How local conditions affect global banking: the case of BBVA and Santander °

This paper explores why Spanish banks internationalise and why Latin America has been the main region for the international expansion of BBVA and Santander. It shows that prior to 1986 Spanish banks had a limited presence abroad, and analyses the main drivers of this initial expansion (remittances and trade connections). However, from 1986 on, there was a confluence of domestic and external factors (economic and regulatory changes in Latin America) that encouraged the international forays of BBVA and Santander. The fact that changes in the Spanish and the Latin American financial sectors occurred just when other transnational banks were turning their attention to other regions created the optimal conditions for the expansion of Spanish banks in Latin America. JEL Classification: G15, G21, N26

The material presented in eabh Papers is property of the authors and should be quoted as such. The views expressed in this working paper are those of the authors and do not necessarily represent the views of the eabh or its members.

Introduction
Since the 1990s, BBVA and Banco Santander have undergone a huge transformation; they have gone from being two of Spain's largest banks to two of the world's biggest. The aim of this paper is to explain why Latin America has been the main region for the international expansion of these two banks. Although we focus on the period of major internationalisation occurring from 1990 onwards, we adopt a long-term perspective to study the previous history of internationalisation, the roots of this expansion and the main changes occurring in the 1990s that encouraged the internationalisation process of both banks. This paper argues that the timing of the changes in the Spanish and Latin American financial sectors, occurring just when other transnational banks were focusing their attention on other regions, created the optimal conditions for the natural but also unexpected (in term of size and scope) expansion of BBVA and Banco Santander in the Latin American region.
Although there are many cross-country studies focusing on different aspects of banking internationalisation, specific case studies could be very helpful in providing greater insights into the general results obtained by those cross-country analyses (Mulder and Westerhuis 2015). Case studies allow the researcher to examine in more detail the factors that encourage foreign banks' entry into new markets: the "follow the customer" hypothesis, regulatory changes, local market opportunities or links between foreign bank entry and other foreign direct investments, among other factors. Taking a long-run perspective, this paper discusses different hypotheses about the factors that affected the pattern and timing of Spanish banks' entry into the Latin American region. In particular, the paper focus on the role of remittance collection and economic integration as proxies for the "follow the client" approach, the local opportunities in the Latin American region and the changes in host country regulation. Some bank characteristics that affect foreign expansion decisions, such as size and profitability along with changes in Spanish regulation, are also considered. Finally, the paper discusses why other multinational banks did not compete (or competed to a lesser extent) for the Latin American market at this point in time, a situation that was exploited by the Spanish banks.
The structure of the paper is as follows. Section 2 presents the main factors identified in the literature as drivers of bank internationalisation. Section 3 explains why the Latin America region is the natural choice for the Spanish banking sector's expansion, while Sections 4 and 5 discuss the different historical phases and features of BBVA's and Santander's banking expansions. Although the Latin America region was definitely the natural choice for Spanish banks' internationalisation, we detail the factors that accelerated and give an unexpected boost to this expansion. Section 6 discusses the strategy of Spanish banks as compared to other transnational banks in the Latin American region. The conclusions are presented in Section 7.

Why banks internationalise
As Mulder and Westerhuis (2015) point out, banks have had an international presence ever since the Middle Ages. In the nineteenth century, there was a process of banking globalisation led by some of the big merchant houses such as Barings, JP Morgan and Rothschild, during which time these banks increased their foreign presence, mainly in the colonies. However, foreign banking investments declined in the interwar period and the process of banking internationalisation only re-emerged in the 1970s, before accelerating in the 1990s (Focarelli and Pozzolo 2005).
When analysing the banking internationalisation process, it should be noted that financial products have different characteristics from manufactured products. For example, financial services and products are information intensive. In order to evaluate their operations, banks need to collect abundant data about their clients/customers and the local markets. Banks do not incur physical transportation costs in offering their products and services, but they do need to invest in some fixed costs, for example in offices (Buch and Lipponer 2006). A bank that decides to expand into other countries has to face certain costs related to the cultural, political, fiscal and economic differences, which mean that it has less knowledge about the host market than local competitors do, and coordinating its activity becomes more complex.
