Spain’s Financial Position in the First Globalization

Abstract

Spain’s financial position during the late nineteenth and early twentieth century has usually been depicted as one of persistent deficit on current account, resulting from the country’s integration into international commodity and factor markets, which would have slowed down economic growth. In this chapter, this proposition is tested on the basis of a reconstruction of the Spanish balance of payments on current account. At odds with this interpretation, during the first globalization, opening up until 1890 allowed a net capital inflow that made it possible to meet the demand for investment boosting economic performance. Conversely, current account reversals in a context of macroeconomic domestic imperfections help explain the economic slowdown at the turn of the century.

This Chapter draws partially on my earlier work, published as L. Prados de la Escosura (2020), “Foreign Capital in 19th Century Spain’s Investment Boom”, European Review of Economic History 24(2): 314–331, and L. Prados de la Escosura (2010), “Spain’s International Position, 1850–1913”, Revista de Historia Económica / Journal of Iberian and Latin American Economic History 28(1): 1–43, but represents a full revision of the methodology and estimates and a full rewriting of the text.

1 Introduction

After the loss of the American empire, the integration into international commodity and factor markets led to a persistent deficit on current account that slowed down Spanish economic growth and deepened the country’s backwardness throughout the nineteenth and early twentieth century.

This thought-provoking but unfounded assertion offers stimulating hypotheses for research. This chapter tests the view that Spain’s international integration hindered growth on the basis of a reconstruction of the balance of payments on current account, and complements the discussion provided in Chap. 7. The main outcome is that the sustained deficit on current account over 1850–1890 highlights the fact that a net inflow of foreign capital made possible to meet the demand for domestic investment and, thus, boosted Spanish economic performance; conversely, current account reversals help explain the economic slowdown at the turn of the century.

The chapter is organized as follows: Sect. 8.2 presents current assessments of Spain’s international financial position. The reconstruction of the balance of payments on current account is discussed in Sect. 8.3, and Sect. 8.4 presents the balance of payments’ main trends and determinants from a “sudden stop” perspective.¹ Section 8.5 discusses the implications for growth of Spain’s financial position. Section 8.6 concludes.

2 Assessing Spain’s International Financial Position

The conventional view of Spain’s position in the international economy until World War I is one of chronic deficit on current account.² This diagnosis stems from the official trade figures (Estadística del comercio exterior), which show a sustained negative commodity trade balance, and from the scattered evidence about the gross inflow of foreign capital into Spain.³

Spain’s trade balance experienced, according to Sardá Dexeus (1948: 277), a sustained deficit for long periods, while Tortella (1974: 122) asserted that the trade balance was negative throughout the late nineteenth century. The persistent deficit represented, in Vicens Vives’s view (1959: 631), a heavy burden that contributed to the economic failure of the Restoration (1876–1923).

This chronic deficit hindered economic growth, according to most historical accounts. Allegedly, the current account deficit historians inferred from the commodity trade deficit set a limit on the growth of demand to which supply had to adjust, slowing down growth.⁴ The acceptance of an external constraint on growth caused by a structural balance of payments deficit has major economic policy implications, as it would require protective measures in order to limit imports, plus a floating exchange rate. Herranz-Loncán and Tirado (1996: 24) observed that the values of income elasticities for imports and exports point to the existence of a constraint on Spain’s economic growth resulting from the trade balance.⁵ Serrano Sanz (1997) estimated the theoretical growth rate for the Spanish economy, which was compatible with the current account equilibrium.⁶ As long as the theoretical rate were above the actual one, there would be no problem. This would have been the case of Spain from 1869 to 1891. However, if the theoretical rate were below the actual rate, as apparently occurred during 1892–1935, the external sector would have hindered long run growth.⁷ In a long-run view of Spain’s external sector, Bajo-Rubio (2010: 115) reached a rather different conclusion: the foreign sector did not represent a constraint on Spain’s long-run growth and only under exceptionally fast growth would an external restriction, resulting from a potentially unsustainable trade deficit, emerge.

Thirlwall’s (1979) ‘external constraint on growth’ is, however, predicated under the assumption that the terms of trade are stable and international capital flows negligible. In the context of the first globalization (1850–1913), such an assumption is far-fetched. Intense international commodity and factor flows took place, while the terms of trade suffered dramatic changes (O’Rourke and Williamson, 1999; Obstfeld and Taylor, 2004; Blattman, Hwang and Williamson, 2007). In fact, the proponents of the ‘external constraint on growth’ view accept that, in the presence of a sustained current account deficit financed through capital inflows, their prediction of the long-term growth rate is inadequate and it would be the pace and magnitude of foreign investment that would set the limit on growth (Thirlwall and Hussain 1982: 501). Thus, before jumping to conclusions about a potential external constraint on growth, it seems necessary to investigate the evolution of the current account over time.

One issue to be considered is the quality of Spain’s trade statistics. Maluquer de Motes (1999: 110–11, 189) argued that exports to the remaining Spanish colonies, Cuba and the Philippines, in particular, were over-exaggerated during the years 1895–1898, as they included supplies for the Spanish troops (military equipment but also foodstuffs, clothing, etc.) that did not involve a commercial transaction.⁸ Were this the case, military supplies should be removed from exports and considered as current Government transfers.⁹ Previously, however, one should prove that no commercial transaction had taken place and that military supplies were sent to the colonies by the Government. If, alternatively, supplies for troops were provided by private firms, they would represent exports. Furthermore, it could be simply the case that, as a result of the increasing number of Spanish troops in the colonies, the demand for foodstuffs and clothing increased. Another important reason behind the increase in exports to the colonies (and to the rest of the world) during the late 1890s is the (real effective) depreciation of the Peseta, which improved the competitiveness of Spanish exports. In fact, the estimated values of export price elasticity suggest that, other things being equal, the depreciation would have triggered a significant rise in the volume of exports.¹⁰ Moreover, an examination of the official trade statistics for 1897 indicates that there were no separate “State trade” records for exports.¹¹ Therefore, I decided not to correct the official exports figures to allow for the hypothetical inclusion of Government supplies to colonial troops.

