International music preferences as a measure of culture: evidence from cross-border mergers

Abstract

This paper introduces the significance of international music preferences as a determinant of cross-border mergers. We argue that international music preferences capture the distance in culture between nations. We find that country pairs whose citizens experience relatively small distance in their music preferences (listen to each other’s music) exhibit more cross-border mergers. Overall, this study highlights that music preferences can measure international similarities in culture.

KEYWORDS:

• International business
• culture
• international music preferences
• cross-border mergers

JEL CLASSIFICATIONS:

• F20
• F30
• Z1
• G34

1. Introduction

Several studies have shown that cultural distance between nations influences managerial decisions. Hofstede (1980) was the first who developed a national culture framework and has naturally dominated this field. Based on surveys of IBM employees between 1967 and 1973, Hofstede developed four cultural dimensions and at a later stage (Hofstede and Bond 1988), further cultural dimensions were added to the list. The first dimension is the power distance which captures the extent to which society thinks that the distribution of power is equally distributed. The second is individualism vs collectivism which measures the extent to which the society is expected to take care of themselves or to be looked after. The third is uncertainty avoidance which measures the extent to which a society tries to avoid change, having formal rules while avoiding new ideas and behaviors. The fourth is masculinity vs femininity indicating the extent to which the values of the society take after characteristics deemed to be masculine or feminine such as the pursuit of money or the level of caring for others. The fifth dimension is long-term vs short-term orientation. Societies with long-term orientation are future-oriented, while societies with short-term orientation tend to show more respect for traditional values. Finally, the indulgence vs restraint dimension measures the extent to which a society regulates and suppresses the satisfaction of human desires.

We explore in this study the significance of international music preferences as a measure of cultural proximity within the field of cross-border mergers. We believe that cross-border mergers offer an ideal context to test the role of culture. Managers tend to select firms in countries that are expected to experience the least potential problems between staff in the two entities after the merger announcement (e.g. Siganos and Tabner 2020¹; Ahern, Daminelli, and Fracassi 2015; Shenkar 2012). The nationality of an overseas firm is also one of the most significant decisions for a firm for the following reasons: approximately one-third of the mergers involve firms from different countries, and this proportion has increased over time with over 50% of the acquisitions being overseas since 2010 (e.g. Erel, Liao, and Weisbach 2012). Cross-border mergers represent most of the foreign direct investment according to the Organization for Economic Co-operation and Development (2007).

Although music has received significant attention in interdisciplinary research such as in sociology, anthropology, and linguistics, there has been very little, if any, exploration in the international business field. We argue in this study that the voluntary selection of individuals’ preferences towards another country’s music reflects the distance in culture between nations, and so such music selections are expected to be related to the number of cross-border mergers. We follow the advice given by Morris, Podolny, and Sullivan (2008) who suggest the measurement of culture by avoiding surveys and self-report values. There is no need to undertake expensive surveys to develop a relevant measure and one could easily update its values with more recent data. The music selections by international participants are also undertaken without much consideration and our measure therefore reflects more accurately participants’ views. The suggested measurement is based on millions of music downloads rather than on costly surveys with relatively few participants.

To elaborate on the theoretical link between international music preferences and culture, music preferences reflect the interaction between nations since music reflects two elements; music production and music reception. The lyrics, melodic structure, musical instruments, and language are all part of someone’s background and way of thinking. Like other arts, music is often used as a way to express and transmit views on the current state of the world. Reyes Schramm (1986), for example, illustrates that Vietnamese refugees expressed the trauma of escape and their experience living in America through their songs. Baker (2005) also highlights that Marco Perkovic’s songwriting focuses heavily on the Balkan War in the 1990s. Music production thus represents one’s national identity and generates a bond among citizens, offering common social ground.

People do not experience music similarly, but their response to music is subject to their experiences, national identity, and emotional state. As an example, lyrics produced reflecting difficult times for a nation would be experienced relatively dispassionately by later generations. According to Demorest and Morrison (2003), listeners are more likely to comprehend, appreciate, and enjoy foreign music to the extent that it is similar to local music. Musical instruments, the tempo of the song, and the lyrics would all make more sense the closer they are to their experiences in life and the type of music produced locally.

Several factors may influence the distance in music preferences between the two nations. Citizens are generally closer to music produced in a nearby nation since their experiences and ways of living are more likely to be similar. For example, citizens living in countries such as Spain, Italy, and Greece are more likely to produce, experience, appreciate, and enjoy similar music. Harnish (2005) indeed demonstrates how exchanges in music through adaptation and innovation gradually made citizens in Bali and the neighboring island of Lombok produce very similar music. It is also likely that the overall sentiment of citizens in nearby countries would be similar. Events such as natural disasters or economic crises that are contagious may influence neighboring countries simultaneously. As a result, citizens of neighboring countries are likely to produce and be interested in listening to similar types of songs. Exchanges in music may also take place through tourism, and the musicality of countries may then be relatively close to nearby countries (Dunbar-Hall 2001). It is thus no surprise that there is a tendency for particular types of music to be prominent in regions. ‘Tsiftenteli', for example, is a well-known type of music in the Middle East but is otherwise not widespread.

The language used can also influence the understanding of song lyrics and therefore the connection with foreign music. Citizens of country pairs that share a language are thus more likely to be attracted to each other’s music. The distance between languages could also influence listeners’ information processing and their perception of overseas music. Linguistic structures, syntax and grammar, speech intonation, the meaning of words, and the sung melody can be interpreted differently among nations (Besson and Schon 2001; Feld and Fox 1994). Lloyd and Mendez (2001), for example, report that although citizens of Botswana speak English, they face difficulties in understanding American lyrics due to the different meanings of the terminology used.

Religious beliefs may also be reflected in songwriting, music style(s), and lyrics. As an example, metal music is often inspired by paganism, esotericism, and Satanism, which has generated a great deal of controversy among admirers and opponents (Moberg 2009). Citizens that share religious beliefs are thus more likely to avoid listening to particular types of music while being attracted to similar music.

