Family Firms, Corporate Governance and Exportzhuc/Family firms.pdf · Family Firms, Corporate...

Family Firms, Corporate Governance and Export

By RAOUL MINETTI†, PIERLUIGI MURRO‡ and SUSAN CHUN ZHU†

†Michigan State University ‡Lumsa University

Final version received 20 June 2015.

This paper investigates the effects of family ownership on export using rich data on Italian firms. We find

that family ownership increases the probability that firms export. This benefit is especially pronounced

when family owners retain control rights and seek the support of external managers. The results suggest

that families better internalize the long-run benefits of internationalization, but that their limited

competencies attenuate this benefit in high-tech industries and in remote and unfamiliar export markets.

Family firms also exhibit some tendency to enter foreign markets in a progressive way (sequential

exporting) and through limited collaborations with foreign firms and intermediaries.

INTRODUCTION

In a global economy, export markets are an important venue for firms to grow. For thisreason, scholars and policymakers intensely debate the determinants of firms’international expansion. There is a growing consensus that firms’ corporate governanceinfluences their ability to export. In recent editorials on the costs and benefits of familyfirms, The Economist (2012, 2013) mentions the successful experience of German andNorthern European family firms in international markets, arguing that these firms haveled the export boom of their countries. According to The Economist, a key benefit offamily owners is their long-termism, that is, their ability to internalize the long-runbenefits of expanding abroad. In line with these arguments, Ward (2006) reports theresults of a survey conducted among 300 executives: 43% of the executives of non-familyfirms acknowledged that their companies under-invested in long-term projects, versus8% of the executives of family firms. However, this positive view about family firms isnot universally shared. For example, it is often argued that family businesses might bereluctant to abandon their initial geographical niche and that this could imply lowerpropensity for international expansion (Casson 2000; Onida 2004). Thus the overallimpact of family ownership on export is ambiguous ex ante and is ultimately an empiricalquestion.

Although family businesses account for a large fraction of economic activity in manycountries, there are very few studies on their internationalization.1 The objective of thispaper is to help fill this gap and investigate whether family firms differ from non-familyones in the probability of exporting (extensive margin) and in the volume of export(intensive margin), conditional on exporting. We are also interested in exploring whetherfamily firms behave differently from non-family firms when expanding abroad. Forexample, we aim at understanding whether family firms have a different pace in theinternationalization process (e.g. a more progressive approach to entry in foreignmarkets) or choose different entry modes (e.g. different degrees of collaboration withforeign firms and intermediaries). To address these questions, we exploit a rich survey ofover 20,000 Italian manufacturing firms conducted by the banking group Capitalia. Thedataset provides unusually detailed information on firms’ export activity based directlyon firms’ responses to survey questions. This includes information on firms’ exportparticipation decisions, foreign sales, and patterns of entry in foreign markets. The

© 2015 The London School of Economics and Political Science. Published by Blackwell Publishing, 9600 Garsington Road,

Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA

Economica (2015) 82, 1177–1216

doi:10.1111/ecca.12156

dataset also contains precise information on firms’ ownership structure, such as the typesand equity stakes of the largest shareholders, the alignment between ownership andcontrol, and the involvement of shareholders in firms’ management. The dataset hasrecently been used as a testing ground for other objectives, such as exploring the impactof financial development on firms’ innovation (see, for example, Benfratello et al. 2008).

We find that family ownership positively affects the probability that firms enterforeign markets (extensive margin of export).2 The effect is sizeable. After controlling forvarious firm characteristics and province fixed effects, family firms are 3.1% more likelyto export than their non-family counterparts.3 We also obtain that family ownershipespecially benefits export when families retain control rights (ownership is aligned withcontrol) and when they hire external managers (ownership is partially separated frommanagement).

The analysis then turns to study the mechanisms through which family ownershipaffects the trade margin. According to the theoretical literature, the possible channelsof influence are the long-termism of families, on the positive side, and the lack ofcompetence, the risk aversion and the narrowness of families, on the negative side.Let us first consider the channels of positive influence. We uncover evidence thatfamily ownership increases the probability of entry in foreign markets especially forolder businesses. This could hint at a role of families’ long-termism, as older firmstypically have a higher survival probability than young ones. More interestingly, weobtain the finding that family firms that issue equity to new investors and that plan togo public have a lower probability of exporting. Since these are the firms in whichfamily owners appear to reduce their involvement, the evidence of reduced probabilityof exporting further suggests the positive impact of families’ long-termism on export.The findings for firms’ pace of internationalization further confirm the importance offamily owners’ long-termism. For example, we find some evidence that family firmsenter foreign markets with an intention to stay, which suggests that it is their long-term perspective that induces them to sustain the high fixed costs associated withentry.

Next, let us consider the channels of negative influence. We find that risk aversiondue to lack of financial diversification does not deter family owners from engaging inexport. However, the results reveal that a more nuanced form of risk aversion, familyowners’ fear of losing control and independence, may influence their entry modes, forinstance inhibiting collaborations with foreign firms and intermediaries in exportmarkets. It is, however, families’ shortage of competence and skills that especiallyappears to attenuate the positive impact of family ownership on firms’internationalization. In fact, the benefit of family ownership for export appears to kick inonly when export does not entail strong knowledge and skills. First, the effect of familyownership appears to be stronger the lower the degree of sophistication and technologycontent in the industry. While family firms are significantly more likely to export intraditional and scale-intensive sectors, they are 9% less likely to participate in exportactivities than their non-family counterparts in high-tech industries. Second, theestimates suggest that family ownership has a positive effect especially for firmsspecialized in niche markets. Third, family firms’ export appears to benefitdisproportionately from the support of a skilled workforce inside the firm and from theassistance provided by domestic institutions (e.g. chambers of commerce and exportconsortia) in foreign markets. In addition, when we investigate firms’ pace ofinternationalization, we uncover some evidence that, possibly because of shortage of

Economica

© 2015 The London School of Economics and Political Science

1178 ECONOMICA [DECEMBER

skills and knowledge, family firms tend to enter foreign markets in a progressive way(‘sequential exporting’), starting with easy, close-by markets and small export volumes.

The above results are robust to using different estimation methods, including OLSand probit with province fixed effects as well as IV techniques (2SLS and bivariateprobit). In particular, the reader may be concerned that family ownership can beendogenous. As we explain in the paper, when doing an IV estimation, we constructinstruments for ownership structure employing information on past regulation of Italianlocal financial markets. The IV results confirm the positive effect of family ownership onexport participation. As we will elaborate, however, our instruments can themselves besubject to some criticisms. Therefore it is important to stress that our analysis constitutesa first step towards studying the relationship between family ownership and exportbehaviour. More work is needed to firmly establish the causal effect of family ownershipon export decisions.

This paper is related to the literature on the impact of ownership structure on firmperformance. Some papers show that in the USA, family firms tend to have higherprofitability than non-family firms (Anderson and Reeb 2003; Villalonga and Amit2006). Recent studies on European countries find that family-owned firms perform betterthan widely held ones (Sraer and Thesmar 2007; Favero et al. 2010; Maury 2006). Thesefindings are often interpreted as supporting the theoretical hypothesis that familyownership reduces classic agency problems between owners and managers, such asmanagers’ short-termism (Fama and Jensen 1983). In contrast with these analyses, otherpapers question the benefits of family ownership. Perez-Gonzalez (2006) shows thatfamily firms are less efficient than widely held firms. The theories on the dynastictransmission of management in family firms offer a possible explanation for this result. Infact, these theories suggest that external professional managers have more competenceand skills than family descendants (Caselli and Gennaioli 2013; Burkart et al. 2003).4

Although family firms play an important role in international markets, there is veryscarce evidence on their internationalization process. Analysing about 400 businesses infive US states, Zahra (2003) highlights the role of family ownership in increasingmanagers’ willingness to expand internationally. By contrast, other papers uncover anegative impact of family ownership on firms’ internationalization. Gallo and GarciaPont (1996) study a sample of 57 companies and find that a focus on the local market andinadequate technologies are the main obstacles to the internationalization of familyfirms. Graves and Thomas (2006) examine the determinants of businesses’ internationalpresence and suggest that family firms tend to have a more local culture.

The remainder of the paper is structured as follows. Section I describes theinstitutional background. Section II discusses the predictions of the theoretical literature.Section III illustrates the data and the empirical methodology. In Section IV, we presentthe results on the relationship between family ownership and export. Section V studiesthe behaviour of family firms in the process of internationalization. Section VI addressespossible endogeneity issues with an instrumental variable approach. Section VIIconcludes.

I. INSTITUTIONAL BACKGROUND

Italy provides an ideal environment for investigating the impact of family ownership onfirm internationalization. In 2000 (roughly the middle year of our sample), among non-state-owned manufacturing firms (82% of the firms), the top shareholder was a family oran individual in 54% of cases, another company in 27% of cases, a foreign firm in 13%

Economica


2015] FAMILY FIRMS, CORPORATE GOVERNANCE AND EXPORT 1179

of cases, and a financial holding in 5% of cases. These figures reveal the key role of familyownership. They also reveal the scanty presence of financial institutions amongshareholders, which is historically due to legal prescriptions introduced in the 1930s thatprevented banks from holding shares in corporations. Although in the early 1990s thelegislation changed (d.lgs. 481/92 and 385/93), the limited role of financial institutions ascorporate owners continues to be a feature of the Italian business sector. Anotherrelevant characteristic of the business sector is the high degree of ownershipconcentration (Bianchi and Bianco 2008).

Turning to export, in the years preceding the crisis, the Italian economy displayed anincrease in export activities: total export went from 17.7 billion dollars in 1995 to 44.9billion dollars in 2008. However, the percentage of exporting firms did not changesignificantly between 1995 and 2008. If we focus on manufacturing, in 2000 thepercentage of exporting firms was 18% (52.7% if we restrict attention to firms with morethan 10 employees). In 2000, four manufacturing industries accounted for more than halfof the total export value of the country: machinery manufacturing (19%), motor vehicles(12%), textiles (10%), and electronic equipment (10%). In 2000, 70% of total export wassold in Europe (57% in the EU-15), 14% in North and South America (11% in the USAand Canada), 11% in Asia, 4% in Africa, and 1% in Oceania.

Italy is the fourth European country in terms of value of export, after Germany,France and the UK. The percentage of Italian manufacturing firms involved in exportactivities is in line with that in other European countries. Bellone et al. (2008) find thatover the 1990–2002 period, 73% of French firms with at least 20 employees engaged inexport. For the UK, Greenaway et al. (2007) document that in a panel of 9292manufacturing firms observed over the 1993–2003 period, almost 70% of firms exportedin at least one year. For Sweden, Hansson and Lundin (2004) obtain that around 89% ofmanufacturing firms with more than 50 employees exported during the 1990–9 period.

II. THEORETICAL PREDICTIONS

To understand the relationship between corporate governance and firminternationalization, it is crucial to keep in mind the challenges associated with exportactivities as well as the advantages and disadvantages of family firms. Let us consider thechallenges of export. First, entering foreign markets entails high fixed costs, some ofwhich are sunk (Melitz 2003; Baldwin 1988; Dixit 1989; Das et al. 2007). Firms need tomodify their existing product lines to satisfy foreign demand. They also need to invest(e.g. in advertising) to increase the awareness of their brand in foreign markets. Andentering foreign markets involves gathering information about the markets and theirregulations (e.g. non-tariff barriers and administrative procedures), which in turnrequires knowledge and skills (Sullivan and Bauerschmidt 1989). Second, firms faceintense competition in foreign markets open to international trade and investment. As aresult, profits generated from foreign sales can be more volatile than profits generatedfrom domestic sales (Vannoorenberghe 2012). Third, because it is difficult for lenders toverify foreign sales and secure collateral assets abroad, it may be hard for firms to obtainfunds to finance their foreign expansion (Chaney 2005; Minetti and Zhu 2011).

These properties of export (high fixed costs, riskiness, and low verifiability) arecrucial for understanding the impact of ownership structure. According to Bertrand andSchoar (2006), theories on family firms can be classified into efficiency-based theories,which regard family ownership as a source of comparative advantage, and the culturalview, according to which strong family values may induce family owners to focus on

Economica



maximizing their utility rather than firm value (see also Burkart et al. 2003). The theorieson the comparative advantage of family firms stress that family owners have a long-termcommitment (due also to families’ attachment to their companies) and a long-termhorizon often spanning several decades (‘patient capital’; see also Sraer and Thesmar(2007) for a discussion). Indeed, founding families may perceive themselves as stewardsof the business for future generations (Villalonga and Amit 2006). Family firms couldthen be more willing to sustain the upfront fixed costs of export in order to reap its long-run returns. By contrast, other types of owners can suffer from short-termism. Forexample, while institutional owners may be efficient in monitoring managers, they mayalso have a short horizon. This can occur for two reasons. First, the typical investmentholding period of some institutional investors is short.5 Second, regulatory prescriptionsmay imply that institutional investors are assessed by courts and by the financial pressbased on their short-run performance. Italy is allegedly a suitable context for capturingthe short-term horizon of institutional investors (OECD 2014). Many Italianinstitutional investors lack the expertise necessary for managing the risks of illiquid long-term investments. This is also due to their limited size, which hampers their ability toexploit economies of scale and engage in complex investment strategies. The difficulty ofinstitutional investors in investing directly in long-term assets is compounded by theunderdevelopment of pooled investment vehicles (e.g. private equity and venture capitalfunds and credit funds). In addition, a number of regulatory restrictions (such asquarterly assessments of asset quality and liquidity) may exacerbate the short-termism ofinstitutional investors. Finally, these problems are made more severe by the fact thatItalian firms are often small and informationally opaque, which makes it difficult forinstitutional investors to assess the risks of long-term investments in the firms.