Tortella (1994) raised objections to the revised figures for nineteenth-century Spanish foreign trade.¹² He noted a contradiction between the corrected trade balance figures—which reduce the commodity trade deficit in the 1850s and 1860s and provide a surplus after 1870—and the inflow of foreign capital.¹³ Tortella (1994: 132) argues that with a positive inflow of capital and emigrant remittances at the end of the nineteenth century, it would be hard to explain the peseta’s depreciation if there had been a trade surplus. This assertion, which seems logical at first sight, is, however, the result of identifying the commodity trade balance with the current account balance, in other words, with the overall balance for goods, services (which includes net income from abroad) and current transfers (including emigrant remittances). Such identification would be only warranted if the balances of services and current transfers were close to equilibrium, or if they cancelled each other out. In the case of the balance of services, such an assumption is inconsistent with the size of both external public debt and foreign investment in the private sector, which involved large interest and dividend payments. Moreover, the identification of gross foreign investment with the (negative of the) current account balance is unjustified, as the latter only records net foreign investment into Spain. Furthermore, migrant remittances, the main component of the balance of current transfers, only became significant at the very end of the period under consideration, as Spanish mass emigration was a comparatively late phenomenon in a European perspective (Sánchez-Alonso, 2000).

Tortella’s argument raises interesting questions. When did emigrant remittances become significant? Why did the depreciation not take place in 1883, as soon as the convertibility of the peseta was suspended? What were the causes behind the delayed, post-1891, depreciation of the peseta? Sardá Dexeus (1948: 219) offered an early diagnosis: ‘the economic causes of this depreciation may be linked to the possible existence of domestic inflation caused by the increase in the quantity of silver and bank notes, with repercussions on prices and the trade balance’, adding, then, ‘it is better to seek the immediate cause in the evolution of the balance of payments’. But it is only the second part of Sardá’s second part of the argument that has enjoyed support in the literature (Gutiérrez et al., 1998; Cubel et al., 1998; Catalan, Sudrià and Tirado, 2001).

Alternative interpretations to Sardá’s have been offered, however. Martín-Aceña (1993: 140–1) underlined the association between macroeconomic stability and a stable exchange rate of the peseta, and Tortella (1981: 131–48) identified Government financial problems as the main cause of the nominal depreciation of the Spanish currency between 1891 and 1905. Later, Sabaté et al. (2006) argued that the Treasury financing needs led to money creation and, hence, to sacrificing a fixed exchange rate.

Unfortunately, the debate is seriously constrained by the lack of quantitative evidence vis-à-vis Spain’s international position. The reconstruction of the balance of payments on current account therefore appears to be an urgent task. The next section of the chapter is devoted to this.

3 A Reconstruction of the Balance of Payments on Current Account

The balance of payments systematically summarizes the economic transactions of an economy with the rest of the world. These are the transactions involving goods, services, and income; financial claims on, and liabilities to, the rest of the world; and transfers (IMF, 1993: 6). I have estimated every item of the balance of payments on current account (commodity and service trade and current transfers). The procedure and sources used in the reconstruction of the main components of balance of payments on current account are summarily discussed in this section, although enough detail is provided for the reader who wants to replicate the computations. Needless to say, these computations are highly tentative and only further research will eliminate the potential errors of my estimates.

3.1 Commodity Trade Balance

Exports and imports of goods

Free on board (f.o.b.) value of goods exported and imported needs to be computed. Estimates on the basis of Spanish official trade statistics and corrected for quantity underestimation, including an estimate of smuggling through Gibraltar and Portugal, and for price biases by Prados de la Escosura (1986) have been used.¹⁴ Cost, insurance, and freight (c.i.f.) imports were converted into f.o.b. imports to comply with balance of payments conventions.

Gold and silver

Quantities of gold and silver recorded in Spanish trade statistics (coins, bars and paste) are considered as monetary gold and silver and, therefore, not included in the estimates of net exports of goods and services.¹⁵

3.2 Service Trade Balance

Three main categories are considered under this label: a) freight and insurance services, b) tourism, emigrants’ funds, passenger services and other services, and c) net income from abroad.

