or may not be costly for the minority owners. The first two examples may not produce negative consequences for them, while the third and fourth do. Thus, private benefits increase the family’s utility of controlling the firm, while the effect on the other owners is neutral or negative.|
|2||We find that groups controlled by families are up to eight levels deep and may contain almost 100 subsidiaries.|
|3||According to this definition, only one family member can own shares in an entrepreneurship.|
|4||Because our data cannot tell how nonfamily directors are related to the owners, the table has no entries for seat identity in nonfamily firms.|
|5||Norwegian boards are very small by international standards. For instance, the average size of 6.2 including employee directors in public firms is among the smallest boards in the world (Bøhren and Strøm, 2010).|
|6||This situation differs widely from the widely held public firm, where large owners are mostly much smaller, owners are much less present on the board, and the CEO is very rarely a large owner.|
|7||The dividend tax increase in 2005 had a very strong effect on dividend policy in both family firms and nonfamily firms. The average payout ratio dropped from about 80% to about 15%, and the fraction of dividend payers dropped from about 50% to 15%. These numbers have increased somewhat during the last five years.|