Songs may even reflect cultural dimensions as developed by Hofstede, such as citizens’ long-term orientation and individualism, and the power distance within a society. Citizens are thus likely to listen to foreign songs from countries that share similar cultures and idiosyncratic characteristics. If the lyrics of a song refer, for example, to unfairness in a society, people in countries that face similar experiences are more likely to be attracted to it. There is also evidence that personality influences music preferences. Soley and Spelke (2016) report that citizens tend to socialize with counterparts who share favorite songs, indicating common interests and personalities. Extroverted people have been shown to prefer cheerful music (Vella and Mills 2017; Rentfrow and Gosling 2003). According to Schwartz and Fouts (2003), people that choose to listen to metal music or even rock music often question authority and tend to disagree with following traditional customs and habits. Interestingly, DeWall et al. (2011) report that the lyrics of US songs show evidence of increases in individualism levels between 1980 and 2007 that, according to the authors, may reflect the increase of the level of individualism in the US over time.

Government decisions could also influence citizens’ music preferences. Simple repetition by media channels of the same song has been shown to change citizens’ views on how appealing it is (Hargreaves, North, and Tarrant 2006). The extent of government control over broadcasting decisions can thus be used as a measure of reducing or maximizing the attraction of particular foreign songs. The regime in North Korea, for example, intervened to reduce the exposure of its citizens to the Gangnam style of music produced in South Korea.² Another example is that between Russia and the United States that has often been some tension. Music education is an alternate approach that a government could use to influence its citizens’ music preferences over the long term. Music curricula impact people’s views at an early stage in their lives, when they are most vulnerable, by, for example, focusing on particular types of songs. To highlight the significance of music education on students’ understanding of international cultures, ‘world music education' has been recently adopted in several developed countries, such as the US and the UK. According to the world music education curriculum, students can better understand other cultures by simply being educated about their local music (Chen-Hafteck 2007; Campbell and Scott-Kassner 2006).

The image of a counterpart country may also influence international citizens’ music preferences. Some citizens may listen to US and UK music simply because these particular countries are typically idolized as representative of Western life. Wang (2005) explores youth culture and music in China, highlighting the significance of ‘coolness' behind the music preferences of Chinese youth. Saldanha (2002) also discusses the reception of Western pop music in Bangalore, highlighting that Western music makes people able to imagine other places while connecting with people of a particular social class. Citizens may then prefer musical styles arriving from countries that are in line with their aspirations of belonging to a certain social group.

We use a novel measure to capture citizens’ international music preferences at a country level, as recently developed by Liu, Hu, and Schedl (2018). This distance metric is based on actual album downloads from the LFM-1b online platform, which offers the nationality of thousands of online users and their historic international music preferences. We find that international music preferences are an important determinant of the number of cross-border mergers in country pairs. This relation is present after controlling for Hofstede’s cultural dimensions and other typical determinants used in the literature. We find that country pairs with relatively low music distance (i.e. where citizens listen to each other’s songs) experience many cross-border mergers. Distance in music preferences is one of the most prominent determinants in our empirical analysis. Overall, these results highlight the significance of international music preferences to explain managerial decisions.

The remainder of this paper is structured as follows: Section 2 presents the data used in this study and the main methodology followed. Sections 3 and 4 report the empirical findings, and finally, Section 5 concludes.

2. Data and methodology

2.1 Data

We use several data resources to undertake our study. We use the estimations by Liu, Hu, and Schedl (2018) as the distance in international music preferences. This is estimated at a national level and shows the absolute distance on the number of downloads of international albums per country pair. Unfortunately, no data or descriptive statistics are available during the period.³ It is a continuous variable based on the album downloads from the online music service Last.fm, LFM-1b, in the post-2000 period. The distance in music preferences is determined by analyzing the historic album selections of 46,619 online users with available nationality identification. The original dataset includes 208 countries, with over 3,000,000 artists. However, only 54 percent of the users identified their nationality and in several countries, only a relative percentage of the total number is available, for this reason, they only include countries with at least 1 percent of the total users in LFM-1b. This covers 11,165,177 unique album selections, and over 2 million unique artists, within twenty countries.⁴ It thus includes the ‘big' players in the global music industry, since nine out of the top ten countries, according to BPI (2017), are included in our sample (apart from South Korea). The shortest distance in music preferences is between the US and the UK (0.18), which makes sense considering the similarities between the two nations. On the counterpart, the largest distance in music preferences is between Japan and Belarus (0.78).

We use Thomson One Banker to access the total number of completed cross-border merger deals per country pair in the period between 2000 and 2018 which corresponds with data availability from music preferences. We average all years into one observation per country pair in line with the main independent variable as discussed earlier. We only include public-to-public merger deals with ‘MA' and ‘IMA' master deal types. Also, only mergers with at least 50% acquisition are included in our sample.⁵ We divide the total number of cross-border mergers in a country pair with the sum of the number of domestic mergers that took place in the counterpart countries in the same period. This adjustment is to ensure that we capture an ‘abnormal' level of merging activity.^{6 7}

Table 1 shows the country pairs with the top number of overseas mergers and also with the top music distance. Figure 1 also shows the histogram of the number of mergers undertaken in our country pairs. We find that the number of mergers is highly volatile from country pairs that have no mergers to the US/Canada pair that had a maximum of 8,513 mergers. As shown in Figure 1 over 80% of the country pairs have had less than 1000 deals, with a few of the country pairs exhibiting a much larger number of merger deals. We find some introductory evidence that the number of cross-border mergers is related negatively to the music distance. Country pairs with high (low) music distance tend to experience fewer (more) mergers.

Figure 1. Histogram of the number of overseas mergers undertaken in our country pairs.

Table 1. Country pairs with top music distance and top number of overseas mergers.