Freight and insurance

Freight income received for exports carried in Spanish ships less freight expenses paid for imports transported in foreign vessels constitutes the first item to be computed under this label. Following North and Heston (1960), the freight-value method, or freight factor, was preferred to the earnings per ton method.¹⁶ Total freight revenues on exports and imports were first computed by applying freight factors to the f.o.b. value of exports and imports and, then, in order to ascertain the freight income on exports (a credit for Spain) the share of tonnage exported carried under Spanish flag was used, while the share of imported tonnage in foreign ships was employed to compute freight expenses on imports.¹⁷ In addition, freight income from carrying trade between foreign ports was assumed, following North (1960) and Simon (1960), to represent a percentage of freight earnings, and a 10% of freight income on exports was accepted.¹⁸ Port outlays by Spanish ships in foreign ports and by foreign ships in Spain’s harbours as payments for port dues, loading and unloading expenses and coal are assumed to represent a fixed share of shipping earnings and expenses.¹⁹ Foreign ships transported more tonnage than Spanish vessels, as they had a more efficient transport capacity ratio.²⁰ I assumed that more fully loaded vessels made smaller outlays per ship and, hence, port outlays by Spanish ships abroad (a debit) were established at 30% of the freight income on exports, while port outlays by foreign ships in Spain (a credit) were fixed at 20% of freight expenses on imports.²¹ Finally, marine insurance income and expenses were computed under the widely shared assumption that underwriting follows the flag and exports in Spanish ships were, therefore, usually insured by Spanish companies, while imports in foreign vessels were insured by foreign companies.²² I arbitrarily assumed that insurance rates were identical by Spanish and foreign companies and accepted those used by Prados de la Escosura (1986), to which I added an extra 2% to include shipping commissions and brokerage.

Tourism, emigrants’ funds, passenger services and other services

Yearly income from tourist services was derived on the basis of expenses per visitor (net of Spanish tourist expenses abroad) calculated for 1931 by Jáinaga (1932), times the annual number of tourists and, then, reflated with a cost of living index to obtain current price estimates.²³ Unfortunately, the total number of tourists is only known since 1929 and was backward projected to 1882 with the rate of variation of passengers arriving by sea, while no tourism was assumed to exist during the period 1850–1881.²⁴

Spain was a net emigration country over the late nineteenth and early twentieth century (Sánchez-Alonso (1995). Emigrants carried small sums with them to cover their arrival expenses. It can be estimated that, by 1931, emigrant funds to America represented, on average, 200 gold pesetas, that is, 400 current pesetas, including the fare and small amounts to cover arrival expenses (Jáinaga, 1932). If the fare represented around 340 current pesetas, 60 pesetas would correspond to emigrant’s funds.²⁵ However, Jáinaga only added ‘a small amount for unavoidable expenses’, to the cost of the passage, and this sum is most likely an underestimate.²⁶ Therefore, I accepted a higher estimate, 100 pesetas for those emigrating to America, and one-tenth, 10 pesetas, for those to Algeria (and to France) for the eve of World War I.²⁷ These average sums times the number of emigrants to America, Algeria and France cast a yearly series of emigrants’ funds that was reflated with Reher and Ballesteros (1993) unskilled wage index.

In addition, revenues and expenses from passenger transport have to be taken into account. Fares paid by tourists carried by Spanish ships and by returning immigrants returning in Spanish vessels are included on the credit side, while fares paid by emigrants to foreign shipping companies represented a debit. The number of migrants provided by Sánchez-Alonso (1995) for 1882–1913 has been completed with an estimate of migrants from 1850 to 1881 on the basis of scattered foreign evidence.²⁸ The share of arrivals and departures in Spanish and foreign ships is provided by official migration statistics from 1911 onwards and shows a stable pattern; roughly one third of emigrants returned home under a Spanish flag and three-fourths left in foreign ships.²⁹ These shares were accepted for the nineteenth and early twentieth century. The fares for trips to Argentina, Cuba and Algeria are derived from Vázquez (1988), Llordén (1988), and official emigration statistics.³⁰

Lastly, Government transactions (credits and debits) taken from official sources were added up (Díaz García, 1976).

3.3 Net Income from Abroad

Due to a dearth of data, only crude estimates of foreign capital incomes, on the debit side, and of Spanish labour returns abroad (wages and salaries), on the credit side, have been carried out. These are assumed to be the main components of net income from abroad, as neither Spanish investments abroad nor foreign labour in Spain were significant during the period considered.

Foreign capital income

Ascertaining the amount of and the returns to each type of capital asset invested abroad and foreign capital invested at home is fraught with difficulties and becomes an all but impossible task in historical terms. Investment, whether domestic or foreign, results from microeconomic decisions of multiple agents, and no statistics exist to register all of them, particularly as we move back in time. Even in nineteenth century Britain, ‘investment was a private matter and the income from abroad was not subject to distinctive report until late in the century, and then only for certain classes of such income’ (Imlah, 1952: 222).

The realization of this intractable problem led Imlah (1952) and Brezis (1995), North (1960) and Simon (1960), Hartland (1960), Lévy-Leboyer (1977), and Gregory (1979), to construct indirect ‘residual’ measures of the capital account balance for the United Kingdom, the United States, Canada, France, and Russia, respectively.

In the indirect, ‘residual’ approach, the net payments to capital from abroad is computed indirectly by applying a rate of interest to the country’s international indebtedness at the beginning of the considered period, which is yearly updated with the net inflow of capital. It requires, then, a benchmark level of international indebtedness plus a representative rate of interest. Unfortunately, this implies arbitrary assumptions (Simon, 1960: 694). The initial amount of a country’s international indebtedness is not accurately computed and ‘informed guesses’ have frequently been used in historical studies (Imlah, 1952: 227; North, 1960: 587). Moreover, the rate of return applied hardly captures the average returns of a wide and changing variety of capital assets and even less with yearly precision (Imlah, 1952: 222). Furthermore, any alteration in either the interest rate applied or the initial estimate of international indebtedness results in far from negligible differences in the current account balance over the long run (North 1960: 574–5).

I have carried out both direct and indirect estimates of the net foreign capital income from abroad, although I find the indirect approach preferable and the discussion will focus on its results.