In contrast with the above positive view about the comparative advantages of familyfirms, the cultural view maintains that strong family values may distort the choices offamily owners (Demsetz and Lehn 1985). Distortions could stem from family owners’desire to remain rooted in the original territory and product niche (Sraer and Thesmar2007). This local orientation and narrowness can deter export and induce firms to enteronly close-by markets or at least adopt a progressive approach to entry in foreignmarkets. Distortions can also arise from family owners’ fear of losing independence andcontrol over their business. This can manifest itself in their desire to retain control of thefirm within the subsequent generations of the family (i.e. to preserve the legacy), even atthe cost of transmitting ownership and positions of responsibility to inefficientdescendants, rather than recruiting more competent professional managers. (For adiscussion of this ‘cultural nepotism’, see Bertrand and Schoar (2006) and Bertrand et al.(2008).) This dynastic transmission of management may induce a shortage of skillsnecessary for entering foreign markets (Burkart et al. 2003). For example, the shortage ofskills and competencies could prevent family firms sustaining the fixed costs associatedwith the acquisition of knowledge about foreign markets (or at least force firms to enterforeign markets in a progressive way in order to build competencies). Family owners’fear of losing control could also manifest itself in reluctance to collaborate with foreignagents (e.g. foreign firms and distributors) that might interfere with family control. Thiscan discourage joint ventures and other forms of collaboration that may facilitate exportactivities.

On top of the risk aversion stemming from the fear of losing control, family ownerscould also be risk averse because of their lower financial diversification relative tocorporate and institutional shareholders. In fact, families generally have a large share oftheir wealth invested in their company (Bolton and von Thadden 1998; Villalonga and

Economica



Amit 2006). This could discourage them from entering risky foreign markets. However,one should note that according to various studies, entering foreign markets could alsoallow diversification of the risk associated with demand fluctuations in the domesticmarket (see, for example, Buch et al. 2009; Di Giovanni and Levchenko 2009).

To summarize, the theoretical literature yields the following testable hypotheses.

Hypothesis 1 (export decision). When the benefits of their long-term horizon prevailover the cost of their limited skills, risk aversion, and narrowness, family firms shouldengage in export more than non-family firms. Their export activities should also benefitfrom the contribution of professional managers and from aligning ownership withcontrol. Due to limited skills, family firms could, however, focus on entering close-byand familiar markets, and be more inclined to export if they operate in traditionalindustries.

Hypothesis 2 (internationalization process).

(i) (Pace of internationalization). Given their limited skills and their risk aversion,family firms could choose to ‘test the ground’ in foreign markets by initiallyexporting little, and later expanding their export volume. They could also start byexporting to close-by and familiar markets, and only subsequently enter moredistant markets. However, once they choose to sustain the high fixed costsassociated with entering a distant and difficult export market, given their emphasison long-run gains, they should enter that market with the intention to stay.

(ii) (Entry modes). Two forces can exert opposite effects on firms’ choice whether tocollaborate with other parties. On the one hand, because of lack of competenciesand knowledge, family firms should value the support of domestic institutions suchas export consortia, chambers of commerce and consulates. On the other hand, thefear of losing control and independence could deter family firms from engaging intight collaborations and partnerships with foreign firms and intermediaries.

III. DATA AND EMPIRICAL STRATEGY

Empirical model

We analyse the differences between family firms and their non-family counterparts inexport decisions. We first examine the extensive margin of trade, that is, the probabilityof exporting. The probability that firm i exports can be written as

PðExporti ¼ 1jOi;ZiÞ ¼ Uða1 þOib1 þ Zic1Þ;ð1Þ

where Φ(�) is the standard normal cumulative density function, Oi is a measure of firm i’sownership structure (e.g. a binary variable that equals 1 if the main shareholder of firm iis an individual or family, 0 otherwise) and Zi is a vector of controls for firmcharacteristics that may affect firm i’s export decision, as well as controls for regionaldifferences and dummy variables indicating the year when firm i was surveyed. In theempirical analysis, we also adopt a linear probability model to characterize the exportparticipation decision as follows:

Economica



PðExporti ¼ 1jOi;ZiÞ ¼ a1 þOib1 þ Zic1:ð2Þ

We instead use the following specification to study the intensive margin of trade, i.e.the value of exports, conditional on exporting:

yi ¼ a2 þOib2 þ Zic2 þ e2i;ð3Þ

where yi is the logarithm of firm i’s value of exports, ɛ2i is the error term that captures theunobserved firm characteristics and any other unknown factor that may affect yi and allthe independent variables are the same as in equation (1) or (2).

One might be concerned that firm internationalization can trigger changes in firmownership structure, that is, the causality may be reversed. However, a distinct feature offamily ownership is its persistence over time, that is, to a large extent family ownership isa structural characteristic of a firm. Moreover, in Italy this persistence is particularlypronounced (Bianco 2003; Bianchi and Bianco 2008). In addition to this, our empiricalspecification controls for a rich set of factors that may affect export decisions, includingfirm-level characteristics and province fixed effects. This should minimize the risk ofomitting factors correlated with both family ownership and export decisions. In spite ofthese considerations, it remains possible that there exist unobserved factors thatsimultaneously affect ownership structure and such decisions. To assuage this possibleconcern, we complement OLS and probit estimates with an instrumental variableapproach. The set of instruments include province-level proxies for the tightness of thebanking regulation introduced in Italy in 1936. As will be detailed, we expect thesevariables to be correlated with the ownership structure but to affect the export decisiononly through the ownership channel. We will further elaborate on the instrumentalvariable approach in Section VI.

Data

Our main data source is the ‘Indagine sulle Imprese Manifatturiere’, a survey carried outby the Italian banking group Capitalia. We use four waves of the Capitalia survey, whichcover three-year periods ending respectively in 1997, 2000, 2003 and 2006. The dataset,directed to manufacturing firms within Italy, includes a representative sample ofmanufacturing firms with 10–500 employees (about 94% of firms in the sample) and theuniverse of manufacturing firms with more than 500 employees. Overall, approximately4500 firms were interviewed in each survey wave. The firms analysed in the surveyrepresent about 9% of the population in terms of employees and 10% in terms of valueadded.

Collected data include: data on export activities, such as markets for the firm’sproducts, the percentage of export in total sales, and details on the internationalizationprocess; information on the largest shareholders, including their type and equity shares,as well as other rich details on the ownership structure. The survey also contains detailsabout balance sheet data, company characteristics, including demographics, data onmanagement and workforce at various organizational levels, participation in groups andconsortia; data on relationships with customers, suppliers and banks, and on sources offinance. Some of these variables are available for each year covered by the survey; somerefer to the time of interview; others refer to the three-year period covered by the survey.Table 1 displays summary statistics.

Economica



TABLE1

SUMMARYSTATISTIC

S

Allfirm

sOwnership

Exportstatus

Mean

S.D

.Family

Non-family

t-test

Exporter

Non-exporter

t-test

Exportparticipationandsales

Exportparticipation

0.66

0.47

0.64

0.71

�9.06

Export/sales

42.22

27.80

41.67

43.54

�2.95

Log(export)

12.62

1.71

12.33

13.33

�24.19

Ownershipstructure

Family

0.75

0.43

0.73

0.79

�9.04

Share_family

0.52

0.40

0.42

�3.83

Financialinstitution

0.09

0.29

0.11

0.06

12.07

Corporate

governance

Familywithcontrol

0.93

0.68

0.73

�6.67

Familywithoutcontrol

0.07

0.05

0.06

�3.54

Externalmanagers

0.48

0.50

0.41

0.66

�26.50

0.54

0.34

21.59

Firmcharacteristics

Log(totalassets)

8.79

1.37

8.52

9.55

�38.97

9.02

8.32

31.56

Log(number

ofem

ployees)

3.72

1.11

3.52

4.31

�37.10

3.92

3.33

38.45

Log(capitalintensity)

5.17

3.01

4.93

5.83

�14.57

5.18

5.15

0.44

Log(labourproductivity)

5.38

2.84

5.18

5.92

�12.53

5.41

5.30

2.09

Age

24.30

17.67

24.09

24.97

�2.68

25.26

22.52

10.14

Corporation

0.93

0.25

0.93

0.94

�1.68

0.95

0.91

10.07

Consortium

0.09

0.28

0.09

0.08

2.49

0.10

0.07

5.76

Leverage

0.00

0.01

0.00

0.00

0.56

0.00

0.00

�1.97

Financialconcentration

0.92

0.18

0.93

0.91

3.59

0.91

0.95

�9.49

ATECO5-digit

0.31

0.46

0.32

0.27

6.47

0.33

0.27

9.51

ATECO4-digit

0.42

0.49

0.42

0.43

�1.67

0.39

0.49

�12.53

Economica



TABLE1

CONTIN

UED

Allfirm

sOwnership

Exportstatus

Mean

S.D

.Family

Non-family

t-test

Exporter

Non-exporter

t-test

North

0.68

0.47

0.67

0.72

�7.04

0.72

0.61

15.20

Centre

0.18

0.38

0.19

0.16

4.20

0.17

0.19

�2.92

South

0.14

0.35

0.14

0.12

4.57

0.11

0.20

�16.40

Pavitt’staxonomy

Traditionalsector

0.49

0.50

0.51

0.42

10.53

0.47

0.52

�6.53

Scale-intensivesector

0.21

0.40

0.19

0.24

�6.76

0.17

0.27

�15.12

Specializedsector

0.26

0.44

0.26

0.27

�2.12

0.31

0.16

23.11

High-techsector

0.05

0.21

0.04

0.06

�5.53

0.05

0.04

1.20

Internationalizationprocess

Localdistributors

0.44

0.50

0.43

0.46

1.57

0.44

0.47

0.70

Interm

ediaries

0.07

0.26

0.06

0.11

4.10

0.08

0.04

�1.88

Assistance

EU

0.04

0.20

0.04

0.05

2.28

0.06

0.01

�18.82

Assistance

non-EU

0.03

0.18

0.03

0.04

3.23

0.05

0.00

�17.31

Notes

Thistable

reportssummary

statisticsofthevariablesusedin

theem

piricalanalysis.Familyisabinary

variable

thatequals1ifthemain

shareholder

isanindividualora

family,0otherwise.Share_familyistheequityshare

heldbythefamily,whichisacontinuousmeasure

ofownership

structure.Financialinstitutionisabinary

variablethat

equals1ifthemain

shareholder

isabankorafinancialinstitution.

Capitalintensity

ismeasuredasfixed

assetsper

worker.Labourproductivityiscalculatedasvalueadded

per

worker.CorporationandConsortium

are

binary

variables

indicatingwhether

afirm

isacorporationorbelongsto

aconsortium.Leverageisdefined

asafirm

’sratiooftotalliabilitiesto

equity.Financialconcentrationiscalculatedas

theHerfindahl-Hirschmanindex

ofsharesoffinancialassetsinvestedbyafirm

.ATECO

istheItalianClassificationofEconomicActivity,whichisthenationalversionof

theEuropeannomenclature,NACE.North,South

andCentreare

binary

variablesindicatingwhether

afirm

isheadquartered

inthenorth,south

orcentreofItaly.

Pavitt’staxonomycategorizesindustrialfirm

sinto

fourtypes:traditional,scale-intensive,specializedandhigh-tech.More

detailisgiven

inSectionIV

.‘A

ssistance

EU’and’A

ssistance

non-EU’are

dummyvariablesthatequal1ifafirm

hasaccessto

assistance

from

consulates,chambersofcommerce,exportconsortia

and

other

such

institutionsforentryinto

EU

marketsandnon-EU

markets,respectively.

See

SectionIIIformore

detailaboutmeasurement.

Economica



The firms are largely located in the north of Italy (68% of the total), while 18% of thefirms are in the centre and 14% are in the south. Using Pavitt’s taxonomy (Pavitt 1984),the distribution among sectors shows the predominance of businesses operating intraditional manufacturing sectors (almost half of the sample). The portion of high-technology firms is relatively low—less than 5%. The average firm size is small tomedium (with an average of 105 employees and a median of 31). We compared thedemographic statistics for the firms in our sample with those for the pooled 1998 and1993 waves of the National Survey of Small Business Finances (NSSBF) conducted bythe US Board of Governors of the Federal Reserve System and the Small BusinessAdministration. On average, the businesses in the pooled NSSBF waves have 30employees (with a median of 6). Thus the businesses in our sample are slightly larger thanthose in the NSSBF, although they are still small or medium-sized. When compared withthe datasets used by a number of studies on family firms in the USA, which examine largepublicly traded businesses (see, for example, Villalonga and Amit 2006), our dataset thusmostly focuses on smaller and privately held firms. This reflects the structure of theItalian business sector, where the number of large publicly traded firms is relatively low.

As for firms’ legal structure, this is not reported in the early waves of the survey.When such information was unavailable, we performed web searches and obtained itfrom firms’ websites. Then we hand-matched this information with the surveys using theVAT identification number. 94% of the firms have limited liability. Among them, 57%are private limited companies (SRL: societ�a a responsabilit�a limitata) and 36% are publiclimited companies (SPA: societ�a per azioni).

To complement the survey, we employ data made available by the Bank of Italy onthe presence of banks in local markets. We use data from the Italian National StatisticsOffice (ISTAT) on the value added and population of provinces. Finally, we employ theindex of external financial dependence put forward by Rajan and Zingales (1998).

Measurement

Ownership structure The survey asks each firm to report the characteristics of the mainshareholders of the firm, such as their types and equity shares. Our measure of familyownership is a binary variable that equals 1 if the main shareholder is an individual or afamily, 0 otherwise (see the Appendix for a detailed definition of all the variables). Thedata confirm the relevance of family firms in the Italian manufacturing sector. In oursample, in 75% of the firms the main shareholder is an individual or a family; in 9% it isa bank or a financial institution; in 16% it is another manufacturing firm or a holdingcompany. The data further inform us about the alignment between ownership andcontrol and between ownership and management. In our sample, in 93% of family firmsthe family has control rights; 41% of family firms have external managers on the boardof directors.

Export The survey provides us with information about whether or not a firm exported inthe year of each survey wave, and about foreign sales if the firm exported. On average66% of the firms in the sample exported. In particular, 64% of family firms and 71% ofnon-family firms exported over the sample period. Conditional on exporting, familyfirms exported less and had a smaller share of foreign sales in total sales (42% vs. 44%).