However, this international expansion also has some benefits. There are a number of key advantages of banking internationalisation often cited in the literature. The first is that it enables banks to grow beyond the limits of the home market and achieve economies of scale and scope (Tschoegl 2004). As Grubel (1977) suggests, banks not only try to serve their corporate clients/customers that move abroad in search of new business (the follow-the-client motive) but also go abroad to capture new clients in countries with underdeveloped financial systems and a low level of financial intermediation (an offensive expansion). Banks seek out new markets that are more profitable or offer higher margins (Aliber 1976). Moreover, there is a two-way relationship between size and internationalisation: banks expand into other countries in order to grow; and large banks are much more international than small ones, because they have a larger and more international customer base, involved in activities that are typically international (Focarelli and Pozzolo 2005). The second advantage is risk diversification. Banks attempt to diversify their portfolio internationally and use foreign markets to compensate for the fluctuations in their home markets. Thirdly, internationalisation allows banks to exploit the resources and capabilities that they have developed in their local markets, and employ them in new emerging markets, at a relatively low cost (Mulder and Westerhuis 2015). Lastly, moving abroad helps banks to respond to the challenge of domestic rivals (a defensive expansion); in this case, the internationalisation strategy depends on the behaviour of competitors (Grubel 1977).
There are certain elements that have acted as push factors in banking globalisation processes, such as capital account liberalisation, financial deregulation, falling communication costs, financial innovation, and trade and financial integration 1 . Cull and Martinez Peria (2010) differentiated between four types of factors that act as drivers of foreign bank entry: 1) the "follow the client " approach or the desire of the bank to continue serving its clients in their overseas operations , b) profit opportunities in the host country, 3) the absence of barriers to foreign bank entry in the host country and 4) other factors that reduce the information cost of doing business abroad, such as a common language or cultural similarities between the home and the host countries. Moreover, the internationalisation process is also influenced by certain characteristics related to the size, efficiency and performance of the banks that expand abroad (Clark et al. 2003). As mentioned above, big banks are more likely to have multinational operations and customers that operate abroad and, therefore, they are more likely to follow their clients abroad. Moreover, banks need a minimum size in order to start their international forays and compete in global markets. Several studies, such as Tschoegl (1983), Grosse and Goldberg (1991) and Focarelli and Pozzolo (2000), have found a positive correlation between bank size and internationalisation. Efficiency and bank performance could also affect banking expansions abroad although the empirical evidence is not conclusive as to whether domestic banks are more efficient than foreign banks, or vice versa (Clarke et al. 2003). Whereas Dermiguc-Kunt and Huizinga (2000) and Claessens et al. (2000) show that foreign banks had lower profitability rates than local banks in developed countries, Kiraly et al. (2000) report that large foreign banks rank are among the most efficient banks, and Focarelli and Pozzollo (2000) find that profitability (measured as the bank's return on assets) is correlated with the international expansion. Moreover, empirical research about the main drivers of banking internationalisation finds that the size of the home country and its level of economic and financial development are two key determinants of banking internationalisation: countries which have a more saturated home market and which have made the most progress in financial development have a higher level of banking internationalisation (Lane andMilessi-Ferretti 2008, Mulder andWesterhuis 2015).
Finally, home country regulations can also determine which banks internationalise.
Local regulatory restrictions or, alternatively, financial deregulation processes that increase competition, could incentivise the bank to increase its presence in other countries. Other noneconomic aspects, such as linguistic and cultural similarities, are also correlated with the internationalisation pattern of banks, because such factors lower the information costs of doing business in a new market (Buch 2003, Focarelli andPozzolo 2005).