Direct approach

Due to the dearth of data only a few major sectors can considered, and returns from banking are, for example, neglected. It can therefore be conjectured that, most probably, the estimates provide a lower bound of the actual returns to foreign capital.

I have distinguished three main items: the external debt service; dividends and interests paid to foreign owned railway shares and debentures; and returns to foreign factors in mining. These three items together represented four-fifths of British portfolio investment in Spain over 1865–1913 (Stone, 1999).³¹

Service payments on the external debt have been computed by applying specific interest rates to each class of Government bonds.³² Some caveats about the volume of external public debt in foreign hands are needed. After the sovereign debt re-scheduling in 1882, which exchanged existing foreign debt for new bonds (at 43.75% of its nominal value), and simultaneously with the abandonment of gold convertibility of Spanish currency in 1883, debt repatriation started as Spaniards found it more secure to invest in bonds serviced in gold pesetas as a shelter against currency depreciation.³³ As of 1891, when the peseta started depreciating, Spanish citizens purchased external debt bonds, while foreign bondholders tried to get rid of them. The Government reacted by introducing the so called ‘affidavit’ in 1898, which implied that only non-resident bondholders would continue receiving their interest payments in gold pesetas (or in French francs), while the rest would be paid in current pesetas (and given the opportunity to convert their external debt bonds into internal debt). As a result, the external public debt fell, in 1903, to 52.7% of its volume in 1898, which implies that Spanish residents had purchased almost half Spain’s external public debt between 1891 and 1898. Hence, only half of the interests paid (52.7%) on external debt should be computed as payments to foreign capital invested in external debt over 1891–1898. I have, then, assumed that the interest payments effectively paid to foreign bondholders from 1891 to 1902 should be computed on the volume of external debt in existence in 1903.³⁴ Moreover, in so far as external debt was serviced in gold pesetas, the amount of interests paid (obtained by applying the interest rate to foreign debt in non-residents’ hands) has to be increased by the depreciation rate of the current peseta with respect to the gold peseta from 1891 to 1914.³⁵

Railway companies were highly concentrated, and the detailed research by Tedde de Lorca (1978, 1980) provides enough evidence to estimate dividends on share capital and interests on debentures paid to non-residents.³⁶ Dividends paid to shareholders and interest payments on debentures issued by the three major railway companies are available from the mid-nineteenth century onwards.³⁷ Both the percentage represented by the three main companies in the total capital invested in railways and the proportion of railways capital in foreign hands have to be ascertained in order to compute the returns to foreign capital invested in Spanish railways. Tedde de Lorca provides total capital shares and bonds held by the three major companies and their proportion in total investment, and, on the basis of Broder’s research, also the participation of French capital in total capital invested in 1867, at the time of network construction and through the nineteenth century.³⁸ Broder’s estimates of foreign investment in railways made it possible, in turn, to re-scale French railways capital to cover all foreign capital.³⁹

Foreign capital in mining was mainly British. On the basis of effective capital invested by British companies and cumulated total foreign investment in mining, it can be suggested that, from 1870 to 1913, more than half of all foreign capital in Spanish mining came from the U.K.⁴⁰ Decadal averages of dividend and interest payments to British companies provided by Harvey and Taylor (1987) were re-scaled to include all payments to foreign capital in Spanish mining for 1851–1913, assuming similar rates of return in non-British foreign investment, and using the estimated British participation in total foreign capital.⁴¹ Estimates of foreign capital returns in mining derived through this procedure were then distributed annually with an index of non-retained value in Spanish mineral exports.⁴²

Indirect approach

The first challenge has been selecting a stock representative of Spain’s international indebtedness at the beginning of 1850. Since private foreign investment in early nineteenth century Spain has been considered negligible (Sardá, 1948: 262), a sensible assumption would be to consider the level of international indebtedness equivalent to the value of external public debt. The nominal value of the external debt by January 1, 1850 can be estimated at 1504.8 million Pesetas (Comín, 1996: 131).⁴³ It is widely acknowledged in the historical literature that that external debt was never traded above half its nominal value in nineteenth-century Spain (Sardá, 1948: 257).⁴⁴ However, interests were paid on the nominal debt, so it is the nominal value of the investment that should be considered when computing interest payments (Tedde, 2015: 174).

As regards the rate of return, a weighted average of specific interest rates paid to each class of external debt bonds may provide a reasonable measure.⁴⁵ The use of the interest rate on nominal external debt may be considered to represent, however, a lower bound for the rate of return on all foreign investment.⁴⁶ Nonetheless, it is worth noting that the higher the interest rate applied to the stock of international indebtedness, the larger the resulting amount of net payments to foreign capital and, hence, the current account deficit.⁴⁷ As this exercise aims to test whether the net capital inflow derived from the direct approach results in an underestimate, biasing the indirect estimates against the hypothesis seems advisable.

Net payments to foreign capital for 1850 can thus be computed by applying the weighted nominal interest rate on external debt on that year to the nominal value of the external debt on January 1st, 1504.8 million pesetas. For subsequent years, the level of international indebtedness has been updated with the net inflow of capital.