The survey also asks the firms about the geographical area(s) where they exportedtheir products. The most popular destination is the EU-15: 62.2% of the businessesexport to the European Union (EU). As for the other markets, 24.5% of the firms export

Economica



to Russia and central-eastern Europe, 23.3% to the USA and Canada, 20.8% to Asiaexcluding China, 12.5% to Central and South America, 10.7% to Africa, 7.1% toOceania, and 6.4% to China. A cross-tabulation between firms’ sector of activity andexport decisions reveals that the majority of firms in traditional and high-tech sectorsengage in export (63.4% and 67.7%, respectively). Moreover, the propensity to export(the ratio between the number of exporters and the total number of firms) is higher in thenorth than in the centre or south.

Regarding the intensive margin of export, on average foreign sales were 1.23 millioneuros, accounting for 42% of the total sales of a firm. Only a few firms engage in FDI oroutsourcing (with the large majority of them concentrated among exporters). This is notsurprising given that the sample median firm size is 31 employees, and typically only largefirms can sustain the sizeable fixed costs associated with FDI or outsourcing.

Control variablesWe now discuss the other explanatory variables. To account for the factthat more productive, larger and more capital-intensive firms are more likely to export(see, for example, Bernard and Jensen 2004), we include labour productivity, measured asthe value added per worker, firm size (the log of total assets), and capital intensity (fixedassets per worker). We also include dummy variables indicating whether a firm is acorporation, and whether it belongs to a consortium. A consortium may allow a firm toshare the distribution network with other firms and thus reduce the cost for enteringforeign markets.

The literature suggests that the probability of export depends on a firm’s ability tocover entry costs (Das et al. 2007). For this reason, we add controls for firm financialconditions, including the leverage ratio (the firm’s ratio of total liabilities to equity) andthe index of external financial dependence proposed by Rajan and Zingales (1998), whichcaptures the different degree of dependence of industrial sectors on external sources offinance. We also include an interaction term between the leverage ratio and the index offinancial dependence to capture the possibility that the effect of external financedependence on export decisions may be stronger for more leveraged firms.6 Moreover, weinclude industry dummy variables to account for other sources of comparative advantageand for the pattern of global demand for goods. We also construct two dummy variablesequal to 1 when the firm is classified in a four- or five-digit ATECO (Italian Classificationof Economic Activity) sector, 0 otherwise. Firms that produce goods in a four- or five-digit ATECO sector are more likely to be specialized in niche markets compared to firmsproducing in a three-digit ATECO sector. In additional tests, we include further proxiesfor the fixed costs that firms can face when entering foreign markets. It is frequentlyargued that firms face relevant costs for complying with non-tariff measures. These costsinclude the difficulty of interpreting procedures and administrative requirements inforeign markets. To the extent that family firms have less sophisticated knowledge aboutinternational markets, they could find it particularly hard to comply with non-tariffbarriers. To measure such barriers, we obtained data from the United NationsConference on Trade and Development database (UNCTAD 2013) on non-tariffbarriers. We then computed the number of non-tariff barriers in each industry and eachsurvey period by summing non-tariff barriers for that industry across the broad exportmarkets identified by the survey.7 We also constructed a weighted measure where theweights are given by the foreign markets’ shares of imports from Italy.

Finally, we control for local socioeconomic conditions using province fixed effects.Alternatively, we control for province-level heterogeneity using provincial GDP growthand area dummies indicating whether a firm is headquartered in the south or centre of

Economica



Italy (the main geographical areas of Italy differ substantially in infrastructure andinstitutions). The inclusion of area dummies is also useful because the north of Italy iscloser to the EU markets where Italian firms mostly export.

IV. FAMILY FIRMS AND EXPORT

In this section, we investigate the relationship between family ownership and theextensive and intensive margins of export. In Section V, we will study how familyownership influences the internationalization process.

Baseline results

Table 2 reports the baseline results. In all the regressions, family ownership is defined as adummy variable that equals 1 if the main shareholder is an individual or a family, 0otherwise. Column (1) shows the OLS estimates of the linear probability model inequation (2). We find that after controlling for various firm characteristics and forprovince fixed effects, family firms are 3.1% more likely to export than non-family firms.Column (2) displays the probit marginal effects of equation (1). The marginal effect offamily ownership is 0.037, which is slightly higher than the OLS estimate, although thedifference is statistically insignificant.

As for the control variables, columns (1) and (2) of Table 2 illustrate that bigger firmsare significantly more likely to export. The coefficient on total assets suggests that adoubling of firm size increases the probability of exporting by more than 10%. Thecoefficient on capital intensity (defined as the ratio of total assets to the number ofemployees) is significantly negative. This probably stems from the fact that we use totalassets to measure firm size; the coefficient turns positive when firm size is measured usingthe number of employees. Being a corporation and belonging to a consortium appear toincrease the likelihood of export. Firms in industries with higher dependence on externalfinance (measured by the Rajan–Zingales index) are less likely to export. Interestingly, wefind that the negative effect of external finance dependence is significantly stronger formore leveraged firms. This is in line with the theoretical predictions that exportinginvolves high entry costs, and since entry costs must be paid up front, only firms withsufficient liquidity can cover them (Manova 2013; Chaney 2005). As noted, we alsoexperiment with inserting additional proxies of the fixed costs for entering exportmarkets. In columns (3) and (4), we add the industry-level measure of non-tariff barriers,as well as its interaction with family ownership. In line with our expectations, theestimates show that the higher non-tariff barriers are, the lower is the probability ofexporting.8 Since the measure of non-tariff barriers suffers from limitations (e.g. it doesnot account for the heterogeneous intensity of different barriers), in the remainder of theanalysis we do not include it in the regressions. Further, we find that firms that specializein a five-digit ATECO sector are significantly more likely to export, which suggests thatproducing in a niche market increases the probability of export.

In addition to influencing corporate governance, family ownership can have animpact on firm size, productivity, capital accumulation and other aspects of a firm, whichin turn affect the decision to export. Thus in columns (1) and (2) of Table 2, we havecontrolled for various firm characteristics to isolate the effect of corporate governancefrom the effect of other aspects of a firm. We experimented with dropping controls forfirm characteristics one by one, and obtained that the results on family ownership arelargely unchanged. The only exception is when we exclude firm size. In columns (5) and

Economica



TABLE2

BASELIN

EESTIM

ATES

Extensivemargin

Intensivemargin

OLS

Probit

OLS

OLS

OLS

Probit

OLS

OLS

OLS

OLS

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Family

0.031***

0.037***

0.032***

0.030***

�0.040***

�0.043***

0.043

0.041

�0.654***

�0.662***

(0.011)

(0.013)

(0.011)

(0.010)

(0.010)

(0.011)

(0.030)

(0.030)

(0.043)

(0.042)

Log(total

assets)

0.101***

0.124***

0.101***

0.101***

1.070***

1.049***

(0.007)

(0.009)

(0.007)

(0.007)

(0.015)

(0.027)

Log(capital

intensity)

�0.025***

�0.029***

�0.026***

�0.026***

0.024***

0.026***

�0.229***

�0.223***

0.351***

0.340***

(0.009)

(0.010)

(0.009)

(0.009)

(0.007)

(0.007)

(0.026)

(0.027)

(0.038)

(0.036)

Log(labour

productivity)

0.009

0.004

0.008

0.008

0.042***

0.047***

0.220***

0.219***

0.458***

0.434***

(0.012)

(0.014)

(0.011)

(0.012)

(0.012)

(0.013)

(0.033)

(0.032)

(0.063)

(0.062)

Age

0.0003

0.000

0.000

0.000

0.001***

0.002***

�0.004***

�0.004***

0.006***

0.006***

(0.0003)

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

(0.001)

(0.001)

(0.001)

(0.001)

Corporation

0.129***

0.144***

0.133***

0.126***

0.126***

0.133***

0.096

0.075

0.136

0.071

(0.027)

(0.029)

(0.026)

(0.026)

(0.026)

(0.027)

(0.066)

(0.078)

(0.108)

(0.134)

Consortium

0.055***

0.062***

0.056***

0.057***

0.052***

0.056***

0.043

0.026

�0.028

�0.063

(0.016)

(0.016)

(0.016)

(0.015)

(0.016)

(0.017)

(0.038)

(0.039)

(0.052)

(0.059)

ATECO

5-digit

0.045***

0.052***

0.046***

0.050***

0.032**

0.036**

0.033

0.023

�0.165**

�0.183***

(0.014)

(0.016)

(0.014)

(0.014)

(0.015)

(0.016)

(0.042)

(0.044)

(0.066)

(0.067)

ATECO

4-digit

0.005

0.009

0.006

0.005

�0.003

�0.002

�0.015

�0.017

�0.078

�0.077

(0.014)

(0.015)

(0.014)

(0.014)

(0.015)

(0.016)

(0.042)

(0.042)

(0.071)

(0.072)

Leverage

1.357**

3.249

1.409**

1.455**

1.218*

3.875

16.774

17.836

13.767

14.254

(0.637)

(3.473)

(0.630)

(0.632)

(0.719)

(3.439)

(15.097)

(15.384)

(27.164)

(27.094)

Leverage

*Rajan-

Zingales

index

�14.649***

�38.707***

�14.573***

�14.649***

�13.504***

�37.432**

�33.66

�29.17

�29.728

�10.74

(4.265)

(14.181)

(4.167)

(4.145)

(4.854)

(16.882)

(42.367)

(43.243)

(63.624)

(70.589)

Rajan-Zingales

index

�0.059***

� 0.063***

�0.059***

�0.059***

�0.051**

�0.044*

0.002

0.016

0.008

0.031

(0.019)

(0.024)

(0.019)

(0.019)

(0.020)

(0.023)

(0.079)

(0.082)

(0.128)

(0.129)

Economica



TABLE2

CONTIN

UED

Extensivemargin

Intensivemargin

OLS

Probit

OLS

OLS

OLS

Probit

OLS

OLS

OLS

OLS

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

NTM

�0.001**

�0.005***

(0.000)

(0.002)

NTM *Family

0.000

0.000

(0.000)

(0.002)

Inverse

Millsratio

�0.161

�0.351

(0.140)

(0.262)

Province

fixed

effects

YY

YY

YY

YY

YY

Industry

fixed

effects

YY

YY

YY

YY

YY

Survey

year

fixed

effects

YY

YY

YY

YY

YY

Observations

12,368

12,368

12,368

12,368

12,368

12,368

5876

5834

5876

5834

R2

0.181

0.183

0.183

0.133

0.679

0.679

0.212

0.212

Notes

Alltheregressionsincludeprovince,industry

andsurvey

yearfixed

effects.In

columns(1)-(6),thedependentvariableisabinary

variablethatequals1ifthefirm

exports,0

otherwise.

Incolumns(7)-(10),thedependentvariable

isthelogarithm

ofthevalueofexports.

TheRajan-Zingalesindex

(1998)capturesthedegreeofdependence

of

industrialsectors

onexternalfinance.NTM

representstheindustry-andyear-specificmeasure

ofnon-tarifftradebarriers.In

column(3),NTM

isunweighted.In

column(4),

NTM

isweightedbytheforeignmarkets’sharesofim

portsfrom

Italy.See

thenotesto

Table1andSectionIIIformore

detailaboutcontrolvariables.

Columns(2)and(6)reporttheprobitmarginaleff

ects.

Inparentheses

are

robuststandard

errors

clustered

attheprovince

level.***,**,*indicate

p<0.01,p<0.05,p<0.1,respectively.

Economica



(6), we show the results when total assets are excluded: now the estimated coefficients onfamily ownership turn negative.

Columns (7) and (8) of Table 2 display the OLS estimates of equation (3) for theintensive margin of trade. As shown in column (7), conditional on exporting, familyownership has no significant effect on the value of export. However, our estimate couldbe biased due to firms’ self-selection into the export market: we can only observe positiveforeign sales for exporters, while for non-exporters foreign sales are zero. To deal withthis selection problem, we use a Heckman-type sample selection model by adding aninverse Mills ratio to equation (3) (see Wooldridge 2002, p. 567). The inverse Mills ratiois estimated from a probit model of export participation decision on the controlsincluded in columns (1) and (2) and discussed in Section III, as well as a dummy variableindicating whether the firm distributed its products through specialized intermediaries(i.e. the excluded instrument). We find that firms that had access to specializedintermediaries for distributing products are significantly more likely to export: theestimated coefficient in the probit model is 0.362, with standard error 0.038. On the otherhand, we find no difference in foreign sales between firms that had access to thismarketing channel and those that did not. Thus the indicator of whether a firmdistributed its products through specialized intermediaries is excluded from theregression of foreign sales, which helps to identify the effect of family ownership onforeign sales. We then estimate (3) by adding the inverse Mills ratio computed using theprobit estimates. As reported in column (8), the estimated effect of family ownership onforeign sales is 0.041, which is almost identical to the estimate reported in column (7). Onthe other hand, the inverse Mills ratio is not statistically significant, which suggests thatthe null hypothesis of no sample selection bias cannot be rejected.9

In columns (9) and (10) of Table 2, we exclude total assets and find that family firmsexport significantly less than non-family firms. However, as discussed above, since familyfirms are significantly smaller than non-family firms, and larger firms export substantiallymore than smaller firms, excluding controls for firm size may prevent us disentangling theeffect of corporate governance from that of firm size on export decisions. Therefore in thefollowing we always control for firm size and other firm characteristics.

Family control, management and export

We expect that family owners better internalize the long-run benefits of export activitieswhen they have control rights, that is, when firm ownership is aligned with control. Bycontrast, the alignment of ownership with management has an ambiguous impact ex ante.On the one hand, when families are directly involved in management, agency problemsbetween owners and managers should be milder, and this could benefit internationalexpansion. On the other hand, families may lack competencies and skills crucial forinternationalization. Thus hiring external professional managers could benefit export.