		Music distance	Overseas mergers
Top 10 country pairs with high music distance
Belarus	Japan	0.78	0
Finland	Japan	0.76	3
Japan	Ukraine	0.75	3
Japan	Sweden	0.75	26
Czech	Japan	0.74	11
Japan	Poland	0.73	5
Germany	Japan	0.71	139
Japan	Norway	0.71	9
Brazil	Japan	0.71	42
Japan	Russia	0.71	21
Top 10 country pairs with high overseas mergers
Canada	US	0.18	8513
UK	US	0.18	7102
Germany	US	0.44	2473
France	US	0.37	1845
Australia	US	0.20	1841
Germany	UK	0.35	1534
France	UK	0.32	1339
Germany	France	0.45	998
Japan	US	0.64	901
Norway	Sweden	0.41	866

We also access the following six dimensions developed by Hofstede (Hofstede, Hofstede, and Minkov 2010; Hofstede and Bond 1988; Hofstede 1980): power distance, individualism, masculinity, uncertainty avoidance, long-term orientation, and indulgence.⁸ We estimate the logarithm of one plus the corresponding absolute distance between the country pairs available in our sample. We use the national logarithm in all our variables that exhibit extreme values to reduce the impact of outliers in the estimated parameter coefficients.⁹

We also control for Facebook connections that indicate the affinity strength amongst countries.¹⁰ A high number of Facebook connections indicates high affinity. We also add several geographic controls. We estimate the flight time in hours between the capital cities of each country pair. Also, whether the country pairs share a border or whether they share a continent. These are dummy variables that take one when this is valid otherwise zero. We use the FreeMap Tools and the Airport Distance Calculator to access relevant data.¹¹

We also add several citizens’ characteristics. We use the CIA World Factbook to access relevant data regarding commonality in language, religion, and law.¹² We use a dummy that takes one if a country pair shares at least one language, the same religion, and the same law otherwise zero. We include all languages per country to capture the extent to which some of the citizens understand the content of the overseas songs. We also use the International Historical Statistics database to access the history of a country and use a dummy variable that equals one when countries share history otherwise zero.

We also access WorldBank¹³ to capture economic characteristics. We estimate for each country to pair the absolute distance in years for compulsory education, tax, GDP per capita, and economic freedom. All four variables are continuous. Education, Tax, and GDP data are collected from the averaging available data per country between 2000 and 2018. We also control for the sum GDP for both countries in the pair and their sum GDP growth. We also download the currency movements in the period since they may make particular firms more or less attractive in their valuation. We estimate the absolute distance in currency movements with countries with high distances that may exhibit more acquisitions. We estimate the average Economic Freedom level per country as available from the 2018 Report from the Economic Freedom website¹⁴ to estimate the average level of economic freedom in a country, where a high level indicates greater economic freedom. We also average the percentage of trade which is the sum of exports plus imports between 2000 and 2018 in relation to their GDP.¹⁵ Finally, we average tourism data between 2000 and 2018 from the World Tourism Organization to measure the distance in the international tourism expenditure in-country pairs.¹⁶

Table 2 reports the descriptive statistics of the variables used in this study. Note that N indicates the number of country pairs available in our dataset (rather than the number of mergers). We get access to data for 190 country pairs for all our variables apart from Hofstede’s and Facebook's connections due to a few missing observations. Hofstede’s country coverage does not fully match the dataset available from international music preferences. No data are available for one country (Russia) in the Facebook connections variable.

Table 2. Descriptive statistics.

	Mean	Median	S.D	Min	Max	N
Number of overseas mergers	251	47.00	856	0.00	8513	190
Number of overseas mergers / domestic mergers	0.03	0.01	0.05	0.00	0.37	190
Music distance	0.48	0.47	0.12	0.18	0.78	190
Ln Power distance	2.72	3.00	1.01	0.00	4.14	153
Ln Individualism distance	2.84	3.04	0.85	0.00	4.13	153
Ln Masculinity distance	3.00	3.22	0.93	0.00	4.51	153
Ln Uncertainty avoidance distance	2.94	3.14	0.88	0.00	4.20	153
Ln Long term orientation distance	3.00	3.19	0.87	0.41	4.22	190
Ln Indulgence distance	2.99	3.28	0.96	0.00	4.43	190
Ln Facebook connections	1.23	1.24	1.23	−1.50	4.86	171
Flight time between capitals	6.81	6.49	6.06	0.21	21.59	190
Shared border	0.11	0.00	0.31	0.00	1.00	190
Shared continent	0.51	1.00	0.50	0.00	1.00	190
At least one shared language	0.21	0.00	0.40	0.00	1.00	190
Shared religion	0.18	0.00	0.38	0.00	1.00	190
Shared history	0.09	0.00	0.29	0.00	1.00	190
Shared law	0.51	1.00	0.50	0.00	1.00	190
Educational distance	1.44	1.00	1.13	0.00	4.00	190
Ln Tax distance	1.98	2.11	0.78	0.01	3.19	190
Ln GDP distance	9.17	9.50	1.21	2.23	10.77	190
Ln GDP	10.98	11.03	0.33	9.88	11.54	190
GDP growth	0.08	0.07	0.02	0.05	0.15	190
Exchange rate distance	0.19	0.07	0.29	0.00	1.13	190
Economic freedom distance	0.79	0.62	0.59	0.00	2.36	190
International trade	0.71	0.68	0.22	0.26	1.33	190
Tourism distance	4.93	5.06	2.93	0.01	12.07	190

Note: This table reports descriptive statistics. N indicates the number of country pairs available in our dataset. We average all years into one observation per country pair.

Table 3 offers the Pearson correlations among our variables, where the star highlights statistical significance at the 0.1 percent level. We find the first evidence that distance in international music preferences is a strong determinant of the number of cross-border mergers. The correlation between the number of cross-border mergers and the distance in music is −0.41, which is the second highest value amongst all the determinants used in this study (behind Facebook connections). Distance in international music preferences makes some sense in comparison to other variables used. We find for example that the distance in international music preferences is relatively low within country pairs whose citizens share a language (−0.42), continent (−0.26), geographical border (−0.25), and affinity as shown in Facebook connections (−0.53). The highest correlation in absolute terms is between countries that share a continent with a low flight time between their capitals (−0.85). Several other variables are highly correlated. We focus in this study on international music preferences and a high correlation between some of the controls is not necessarily a problem since we are interested in whether the main relation holds after all relevant controls. Still, we report univariate results and we also report in Table 7 the main parameter coefficient after orthogonalizing variables with correlations higher than 0.5 to ensure that our conclusions are not driven by multicollinearity.