Most of the research on the factors that encourage banks' entry into foreign markets is based on cross-country evidence and shows that economic, political and regulatory conditions are crucial to an understanding of the pattern and timing of banking internationalisation (Barth et al. 2001Clarke et al. 2003, Cerutti, Dell'Ariccia and Martínez-Peria 2007. However, although the advantages of cross-country analysis are evident, it is also important to consider some specific case studies to obtain evidence about how local conditions in the host countries could act as incentives for attracting foreign banking investments and to contrast whether the general conclusions stemming from cross-country analyses hold true for a specific case study.

The internationalisation of Spanish banks: LatAm as the "natural" choice for their international expansion
To analyse the main factors behind the international expansion of Spanish banks, it is necessary to draw a general picture of the main phases of Spanish bank internationalisation and thereby explore the changing drivers that have operated in each of these phases. In his study of the Santander expansion, Martín-Aceña (2010) made a distinction between the pre-1980 period and the period from 1980 onwards. In their paper about the BBVA and Santander international expansion, López-Morell and Bernabé Pérez (2018) differentiated between the pre-1980 period, the 1990s and the twenty-first century. Sierra (2007) identified two separate periods: a) from 1900 to 1992 and b) from 1993 to 2007. For the most recent banking globalisation period, Parada et al. (2009) distinguished between the late 1980s, the 1990s and from 1999 onwards.
Each of these phases shows different characteristics and, as we explain in the following pages, the drivers that have influenced Spanish banking globalisation have historically changed alongside the shifts in international and external conditions. In order to study the timing and drivers of Spanish banking expansion, we differentiate between the pre-and post-1986 periods. In our empirical analysis of the factors that have encouraged the international expansion of Santander and BBVA, we concentrate on a number of aspects that the theoretical and empirical literature have identified as the main determinants of banking internationalisation: a) the economic integration between the home (Spain) and the host countries as a proxy for the "follow the client" approach, b) the profit opportunities in the host countries, 3) the host country regulations, 4) other factors that could affect the pattern of expansion such as a common language or cultural similarities between the home and the host countries, 5) the bank characteristics that affect their foreign expansion decisions, in particular, size and profitability, and 6) the changes in regulation in the home country. As we show in the next section, Spanish banking internationalisation was a dynamic process motivated by multiple domestic and external factors (economic, political and regulatory) that made LatAm the main region for Spanish banks' expansion. We maintain that the approaches of other competitors (the US, British, French, German or Japanese banks) left Spanish banks with greater room for manoeuvre in their globalisation process during the 1980s and 1990s.

Before 1986: the initial phase
In the pre-1986 period, particularly the period up until 1970, Spanish banks had a limited presence abroad. They extended their overseas activities through representative offices, but these offices did not deal with local clients, the volume of investment abroad was low, banks took small risks, and the profits linked to these foreign operations did not represent a substantial share of their total profits. An important question is why Spanish banks did not adopt a more vigorous international strategy during this period. There are a number of potential explanations. Firstly, internal conditions in Spain did not favour overseas expansion. Throughout most of these years (especially after the Spanish Civil War), the Spanish banking sector had high barriers to entry, was strictly regulated and had a relatively low level of competition. In this context, banks achieved high rates of profitability and interest margins and thus were not incentivised to look for new markets 3 (Pons 2002, Pueyo 2003. Therefore, until the 1980s, the very protective domestic regulation, which ensured high profits, discouraged foreign banking expansion. Secondly, the size of Spanish banking firms was not favourable. As Canals (1991) and 3 The average rate of profitability of Spanish banks in nominal terms between 1960 and 1975 was around 20 per cent, nearly 12 per cent in real terms (Pueyo 2003) Cardone-Riportella and Cazorla-Papis (2001) mention, the small size of Spanish firms prevented them from expanding internationally any earlier 4 (see Table 1). Only when Spanish banks had reached a sufficient size did they have enough financial resources to operate in international markets. External conditions (or local conditions from the perspective of the host countries) did not stimulate Spanish banking investment overseas. Until the 1980s, most of the potential host countries (such as those in LatAm) had high barriers to entry that deterred foreign investment, and there were major interventions in their financial systems (Table 2). Although in the last decades of the nineteenth century and the first decades of the twentieth century, foreign banks had a strong presence in certain LatAm countries such as Brazil and Mexico, the situation changed over the course of the twentieth century, especially after the Great Depression. In Mexico, for example, following the Mexican revolution, and also as a consequence of the changes in the financial context linked to the First World War and the monetary instability during the 1920s, there was an increase in risk and a decline in banking profitability that made foreign stockholders lose interest in the Mexican business. Moreover, the situation deteriorated after the 1931 Banking Law, which subjected foreign commercial banks to severe legal constraints (Marichal and Trainer 2001). In fact, from 1931 to 1981, there were virtually no foreign banks in Mexico. This situation was typical in most LatAm countries, which, from the 1940s until the 1980s, had very interventionist banking regulations. This was the case not only in Mexico but also in Argentina and Brazil, among others. In Argentina, for example, the state interventionism in the banking sector that began in the 1940s reduced foreign banking activity 5 .