In Table 3, we investigate the impact of family control and management on exportdecisions. For this purpose, we first examine the consequences of separation betweenownership and control, and then turn to the effects of separation between ownership andmanagement. In column (1), we distinguish family firms in which the main shareholderhas control over the firm and family firms in which the main shareholder has no controlrights (thus the omitted group is firms in which the main shareholder is not a family). Wefind that firms in which families have control rights have a significantly higher probabilityof exporting than their non-family counterparts. However, those in which families do notretain control rights are not significantly different from non-family counterparts in their

Economica



export participation decisions. This result shows the alignment of families’ control rightswith their cash-flow rights benefits export. Separating control rights from cash-flowrights increases agency costs, and in particular it can distort project selection (Shleiferand Vishny 1997). Claessens et al. (2002) argue that the agency problems due to themisbehaviour of a large shareholder are more severe when there is a divergence betweencontrol rights and cash-flow rights, because the large shareholder does not fullyinternalize the consequences of his decisions. Grossman and Hart (1988) show thatseparating ownership from control can lower shareholders’ value.

Next, we study the effect of separation between ownership and management. The lasttwo waves of the Capitalia survey ask each firm whether it has external managers on itsboard.10 In column (2) of Table 3, we restrict the analysis to these two waves andexamine the impact that outside managers have on the extensive margin of export. Wefind that family firms with external managers are 4.5% more likely than non-family firmsto export, whereas those without external managers are not significantly different fromnon-family firms in the probability of export participation. This result corroborates theidea that family ownership is especially beneficial to export when families rely on theskills and competence of external managers (The Economist 2012).

In columns (3) and (4) of Table 3, we examine the role of family control andmanagement in determining the value of export. The signs of the coefficients areconsistent with those for export participation decisions, although the estimatedcoefficient turns out to be significant only for firms with external managers.

Disentangling the ownership–export links

In what follows, we study the channels through which family ownership affects export.As noted in Section II (see the discussion of the hypotheses), the possible channels of

TABLE 3FAMILY CONTROL AND MANAGEMENT

Extensive margin Intensive margin

(1) (2) (3) (4)

Family with control 0.033*** 0.043

(0.011) (0.030)Family without control 0.008 0.039

(0.020) (0.049)Family with external managers 0.045*** 0.085**

(0.010) (0.037)Family without external managers �0.016 0.017

(0.015) (0.036)

Observations 12,368 8600 5876 49540.181 0.197 0.679 0.692

NotesAll of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index and the interaction of leverageand the Rajan-Zingales index. In columns (1) and (2), the dependent variable is a binary variable that equals 1 ifthe firm exports, 0 otherwise. In columns (3) and (4), the dependent variable is the logarithm of the value ofexports. See the notes to Table 1 and Section III for more detail about the control variables.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



influence are the long-termism of families, on the positive side, and the risk aversion, thelack of competence and the narrowness of family owners, on the negative side. Thedataset provides rich information on firm and industry characteristics that are suitablefor isolating these channels. To ease the interpretation of the results, in Table 4 wesummarize these channels and the way we test for their presence.

Family firms and long-termism Capturing the possible long-termism of family owners isadmittedly a difficult task. As it is often the case in other studies, the dataset does notinclude a precise proxy for the length of firms’ horizon. The literature on firms’ survivalconsistently finds that older firms have a higher probability of survival and hence alonger-term horizon. In Table 5, columns (1) and (2), we split the sample based on firmage. The results show that family ownership has a positive effect on export participationfor older firms, i.e. those with more than 21 years of operation (the sample median). Thecoefficient on family ownership is 0.034 and significant at the 1% level. By contrast,although family ownership has a positive coefficient for younger firms, the effect isstatistically insignificant. Thus the estimated positive effect of family ownership onexport can be picking up the fact that family owners have a relatively bigger incentive toundertake long-term export projects in old firms.11 The reader might, however, beconcerned that the stronger effect for older firms stems somewhat mechanically from thefact that entering foreign markets requires time. We then turn to a more precise proxy forthe length of firms’ horizon.

The literature stresses that measures of shareholders’ stock turnover provide usefulinformation on the length of shareholders’ horizon (see, for example, Roe (2013) for areview). Based on this intuition, we study whether changes in ownership structure thatshorten the decision horizon of family owners trigger changes in export decisions—entryinto new markets or exit from old markets. The survey provides information on whetherin the years prior to the survey financial institutions subscribed new shares of the firmand on whether the firm intended to go public in the following years. The decision of afamily owner to sell shares to a financial institution can be interpreted as a sign that the

TABLE 4FAMILY OWNERSHIP AND EXPORT BEHAVIOUR-ROADMAP OF TESTS

Channel Family ownership and export Family ownership and(extensive and intensive margin) internationalization process

Long-termism Firm age Pace of internationalization

(export persistence)Financial institutionssubscribing shares

and intention to go publicRisk aversion Financial diversification Collaboration with other parties

(foreign agents)Lack of competence

and knowledge

Presence of external managers Number of markets

Industry sophistication Pace of internationalization (sequentialexporting)

Export market sophistication Collaboration with other parties

(domestic agents)Skilled workforce

Excess of narrowness Niche markets

Economica



TABLE5

FAMIL

YFIR

MSANDLONG-T

ERMISM

Extensivemargin

Intensivemargin

Enteringnew

markets

Exitingfrom

old

markets

Old

Young

Old

Young

(2003-6)

(2003-6)

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Family

0.034***

0.026

0.070*

0.028

0.055*

0.058**

0.054*

�0.033

�0.04

�0.041

(0.011)

(0.016)

(0.039)

(0.041)

(0.028)

(0.028)

(0.029)

(0.047)

(0.036)

(0.035)

Fin.inst.

subscribe

(2001-3)

�0.192***

0.228**

(0.054)

(0.088)

Intentionto

gopublic

(2001-3)

�0.141

�0.07

(0.136)

(0.091)

Number

of

exportmarkets

(2003)

�0.056***

�0.056***

0.174***

0.171***

(0.008)

(0.007)

(0.007)

(0.007)

Observations

6754

5614

3437

2439

778

776

745

778

776

745

R2

0.209

0.175

0.690

0.682

0.127

0.131

0.136

0.199

0.201

0.217

Notes

AlloftheOLSregressionsincludeprovince,industry

andsurvey

yearfixed

effects,andcontrolforfirm

characteristics

includinglog(totalassets),log(capitalintensity),log

(labourproductivity),age,

corporation,consortium,ATECO

5-digit,ATECO

4-digit,leverage,

theRajan-Zingalesindex,andtheinteractionofleverageandtheRajan-

Zingalesindex.In

columns(1)and(2),thedependentvariableisabinary


exports,0otherwise.In

columns(3)and(4),thedependentvariable

isthelogarithm

ofthevalueofexports.See

thenotesto

Table1andSectionIIIformore

detailaboutmeasurementofthecontrolvariables.

Incolumns(5)-(10),weexaminewhether

changes

inownership

structure

could

trigger

changes

inexport

decisionsin

term

sofentryinto

new

marketsorexitfrom

old

markets.Changes

inownership

structure

are

capturedbyvariablesindicatingwhether

financialinstitutionssubscribed

new

sharesofafirm

in2001-3,orwhether

afirm

has

anintentionto

gopublicin

2001-3.

Inparentheses

are

robuststandard

errors

clustered

attheprovince

level.

***,**,*indicate

p<0.01,p<0.05,p<0.1,respectively.

Economica



family owner is progressively reducing his involvement in the firm. Indeed, financialinstitutions that subscribe shares often consist of equity funds and other institutionswhose main objective is to help the firm to progressively make the transition from thefamily owner to a broader pool of shareholders and eventually be listed on the stockmarket. Similarly, the intention to go public is likely to signal a plan of the family ownerto progressively reduce his involvement in the firm. Thus both variables indicate thatfamily owners have a shorter horizon in their decisions.

In practice, we exploit the panel dimension of our dataset focusing on the last twowaves of the survey (2001–3 and 2003–6), and examine the impact of changes inownership structure on market switching between these two periods. In our sample, 36%of exporters sell to a single foreign market (93% of which choose the EU market),another 27% of exporters serve two foreign markets, and the remaining 37% export toat least three markets. The independent variables are measured using the 2001–3 wave ofthe survey. The results are reported in columns (5)–(10) of Table 5. As shown in column(5), family firms are more likely to expand into new export markets. The coefficient onfamily ownership is 0.05 and is significant at the 10% level. In columns (6) and (7), westudy the impact of changes in ownership structure that tend to shorten the horizon offamily owners. Column (6) shows that firms in which financial institutions subscribednew shares in 2001–3 are 19.2% less likely to expand into new markets in 2003–6.Column (7) reports that firms with an intention to go public in 2001–3 are 14.1% lesslikely to expand into new markets in 2003–6, although the effect is not statisticallysignificant. Both of these results support the hypothesis that the positive effect of familyownership on export is driven by the long-term horizon of family owners. In columns(8)–(10), we also examine the decision about exiting from old markets. The coefficientson family ownership are negative, indicating that family firms are less likely to exit fromold markets. In addition, firms in which financial institutions subscribed new shares in2001–3 are significantly more likely to exit from existing markets in 2003–6.12

Family firms and risk aversion A channel through which family ownership can negativelyaffect the decision to export is family owners’ risk aversion. In Table 6, we split thesample based on a firm’s financial diversification. In fact, if the financial portfolio of afamily firm is not diversified, then the family owner could be reluctant to engage in riskyprojects such as export (Bolton and von Thadden 1998). Our measure of financialdiversification is based on a question asking firms about the allocation of their financialinvestments among equity participation in Italian companies, equity participation inforeign companies, short-term Italian bonds, medium- and long-term Italian bonds,foreign bonds, other financial instruments. The rate of response to this question is about35%. We measure the diversification of firms’ financial portfolio using the Herfindahl–Hirschman index of the various asset shares. Firms with an index of 1 are considered asless diversified, while firms with an index lower than 1 are considered as more diversified.We report the results separately for these two groups of firms in columns (1) and (2) ofTable 6. Interestingly, although family ownership has a positive effect on exportparticipation for both groups, the effect is statistically significant only for less diversifiedfirms. This suggests that risk aversion, due to lack of financial diversification, does notdeter families from promoting export.

As noted in Section II when discussing the theoretical predictions, an alternativehypothesis is that family firms use export to foreign markets as a way to diversify therisks incurred in domestic markets. We expect such an incentive to be stronger for firmsthat hold their whole wealth invested in domestic assets and hence do not already

Economica



diversify their financial portfolio by holding foreign assets. Thus in columns (3) and (4) ofTable 6 we re-estimate the above regressions by restricting attention to firms that declarethat they hold all their wealth in Italian assets (bonds or equity). The results carrythrough, which suggests that export not only is not inhibited by family owners’ riskaversion but also is actually perceived by family owners as a way to diversify risk.

A possible concern with the tests conducted using information on financialdiversification is that we observe the degree of diversification of the financial portfolio ofthe firm, but not of the portfolio of the main shareholder. As suggested by Onida (2004),for example, for small and medium-sized Italian firms the distinction between the twoportfolios is very often blurred. Thus we check whether the results in columns (1) and (2)of Table 6 carry through for smaller firms. Columns (5) and (6) report the results forfirms with total assets less than 17.8 million euros (the 80th percentile of the sample). Theresults suggest that family ownership has no significant effect on export participation forsmaller and less diversified firms.

Family firms and skills As noted in Section III on the theoretical predictions, besides riskaversion, a second channel through which family ownership could discourage export isthe shortage of skills of family owners (as induced, for example, by the tendency of

TABLE 6FAMILY FIRMS AND RISK AVERSION

All firms Domestic assets

Totalassets< 2 \euro17.8

million

Fin.conc. = 1

Fin.conc. < 1

Fin.conc. = 1

Fin.conc. < 1

Fin.conc. = 1

Fin.conc. < 1

(1) (2) (3) (4) (5) (6)

Panel A: Extensive margin

Family 0.043** 0.011 0.051*** 0.011 0.036 0.014(0.017) (0.027) (0.019) (0.028) (0.022) (0.053)

Observations 3307 887 3043 880 2390 411

R2 0.210 0.303 0.207 0.303 0.223 0.373

Panel B: Intensive margin

Family 0.062 0.056 0.070 0.055 0.043 0.004(0.055) (0.091) (0.065) (0.093) (0.073) (0.221)

Observations 1797 527 1617 520 1148 200

R2 0.669 0.791 0.656 0.793 0.433 0.699

NotesAll of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index, and the interaction ofleverage and the Rajan-Zingales index. In panel A the dependent variable is a binary variable that equals 1 if thefirm exports, 0 otherwise. In panel B the dependent variable is the logarithm of the value of exports. See thenotes to Table 1 and Section III for more detail about measurement of the control variables.Financial concentration is calculated as the Herfindahl-Hirschman index of the shares of financial assetsinvested by firms. Firms with an index of 1 are considered as less diversified, and firms with an index less than 1are considered as more diversified. In columns (3) and (4), the sample includes firms that hold all their wealth inItalian assets. Columns (5) and (6) report the results for firms with total assets less than 17.8 million euros.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



family owners to transmit ownership and control to their descendants rather thaninvolving skilled outsiders). In particular, family owners could lack the knowledgenecessary to internationalize their business. The results on the benefit of externalmanagers in Table 3 suggest that the lack of competence of family owners could indeedhinder export activities. We perform various tests to further isolate this channel. First, weexamine whether the positive impact of family owners on export is weaker in sectors andmarkets characterized by a higher level of sophistication and complexity.

In Table 7, we split the sample into four types of industries based on the Pavitttaxonomy (Pavitt 1984). The four types of industries can be ranked according to the levelof sophistication and technological content of production.13

• Traditional or supplier-dominated sectors (such as textiles, food, tobacco, paper) arecharacterized by highly standardized processes and established technologies. Most oftheir innovations are acquired from external sources.