Table 3. Pearson correlation matrix (* indicates Pearson correlation coefficients that are significant at the 0.1% level).

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)
(1) Number of overseas mergers / domestic mergers
(2) Music distance	−0.41*
(3) Ln Power distance	−0.29*	0.25
(4) Ln Individualism distance	−0.31*	0.21	0.40*
(5) Ln Masculinity distance	−0.19	0.37*	0.08	−0.05
(6) Ln Uncertainty Avoidance Distance	−0.11	0.13	0.47*	0.21	0.09
(7) Ln Long term orientation distance	−0.16	0.24*	−0.04	0.06	−0.05	0.01
(8) Ln Indulgence distance	−0.21	0.34*	0.42*	0.34*	−0.01	0.40	0.23
(9) Ln Facebook connections	0.52*	−0.53*	−0.35*	−0.48*	−0.04	0.00	−0.18	−0.32*
(10) Flight time between capitals	−0.37*	0.21	−0.02	0.25	−0.06	−0.09	0.26*	0.08	−0.31*
(11) Shared Border	0.46*	−0.25*	−0.10	−0.19	−0.19	−0.16	−0.13	−0.25*	0.51*	−0.4*
(12) Shared continent	0.36*	−0.26*	0.01	−0.26*	0.00	0.08	−0.24	−0.07	0.34*	−0.85*	0.34*
(13) At least one shared language	0.25*	−0.42*	−0.26	−0.17	−0.11	−0.15	−0.10	−0.31*	0.23	0.05	0.25*	−0.07
(14) Shared religion	0.41*	−0.23	−0.03	−0.01	−0.14	−0.31	−0.20	−0.08	0.06	−0.09	0.24*	0.08	0.20
(15) Shared history	0.32*	−0.17	0.06	−0.09	−0.05	−0.09	−0.10	−0.10	0.34*	−0.21	0.65*	0.18	0.24
(16) Shared law	0.14	0.12	0.05	−0.19	0.06	−0.20	−0.08	−0.035	0.01	−0.12	0.13	0.17	0.08
(17) Educational distance	0.10	−0.04	−0.05	0.01	0.09	−0.02	0.04	−0.02	−0.02	−0.10	−0.09	0.05	−0.04
(18) Ln Tax distance	−0.01	0.05	0.03	−0.04	−0.01	−0.07	0.05	−0.03	−0.08	0.07	0.00	−0.09	−0.17
(19) Ln GDP distance	−0.17	0.09	0.33*	0.21	−0.07	0.09	−0.04	0.20	−0.06	−0.10	−0.05	0.07	−0.06
(20) Ln GDP	0.25*	−0.13	−0.31*	−0.22	0.21	0.08	−0.04	−0.23	0.10	−0.05	−0.12	−0.01	−0.01
(21) GDP growth	−0.22	0.04	0.33*	0.07	−0.06	0.07	0.04	0.19	0.09	−0.20	0.24*	0.26*	−0.17
(22) Exchange rate distance	−0.09	0.07	0.22	0.23	−0.05	0.04	0.07	0.29*	−0.11	0.19	−0.09	−0.22	0.01
(23) Economic freedom distance	−0.35*	0.19	0.52*	0.53*	−0.14	0.26	0.09	0.40*	−0.22	0.08	−0.14	−0.06	−0.01
(24) International trade	0.05	−0.06	−0.10	−0.22	0.10	−0.01	−0.05	0.08	0.19	−050*	0.14	0.47*	−0.13
(25) Tourism distance	−0.04	−0.15	0.04	−0.04	0.01	−0.10	−0.07	−0.09	0.01	0.06	0.05	0.00	0.04
	(14)	(15)	(16)	(17)	(18)	(19)	(20)	(21)	(22)	(23)	(24)
(15) Shared history	0.18	1.00
(16) Shared law	0.24*	0.14	1.00
(17) Educational distance	0.10	−0.04	0.12	1.00
(18) Ln Tax distance	−0.09	−0.02	−0.09	0.06	1.00
(19) Ln GDP distance	0.00	−0.05	−0.04	−0.02	−0.09	1.00
(20) Ln GDP	−0.13	−0.13	−0.35*	0.13	−0.02	−0.11	1.00
(21) GDP growth	−0.12	0.26*	0.15	−0.11	0.09	0.23	−0.60*	1.00
(22) Exchange rate distance	0.11	−0.02	0.05	0.10	−0.05	0.04	−0.22	−0.16	1.00
(23) Economic freedom distance	−0.19	−0.10	−0.11	−0.12	−0.01	0.54*	−0.38*	0.33*	−0.01	1.00
(24) International trade	0.11	0.03	0.04	0.22	−0.02	0.14	−0.17	0.33*	−0.05	0.01	1.00
(25) Tourism distance	−0.05	0.05	−0.21	−0.11	−0.02	0.18	0.05	0.28*	−0.16	0.02	−0.04

2.2 Methodology

For the main analysis, we undertake the following Poisson estimation since the dependent variable is non-integer and cannot take negative values. We use robust standard errors across all estimations.

NumberofOverseasMergers/DomesticMergers_i = constant + b₁ MusicDistance_i+ b₂ lnPowerDistance_i+b₃ lnIndividualismDistance_i+ b₄ lnMasculinityDistance_i+ b₅ lnUncertaintyAvoidance_i+ b₆ lnLongTermOrien-tationDistance_i +b₇ IndulgenceDistance_i + b₈ lnFacebookConnections_i + b₉ FlightTime_i + b₁₀ SharedBorder_i +b₁₁ SharedContinent_i + b₁₂ AtLeastOneSharedLanguage_i + b₁₃ SharedReligion_i + b₁₄ SharedHistory_i + b₁₅SharedLaw_i + b₁₆ EducationDistance_i + b₁₇ lnTaxDistance_i + b₁₈ lnGDPDistance_i + b₁₉ lnGDP_i + b₂₀ lnGDP-Growth_i + b₂₁ ExchangeRateDistance_i + b₂₂ EconomicFreedomDistance_i + b₂₃ InternationlTrade_i + b₂₄ TourismDistance_i + u_i (1)

Our dependent variable is the number of cross-border mergers adjusted by the domestic mergers for country pair i. The main variable under consideration is the Music Distance. We expect that the relevant parameter coefficient is significantly negative. This result would indicate that country pairs with little music distance exhibit many cross-border mergers. There should be a negative relationship between cultural distance indicators and the number of cross-border mergers to the extent that managers consider that low cultural distances would potentially generate fewer problems in the process of merging the two firms after the merger announcement (e.g. Stahl and Voigt 2008).