Moreover, until the 1990s, the state-owned banks in LatAm held more than 50 per cent of total banking assets 6 (Tovar 2007). In Brazil, the banking system was nationalised during the 1920s, after which time the activity of foreign banks was extremely limited (Cortes and Marcontes 2016). Other countries nationalised banks during the 1940 such as the abovementioned Argentinian case in 1946 or Costa Rica, which nationalised the banking sector in 1948. In the 1980s, there was a new wave of nationalisations, such as that in El Salvador (1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990) or Mexico (1982Mexico ( -1990 (Del Angel and Martinelli 2013). In this context, with very restrictive regulation and a strong presence of state-owned banks 7 , the incentives for foreign banks to enter the region were weak. Most LatAm financial systems in the second half of the twentieth century can be characterized as repressed by severe institutional arrangements that 5 In Argentina, private banks were nationalised in 1946, during Perón's first presidential term; later they were denationalised during the "Revolución Libertadora" before being re-nationalised in 1973 (Rapaport, 2010). 6 However, there were substantial differences among countries. For example, whereas in Brazil the share of stateowned banks in 1994 was 48 per cent, in Chile it was only 14 per cent (Goldstein and Turner, 1996). 7 There was also an attempt to nationalise the banking sector in Peru (1987) but it was declared unconstitutional (Machuca 2017). In other countries there were short periods of nationalisation, such as that in Chile between 1970-73, and 1981-82, or in Venezuela in 1995 as a consequence of the banking crisis, but Venezuelan banks were re-privatised in and 1997(De Krivoy 2003. forced public and private agents to access foreign debt markets in order to meet their financing needs. Hence, financial repression acted as a restrictive factor for financial sector development.  Marichal and Trainer (2001); for Brazil, Cortes and Marcondes (2016); for Argentina, Rapoport (2010) and Jones (1995); for Chile, Drake (1984) and Gozzi and Tappatá (2010); for Colombia, Forero-Laverde (2017), Kalmanovitz (2010), Ocampo (2015), and Gozzi and Tappatá (2010). Zegarra (2014) affirms that the low level of development of the LatAm financial system and the minimal presence of foreign banks in this region was also linked to the low demand for banking services and the underdevelopment of the financial system (Table 3). The economic and political instability of this region throughout most of the period may also have been a limiting factor for Spanish banks, in particular the economic difficulties of LatAm countries in the 1970s and the "lost decade" of the 1980s (Batiz-Lazo et al. 2007). In fact, during the 1980s, there were systemic banking crises in Argentina, Chile, Colombia and Uruguay, and there was also banking distress in Ecuador, Bolivia and Mexico (Morris et al., 1990).  There were two main drivers of banks' internationalisation in this period: to provide a service to their clients overseas and to use the experience accumulated in the domestic and first experiences in the international markets in the later expansion stages. To analyse the followthe-client motive, we consider the importance of emigration and trade conditions. In the pre-1986 period, one of the key drivers of Spanish banking internationalisation was the remittances sent home by Spanish emigrants (Sierra 2007 However, during the Civil War and the years immediately after, there was significant migration for political reasons (the migrants known as "exiliados"), and LatAm was again the main destination, particularly Mexico, the Dominican Republic and Chile (Abellán 1983) 9 . There was also a new wave of emigration to LatAm from 1946 to 1958, this time motivated by economic conditions, with Argentina (35.1 per cent of total migrants) and Venezuela (32 per cent) being the most attractive destinations, followed by Brazil (10.9 per cent). The total number of migrants in this period was estimated at more than 560,000 (Sallé 2009 According to Esteve and Khoudour-Costeras (2009) Unfortunately, as discussed above, although we have qualitative evidence on the importance of emigration for banking expansion in LatAm, we cannot quantify Spanish banks' remittance activities.