TABLE 7FAMILY FIRMS AND SKILLS

IndustriesNon-EU

All Traditional Scale-intensive Specialized High-tech markets(1) (2) (3) (4) (5) (6)

Panel A: Without control for skillsFamily 0.031*** 0.038* 0.049** 0.017 �0.090** 0.000

(0.011) (0.020) (0.020) (0.016) (0.041) (0.011)Observations 12,368 6064 2554 3183 546 8168

R2 0.181 0.171 0.262 0.164 0.299 0.161

Panel B: With control for skillsFamily 0.002 0.012 0.012 0.004 �0.116** 0.009

(0.013) (0.021) (0.037) (0.020) (0.056) (0.016)Family

* Skilledworkforce

0.049*** 0.047** 0.065* 0.015 0.058 0.048**

(0.014) (0.018) (0.037) (0.019) (0.078) (0.018)

Skilled

workforce

0.009 0.046*** �0.027 �0.001 �0.049 0.018

(0.012) (0.017) (0.037) (0.019) (0.055) (0.016)Observations 10,367 5121 2160 2634 452 6370R2 0.191 0.185 0.277 0.181 0.330 0.328

NotesAll of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index, and the interaction ofleverage and the Rajan-Zingales index. See the notes to Table 1 and Section III for more detail aboutmeasurement of the control variables.In columns (2)-(5), firms are respectively categorized into traditional, scale-intensive, specialized and high-techbased on the Pavitt taxonomy. See Section IV for more detail. In columns (1)-(5), the dependent variable is abinary variable that equals 1 if the firm exports, 0 otherwise. In column (6), the dependent variable is a dummythat equals 1 if the firm exports to non-EU markets (in addition to EU markets), and 0 if the firm exports onlyto EU markets.In panel B, ’Skilled workforce’ is a dummy variable that takes the value 1 if the percentage of employees with atleast a secondary school degree exceeds the industry median, and 0 otherwise. It is used to capture the degree ofemployees’ education and training.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



• Scale-intensive sectors (e.g. iron, glass, car manufacturing, metal products) arecharacterized by a level of sophistication somewhat higher than traditional industries.They typically import innovations from external sources but sometimes also developthem internally.

• Specialized industries (such as mechanical machinery, electronics, telecommunicationsappliances) can be positioned on an even higher level of sophistication andtechnological content. Firms in these industries typically produce machinery andsoftware for other industries; their innovations often arise from complex interactionswith the users of their products.

• Finally, high-tech industries (e.g. chemical and bioengineering) feature the highestdegree of sophistication, technological content, and R&D intensity.

The results in columns (2)–(5) of panel A of Table 7 show that the positive impact offamily ownership on export participation is stronger the lower the degree ofsophistication and technological content in the industry. Specifically, the coefficients onfamily ownership are 0.038 and 0.049 for the traditional and scale-intensive sectors,respectively. By contrast, the coefficient on family ownership for the specialized sectors ismuch smaller (0.017) and statistically insignificant. And for the high-tech sectors, familyfirms are significantly less likely to export: the probability of exporting is 9% lower forfamily firms. These results thus support the hypothesis that family owners are beneficialto export especially in sectors in which internationalization does not require highcompetence and skills.14

An additional way to test for an impact of the limited competencies of family ownersis to check whether the effect of family ownership on export is weaker for businesses witha less educated and trained workforce. An unskilled workforce could particularly hinderthe internationalization effort of family firms because it could compound the limitedcompetencies of family owners. Moreover, family owners could have to monitor lesscompetent employees more intensely because such employees face lower costs if they arecaught shirking, and also tend to have lower motivation (Rebitzer 1995). To capture thelevel of employees’ education and training, we construct a dummy that takes the value 1if the percentage of employees with at least a secondary school degree exceeds theindustry median, 0 otherwise. In panel B of Table 7, we then insert this dummy and itsinteraction with the dummy for family ownership. The results indicate that the positiveeffect of family ownership on export is stronger for firms with a more skilled workforce,suggesting that for such firms the detrimental effect of the limited skills of family ownerscould be attenuated.15 To further probe this point, in column (6) we test whether theinteraction of family ownership with skilled workforce has a stronger effect on entry intodifficult non-EU markets (conditional on exporting to the EU). The results support thishypothesis, which is consistent with the idea that skills and knowledge are especiallyuseful for entry into unfamiliar markets.

Family firms’ narrowness The theoretical literature predicts that besides risk aversion andlack of skills, a third disadvantage of family firms can be their tendency to defendconservatively their position in niche markets. In columns (1) and (2) of Table 8, westudy whether family firms underperform in niche markets. We split the sampleaccording to whether or not firms are classified in a five- or four-digit ATECO sector. Afirm active in a five- or four-digit ATECO sector is more specialized than a firm thatproduces in a three-digit ATECO sector. The estimates reveal that the effect of familyownership on export participation is positive and significant for firms that produce in

Economica



niche markets, but not for firms that span their production in multiple sectors. Thus wefind no evidence of an excess narrowness of family firms.

Further discussion When reading through the results on the channels of influence offamily ownership on export, the reader may have wondered whether such channels canbe expected to be stronger for founding families or for subsequent generations of familymembers (see, for example, Ellul et al. (2010) for the importance of distinguishingbetween founders and subsequent generations). Information on whether a shareholderwas the firm founder is rarely available, so very few studies can identify founding families(for an exception, see, for example, Villalonga and Amit 2006). Our dataset features thesame limitation and does not allow us to separate founding families from other familyowners. It is, however, interesting to discuss how founders could differ from subsequentgenerations in export activities. In the above-mentioned cultural view of family firms(Bertrand and Schoar 2006), founders are sometimes perceived to be more authoritarianand more inclined to have close ties to their business than subsequent generations. Thisimplies that they can have stronger commitment to the firm and longer investmenthorizons than non-founding families. But it also implies that they could have more fearof losing control over the firm and hence be more risk averse. By contrast, subsequentgenerations should suffer less from a problem of excess attachment and be less concernedabout losing control over the firm. Indeed, subsequent generations may even perceiveinternationalization as a way of creating space for themselves inside the business (‘born-again global companies’). Finally, it is sometimes stressed that subsequent generationsmay lack the skills of founders and that this may hinder internationalization.

V. FAMILY FIRMS AND THE INTERNATIONALIZATION PROCESS

Family ownership can influence not only a firm’s decision to export and the volume ofexport but also the firm’s internationalization process. This includes the number of

TABLE 8FAMILY FIRMS’ NARROWNESS

Extensive margin Intensive margin

5- or 4-digit 3-digit 5- or 4-digit 3-digit

ATECO (1) ATECO (2) ATECO (3) ATECO (4)

Family 0.033*** 0.023 0.032 0.040(0.012) (0.019) (0.040) (0.070)

Observations 8985 3383 4161 1715R2 0.196 0.202 0.675 0.719

NotesAll of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index, and the interaction ofleverage and the Rajan-Zingales index. In columns (1) and (2), the dependent variable is a binary variable thatequals 1 if the firm exports, 0 otherwise. In columns (3) and (4), the dependent variable is the logarithm of thevalue of exports. See the notes to Table 1 and Section III for more detail about measurement of the controlvariables.A firm that is active in a 5- or 4-digit ATECO industry is more specialized than a firm that produces in a 3-digitATECO industry.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



foreign markets that a firm chooses to serve. It also includes the pace at which a firmenters foreign markets. For example, it is frequently observed that firms engage in‘sequential exporting’, that is, they start by exporting small quantities or entering ‘easy’,close-by markets and subsequently expand their export volume and enter less familiarmarkets (see, for example, Eaton et al. 2008; Albornoz et al. 2012).16 Firms could alsoenter a foreign market and remain there permanently, or instead exit from that marketafter a short period, possibly switching to new export destinations or simply giving upexport. Indeed, various studies find that after entering a foreign market, many firms ceasetheir export to that market in less than a year (Eaton et al. 2008).

Finally, the internationalization of a firm can entail different degrees of collaborationwith foreign intermediaries, distributors and firms, or with domestic institutions thatassist exporters in foreign countries, such as chambers of commerce, export consortiaand consulates.

Export markets

Let us first consider the decision on the set of foreign markets to serve. Several empiricalstudies find that most exporting firms sell to only one foreign market (Eaton et al. 2008).If family owners have limited skills and competencies, then they could find it particularlycostly to export to multiple markets. We first examine the choice between exporting to asingle market and selling to the domestic market only (see column (1) of Table 9). Thepositive effect of family ownership is confirmed. The coefficient on family ownership isequal to 0.041 and significant at the 1% level. In column (2) of Table 9, we present theresults for the choice between exporting to multiple markets and selling to the domesticmarket only (the estimation excludes firms that export to a single market). Enteringmultiple markets may allow exporters to diversify demand risk, but in principle mayinvolve extra entry costs in terms of acquisition of knowledge and competencies. We findthat family ownership still has a statistically positive effect on export participation, butthe magnitude of the effect is smaller than that for single-market exporters. Therefore,consistent with arguments that stress the limited competencies of family firms, familyownership has a stronger impact on single-market exporters than on multiple-marketexporters.

Pace of internationalization

As detailed in Section III when discussing the theoretical predictions, besides the set ofexport markets, family ownership can influence the pace of the internationalizationprocess. Various studies find that after entering a foreign market, several firms abandonthat export market after a short period. Thus a first interesting aspect that we canexamine is whether family firms tend to be more persistent exporters than non-familyones. To investigate this point, we exploit the panel dimension of the dataset focusing onthe last two survey waves (2001–3 and 2003–6). We then construct a dummy equal to 1 ifin the 2001–3 survey a firm reports that it has entered a new foreign market and in thefollowing survey wave the firm reports that it has continued to export to that market.17

We regress this dummy on family ownership and on controls. The results are gathered incolumns (3) and (4) of Table 9. While the estimates for the whole set of firms do notreveal a higher export persistence of family firms, when we allow the coefficient on familyownership to vary according to the type of industry, we find that family firms operatingin high-tech industries are significantly more persistent exporters (see column (4)). This

Economica



may suggest that family firms sustain the high fixed costs associated with starting toexport a sophisticated (high-tech) product to a foreign market only when, because oftheir long-term horizon, they plan to remain in that market. To further probe this point,we also test whether family firms tend to be more persistent exporters when they employa more skilled workforce, which again could signal that they export relatively

TABLE 9EXPORT MARKETS AND PACE OF INTERNATIONALIZATION

Export markets Pace of internationalization

One

market

Multiple

marketsExport persistence

Export

growth

Non-EU

markets Oceania(1) (2) (3) (4) (5) (6) (7)

Panel A: Without control for skills

Family 0.041*** 0.029** 0.038 0.015 0.037 0.000 0.014*(0.015) (0.013) (0.035) (0.041) (0.051) (0.011) (0.007)

Family

* High-techindustry

0.378**

(0.160)

High-techindustry

�0.395***(0.124)

Observations 6445 9818 778 758 674 8168 8168R2 0.160 0.245 0.163 0.173 0.174 0.161 0.071

Panel B: With control for skillsFamily 0.031** �0.007 �0.115 �0.216** 0.009 �0.002

(0.015) (0.016) (0.083) (0.105) (0.016) (0.011)

Family* Skilledworkforce

0.023 0.055*** 0.302* 0.582** 0.048** 0.034*(0.026) (0.015) (0.157) (0.242) (0.018) (0.017)

Skilledworkforce

0.019 0.011 �0.220* �0.357* 0.018 �0.025*(0.021) (0.015) (0.132) (0.193) (0.016) (0.015)

Observations 6433 8385 678 600 6370 8156

R2 0.161 0.248 0.200 0.208 0.328 0.071

Notes All of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index, and the interaction ofleverage and the Rajan-Zingales index. See the notes to Table 1 and Section III for more detail aboutmeasurement of the control variables.In column (1), the dependent variable is a dummy that equals 1 if the firm exports to one market only, and 0 ifthe firm sells to the domestic market only. In column (2), the dependent variable is a dummy that equals 1 if thefirm exports to multiple markets, and zero if the firm sells to the domestic market only. In columns (3) and (4),the dependent variable is a dummy that equals 1 if, after entering a foreign market, the firm continues to exportto that market in the following survey wave, and 0 otherwise. In column (5), the dependent variable is a dummythat equals 1 if in a given survey a firm reports it has entered a new foreign market and in the following survey itreports an increase in its export volume, 0 otherwise. In column (6), the dependent variable is a dummy thatequals 1 if the firm exports to non-EU markets (in addition to EU markets), and 0 if the firm exports only to EUmarkets. In column (7), the dependent variable is a dummy that equals 1 if the firm exports to Oceania (inaddition to EUmarkets), and 0 if the firm exports only to EUmarkets. See Section V for more detail.In panel B, ’Skilled workforce’ is a dummy variable that takes the value 1 if the percentage of employees with atleast a secondary school degree exceeds the industry median, and 0 otherwise. It is used to capture the degree ofemployees’ education and training.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



sophisticated products. Consistent with the previous findings, family firms with moreskilled employees tend to be more persistent exporters (see column (3) in panel B ofTable 9).

A second relevant aspect in the analysis of firms’ pace of internationalization iswhether firms internationalize their activities by first ‘testing the ground’ (see, forexample, Eaton et al. 2008; Albornoz et al. 2012). For example, firms could choose toinitially export a relatively small volume and later expand their export. To test whetherfamily firms especially exhibit this export behaviour, again using the last two surveywaves (2001–3 and 2003–6), we construct a dummy equal to 1 if in the 2001–3 survey afirm reports that it has entered a new foreign market and in the following survey wave itreports an increase in its export volume. The estimates are in column (5) of Table 9.Although overall family firms do not differ significantly from non-family firms (panel A),we find that family firms with a higher share of skilled employees are significantly morelikely to experience sustained export growth after entering a new foreign market, asshown in panel B. Firms could also test the ground by using familiar, EU markets as astepping stone toward less familiar, non-EU markets. In column (6) of Table 9, we thenlook at how family ownership affects the decision whether or not to enter non-EUmarkets by firms that have already exported to the EU. The estimate shows that amongfirms that have already exported to the EU, family ownership does not seem to promotethe entry into a second foreign market. However, interestingly, when we examine theeffect of family ownership across various destinations, family firms that have alreadyexported to the EU turn out to be 1.4% more likely to export to Oceania (a very distantmarket) than their non-family counterparts (see column (7)).18

Overall, the results thus suggest that, possibly because of their limited skills andcompetencies, family firms exhibit some tendency to enter foreign markets in aprogressive way. At the same time, when they choose to enter a foreign market, they doso with the intention to stay in that market.