We also expect that there are more cross-border mergers in country pairs with lower distances and in-country pairs that share a continent or border (e.g. Ahern, Daminelli, and Fracassi 2015; Chakrabarti and Mitchell 2013). In line with the literature (e.g. Makridis 2022; Bailey et al. 2021; Siganos and Tabner 2020), we expect that country pairs with high Facebook connections exhibit a high number of cross-border mergers.

We further expect that more mergers within country pairs share a language, religion, law, and history (e.g. Feito-Ruiz and Menendez-Requejo 2011). Country pairs with small differences in education, GDP, economic freedom, and tourism levels are more likely to have similar levels of economic development and attitude towards foreign entities and therefore are expected to experience many cross-border mergers. Instead, country pairs with large differences in taxation may experience more cross-border mergers for managers to transfer profits between countries to reduce the tax bill (e.g. Erel, Liao, and Weisbach 2012; Huizinga and Voget 2009; Pablo 2009). We expect countries with high GDP and high GDP growth to exhibit more mergers. We estimate the absolute distance in currency movements with countries with high distances that may exhibit more acquisitions. Note that we cannot add country-fixed effects since we have one observation per country pair.

3. Determinants of international music preferences

We first offer some empirical validation of international music preferences. We run OLS regressions where the dependent variable is music distance and the independent variables are those discussed in the previous section. Table 4 reports the results. Column 1 shows the results when adding Hofstede’s cultural dimensions as independent variables. We find that five (out of six) parameter coefficients are positive. Three of them are significant. This result offers some indication that citizens in countries with close cultural distance are more likely to listen to each other’s songs.

Table 4. Determinants of music distance.

	Music distance
Dependent variable:	(1)	(2)	(3)	(4)	(5)	(6)
Ln power distance	0.016					0.005
	(1.59)					(0.35)
Ln individualism distance	0.016*					0.004
	(1.79)					(0.39)
Ln masculinity distance	0.046***					0.046***
	(5.63)					(5.65)
Ln uncertainty avoidance distance	−0.007					0.008
	(−0.55)					(0.72)
Ln long term orientation distance	0.026***					0.022**
	(2.83)					(2.24)
Ln indulgence distance	0.017					0.017*
	(1.46)					(1.74)
Ln Facebook connections		−0.051***				−0.038***
		(−7.78)				(−2.88)
Flight time between capitals			−0.001			−0.001
			(−0.36)			(−0.28)
Shared border			−0.069***			0.106**
			(−2.80)			(2.15)
Shared continent			−0.058			−0.024
			(−1.61)			(−0.77)
At least one shared language				−0.108***		−0.077***
				(−4.87)		(−3.36)
Shared religion				−0.055***		−0.034
				(−2.97)		(−1.48)
Shared history				−0.033		−0.008
				(−1.13)		(−0.17)
Shared law				0.049***		0.019
				(3.08)		(0.98)
Educational distance				−0.006		0.001
				(−0.93)		(0.19)
Ln tax distance					0.007	−0.007
					(0.70)	(−0.70)
Ln GDP distance					0.003	−0.001
					(0.32)	(−0.14)
Ln GDP					−0.017	−0.093
					(−0.55)	(−1.17)
GDP growth					0.187	−0.033
					(0.36)	(−0.03)
Exchange rate distance					0.015	−0.069**
					(0.58)	(−2.05)
Economic freedom distance					0.029	−0.027
					(1.41)	(−0.87)
International trade					−0.046	−0.067
					(−1.05)	(−1.33)
Tourism distance					−0.007**	−0.001
					(−2.27)	(−0.24)
Constant	0.132**	0.543***	0.522***	0.497***	0.648*	1.383
	(2.48)	(48.02)	(12.52)	(36.72)	(1.82)	(1.52)
N	153	171	190	190	190	136
R2 adjusted	0.261	0.283	0.097	0.24	0.076	0.576

Note: This table explores determinants of music distance. T-statistics are shown in parentheses. *, **, and *** indicate statistical significance at the 10, 5, and 1% levels, respectively.

Column 2 only reports the relationship between Facebook connections and music distance. The parameter coefficient is significantly negative showing that citizens that share affinity exhibit less music distance, and so are more likely to listen to overseas songs in counterpart countries. Column 3 explores the relationship between geographical measurements and music distance. We find that all three relevant parameter coefficients are negative, and one is significant. So again some indication that the citizens are more likely to listen to songs from nearby countries is in line with the existing literature (e.g. Harnish 2005; Dunbar-Hall 2001). Column 4 shows the relationship between citizens’ similarities (i.e. shared history) and international music preferences. Once again the results indicate reasonable relations considering the literature (e.g. Moberg 2009; Lloyd and Mendez 2001; Besson and Schon 2001; Feld and Fox 1994). We find that all relevant parameter coefficients are negative with three (out of five) being significant. So citizens that share characteristics exhibit less music distance, i.e. more likely to listen to each other’s songs.

Column 5 tests whether there is a relation between economic variables and music distance. As expected no relations appear other than that of significantly negative tourism distance and so citizens that are more likely to visit counterpart countries for holiday are more likely to listen to their songs (e.g. Harnish 2005; Dunbar-Hall 2001). For completion purposes, we report results in column (6) when adding all the independent variables. Some of the relations do not hold in this test but this is not necessarily a concern due to multicollinearity.

Overall, these results indicate that music distance offers some reasonable relations showing that we can test next its relation with cross-border mergers.