Another possible determinant of financial expansion is trade integration. As Lane and Milessi-Ferretti (2008) indicate, trade openness and financial integration go hand in hand.
Until the 1960s, Spain was a closed economy and the low level of commercial integration did not stimulate foreign banking activity.  Tena (2005) Although Spanish banking investments abroad were concentrated in LatAm, from 1900 to 1984 this region's share of total Spanish trade fell in comparison to other regions (Table 5).
In their early years, some Spanish banks, such as Banco Hispano Americano and Banco Santander, were linked to LatAm and Spain's ex-colonies (serving the indianos or Spaniards who made their fortune across the Atlantic) (Tortella and García Ruiz 1993) but later on this market became less commercially attractive. As can be seen in Table 5 that allowed them to achieve the size they needed to expand abroad. Both banks adopted a strategy of entry via acquisition rather than via greenfield operations. The main difference between the two was that Banco Santander took majority stakes in the banks in question, with the objective of ensuring full managerial control 12 , whereas BBVA started with minority stakes, which it gradually increased over time (Guillén and Tschoegl 1999). Banco Santander and BBVA have specialised in retail banking, in contrast to other foreign banks such as Citibank, which focus on a minority of the population: high-net-worth households and investors (Citi Report, 2017).
The geographical pattern of Santander and BBVA locations in LatAm in the post-1986 period is shown in Table 6. Santander's activity is distributed among different countries (Argentina, Brazil, Chile and Mexico) whereas BBVA's activity is more concentrated in Mexico, although is has also presence in Argentina and Brazil. It is interesting to observe how Banco Santander and BBVA show different patterns of geographical distribution, although they do compete in some countries.  Martín-Aceña (2007, 2008, Calderón and Casilda (2000), and Annual reports of Santander and BBVA Several factors facilitated this international expansion. The Spanish banking system underwent a process of domestic de-regulation in the late 1980s. As mentioned above, there was limited competition among Spanish banks until the 1970s. Interest rates were regulated and banks used other mechanisms to compete, such as free commissions or opening more branches. Moreover, during the Franco period, regulation limited the entry of foreign banks.
Banking liberalisation started in the 1970s with the elimination of interest rate controls, branch liberalisation, and the abolition of the compulsory investment coefficients. As a result, competition intensified. This process was first accelerated by Spain joining the European Economic Community (EEC) in 1986 and later by its membership of the euro area, as this meant that Spain had to adapt to European regulations. In 1989, there was a complete liberalisation of interest rates and commissions, a number of stock market reforms were introduced, and new measures were passed to regulate pension funds. The entry of foreign banks was liberalised, although these banks faced serious difficulties in expanding because of the large number of branches of Spanish banks. Moreover, competition increased not only among banks in general but specifically between banks and saving banks (Caminal et al.1990;Salas and Saurina 2003). The problem was that this increased competition took place in a saturated banking market (see Table 7) highlighting an evident need to look for new markets.  This increase in competition triggered a necessary reaction from Spanish banks, which were forced to adopt a more market-oriented business strategy to increase their profits. The first banking reaction was to increase bank size through mergers 14 . Secondly, as the adoption of the euro eliminated some traditional sources of revenues, such as currency trading or currency exchange commissions, banks were forced to offer new products other than deposits.