Export and collaboration with other parties

A third relevant aspect of firms’ internationalization process is their degree ofcollaboration with other agents. A first group of such agents consists of foreigndistributors, intermediaries and firms. Because of their limited competencies andknowledge, family firms could benefit from collaborating with such agents, but as noted,they could also fear losing control and independence. In Table 10, we test whether familyownership has any influence on activities of cooperation with foreign agents in exportmarkets. We find no evidence of an impact on collaborations with foreign distributors orintermediaries, or on joint ventures with foreign firms (see panel A). These results suggestthat the risk aversion associated with the fear of losing control could inhibit suchcollaborations, in spite of their potential advantages.

A second way in which firms can try to overcome limited competencies consists ofseeking help from domestic institutions, such as export consortia, consulates andchambers of commerce. For example, export consortia can pool knowledge on foreignmarkets and hence facilitate firms’ export. And chambers of commerce allegedly play animportant role in providing information to exporters, especially when these enter non-industrialized countries. Unlike for collaborations with foreign intermediaries and firms,instead, we do not expect family owners to fear losing control over firms when obtainingassistance from such domestic agents. The survey asks the firms whether they had accessto assistance from consulates, chambers of commerce, export consortia and other such

Economica



institutions in export markets. We then re-estimated the baseline regressions inserting adummy equal to 1 if the firm declares it had access to such assistance, as well as aninteraction term between this dummy and family ownership. The results suggest that theexport of family firms benefits more from the assistance of these agents than the export ofnon-family firms (see panel B of Table 10). Moreover, this benefit appears to be especiallystrong when family firms receive assistance in non-EU markets. This corroborates thehypothesis that such collaborations are especially beneficial for family firms because theyalleviate their lack of knowledge about foreign markets without interfering with theirownership and control.19

VI. ENDOGENEITY OF FAMILY OWNERSHIP

The OLS and probit estimates might be affected by reverse causality problems. One maywonder whether internationalization triggers changes in firm ownership structure.

TABLE 10EXPORT AND COLLABORATION WITH OTHER PARTIES

Local distributors Intermediaries Joint ventures(1) (2) (3)

Panel A: Collaboration with foreign partiesFamily 0.005 �0.01 �0.001

(0.027) (0.014) (0.008)Observations 2235 2097 4113R2 0.108 0.112 0.615

EU Non-EU Chambers of commerce(1) (2) (3)

Panel B: Assistance of domestic partiesFamily 0.037*** 0.034*** 0.037***

(0.012) (0.012) (0.012)

Family * Assistance 0.037 0.101*** 0.044*(0.030) (0.029) (0.023)

Assistance 0.190*** 0.089*** 0.176***(0.022) (0.023) (0.021)

Observations 9900 9900 9971R2 0.205 0.200 0.208

NotesAll of the OLS regressions include province, industry and survey year fixed effects, and control for firmcharacteristics including log(total assets), log(capital intensity), log(labour productivity), age, corporation,consortium, ATECO 5-digit, ATECO 4-digit, leverage, the Rajan-Zingales index, and the interaction ofleverage and the Rajan-Zingales index. See the notes to Table 1 and Section III for more detail about the controlvariables.In panel A, the dependent variable is a dummy that equals 1 if a firm collaborates with local distributors orforeign intermediaries, or forms joint ventures with foreign firms, and 0 otherwise.In panel B, the dependent variable is a dummy that equals 1 if a firm exports, and 0 otherwise. In columns (1)and (2), ’Assistance’ is a dummy that equals 1 if a firm has access to assistance from consulates, chambers ofcommerce, export consortia and other such institutions for entry into EU markets and non-EU markets,respectively. In column (3), ’Assistance’ is a dummy that equals 1 if a firm has access to assistance fromchambers of commerce in export markets.See Section V for more detail.In parentheses are robust standard errors clustered at the province level.***, **, * indicate p < 0.01, p < 0.05, p < 0.1, respectively.

Economica



However, ownership structure is stable for 70–80% of the family firms in our sample.Another concern is that although our empirical specification controls for various factorsthat may affect firm export decisions, it is possible that there exist some unobservedfactors that simultaneously affect ownership structure and firm export. The direction ofthis bias is unclear a priori. To deal with the possible endogeneity of family ownership, weconstruct instruments that we expect to directly influence firm ownership structure, butto have no direct impact on firm export.

Instruments for ownership structure

Following Guiso et al. (2003, 2004) and Herrera and Minetti (2007), our instrument setconsists of provincial data on the number of savings banks and the number ofcooperative banks in 1936 (per 1000 inhabitants). To understand the choice of theseinstruments, we need to discuss the Italian banking regulation. In 1936 the ComitatoInterministeriale per il Credito e il Risparmio (CICR) enacted strict norms for the entryof banks into local credit markets: from 1938 each credit institution could open branchesonly in an area of competence (one or multiple provinces) determined on the basis of itspresence in 1936. Banks were also required to shut down branches outside their area ofcompetence. While the regulatory prescriptions were uniform across Italy, theconstrictiveness of regulation varied across provinces and depended on the relativeimportance of different types of banks in the local market in 1936. For example, savingsbanks were less constrained by the regulation, while cooperative banks were moreconstrained. Thus in provinces with a higher share of savings banks, access to externalfunds was easier. Guiso et al. (2003, 2004) demonstrate empirically that the 1936regulation had a profound impact on the local supply of banking services (creation andlocation of new branches) and hence on the ability to obtain credit.

We expect that the 1936 regulation had a long-lasting impact on firms’ ownershipstructure, leading to substantial variation in the incidence of family ownership acrossItalian provinces. The key channel though which this could have occurred in the Italiancontext is the effect of local financial market development on ownership transfers andhence on the intergenerational transmission of ownership in family firms (‘dynasticmanagement’). When restrictions on the local supply of credit are more severe, it can bemore difficult for potential acquirers to obtain the liquidity necessary to purchase sharesof firms. Caselli and Gennaioli (2013) demonstrate theoretically that credit marketfrictions prevent investors from borrowing and acquiring firms’ equity. Caselli andGennaioli (2013) also show that less developed credit markets increase the incentive ofuntalented heirs to retain ownership of their firms rather than transferring it to newowners. Thus one should observe that the inclusion of new shareholders and the transferof ownership are more frequent when there are fewer credit frictions.20

The arguments of Caselli and Gennaioli (2013) are particularly relevant for a countrylike Italy. Several studies on the Italian business sector maintain that the lack of banks’support hinders the functioning of the market for corporate control (for a detaileddiscussion, see, for example, Onida 2004; Barca et al. 1997). Indeed, Barca et al. (1997)provide evidence from a survey that shows that many Italian banks have remained toounsophisticated to actively participate in the market for corporate control. This lack ofsupport manifests itself in lack of funding for acquisitions of firms’ shares and also inlack of advice and information provision to firms interested in ownership transfers. Forexample, several firms claim that banks do not provide help for searching buyers, carryout evaluations of firms or supply fiscal or legal advice for bargaining procedures. To

Economica



make an example of the expected effect of the regulatory restrictions, suppose that thepurchase of the shares of a firm involves the financing of potential buyers outside theprovince. A local bank constrained by the regulation could have not been accustomed toserving clients outside the province and hence could have been reluctant to finance suchbuyers.

A second channel through which the limited local financial development could haveinhibited the transfer of ownership is the lack of signals from the credit market topotential shareholders. For example, Shockley and Thakor (1992) find that the existenceor renewal of a loan is a positive signal to potential shareholders. For this reason,regulatory restrictions on the local supply of loans may affect a firm’s ability to issue newequity. The mechanisms discussed imply that in provinces where the regulation wastighter, a family owner could have been forced or induced to retain the main share of thefirm with a higher probability.

The ownership structure is a highly persistent firm characteristic, and in Italy itspersistence is very pronounced (Bianco 2003; Bianchi and Bianco 2008). Thus we expectthat the constrictiveness of the regulation in a province, as determined by the relativeimportance of the different types of bank in the province, shaped firms’ ownershipstructure during the decades in which it was in place, and that this impact persisted forseveral years after the deregulation at the end of the 1980s. We then expect the 1936regulation to be correlated with the current ownership structure. On the other hand, asshown by Guiso et al. (2003, 2004), the distribution of types of banks across provinces in1936, and hence the constrictiveness of regulation in a province, stemmed from ‘historicalaccident’ and in particular reflected the interaction between previous waves of bankcreation and the history of Italian unification.21 Moreover, the different limits ondifferent types of bank stemmed from different connections of the various types of bankwith the Fascist regime, and thus were not correlated with structural characteristics of theprovinces. Therefore the regulation is unlikely to have any direct impact on the morerecent export decisions by firms. A further concern is that the regulation could haveaffected export decisions in the 1980s, and in turn this could have had an impact onexport in the years covered by our dataset. However, this argument hinges on the degreeof export persistence. As documented by Roberts and Tybout (1997), while exportexperience in the previous year increases the probability of exporting, the export historythree years earlier has no predictive power for current export.

Finally, the 1936 banking law is unlikely to have affected credit supply conditions forlong after its complete removal in the late 1980s. For example, consider a firm seekingcredit in 2000. We do not expect that its probability of obtaining funds or its collateralrequirement were significantly affected by a regulation that was removed more than tenyears earlier. Therefore our instruments are unlikely to pick up any direct effect on exportof credit market conditions in our sample period. However, the reader could remainsomewhat concerned that the exclusion restriction might be violated: while it is plausiblethat the effects of the credit market regulation progressively faded over the 1990s, andhence that credit market conditions faced by the firms in the sample were not influencedby the regulation, it has also been observed that financial development exhibited somepersistence in Italy. This implies that the 1936 regulation could affect export decisionsthrough its long-run effect on financial development. To control for this possible channel,our specifications have controlled directly for proxies of credit market conditions (seeSection 3). Nonetheless, some residual concerns could linger about the exclusionrestriction, hence it is useful to discuss how the results could be affected if this restrictionwere violated. If our additional controls fail to fully capture local credit market

Economica



conditions, the possible long-run persistence of financial development that was shaped bythe 1936 regulation could affect export decisions. Specifically, firms in more financiallydeveloped provinces could have better export performance (i.e. have a higher probabilityof exporting and have a higher export value conditional on exporting). At the same time,more financially developed provinces could have a lower share of family firms. Thus ifthe 1936 regulation affects export decisions through its long-run effect on financialdevelopment other than through its impact on firm ownership structure, then we wouldunderstate the effect of family ownership on export performance.

To address the possible endogeneity and identify the effect of family ownership onexport decisions, we use the provincial data on the number of savings banks in 1936 as aninstrument for family ownership. Since the 1936 regulation was tighter in provinces witha lower share of savings banks (Guiso et al. 2003), we expect the share of savings banksto be negatively correlated with family ownership. Because cyclical variations in theeconomic activity of a province after deregulation could be correlated with ourinstrument and with firms’ export decisions, we also control for the average growth rateof the value added of the province in 1991–8. Given that the instrument is at the provincelevel, we replace province fixed effects with area dummies and provincial GDP growth tocontrol for province-level socioeconomic conditions.

IV estimates

Column (1) of Table 11 reports the OLS estimates of the baseline specification inequation (2) when province fixed effects are replaced by area dummies and provincialGDP growth. The estimated coefficient on family firms is 3.3%, which is almost identicalto the estimate reported in column (1) of Table 2. In column (2) of Table 11, we reportthe 2SLS estimates of the linear probability model. To conserve space, the bottom ofcolumn (2) reports only the coefficient on the number of savings banks from the first-stage regression. Consistent with our expectation, the probability that the mainshareholder is a family decreases in the number of savings banks in the province in 1936.Based on the prescriptions of the 1936 Italian banking regulation (see, for example,Guiso et al. 2003), provinces with a larger number of savings banks should have sufferedless from the regulatory freeze. Our result supports the hypothesis that less bindingregulation implies lower probability that the main shareholder is a family, which is in linewith the theoretical predictions in Caselli and Gennaioli (2013). Further, column (2)shows that in the second-stage regression, family ownership has a negative butstatistically insignificant impact on a firm’s export participation decision. Theinsignificant result for family ownership likely arises from the relatively weak partialcorrelation between the number of savings banks and family ownership in the first-stageregression.