4. International music preferences as a determinant of mergers

4.1 Univariate results

We explore in this section the relationship between international music preferences and cross-border mergers. We first report the univariate results as shown in Table 5. In line with the developed theory, we find that the parameter coefficient on music distance is significantly negative at the 1% level. A small distance in international music preferences is thus related to a relatively large number of cross-border mergers. The Chi-square and pseudo-square values are relatively high (5.3%). International music preferences explain well managerial decisions. The only variable that can explain better managerial decisions is Facebook connections (8.6%).

Table 5. Determinants of cross-border mergers: Univariate results.

	Number of overseas mergers / domestic mergers
Dependent variable:	beta coeff.	Constant	N	Chi-square	Pseudo R-square
Music distance	−5.616***	−1.034***	190	104.327	0.053
	(−10.21)	(−4.22)
Ln Power distance	−0.357***	−2.449***	153	8.653	0.020
	(−2.94)	(−7.39)
Ln Individualism distance	−0.452***	−2.145***	153	14.657	0.024
	(−3.83)	(−6.21)
Ln Masculinity distance	−0.279**	−2.544***	153	4.814	0.010
	(−2.19)	(−6.42)
Ln Uncertainty avoidance distance	−0.165	−2.874***	153	1.809	0.003
	(−1.34)	(−7.74)
Ln Long term orientation distance	−0.271***	−2.721***	190	10.704	0.008
	(−3.27)	(−10.86)
Ln Indulgence distance	−0.299***	−2.656***	190	13.144	0.013
	(−3.63)	(−11.19)
Ln Facebook connections	0.676***	−4.624***	171	126.688	0.086
	(11.26)	(−30.10)
Flight time between capitals	−0.142***	−2.829***	190	31.398	0.054
	(−5.60)	(−18.62)
Shared border	1.433***	−3.795***	190	34.639	0.046
	(5.89)	(−34.44)
Shared continent	1.302***	−4.361***	190	34.871	0.043
	(5.91)	(−24.33)
At least one shared language	0.797***	−3.728***	190	10.397	0.017
	(3.22)	(−28.97)
Shared religion	1.233***	−3.866***	190	35.011	0.041
	(5.92)	(−25.74)
Shared history	1.127***	−3.686***	190	13.478	0.024
	(3.67)	(−32.88)
Shared law	0.466*	−3.770***	190	3.591	0.007
	(1.90)	(−18.43)
Educational distance	0.137	−3.716***	190	2.283	0.003
	(1.51)	(−18.13)
Ln Tax distance	−0.029	−3.449***	190	0.037	0.000
	(−0.19)	(−11.72)
Ln GDP distance	−0.176***	−1.917***	190	12.704	0.008
	(−3.56)	(−4.28)
Ln GDP	1.585***	−21.012***	190	10.723	0.024
	(3.27)	(−3.94)
GDP growth	−21.911***	−1.858***	190	16.395	0.019
	(−4.05)	(−4.35)
Exchange rate distance	−0.56	−3.410***	190	1.986	0.003
	(−1.41)	(−24.88)
Economic freedom distance	−1.284***	−2.724***	190	31.322	0.046
	(−5.60)	(−15.01)
International trade	0.385	−3.781***	190	0.805	0.001
	(0.90)	(−11.35)
Tourism distance	−0.022	−3.397***	190	0.678	0.001
	(−0.82)	(−19.57)

Note: This table reports the univariate results of the determinants of the number of overseas mergers divided by the number of domestic mergers. The main independent variable of interest is Music Distance. Z-statistics are shown in parentheses. *, **, and *** indicate statistical significance at the 10, 5, and 1% levels, respectively.

Several other determinants used in this study support their significance. The affinity variable as captured by Facebook connections is significantly positive. Five out of the six cultural distance measurements from Hofstede’s framework are significantly negative. The parameter coefficients for flight distance, shared border, and shared continent are all in the direction of our expectations. The parameter coefficients for language, religion, law, and history are also significantly positive, highlighting the significance of citizens’ similarities behind managerial selections. The parameter coefficients of GDP distance and economic freedom distances are also negative, showing the significance of economic variables. The parameter coefficient of GDP is also significantly positive showing that more acquisitions take place in large economies. We find that the parameter coefficients of 18 out of the total of 24 determinants are significant within the univariate framework.

4.2 Multivariate results

We then undertake multivariate analysis in line with Equation (1) to explore whether the main relation holds within a multivariate framework. Table 6 shows the empirical results. Column 1 reports for easiness once again the univariate result of the main independent variable (music distance), column 2 shows results with the addition of Hofstede's six cultural dimensions, column 3 with the addition of non-cultural dimensions, and finally column 4 with the use of all control variables.

Table 6. Determinants of cross-border mergers: Multivariate results.

	Number of overseas mergers / domestic mergers
Dependent variable:	(1)	(2)	(3)	(4)
Music distance	−5.616***	−3.766***	−2.911***	−3.336***
	(−10.21)	(−3.33)	(−4.12)	(−5.52)
Ln power distance		−0.204		−0.154**
		(−1.22)		(−2.32)
Ln individualism distance		−0.299**		−0.147**
		(−2.42)		(−1.98)
Ln masculinity distance		−0.097		−0.09
		(−0.71)		(−1.47)
Ln uncertainty avoidance distance		0.028		0.108
		(0.14)		(1.02)
Ln long term orientation distance		−0.191*		0.14
		(−1.84)		(1.46)
Ln indulgence distance		0.178		0.093
		(1.38)		(1.29)
Ln Facebook connections			0.459***	0.353***
			(4.49)	(4.38)
Flight time between capitals			−0.081***	−0.099***
			(−3.79)	(−5.54)
Shared border			0.052	0.128
			(0.23)	(0.49)
Shared continent			−0.048	−0.063
			(−0.17)	(−0.25)
At least one shared language			−0.05	−0.061
			(−0.29)	(−0.40)
Shared religion			0.950***	0.978***
			(6.10)	(5.95)
Shared history			0.27	0.189
			(1.28)	(0.78)
Shared law			0.173	0.316**
			(0.95)	(2.24)
Educational distance			0.079	0.036
			(1.50)	(0.69)
Ln Tax distance			0.09	0.146**
			(0.90)	(2.00)
Ln GDP distance			−0.04	−0.049
			(−0.64)	(−0.91)
Ln GDP			0.578	0.042
			(1.54)	(0.08)
GDP growth			−6.643	−5.115
			(−0.81)	(−0.58)
Exchange rate distance			0.561*	0.361
			(1.77)	(1.10)
Economic freedom distance			−0.051	0.209
			(−0.21)	(0.89)
International trade			−1.193***	−1.031**
			(−2.75)	(−2.52)
Tourism distance			0.013	0.039
			(0.51)	(1.64)
Constant	−1.034***	−0.146	−8.112*	−2.059
	(−4.22)	(−0.19)	(−1.78)	(−0.32)
N	190	153	171	136
Chi-square	104.327	43.735	646.429	734.807
Pseudo R-square	0.053	0.06	0.166	0.154