These included investments, pension funds or, later on, highly complex financial products, such as derivatives, which increased the cost of funds and had a negative impact on profitability. They also tried to find other sources of revenue, such as commission for services rendered, or reallocation of loan activities to more profitable sectors (Saurina 2012), as well as restructuring their branch networks (Casilda 2016). Lastly, they invested in R&D to increase efficiency by introducing new services (cash machines, phone banking…) and also to manage data, information, risk and marketing. 15 .
The main reaction of Spanish banks to this increasing competition was international expansion. The merger processes facilitated the banking sector's international expansion. For example, the acquisition of Banesto by Banco Santander in 1994 put the resulting bank in a good position to expand overseas. Although Spain is in Europe and the other European countries are its main trading partners-even more so after Spain joined the EEC and later the 13 The average net interest margin from 1990 to 1994 was 3.1 per cent in Spain whereas it was 9.2 per cent in Argentina, 8.3 per cent in Colombia, 8.1 per cent in Venezuela and 6.8 per cent in Brazil (Guillén and Tschoegl 1999). 14 The result was the creation of big banks such as BBVA in 1991 or Banco Santander Central Hispano (BSCH) in 1999. 15 Banco Santander led the way in these new strategies by offering the Super Account, which was launched in 1989 with a massive advertising campaign, and which paid 11 per cent interest -more than double the interest rate paid by other banks at that time. As Parada et al. (2009) explain, Banco Santander also developed a system for better monitoring credit risk that allowed prompt identification of bad debts, and the bank made an extraordinary effort to strengthen its retail businesses, for example, by encouraging clients to directly deposit their wages in their account and by identifying its clients' needs though information about their spending and income patterns.

European Union (EU)-the main receiving region for Spanish banks' investment abroad was
LatAm (Tables 8 and 9). In some countries, such as Colombia and Chile, Banco Santander and BBVA have a large presence. Source: Calderon and Casilda (2000) -Papis, 2001;Casilda, 2016). Thirdly, in the post-1986 period and especially at the turn of the century, certain market conditions made LatAm countries a favourable location for Spanish banking expansion.

Potential growth of the region and economic cycle
After the lost decade of the 1980s, the LatAm region experienced a relative improvement in terms of growth (see Table 10). Although the growth rates were not sustained and there were differences among countries, they came in marked contrast to the situation in the previous period (Minda, 2007). Banco Santander redefined its strategy from 1986 on by selling four entities in Costa Rica, Guatemala, Ecuador and Panama, and concentrating its activity in Argentina, Uruguay, Chile and Puerto Rico. In the second half of the 1990s, in a better macroeconomic climate than in the previous decade, Banco Santander, led by Santander Investment, accelerated its expansion in LatAm. Its investments in this region rose from 9 per cent of total assets in 1990 to 24 per cent in 1998, representing 30.6 per cent of total profits.
From 1994 to 1998, Banco Santander invested 5 billion euros into the LatAm region. BBVA also achieved a major business expansion in LatAm in the 1990s, with Mexico being the country that received the biggest investments from BBVA in this decade. At the turn of the twenty-first century, economic conditions in LatAm were more favourable, with more positive expectations 16 . Growth was accompanied by greater macroeconomic stability. Expectations were also positive about the expanding middle class.