Spanish labour returns from abroad

Assessing returns to Spanish labour employed abroad is a complex task, as labour incomes (wages and salaries), which belong to the balance of services, have to be distinguished from emigrants’ remittances, which belong to the balance of unilateral transfers. Actually, the distinction can only be made since 1917 and I accepted that only 5% of those emigrating to America and 60% of those migrating to Algeria returned within the year from 1850 to 1913.⁴⁸ The next step was to assess the amount that, on average, was brought home by Spanish workers returning after 1 year, or less, away from home. I computed an average sum that was taken home by the temporary emigrant or sent annually by the long-term emigrant to their relatives and friends. Estimates of total sums sent home by emigrants have been gathered in recent historical research for the early twentieth century.⁴⁹ García López (1992) presents the most comprehensive estimates for the years prior World War I, 250–300 million pesetas as an annual average over 1906–1910, which amounts to around 340–400 pesetas per emigrant (either returning home or sending remittances). I accepted 400 pesetas per emigrant as a benchmark that was then projected backwards and forwards with a nominal wage index constructed for the destination countries, and adjusted for the exchange rate between the peseta and each destination country’s currency over 1850–1913.⁵⁰ Finally, returns to Spanish labour abroad were obtained by multiplying the annual sum per head by the number of emigrants returning home within their first year abroad.

Once net payments to foreign capital were obtained, they were added to net payments to domestic labour to derive the balance on payments to foreign factors.

3.4 Current Transfers Balance

Emigrants’ remittances constituted the main historical component in Spain prior to 1913. Not all emigrants sent money home while abroad. In historical estimates, it is usually accepted that most of those who established themselves abroad stopped sending money after 5 or 6 years, either because they had already paid for their debts or because they planned to invest in the receiving country. I discretionally assumed that emigrants only sent money home within their first 5 years, and computed emigrants’ remittances by multiplying the estimated average sum per emigrant by the cumulative figure of emigrants arrived in the last 5 years, after deducting those migrants who returned home within 1 year.⁵¹

3.5 The Balance of Payments on Current Account and the Net Inflow of Capital

Adding up the balances of goods, services—including net payments to foreign factors—, and current transfers, the balance of payments on current account is obtained.

The capital account balance is obtained by subtracting the current account balance from the net change in reserves. The estimates of net changes in reserves result from adding the net imports of gold and silver to the annually minted figures.⁵²

4 Trends in Spain’s International Financial Position

Let us begin by looking at the commodity trade balance. Two clearly defined periods can be distinguished: one of deficit, from 1850 to the 1866 crisis, except for 3 years, 1854–1856, followed by one of surplus up to the eve of World War I, except for 1876 (Fig. 8.1). If we now turn to the balance of services, a persistent deficit is observed. Transport, tourism and intergovernmental transactions show a negative sign (Fig. 8.2) and, more importantly, the balance of services main item, the net income from abroad, too (Fig. 8.3).⁵³ The service of the public debt dominated net income from abroad until the beginning of the Restoration (1876). After the sovereign debt re-scheduling (1882) and, especially, from 1890 onwards, this situation changed with net returns to foreign capital in railways and mining gaining weight. The results confirm that the direct computation of net payments to foreign capital produces lower levels than the indirect approach. Such a difference derives, as suggested in the historical literature (Goldsmith, 1955), from its incomplete coverage of investment from abroad. Emigrant remittances became increasingly important from the late 1880s and increased dramatically from 1904 onwards, partially offsetting the net payments to foreign factors (Fig. 8.4).

Fig. 8.1

Commodity trade balance (million Pesetas). Sources: See text and Table 8.1

Fig. 8.2

Services trade balance (excluding net income from abroad) (million Pesetas). Sources: See text and Table 8.1

Fig. 8.3

Services trade balance: net income from abroad (million Pesetas). Indirect estimates. Sources: See text and Table 8.1

Fig. 8.4

Current transfers balance (million Pesetas). Sources: See text and Table 8.1

Adding up the commodity, services, and current transfers balances provides the current account balance. Two distinctive phases, with 1891 as a turning point, can be distinguished. A sustained current account deficit was in place throughout the period 1850–1890. Then, from 1891 up to World War I, a surplus prevailed, with the exception from 1899–1904 (Fig. 8.5).⁵⁴

Fig. 8.5

Current account balance: indirect estimates (million Pesetas). Sources: See text and Table 8.1

The net inflow of capital, higher when computed indirectly,⁵⁵ provides the mirror image of the current account balance, although it presents a higher absolute level from 1859 to 1874 and, especially, in the wake of Cuba’s (and Puerto Rico’s and the Philippines’) independence (1898), when capital repatriation presumably took place, and a lower level during the 1880s (Fig. 8.6).

Fig. 8.6

Net capital inflow: indirect estimates (million Pesetas). Sources: See text and Table 8.1

Finally, crude estimates of the annual balance of Spain’s international indebtedness are presented in Fig. 8.7.⁵⁶ A previous clarification is needed though. The nominal level of public debt at the beginning of the period under study (1st January 1850) was assumed to represent the level of international indebtedness to which the weighted nominal interest rate was applied to obtain net payments to foreign capital for 1850 and, thus, the level of international indebtedness at the end of the year, which was yearly updated with the net inflow of capital. However, if we want to estimate the actual level of Spain’s international indebtedness, a more realistic way to achieve this is to accept the effective, rather than the nominal, value of Spain’s external debt on 1st January 1850, about 50% of its nominal value, 752.4 million, and then updating it with the yearly estimates of the net capital inflow. It appears that international indebtedness grew up to the mid-1870s, before stabilizing until the mid-1900s, except for episodes of decline in the early 1880s and during Cuba’s War of Independence (1895–1898), and declined steadily thereafter.

Fig. 8.7

Effective international indebtedness: indirect estimates (million Pesetas). Sources: See text and Table 8.1

A sharp contrast results between the commodity and current account balances. The commodity trade balance is positive in 49 out of the 64 years, with deficit concentrated in the years 1857–1866—in which large imports associated to railways construction took place—; meanwhile the current account was in deficit for most of the time except for three episodes of rising levels, 1880–1882, 1895–1898, and 1906–1913.⁵⁷ These three periods and 1857–1866 are the only ones in which the signs of the two balances coincide.