Since the linear probability model does not account for the fact that both exportparticipation decision and family ownership are binary variables, we take an alternativeapproach and estimate a bivariate probit model. The ownership equation can bemodelled using the probit

PðOi ¼ 1jIp;ZiÞ ¼ PðIpdþ Zikþ mi [ 0Þ ¼ UðIpdþ ZikÞ;ð4Þ

where Oi is a binary variable that equals 1 if the main shareholder of firm i is anindividual or family, 0 otherwise; Ip is the set of instruments that capture the tightness of

Economica



TABLE11

IVESTIM

ATES

Extensivemargin

Intensivemargin

OLS

2SLS

Probit

Biv.probit

OLS

2SLS

(1)

(2)

(3)

(4)

(5)

(6)

Family

0.033***

�0.678

0.039***

0.188***

0.042

1.609

(0.011)

(0.816)

(0.013)

(0.062)

(0.030)

(1.496)

Log (total

assets)

0.100***

0.016

0.120***

0.025***

1.064***

1.256***

(0.007)

(0.098)

(0.009)

(0.008)

(0.014)

(0.185)

Log (capital

intensity)

�0.027***

�0.014

�0.031***

0.013***

�0.235***

�0.273***

(0.009)

(0.019)

(0.010)

(0.008)

(0.028)

(0.053)

Log (labour

productivity)

0.016

0.014

0.011

0.006

0.245***

0.253***

(0.012)

(0.012)

(0.014)

(0.011)

(0.037)

(0.049)

Age

0.000

0.001

0.000

0.001

�0.004***

�0.007***

(0.000)

(0.001)

(0.000)

(0.0003)

(0.001)

(0.003)

Corporation

0.138***

0.126***

0.150***

0.117***

0.087

0.058

(0.027)

(0.033)

(0.029)

(0.023)

(0.065)

(0.093)

Consortium

0.055***

0.063***

0.060***

0.054***

0.049

�0.012

(0.015)

(0.021)

(0.015)

(0.017)

(0.038)

(0.077)

ATECO

5-digit

0.044***

0.048***

0.050***

0.049***

0.044

0.042

(0.014)

(0.018)

(0.014)

(0.016)

(0.044)

(0.043)

ATECO

4-digit

0.002

0.000

0.006

0.004

�0.008

�0.005

(0.013)

(0.018)

(0.015)

(0.016)

(0.043)

(0.043)

Leverage

1.106*

1.335***

1.833

1.255

15.873

6.969

(0.633)

(0.382)

(2.753)

(8.879)

(12.211)

(16.142)

Leverage

*Rajan-Zingales

index

�14.262***

�11.551***

�35.133***

�53.895**

�27.071

�69.74

(4.163)

(3.816)

(13.451)

(12.091)

(38.504)

(54.572)

Economica



TABLE11

CONTIN

UED

Extensivemargin

Intensivemargin

OLS

2SLS

Probit

Biv.probit

OLS

2SLS

(1)

(2)

(3)

(4)

(5)

(6)

Rajan-Zingales

index

�0.057***

�0.060*

�0.062***

�0.044***

0.005

0.092

(0.019)

(0.031)

(0.022)

(0.029)

(0.079)

(0.126)

Centre

�0.056**

�0.039

�0.067**

�0.033***

0.013

�0.043

(0.024)

(0.027)

(0.027)

(0.023)

(0.077)

(0.086)

South

�0.127***

�0.095**

�0.144***

�0.094***

�0.327***

�0.423***

(0.019)

(0.039)

(0.020)

(0.022)

(0.045)

(0.097)

ProvincialGDP

growth

0.014

0.066

�0.018

0.004

�0.001

�0.153

(0.127)

(0.134)

(0.139)

(0.109)

(0.329)

(0.405)

Instrumentalvariable

Number

ofsavings

banksin

1936

�0.034**

�0.104*

�0.044**

(0.016)

(0.062)

(0.019)

Province

fixed

effects

NN

NN

NN

Industry

fixed

effects

YY

YY

YY

Survey

yearfixed

effects

YY

YY

YY

Observations

12,368

12,368

12,368

12,368

5876

5876

R2

0.163

0.666

Notes

Alltheregressionsincludeindustry

andsurvey

yearfixed

effects.In

columns(1)-(4),thedependentvariableisabinary


exports,0otherwise.

Incolumns(5)and(6),thedependentvariable

isthelogarithm

ofthevalueofexports.See

thenotesto

Table

1andSectionIIIformore

detailaboutmeasurementofthe

controlvariables.

Because

theinstrument(thenumber

ofsavingsbanksin

1936)is

attheprovince

level,wereplace

province

fixed

effects

withprovincialGDPgrowth

andareadummy

variables(C

entreandSouth)to

controlforprovince-levelsocioeconomicconditions.

Columns(3)and(4)reportthemarginaleff

ects.In

parentheses

are

robuststandard

errors

clustered

attheprovince

level.

***,**,*indicatep<0.01,p<0.05,p<0.1,respectively.

Economica



the 1936 banking regulation at the provincial level; the Zi are control variables inequation (1); and mi is a normally distributed random error with zero mean and unitvariation. Equations (1) and (4) constitute a recursive bivariate probit model. The effectof ownership on the probability of exporting can be identified under the assumption thatthe set of instruments Ip is excluded from equation (1). Although Oi enters equation (1) asan endogenous variable, we can estimate (1) and an equation of family ownership usingstandard bivariate probit software (Greene 2002, pp. 715–16). Moreover, since theinstruments are at the province level, we cluster standard errors by province.

For the purpose of comparison, in column (3) of Table 11 we report the estimatedmarginal effects of equation (1) in which area dummies and provincial GDP growth areused to control for province-level socioeconomic conditions. The marginal effect offamily ownership is 0.039, which is very close to that displayed in column (2) of Table 2.Column (4) displays the results for the bivariate probit model of export participation andfamily ownership. Unlike the 2SLS estimate, the estimated coefficient on familyownership is now significantly positive. The marginal effect of family ownership is 0.188,which is substantially larger than the probit marginal effect of 0.039 in column (3).Unlike the 2SLS estimate, identification of the effect of family ownership in the bivariateprobit model can also be based on the non-linearity of the functional form. Thusalthough the bottom of column (4) shows a relatively weak correlation between theexcluded instrument (the number of savings banks) and family ownership, we obtain asignificantly positive coefficient on family ownership.22

In columns (5) and (6) of Table 11, we report the OLS and 2SLS estimates of theintensive margin of trade equation (3). Both estimates suggest that family ownership hasno significant effect on the intensive margin. Overall, Table 11 shows that the IVestimates are largely consistent with the baseline results reported in Table 2.

VII. CONCLUSION

This paper investigates the impact of family ownership on firms’ internationalizationusing an unusually rich sample of Italian firms. The theoretical literature yieldsambiguous predictions on whether family firms have more incentives and ability toexport than non-family ones. We find that family firms are significantly more likely toexport than non-family firms, and that this positive effect is especially pronounced whenfamily owners retain control rights and hire external managers. Collectively, the resultssupport the hypothesis that family owners promote export because they have a longer-term horizon and better internalize the long-run benefits of internationalization. At thesame time, they also suggest that family owners’ lack of competence and skills as well asthe risk aversion induced by their fear of losing control and independence couldattenuate this positive effect. The analysis further reveals that family ownership alsoinfluences a firm’s internationalization process. In particular, again possibly because oflimited skills and knowledge of international markets, family firms exhibit some tendencyto enter foreign markets in a progressive way and to benefit from the support of domesticinstitutions in their international expansion. At the same time, due to their emphasis onlong-term returns, once family firms choose to sustain the relevant fixed costs for entry indifficult export markets, they turn out to be more persistent exporters than non-familyfirms.

A relevant point worth discussion is the applicability of our results to other countries.Our sample comprises a large number of privately held firms as well as the universe ofpublicly listed Italian companies. As noted, in line with the structure of the Italian

Economica



business sector, most firms in our sample are effectively small and medium-sizedbusinesses. Many countries feature a structure of the business sector similar to that ofItaly, with a significant importance of privately held small and medium-sized firms, andwidespread family ownership among privately held firms. Various countries also exhibitproblems analogous to those emphasized in our analysis, such as the excessive tendencyof families to transmit ownership and control to subsequent generations, but also a morepronounced short-termism of institutional investors relative to family owners (for adiscussion, see, for example, OECD 2014). Thus overall our study can yield usefulinsights into the impact of family firms on the export performance of several countries. Anumber of studies in prior literature on family firms examine the case of the USA, whichfeatures a relatively strong presence of publicly traded firms. Because of this structure ofthe US business sector, and due also to data availability, such studies focus on publiclytraded firms controlled by families. Thus one should use caution in applying the lessonsfrom Italy to the US case. Nonetheless, our analysis can help in understandingmechanisms that may also be relevant in the US experience. These include, for example,the effects of the long-termism of family owners on international expansion, and on thenegative side, the impact on internationalization of families’ fear of losing control overfirms, which is allegedly a recurrent issue in large, publicly traded companies.

This analysis represents a first step in a potentially fruitful line of research. Whilethere is established evidence that corporate governance significantly affects firms’performance, we still know little on the channels through which this influence unfolds.This paper uncovers a key role of corporate governance in firms’ internationalization.

APPENDIX: DATA SOURCES AND VARIABLE DEFINITIONS

Table A1 describes the definitions of the variables used in the paper. Three main data sources areused in the empirical analysis:

• four waves of the Capitalia Survey of Italian Manufacturing Firms (SIMF), whichcover three-year periods ending respectively in 1997, 2000, 2003 and 2006;

• the province-level database of the Italian National Statistics Office (ISTAT);

• the book Struttura Funzionale e Territoriale del Sistema Bancario Italiano 1936–1974(SFT) by the Bank of Italy.

TABLE A1VARIABLE DEFINITIONS

Variable Definition and source (in parentheses)

Main dependent variablesExport participation Dummy that takes the value 1 if the firm exports in the year of

the survey, 0 otherwise. (SIMF)

Log(export) Logarithm of foreign sales. (SIMF)

Ownership structure

Family The survey asks each firm to report the characteristics of themain shareholder of the firm. Family is a dummy that takesthe value 1 if the main shareholder is a family or an

individual. (SIMF)Share_family Equity share held by the main shareholder, if the main

shareholder is a family or an individual. (SIMF)

Economica



†CONTINUED


Financial institution The survey asks each firm to report the characteristics of the

main shareholder of the firm. Financial institution is adummy that takes the value 1 if the main shareholder is afinancial institution. (SIMF)

Corporate governanceFamily with control Dummy that takes the value 1 if a family is the main

shareholder and reports to have control over the firm, 0otherwise. (SIMF)

Dummy that takes the value 1 if a family is the main

shareholder and reports to have no control over the firm, 0otherwise. (SIMF)

External managers The percentage of external managers on the board of the firm.

(SIMF)

Control variables

Total assets, sales,current assets andinventories

These variables are balance sheet data. They are available foreach year covered by the survey. We use the average over thethree years of the survey. (SIMF)

Number of employees Total number of employees in the year of the survey. (SIMF)Age of the firm Number of years since inception. (SIMF)Corporation (businesstype)

The survey asks each firm whether it is publicly listed. In thesurvey, the information on whether the firm is a private

limited company (LTD) or a public limited company (PLC) isavailable only for the 2003 and 2006 surveys. For the otheryears, the information, which is publicly available on firms’

websites, has been imputed by hand using the VATidentification number. Corporation is a dummy that takes thevalue 1 if the firm is an LTD or PLC. (SIMF)

Consortium The survey asks each firm to report whether it belongs to aconsortium. The dummy for participation in a consortiumtakes the value 1 if the firm answers ’yes’ to this question,0 otherwise. (SIMF)

Leverage For each firm and year of the survey, we calculate the ratio oftotal liabilities to equity; then we compute the average overthe three years for the survey. (SIMF)

Financial concentration Concentration of the firm’s financial portfolio, measured as theHerfindahl-Hirschman index of the various asset shares.The survey asks each firm to report the allocation of its

financial investments among equity participation in Italiancompanies, equity participation in foreign companies,short-term Italian bonds, medium- and long-term Italian

bonds, foreign bonds, other financial instruments. (SIMF)Skilled workforce Dummy that takes the value 1 if the share of employees with at

least a secondary school degree exceeds the industry median,0 otherwise. (SIMF)

TABLE A1

CONTINUED

Economica



†CONTINUED


North Dummy that takes the value 1 if the firm is located in a

northern province, 0 otherwise. (SIMF)Centre Dummy that takes the value 1 if the firm is located in a central

province, 0 otherwise. (SIMF)

South Dummy that takes the value 1 if the firm is located in asouthern province, 0 otherwise. (SIMF)

ATECO n-digit Dummy that takes the value 1 if the firm reports its ATECO

classification as an n-digit number, 0 otherwise. (SIMF)Sector of activity The survey reports the sector of activity of firms (ATECO

code). Based on this information, firms are classified as

traditional, scale-intensive, specialized, and high-tech usingthe Pavitt taxonomy. Traditional sectors include producers ofapparel and textiles, food and beverages, tobacco and leather,among others. Scale-intensive firms include producers of

paper and allied products, petroleum and coal, stone, clay,glass and concrete products, among others. Specializedsectors include producers of electric and electronic

equipment, mechanical machinery, radio and TV equipment,among others. High-tech sectors include producers ofmedical and orthopaedic appliances, pharmaceuticals and

agricultural chemicals, among others. (SIMF)Rajan-Zingales index We use the measure of external financial dependence proposed

by Rajan and Zingales (1998). This captures the differentdependence of industrial sectors on external sources of

finance.Provincial GDP growth Average growth rate of the value added of the province where

the firm is located over the years 1985-94. (ISTAT)

Instrumental variablesSavings banks in 1936 Number of savings banks in the year 1936 in the province, per

100,000 inhabitants. (SFT)Cooperative banks in1936

Number of cooperative banks in the year 1936 in the province,per 100,000 inhabitants. (SFT)

Other variablesFinancial institutionsubscriber 2003

Dummy that takes the value 1 if a financial institutionunderwrote new shares of the firm in the years 2001-3; 0otherwise. (SIMF)

Intention to go public2003

Dummy that takes the value 1 if in 2003 the firm plans to gopublic in the following year, 0 otherwise. (SIMF)

Collaboration with local

distributors,intermediaries; jointventures

Dummies that take the value 1 if in foreign markets the firm

engaged in collaborations with local distributors orintermediaries, or in joint ventures with foreign firms. (SIMF)

Assistance of domesticparties

Dummy that takes the value 1 if the firm has activities oftechnical cooperation with foreign partners. (SIMF)

TABLE A1

CONTINUED

Economica



ACKNOWLEDGMENTS

We thank the Editor, Gianluca Benigno, and the anonymous referees for their comments. We alsothank several seminar and conference participants for helpful comments and conversations. Allremaining errors are ours.

NOTES

1. In the USA, about one-third of the S&P 500 firms are controlled by families (Anderson and Reeb 2003). Incontinental Europe, the majority of publicly held firms remain family controlled (La Porta et al. 1999;Faccio and Lang 2002). In East Asia, a small number of families control firms that account for the majorityof stock market capitalization (Claessens et al. 2000).

2. Family ownership appears to have no significant effect on the value of foreign sales, conditional onexporting (the ‘intensive margin’).

3. Italian provinces are geographical entities similar in size to US counties.4. For other analyses of the impact of family ownership, see, for example, Schulze et al. (2001), and for

emerging markets, see Luo and Chung (2013). Fahlenbrach (2009) obtains evidence of a founder-CEOpremium. Other studies investigate the relationship between family ownership and the protection ofminority investors across countries (La Porta et al. 1999) and the impact on firm value of families’ excesscontrol over ownership (Claessens et al. 2002).