LatAm witnessed significant population growth, with a very young population, unlike developed countries with their aging populations (Calderón and Casilda 1999). As mentioned, Spanish banks focused on the mass market (lower-and middle-income segments), in contrast to other foreign banks such as Citibank, which concentrated on the upper-income market. Moreover, the region had a very underdeveloped financial system (Table 11) with considerable potential for the expansion of bank branches and introduction of financial products. The percentage of the population with a savings account, credit card, or personal loan was very low. At that time, the banking penetration rate was 95 per cent in Spain and 35 per cent on average for LatAm (see Table 8). This also implied a relatively low level of competition in the LatAm banking sector (Guillén and Tschoegl 2007). Moreover, there were some profitable business niches such as the mortgage market and pension fund management  Finally, in terms of diversification, the LatiAm region represented a complementary area for the Spanish banks: whereas the Spanish economic cycle is highly correlated with that of the euro zone, it was negatively correlated with LatAm. A clear example of this complementarity is the positive evolution of Spanish banks' profits in LatAm after the 2008 crisis, a crisis which negatively impacted the Spanish market. Therefore, Spanish banks were attracted not only by the potential growth and opportunities the region offered, but also because it had a complementary cycle with respect to Spain. These reforms also included the privatisation of the state-owned banks (Chile 1985-89, Mexico 1991-92, Argentina 1994, Brazil 1998-1999. In this sense, case studies can provide interesting insights into how banking privatisation facilitated the entry of foreign banks in those years and the differences in terms of banking privatisation among LatAm countries. For example, Brazil's privatisation process was less intense than other countries in the region;

Capital controls
indeed, the two largest banks in terms of assets, Banco do Brasil and Caiza Economica Federal, were still public in 1995 and as late as 2015, respectively (Cortés and Marcondes 2016 This de-regulatory process also coincided, however, with the withdrawal of the US banks from the LatAm market, partly due to the impact of prudential limits on bank exposure to specific risk in some countries such as Brazil, Argentina and Mexico. Later on, there was a wave of "second generation reforms" that were introduced in the second half of the 1990s as a consequence of the problems of instability that emerged in those years. These reforms were mainly introduced to strengthen the mechanisms of control and supervision and to facilitate new areas of banking activities and operations. These regulatory changes encouraged the entry of foreign banks in LatAm (Correa 2004).
Finally, although the situation differs substantially among countries, the banking deregulatory process in the 1980s coincided with a rule of law movement that increased legal certainty. Although legal reforms failed on many fronts and overall progress was modest at best, a governance index (which includes rule of law, the expropriation risk and the repudiation of contracts by government, among other variables) shows a decline in values from 1985 to the late 1980s, but a recovery from 1990 to 1995 in countries such as Argentina, Chile, Ecuador and Peru (Loayza and Palacios 1997).

Price and returns
The pattern of internationalisation was also influenced by a price factor: the valuations of European banks were much higher than those of LatAm. Therefore, buying LatAm banks was a cheaper way to increase the size of the parent bank than buying European ones (Guillén and Tschoegl, 2000).
Additionally, the LatAm region offered increasing returns. The net interest margins were considerably higher than in Spain. For example, from 1990 to 1994 they were above 5 per cent in LatAm (5.1 per cent in Mexico and 9.2 per cent in Argentina) and only 3 per cent in Spain (Goldstein and Turner 1996)

Other Spanish foreign investments
The LatAm region was not only the main destination for banking investments but also a key target of other sectors. The first big Spanish investment in LatAm was by Telefonica, followed by Iberia (via privatisations), while other firms such as Endesa, Iberdrola and Repsol also expanded in this region. The massive entry of multinationals into LatAm during the 1990s acted as an incentive to the banks that followed them to provide financial services (Minda 2007). 20 The countries with the highest interest margins between 1990 and 1994 were Argentina (9.2), Colombia (8.3) and Venezuela (8.1), followed by Brazil (6.8), Chile (6.1) and Mexico (5.1) (Goldstein and Turner 1996).

Other factors: some recent events
The entry into LatAm countries provided a stepping-stone for expansion into other countries. Firstly, it allowed Spanish banks to increase their size in order to compete in the international arena. Secondly, it offered a geographical advantage in that it facilitated the later expansion in the US market, particularly in Hispanic-majority areas (where many people, moreover, knew the Spanish banks from their countries of origin). However, after the Great The 2008 crisis has, therefore, forced banks to become more selective in terms of international expansion and portfolio decisions. As Claessens and van Horen (2012) point out, banks in advanced economies (including BBVA and Santander) are less likely to engage in particularly active foreign expansion in the future, whereas banks from emerging economies are more likely to play an increasingly important role as foreign investors. The forthcoming challenges for the banking industry in terms of stricter regulation, technological change and increasing competition will require a change in banks' strategies. The future evolution of the LatAm market will be a crucial issue as BBVA and Santander redefine their strategies.