The divergent evolution of the various components of the balance of payments enables us to reconcile the positions of those who maintained that, from 1870 onwards, the commodity trade deficit resulting from the official figures was implausible (Prados de la Escosura, 1986) and those who stressed that Spain’s international position was one of deficit (Sardá Dexeus, 1948; Tortella, 1994). The explanation for the apparent contradiction between the two balances is to be found in the growing role played by net income from abroad that was not counter-balanced by current transfers, as emigrant remittances only became significant from 1904 onwards. Thus, the current account deficit appears to be associated with the costs of investing in new infrastructures and exploiting natural resources.

How could the current account surplus for the years 1895–1898 and 1906–1913 be explained? The reasons behind the change from a current account deficit to a surplus can be explored in the light of ‘sudden stops’. Edwards (2004: 19) has defined a ‘sudden stop’ episode as ‘an abrupt and major reduction in capital inflow to a country that up to that time had been receiving large volumes of foreign capital’. Sudden stops are, thus, sharp reversals in capital inflows which constrict domestic consumption smoothing.⁵⁸ During the first wave of financial globalization that took place in the late nineteenth and early twentieth century, the main effects associated with sudden stops are drops in the exchange rate and deceleration of economic activity.⁵⁹

Exogenous forces conditioned sudden stops. Monetary tightening in advanced countries (rise in central bank discount rates, for example) represented a significant exogenous element in the reversal of capital inflows. In addition, international crisis irradiating from capital importer countries, such as Argentina during the Baring crisis in the early 1890s, constituted an exogenous force to be considered. However, the fact that not all capital importers suffered a given sudden stop to the same extent suggests that endogenous factors also mattered.

Let us examine the Spanish experience in the light of sudden stops (SS, hereafter). In Fig. 8.8, international capital flows, proxied by British, French and German aggregated current account (excluding all gold flows), are confronted with the net capital inflow in Spain, both expressed in Sterling.⁶⁰ Several slowdown episodes in international capital flows are observed, starting in 1860, 1866, 1873, 1890, and 1896, of which those of 1873 and 1890 appear to have special intensity. In Spain, sudden stops can be observed in 1866–1869, 1876–1880, 1890–1896, and 1904–1907, with particular intensity in the last two episodes. Interestingly, the last sudden stop, at odds with the previous ones, occurred during the expansion of international capital exports prior to World War I.

Fig. 8.8

International net capital flow and Spain’s net capital inflow (indirect estimates) (million £). Sources: See text and Table 8.1, and Jones and Obstfeld (1997)

Which of the predicted effects of the SS are observed in the Spanish experience? To begin with, currency crashes occurred during 1891–1893 and 1896–1898, but not in earlier SS, or in 1904–1907, when the opposite happened and the peseta returned to its position in 1891 (Fig. 8.9).⁶¹ Why such an uneven response to different SS? The consequences of two exogenous events, the Baring crisis (1890) and the Cuban War of Independence (1896–1898) are, no doubt, part of the explanation. According to Catão (2007: 266–9), during the first wave of financial globalization, countries that experienced rapid monetary expansion and had a lax fiscal behaviour were more prone to currency crashes.⁶² In fact, money supply (M2) appears to have grown faster than GDP in Spain during the cyclical upswings 1885–1889 and 1896–1898 (Fig. 8.10). Meanwhile, the public debt/GDP ratio, sharply reduced as a result of the 1882 sovereign debt re-scheduling, experienced a sustained increase over 1893–1896 and went up further in the aftermath of the Cuban War of Independence (1899–1903) (Fig. 8.11).⁶³

Fig. 8.9

Spain’s current account balance indirect estimates (% GDP) and nominal and real effective exchange rate (1913=100), 1870–1913. Sources: See the text and Table 8.2

Fig. 8.10

Ratio M2/GDP (1874–1913). Sources: M2, Martín-Aceña and Pons (2005); GDP, Prados de la Escosura (2017)

Fig. 8.11

Public Debt/GDP (1884–1913). Sources: Public Debt, Comín and Díaz Fuentes (2005); GDP, Prados de la Escosura (2017)

The simultaneity of sudden stops and exchange rate drops during the 1890s tends to downplay the suspension of the gold convertibility of the peseta (1883) suggesting that, during the 1880s, as long as an inflow of foreign capital continued, the Spanish currency remained stable, regardless of whether the exchange rate floated (Fig. 8.9).

This result has implications for the debate between those who emphasise the advantages of a floating exchange rate for a developing economy, due to the high opportunity cost of maintaining gold reserves, as well as the shock absorber role of the exchange rate (Sardá Dexeus, 1948; Tortella, 1974: Flandreau and Zumer, 2004), and those who stress that belonging to the Gold Standard sent the right signal of compliance with orthodox financial practice to capital markets (Martín-Aceña, 1993; Bordo and Rockoff, 1996). To the extent that it could be factored out, macroeconomic stability rather than belonging to the Gold Standard seems to have been the relevant signal for international investors.