Note: This table reports the multivariate results of the determinants of the number of overseas mergers divided by the number of domestic mergers. The main independent variable of interest is Music Distance. Z-statistics are shown in

Our results offer strong empirical support for the significance of distance in international music preferences as a determinant of the number of cross-border mergers. As expected, we find that the parameter coefficient on music distance remains significantly negative at the 1% level in all columns. The (log) number of overseas mergers in relation to domestic mergers is expected to decrease by 3.336 units with the increase by a unit in the distance in international music preferences, holding all other variables in the model constant. The estimated incidence rate ratio decreases by a factor of 0.036. A one-unit decrease in the distance in international music preferences is thus linked with an increase in the number of cross-border mergers over domestic deals by 96% [(0.036–1)*100]. This relation is thus economically important.

The magnitude of the relation, as expected, reduces in the multivariate framework from −5.616 to −3.336 with the use of all the control variables. Many of the remaining determinants fail to matter that may be driven by multicollinearity. Only eight of the remaining 23 parameter coefficients are now significant: power distance, individualism, Facebook connections, Flight time between capitals, shared religion, shared law, tax distance, and international trade.

4.3 Robustness tests

We undertake several robustness tests for the main relation. The results are reported in Table 7. We do not tabulate hereafter the parameter coefficients of the control variables for space consideration. We first report results when using an alternate estimation from the Poisson used earlier. Column 1 reports results with the use of a negative binomial that has one parameter more than the Poisson estimation by adjusting the variance independently from the mean. We find that the parameter coefficient of the music distance remains significantly negative in this alternate estimation.

Table 7. Robustness tests.

Dependent variable:	Number of overseas mergers / domestic mergers	Abnormal number of overseas mergers / domestic mergers	Ln Number of overseas mergers	Number of overseas mergers / domestic mergers only over $1m deals	Number of overseas mergers / domestic mergers	Number of overseas mergers / domestic mergers	Number of overseas mergers / domestic mergers	Number of overseas mergers / domestic mergers
	Negative
Estimation:	binomial	Poisson	Poisson	Poisson	Poisson	Poisson	Poisson	Poisson
Different treatments on independent variables:	None	None	None	None	Ln Music distance	Add: Ln Number of listed firms	No logarithm used	Orthogonalized
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Music distance	−3.336***	−0.957***	−1.037***	−1.420***		−3.779***	−3.233***	−0.609***
	(−5.52)	(−7.04)	(−4.06)	(−4.05)		(−4.62)	(−4.51)	(−7.82)
Ln music distance					−1.166***
					(−5.05)
Ln number of listed firms						−0.099
						(0.39)
Previous controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−2.059	−2.413**	−6.318**	−5.29	−3.92	−1.516	−2.248*	−4.928
	(−0.32)	(−2.40)	(−2.41)	(−1.49)	(−0.58)	(0.82)	(−1.75)	(−0.77)
N	136	136	136	136	136	136	136	136
Chi-square	734.807	904.275	598.370	406.830	693.621	723.560	665.809	734.807
Pseudo R-square	0.154	0.065	0.136	0.182	0.153	0.154	0.151	0.154

Note: This table reports several robustness tests for the main relation as shown earlier in Table 6. Z-statistics are shown in parentheses. *, **, and *** indicate statistical significance at the 10, 5, and 1% levels, respectively.

In columns 2 and 3, we measure our dependent variable differently. Column 2 uses the difference in cross-border mergers per country pair i.e. the abnormal number of cross-border mergers divided by the equivalent number of the number of domestic mergers; 2000–2018 minus 1980–1994. This test intends to ensure that the relationship is not the outcome of previous agreements and the easiness of undertaking cross-border mergers within any particular country pair (e.g. Liu and Maula 2015). Instead, column 3 uses the natural logarithmic number of cross-border mergers without the adjustment on the number of domestic merger deals. We find that our conclusions remain unchanged.

In column 4, we undertake the analysis in a subsample of our sample. We analyze here only those mergers that we know are large, with an equity value in the merger announcement of at least $1m. The data source does not offer the size of the bidder and target firms on many occasions. It is also often not available the magnitude in $ terms of the merger announcements. For these reasons, we have included all available merger deals in the analysis earlier. For robustness, we report here the results for deals that we know are of a significant size (at least $1m). We find that the parameter coefficient of music distance remains significantly negative in this subsample of mergers.

In columns 5–7, we use alternative transformations of our independent variables and also add one more control variable in the estimation. Column 5 shows results when estimating the main relationship with the use of the natural logarithm on Music Distance. We previously did not make a relevant adjustment because music distance does not exhibit extreme values. We find that the parameter coefficient on lnMusic Distance is significantly negative. Column 6 reports results when we also control for the total average number of listed firms in each country pair.¹⁷ We are based on the analysis of public-to-public deals considering that these are the most important deals, especially within the cross-border context of this study. However, this may lead to a deal count that will be higher for economies with a more developed stock market that may generate our results. We invalidate here this explanation by showing that the parameter coefficient on Music Distance remains significantly negative after controlling for the number of listed firms in our country pairs. Column 7 also reports results when using no natural logarithm in none of our independent variables. We used earlier logarithms on variables that exhibit extreme values to ensure that outliers do not drive our results. Once again the parameter coefficient on music distance remains significantly negative. These results indicate that data transformations do not influence previous conclusions.