All of the abovementioned factors that fostered Spanish banking expansion in LatAm occurred at the same time as other transnational banks were focusing their attention on other regions. Given the importance of this aspect to the Spanish banking expansion strategy, we discuss it in detail in Section 6.

Spanish banks versus transnational banks in LatAm
Most of the abovementioned changes in LatAm conditions incentivised Spanish banks' entry into the market, but a key question is why all these factors did not attract other transnational banks in the same way. As Galindo, Izquierdo and Rojas-Suarez (2013)  on other potential markets. This cleared the way for Santander and BBVA to expand in the region.

Conclusions
Although the internationalisation process of BBVA and Santander intensified from the 1990s onwards, some of the banks that had been merged to form these two banks had a previous history of internationalisation. Santander and BBVA started their international expansion after: a) consolidating their positions in Spain and b) a process of mergers that allowed them to achieve the size they needed to expand abroad. In both cases, Latin America was the first area they expanded into, and this region has been crucial for their internationalisation. However, Santander's process of internationalisation started before that of BBVA (or the banks that formed BBVA). In the pre-1986 period, Santander had a much more notable presence in Latin America than BBVA. Santander's initial entry was aimed at providing its clients with the necessary services to carry out their operations abroad ("follow-the-client") but from the 1960s and especially in the 1970s, the bank looked for new growth opportunities and acquired some local banks in Latin America. There is fairly limited quantitative evidence on the activities carried out by the banks in Latin America in the pre-1986 years; in particular, although there is qualitative evidence about the importance of remittances for Santander, there are no precise data on this. The situation is similar with trade, in that there is little available information about how banks helped Spanish exporters in their trade with Latin America.
As shown in the paper, in the pre-1986 period, the banks that would later form BBVA had a relatively limited presence abroad. There are several reasons why Spanish banks did not adopt a more vigorous internationalisation strategy during this period. Firstly, the lack of competition in the Spanish banking sector, with regulation that ensured profits. Secondly, the small size of the Spanish banks. Thirdly, external conditions did not encourage Spanish investment overseas because most of the Latin American countries had barriers to entry and other regulatory constraints. Finally, the low demand for banking services in the Latin American region also acted as a limiting factor for Spanish banks.
In the post-1986 period, Santander Group and BBVA have reinforced their presence in Latin America. In 1999, Banco Santander had more foreign assets as a percentage of total assets (29.92 per cent) than BBVA (21.03 per cent) but neither of these two banks reported a ratio above 30 per cent. However, the main difference arose between 1999 and 2006, when BBVA registered an increase in this percentage from 21.03 to just 26.14 per cent, whereas Santander rose from 29.92 to 62.36 per cent. Both banks adopted a strategy of entry via acquisition rather than via greenfield operations. The main difference was that whereas Banco Santander took majority stakes with the objective of ensuring full managerial control, BBVA started with minority stakes, which it gradually increased over time (Guillén and Tschoegl 1999  There are several key factors that facilitated the Spanish banking sector's international expansion in the post-1986 period. Firstly, the increase in domestic competition in a saturated market forced Spanish banks to expand abroad. Secondly, there was an increase in bank size. It was in this period when, thanks to mergers, Spanish banks became big enough to accelerate internationalisation. Finally, certain conditions made Latin American countries a favourable location for Spanish banking expansion: the potential growth of the region, the elimination of barriers to entry for foreign capital, a process of deregulation and liberalisation in the region, the lower purchase prices of the Latin American banks, the higher interest margins, and the mass entry of multinationals into Latin America during the 1990s, which prompted Spanish banks to follow their clients that were investing in Latin America at that time. The fact that other transnational banks (and direct competitors) focused their attention on regions other than Latin America gave Spanish banks plenty of room to manoeuvre in their international forays.