When macroeconomic discipline was abandoned at the time of the Baring crisis and, then, again, by the need to finance the Cuban War of Independence, the control mechanism which stopped the peseta from falling disappeared.⁶⁴ Macroeconomic instability, especially inflation, which soared over 1895–1904 (Fig. 8.12), had negative effects on the reputation of Spain’s economy, making it less attractive to international capital, as suggested by the spread between the discount rate of the Bank of Spain and those of the central banks in the main capital investing countries during the 1890s (Fig. 8.13).⁶⁵

Fig. 8.12

Rate of Inflation (%) (1883–1913). Sources: GDP deflator from Prados de la Escosura (2017)

Fig. 8.13

Spain’s interest rate spread with Britain, France, and Germany (1884–1913). Sources: Central banks discount rates, Spain, Martín-Aceña and Pons (2005); Britain, France, and Germany, Homer and Sylla (1991)

After the independence of Cuba, Puerto Rico and the Philippines, a current account deficit reappeared between 1899 and 1904, which could be associated with the repatriation of capital from the former colonies in the sound economic environment provided by Fernández-Villaverde’s stabilization plan (Comín, 1999).

Why, then, the sudden stop of 1904–1907, at the time of an international expansion of capital flows, and why was the current account reversal accompanied by an improvement in the exchange rate of the peseta? There are good reasons to explain why the exchange rate did not drop. The fact that, for most of the period up to World War I, inflation remained moderate and the public debt/GDP and M2/GDP ratios continued to fall, as opposed to what had happened in the 1890s, helps explain why a currency crash was avoided. Furthermore, no exogenous events such as the Cuban War of Independence took place, while emigrant remittances played an important offsetting role in the current account balance (Fig. 8.4). However, why was Spain omitted from the international wave of investment prior to World War I? It is noteworthy that the Italian and Portuguese current account balances also experienced a surplus during this period (Bordo et al., 2010; Catão, 2007). Meanwhile Argentina, Brazil, Canada, and Russia became the main capital importers. Why were south-western European countries excluded? Was it because investment opportunities had dried up, or because the opportunity cost was too high? It would be worth investigating the extent to which the decline in a sustainable current account deficit results from a reduction in foreigners’ demand of an emerging country’s assets (Edwards, 2004). In Spain, by the end of the nineteenth century, those sectors that had attracted most foreign capital were already developed: the railway network was completed and mining resources fully exploited. Perhaps this fact helps explain why, in the absence of new investment opportunities, international capital inflow into Spain slowed down.⁶⁶ This is, no doubt, an avenue for further research.

To sum up, the idea that the suspension of the convertibility of the peseta in 1883 and its delayed effect in terms of a currency crash in the 1890s was the result of endemic balance of payment problems is not supported by the evidence presented here. In fact, it is the sudden stops, in a context of domestic financial imperfections, that were to blame.

5 Did International Integration Hold Back Growth?

The traditional view among Spanish economic historians, reinforced by those who argue in terms of the ‘external constraint on growth’ approach, associates a current account deficit with a deterioration of the economic situation or to a threat to growth. Conversely, a current account reversal—that is, a surplus on current account—will imply, according to the sudden stop literature, a decline in investment and, thus, in economic growth that tends to intensify if the country is less open (Edwards, 2004; Bordo et al., 2010).

No consensus has been reached with regard to the importance for growth of a large and resilient current account deficit. The optimistic view emphasizes the intertemporal nature of the current account, arguing that insofar as they reflect a rise in investment, there is no reason for concern (Sachs, 1981; Corden, 1994). The opposite, pessimistic view, epitomized in Thirlwall’s approach (1979), has a more recent expression in Fischer (1988), for whom the first sign of a crisis is the current account deficit. In this context of uncertainty, historical research can make a useful contribution.

How did the interruption of foreign capital inflow affect economic growth in Spain?

Let us begin with the current account identity:

$$ CAB=X\hbox{--} M+ NCT+ NY=S\hbox{--} I $$

(8.1)

Where CAB is the current account balance; X and M are exports and imports of goods and services, respectively; NCT, net current transfers; and NY, net income from abroad; while CAB equals the difference between gross domestic saving (S) and investment (I).⁶⁷ Here we can normalize with respect to GDP (Y), to provide an idea of the relative size of each item,

$$ CAB/Y=S/Y\hbox{--} I/Y $$

(8.2)

Two distinctive phases can be observed in the relationship between investment and saving, with 1890 as the turning point (Fig. 8.14).⁶⁸ Domestic investment was above saving between 1850 and 1890 (except for 1880–1882), which means that foreign capital supplemented domestic saving to meet investment demand. The gap between investment and saving was particularly noticeable from 1850 to 1866. This upsurge of investment, which reached 10% of GDP in the early 1860s, was associated with the construction of the railway network, in which foreign capital played a significant role. From 1890 to World War I, investment depended almost exclusively on domestic saving, as a current account surplus prevailed (exception made of the years 1899–1904). All in all, investment and saving followed the same trend, with investment remaining below 8% of GDP up to 1898, except for the years of the railway construction boom.

Fig. 8.14

Gross investment and saving (indirect estimates) (% GDP). Sources: Investment, Prados de la Escosura (2017); Saving, see the text and Table 8.2

The small size of investment and saving, in terms of GDP, conceals the relative importance of foreign investment in Spain’s gross domestic capital formation. Starting from expression (8.2) we can decompose gross domestic investment into gross domestic saving (private, Sp/Y, and Government, Sg/Y, saving) and the (negative of the) current account balance (Fig. 8.15).⁶⁹

$$ I/Y=S/Y\hbox{--} CAB/Y= Sp/Y+ Sg/Y\hbox{--} CAB/Y $$

(8.3)

Government saving was negative up to 1892, especially between 1861 and 1873, and was not offset by a rise in private saving but by a current account deficit financed through a net inflow of foreign capital. This way, the decrease in government saving did not imply a constraint on the investment ratio. This finding implies that the view of a decline of investment derived from a decrease in government saving—the crowding out hypothesis occasionally discussed in the literature—, is not confirmed by the evidence.