Finally, column 8 reports results when orthogonalizing variables with correlations higher than 0.5 (Lee and Makhija 2009). This is to ensure that multicollinearity does not influence significantly our results. Once again, we find that the parameter coefficient of music distance remains significantly negative at the 1% level. Overall, the main relation holds within various tests.

4.4 Placebo tests

Finally, we undertake two placebo tests to offer further assurance of our results. Results are available in Table 8. Column 1 explores the relation between music distance that is based on data after 2000 in relation to the number of overseas mergers adjusted by the number of domestic deals on data between 1980 and 1994. Since the timing of these two variables does not coincide, we expect no relation. Column 2 explores the relation between music distance and the natural logarithm on the total number of domestic merger deals for each country pair (i.e. without considering the number of their overseas mergers). Once again no relation is expected. We indeed support our expectations. We find that the parameter coefficient on music distance is insignificant in both estimations.

Table 8. Placebo tests.

	Number of overseas mergers /
	domestic mergers	Ln number of
	1980–1994	domestic mergers
Dependent variable:	(1)	(2)
Music Distance	0.247	−0.108
	(0.29)	(−1.32)
Previous Controls	Yes	Yes
Constant	−6.929	−2.037***
	(−0.90)	(−3.38)
N	136	136
Chi-square	1005.686	747.689
Pseudo R-square	0.178	0.033

Note: This table undertakes two placebo tests for the main relation shown earlier in Table 6. Z-statistics are shown in parentheses. *** indicates statistical significance at the 1% level.

5. Conclusion

We argue in this study that international music preferences capture the cultural distance between nations. Music preferences reflect both music production and music reception that are influenced by the producers’ and citizens’ current experiences and ways of life. The lyrics, melodic structure, musical instruments, and language are all part of someone’s background and way of thinking. People also do not experience music similarly, but their response to music is subject to their experiences, national identity, and emotional state. Citizens are likely to listen to songs from countries that share a similar culture. To develop international music preferences, there is no need for expensive surveys and thus one could easily update its values, especially with the continuous improvement of technology. The music downloads are undertaken without much consideration and are based on millions of music downloads.

We use the context of the cross-border merger and explore the role of international music preferences as a determinant. As hypothesized, we find that international music preferences are a major determinant of managerial decisions in cross-border mergers. Citizens in country pairs that listen to each other songs exhibit relatively more cross-border mergers highlighting the role of music in international business.

Unfortunately, we do not have access to the demographics of the participants who use LFM-1b, being likely young individuals. Their age may not necessarily be a problem for the validity of our study, though. Sensitivity to music starts in one’s early years (Schellenberg 2005). As argued by Chatman and Flynn (2001), young people are also the ideal target group to explore a group’s identity before life experiences make participants differentiate themselves from the group. Also, international music preferences are not likely the best to explore short-term relations since a piece of music ‘hit’ may influence downloads without necessarily a real change in a country’s culture. This may not be a concern considering that culture normally changes slowly over time in most countries (e.g. Kirkman, Lowe, and Gibson 2006), and so there is less of a need for high-frequency analysis in this field.

Disclosure statement

No potential conflict of interest was reported by the author.

Data availability statement

The data that support the findings of this study are available from the author upon reasonable request.

Additional information

Notes on contributors

Antonios Siganos

Antonios Siganos is Professor in Finance who has undertaken research work mostly in the fields of international finance and behavioral finance.

Notes

1 Siganos and Tabner (2020) empirically support the significance of societal affinity on cross-border merger decisions by using the abnormal voting bias in the Eurovision Song Contest. Voting bias in the Eurovision Song Contest is not notionally related to international music preferences that this study explores. The UK music industry is for example dominant globally and so in line with this study’s storyline exhibit high international music preferences. However UK songs tend to perform poorly the last 20 years in the Eurovision Song Contest as a result of low societal affinity.

2 https://www.telegraph.co.uk/news/2018/03/26/gangnam-style-soft-diplomacy-much-north-korea/

3 Theoretically, international music preferences should not change significantly every year. The use of average music preferences over the sample period thus reduces the impact of any album ‘hits’ in any particular year. This is to an extent a limitation of this measure. This is a common drawback to almost all previously developed cross-border determinants.

4 Australia, Belarus, Brazil, Canada, the Czech Republic, Finland, France, Germany, Italy, Japan, Mexico, the Netherlands, Norway, Poland, Russia, Spain, Sweden, Ukraine, the United Kingdom, and the United States. Note that for the main analysis later that requires data availability for all variables, Belarus, Russia, and Ukraine are excluded due to missing observations mainly on Hofstede’s cultural dimensions.

5 We also explore results at Table 7 later for merger deals with an equity value at announcement of at least $1m. We find that our results hold in this subsample.

6 At Table 7 we also report results when using the number of cross-border mergers (without the adjustment of the number of domestic deals). We find that our conclusions hold.

7 7% of our country pairs exhibit no mergers during the sample period.

8 https://geerthofstede.com/research-and-vsm/dimension-data-matrix/

9 We do report the main results without the use of log at Table 7 later. We find that our conclusions are unchanged.

10 https://data.humdata.org/dataset/social-connectedness-index

11 https://www.freemaptools.com/how-far-is-it-between.htm,

https://www.airportdistancecalculator.com/flight-time-between-two-cities.php#.Y-JarC_P2Ul

12 https://www.cia.gov/library/publications/the-world-factbook/

13 https://data.worldbank.org/indicator/IC.TAX.PRFT.CP.ZS, https://data.worldbank.org/indicator/ny.gnp.pcap.pp.cd

14 https://www.heritage.org/index/

15 https://ourworldindata.org/trade-and-globalization#two-centuries-of-trade-country-by-country

16 https://datacatalog.worldbank.org/public-licenses#cc-by

17 Relevant data is collected from the World Bank by averaging the total number of listed firms per country pair between 2000 and 2018. https://data.worldbank.org/indicator/CM.MKT.LDOM.NO.