5. See, for example, Bushee (1998) for evidence on the short-termism of institutional investors due to short-term trading strategies and regulatory restrictions.

6. Both variables are expressed as deviations from the sample mean.7. We downloaded the data from http://www.wto.org/english/res_e/statis_e/itip_e.htm (accessed 22 June

2015). We include the non-tariff measures that are applied to all WTO members or EU countries (includingItaly) during our sample period. The measures include technical ones (such as technical barriers to trade)and non-technical ones (such as anti-dumping duties). To construct the frequency index across three-digitATECO industries and over our sample period, we sum up non-tariff measures across all countries thatimposed measures in a particular industry and in a particular year. See UNCTAD (2013) for more detailson the measures.

8. The coefficient on the interaction term between non-tariff barriers and family ownership is statisticallyinsignificant.

9. Columns (7) and (8) of Table 2 show that the value of exports is higher for bigger and more productivefirms. By contrast, older firms export less.

10. The rate of response to this question is around 40% of the sample.11. As shown in column (3) of Table 5, older family firms export significantly more than their non-family

counterparts by 7%.12. The reader might be concerned that the negative coefficient on the variable capturing whether financial

institutions subscribed new shares could be the mechanical outcome of those firms already exporting tovarious markets (hence being less able to enter new markets). To assuage this possible concern, we insert thenumber of export markets as an additional regressor.

13. Notice that although these four categories of industries have different distributions of firm size, this shouldnot have a confounding effect because we control for firm size in all the regressions.

14. Consistent with the results for export participation, we also find that the positive effect of family ownershipon the intensive margin of trade is stronger for the traditional and scale-intensive sectors, but much weakerfor the high-tech sectors. Except for the scale-intensive sectors, the estimated effect of family ownership islargely statistically insignificant.

15. Columns (2)–(5) of Table 7 show that a higher share of skilled employees strengthens the positive effect offamily ownership on export especially in low-tech industries. It is sometimes argued that a higher number ofmanagers is a further way to overcome the limited competencies of family owners. In fact, a moresophisticated governance with a larger management team could help a firm to face the difficulty ofinvestments such as export. In unreported tests (available from the authors), we re-estimated the baselineregressions adding a dummy equal to 1 if in the firm the share of managers over total employees is higherthan the industry median, 0 otherwise, as well as an interaction term between this dummy and familyownership. The estimates provide some evidence that the export activities of family firms benefitparticularly from a larger share of managers.

16. Firms can have scarce knowledge about the profitability of operating in a foreign market. Sequentialexporting can allow firms to learn from experience (Albornoz et al. 2012).

17. We have data on the value of export in the year of interview but not on the value of export in each year.Thus we can conduct this analysis only by looking at changes in the value of export across surveys.

18. Panel B of Table 9 also shows that having a more skilled workforce significantly increases the probability ofentering non-EUmarkets, including Oceania.

Economica



http://www.wto.org/english/res_e/statis_e/itip_e.htm

19. In column (3) of panel B of Table 10, we further restrict attention to assistance received from chambers ofcommerce and offices of the Italian Institute for Commerce. The results carry through.

20. In a similar vein, Quintin (2008) investigates the impact of imperfect enforcement in an environment withinheritance of wealth. Becht and DeLong (2005) show that a deeper stock market led to a decline in thehegemony of families in the USA.

21. For instance, the strong presence of savings banks in the north-east and the centre stemmed from the factthat this institution originated in Austria and started to operate first in the provinces dominated by theAustrian Empire (Lombardia and the north-east) and in close-by states (especially Tuscany and the PapalStates).

22. We also experimented with using both the number of savings banks and the number of cooperative banks asinstruments for family ownership. The results are very similar to those reported in Table 11.

REFERENCES

ALBORNOZ, F., CALVO PARDO, H. F., CORCOS, G. and ORNELAS, E. (2012). Sequential exporting. Journal of

International Economics, 88, 17–31.ANDERSON, R. C. and REEB, D. M. (2003). Founding-family ownership, corporate diversification, and firm

leverage. Journal of Law and Economics, 46, 653–84.BALDWIN, R. (1988). Hysteresis in import prices: the beachhead effect. American Economic Review, 78, 773–85.BARCA, F., FERRI, G. and PESARESI, N. (1997). Banks and corporate governance in Italy: a two-tier model. In

M. Balling, E. Hennessy and R. O’Brien (eds), Corporate Governance, Financial Markets and Global

Convergence. Berlin: Springer, pp. 17–40.BECHT, M. and DELONG, J. B. (2005). Why has there been so little blockholding in America? In R. K. Morck

(ed.), A History of Corporate Governance Around the World: Family Business Groups to Professional

Managers. Chicago: NBER/University of Chicago Press.

BELLONE, F., MUSSO, P., NESTA, L. and QUERE, M. (2008). The U-shaped productivity dynamics of French

exporters. Review of World Economics, 144, 636–59.BENFRATELLO, L., SCHIANTARELLI, F. and SEMBENELLI, A. (2008). Banks and innovation: microeconometric

evidence on Italian firms. Journal of Financial Economics, 90, 197–217.BERNARD, A. B. and JENSEN, B. J. (2004). Exceptional exporter performance: cause, effect, or both? Journal of

International Economics, 47, 1–25.BERTRAND, M. and SCHOAR, A. (2006). The role of family in family firms. Journal of Economic Perspectives, 20,

73–96.————, JOHNSON, S., SAMPHANTHARAK, K. and SCHOAR, A. (2008). Mixing family with business: a study of

Thai business groups and the families behind them. Journal of Financial Economics, 88, 466–98.BIANCHI, M. and BIANCO, M. (2008). L’evoluzione della Corporate Governance in Italia: Meno Piramidi Pi�u

Coalizioni? Bologna: Il Mulino.

BIANCO, M. (2003). L’industria Italiana. Bologna: Il Mulino.

BOLTON, P. and VON THADDEN, E. L. (1998). Blocks, liquidity, and corporate control. Journal of Finance, 53,

1–25.BUCH, C. M., D€OPKE, J. and STROTMANN, H. (2009). Does export openness increase firm-level output volatility?

The World Economy, 32, 531–51.BURKART, M., PANUNZI, F. and SHLEIFER, A. (2003). Family firms. Journal of Finance, 58, 2167–207.BUSHEE, B. (1998). The influence of institutional investors on myopic R&D investment behavior. Accounting

Review, 73, 19–45.CASELLI, F. and GENNAIOLI, N. (2013). Dynastic management. Economic Inquiry, 51, 971–96.CASSON, M. (2000). The family firm: an analysis of the dynastic motive. In M. Casson (ed.), Enterprise and

leadership: Studies on Firms, Markets and Networks. Cheltenham: Edward Elgar.

CHANEY, T. (2005). Liquidity constrained exporters. Mimeo, University of Chicago.

CLAESSENS, S., DJANKOV, S., FAN, J. and LANG, L. (2002). Disentangling the incentive and entrenchment

effects of large shareholdings. Journal of Finance, 57, 2741–71.————, ———— and LANG, L. H. P. (2000). The separation of ownership and control in East Asian

corporations. Journal of Financial Economics, 58, 81–112.DAS, S., ROBERTS, M. J. and TYBOUT, J. R. (2007). Market entry costs, producer heterogeneity, and export

dynamics. Econometrica, 75, 837–73.DEMSETZ, H. and LEHN, K. (1985). The structure of corporate ownership: causes and consequences. Journal of

Political Economy, 93, 1155–77.DI GIOVANNI, J. and LEVCHENKO, A. A. (2009). Trade openness and volatility. Review of Economics and

Statistics, 91, 558–85.

Economica



DIXIT, A. (1989). Entry and exit decisions under uncertainty. Journal of Political Economy, 97, 620–38.EATON, J., ESLAVA, M., KUGLER, M. and TYBOUT, J. (2008). Export dynamics in Colombia:

transactions level evidence. Borradores de Economia Working Paper no. 522, Banco de la Rep�ublica

de Colombia.

ELLUL, A., PAGANO, M. and PANUNZI, F. (2010). Inheritance law and investment in family firms. American

Economic Review, 100, 2414–50.FACCIO, M. and LANG, L. H. P. (2002). The ultimate ownership of Western European corporations. Journal of

Financial Economics, 65, 365–95.FAHLENBRACH, R. (2009). Founder-CEOs, investment decisions, and stock market performance. Journal of

Financial and Quantitative Analysis, 44, 439–66.FAMA, E. F. and JENSEN, M. C. (1983). Agency problems and residual claims. Journal of Law and Economics,

26, 327–49.FAVERO, C. A., PAGANO, M. and VON THADDEN, E. L. (2010). How does liquidity affect government bond

yields. Journal of Financial and Quantitative Analysis, 45, 107–34.GALLO, M. A. and GARCIA PONT, C. (1996). Important factors in family business internationalization. Family

Business Review, 9, 45–59.GRAVES, C. and THOMAS, J. (2006). Internationalization of Australian family businesses: a managerial

capabilities perspective. Family Business Review, 19, 207–24.GREENAWAY, D., GUARIGLIA, A. and KNELLER, R. (2007). Financial factors and exporting decisions. Journal

of International Economics, 73, 377–95.GREENE, W. (2002). Econometric Analysis.Upper Saddle River, NJ: Prentice Hall.

GROSSMAN, S. and HART, O. (1988). One-share, one-vote, and the market for corporate control. Journal of

Financial Economics, 20, 175–202.GUISO, L., SAPIENZA, P. and ZINGALES, L. (2003). The cost of banking regulation. Working Paper, Chicago

Graduate School of Business.

————, ———— and ———— (2004). Does local financial development matter? Quarterly Journal of Economics,

119, 929–69.HANSSON, P. and LUNDIN, N. N. (2004). Export as an indicator on or promoter of successful Swedish

manufacturing firms in the 1990s.Weltwirtschaftliches Archiv/Review of World Economics, 140, 415–45.HERRERA, A. M. and MINETTI, R. (2007). Informed finance and technological change: evidence from credit

relationships. Journal of Financial Economics, 83, 223–69.LA PORTA, R., LOPEZ-DE-SILANES, F. and SHLEIFER, A. (1999). Corporate ownership around the world.

Journal of Finance, 54, 471–517.LUO, X. and CHUNG, C. N. (2013). Filling or abusing the institutional void? Ownership and management

control of public family businesses in an emerging market.Organization Science, 24, 591–613.MANOVA, K. (2013). Credit constraints, heterogeneous firms, and international trade. Review of Economic

Studies, 80, 711–44.MAURY, B. (2006). Family ownership and firm performance: empirical evidence from Western European

corporations. Journal of Corporate Finance, 12, 321–41.MELITZ, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity.

Econometrica, 71, 1695–725.MINETTI, R. and ZHU, S. C. (2011). Credit constraints and firm export: microeconomic evidence from Italy.

Journal of International Economics, 83, 109–25.OECD (2014). Report on effective approaches to support implementation of the G20/OECD high-level principles

on long-term investment financing by institutional investors (annex). Paris: OECD.

ONIDA, F. (2004). Se il Piccolo non Cresce. Bologna: Il Mulino.

PAVITT, K. (1984). Sectoral patterns of technical change: towards a taxonomy and a theory. Research Policy, 13,

343–73.PEREZ-GONZALEZ, F. (2006). Inherited control and firm performance. American Economic Review, 96, 1559–88.QUINTIN, E. (2008). Limited enforcement and the organization of production. Journal of Macroeconomics, 30,

1222–45.RAJAN, R. and ZINGALES, L. (1998). Financial development and growth. American Economic Review, 88, 559–

86.

REBITZER, J. B. (1995). Is there a trade-off between supervision and wages? An empirical test of efficiency wage

theory. Journal of Economic Behavior & Organization, 28, 107–29.ROBERTS, M. J. and TYBOUT, J. R. (1997). The decision to export in Colombia: an empirical model of entry with

sunk costs. American Economic Review, 87, 545–64.

Economica



ROE, M. J. (2013). Corporate short-termism—in the boardroom and in the courtroom. The Business Lawyer, 68,

977–1006.SCHULZE, W., LUBATKIN, M., DINO, R. and BUCHHOLTZ, A. (2001). Agency relationships in family firms:

theory and evidence.Organization Science, 12, 99–116.SHLEIFER, A. and VISHNY, R. (1997). A survey of corporate governance. Journal of Finance, 52, 737–83.SHOCKLEY, R. and THAKOR, A. (1992). Information content of commitments to lend in the future: theory and

evidence on the gains from relationship banking. Indiana University Discussion Paper no. 523.

SRAER, D. and THESMAR, D. (2007). Performance and behaviour of family firms: evidence from the French

stock market.Journal of the European Economic Association, 5, 709–51.SULLIVAN, D. and BAUERSCHMIDT, A. (1989). Common factors underlying barriers to export: a comparative

study in the European and U.S. paper industry.Management International Review, 29, 17–32.THE ECONOMIST (2012). The big engine that couldn’t. 19 May.

THE ECONOMIST (2013). Global niche players. 2 February.

UNCTAD (2013). A preliminary analysis on newly collected data on non-tariff measures. Policy issues in

international trade and commodities. Study Series no. 53.

VANNOORENBERGHE, G. (2012). Firm-level volatility and exports. Journal of International Economics, 86,

57–67.VILLALONGA, B. and AMIT, R. (2006). How do family ownership, control and management affect firm value?

Journal of Financial Economics, 80, 385–417.WARD, J. (2006). Managing for the long-term. Families in Business, 26, 72–3.WOOLDRIDGE, J. M. (2002). Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press.

ZAHRA, S. A. (2003). International expansion of U.S. manufacturing family businesses: the effect of ownership

and involvement. Journal of Business Venturing, 18, 495–512.

Economica



Family Firms, Corporate Governance and Exportzhuc/Family firms.pdf · Family Firms, Corporate...

Documents

Transcript of Family Firms, Corporate Governance and Exportzhuc/Family firms.pdf · Family Firms, Corporate...