Fig. 8.15

Decomposing gross investment: private and government saving and the (negative of the) current account balance (indirect estimates) (% GDP). Sources: Government Saving, Comín and Díaz Fuentes (2005); for the rest, see the text and Table 8.2

The relative importance of the net capital inflow contribution to capital formation is reflected by its share of gross investment (Fig. 8.16).⁷⁰ Between 1850 and 1890, foreign capital financed nearly 30% of domestic investment, rising to almost a half during the third quarter of the nineteenth century. Conversely, in the years 1891–1913, the net outflow of capital implied a contraction of domestic investment (11%), which reached 15% over the period 1891–1897, but experienced a reversal episode over 1899–1903, in which the net inflow of foreign capital represented more than 15% of gross capital formation.

Fig. 8.16

Net capital inflow as a proportion of gross investment (indirect estimates) (%). Sources: See the text and Table 8.2

Thus, it can be suggested that, during the late nineteenth century, as foreign capital complemented domestic saving to meet a growing investment demand, economic growth intensified and, although difficult to quantify, improvements in the quality of capital and embodied technology in new capital goods, whose acquisition was funded by foreign capital, most probably made the economy more efficient. Conversely, the sudden stops at the turn of the century slowed down growth, as the increase in capital accumulation decelerated, and, presumably, the efficiency of the economy declined. Hence, sudden stops, by causing current account reversals and currency drops in a context of domestic macroeconomic imperfections, had a clearly negative effect on Spain’s long-run growth.

6 Concluding Remarks

Between 1850 and 1890, economic expansion coincided with a significant current account deficit, while between 1891 and 1913, growth slowed down at a time of positive current account balances. This inverse correlation between the current account surplus and economic growth casts serious doubts on the widespread view of an external restriction to Spain’s growth during the nineteenth century. This chapter’s results suggest an alternative interpretation: the balance of payments reacted to changes in the equilibrium between saving and investment. Thus, the current account deficit resulted from an inflow of capital which allowed the rate of investment to rise and, in turn, to contribute to more rapid growth. Only when isolation from the international economy increased since 1891, did investment demand have to rely on domestic saving. In the context of globalization that characterized the classical gold standard era, there was no reason why an open economy should not enjoy sustained access to international capital markets and break the link between investment and domestic saving. From this perspective, the persistence of the current account deficit between 1850 and 1890 is better understood.

At the turn of the century, domestic macroeconomic imperfections exacerbated the current account reversals that had been provoked by sudden stops, undermining the confidence of foreign investors in the Spanish economy and encouraging the flight of foreign capital. Furthermore, as Sánchez-Alonso (2000) has shown, the migration push of the 1891 protectionist tariff was largely offset by the microeconomic consequences of the currency crash, preventing individuals from migrating for one and a half decades.

The view that Spanish integration in international markets contributed to a slowing down of economic growth appears to be incorrect. One might suggest that without the current account deficit—that is, without an inflow of foreign capital—Spain would have grown at a slower speed during the second half the nineteenth century. As the inflow of capital dried up, investment had to rely exclusively on domestic savings, slowing down capital accumulation and economic growth.

Appendix

1.1 A.1 The Metallic Stock

In the construction of the metallic stock, setting its initial level represents the first step. Tortella (1981: 124) provides an estimate for 1865. Then, its annual level from 1866 onwards would result from adding to the initial level the yearly gold and silver coinage (newly minted, less re-minted, plus illegal coinage) and the net imports (legal and illegal or unregistered) of gold and silver coin, and subtracting gold and silver coin hoarded, lost or destroyed.

Alas, re-minting is an unknown, as are illegal minting and gold and silver hoarded or lost, even though re-minting was a very small proportion of total coinage in the late 1860s (when data are available) (Anes, 1974: 111; Tortella, 1974: 120).

In addition, the available data on the net imports of gold and silver also raise objections. Trade in gold and silver is poorly covered in most countries’ historical statistics. Spanish official gold and silver trade statistics have been deemed incomplete due to underreporting (Tortella 1974: 121–122; Moro et al., 2015, suppl. 2). Specifically, official statistics do not record any imports of gold and silver between 1850 and 1882.

Fortunately, however, trade statistics of Spain’s main trading partners offer an alternative source. The U.K. trade statistics provide gold and silver trade (imports from Spain only since 1858) between the United Kingdom and Spain and Gibraltar (the latter as a proxy for smuggling), and the United States trade statistics supply the value of gold and silver exported to and imported from Spain. Moreover, Tedde (2015: 181) presents the Bank of Spain’s imports of gold and silver, mainly from France but also from Britain, during 1859–1874. Tedde (2015) also provides smaller purchases of silver during 1849–1855 that I have assumed came from France too and distributed them evenly through 1850–1855. All this information allowed me to revise, at least partially, the official figures. The revised series of gold and silver trade result from replacing official figures of exports and imports by those from the statistics of Spain’s main trading partners but only for those years in which the latter exceeded the former. It is worth stressing that most of the correction of the official figures of gold and silver trade corresponds to imports.

Thus, crude estimates of the metallic stock for 1866 and successive years would be derived by adding the annually minted gold and silver and the net imports of gold and silver coin to the stock in the previous year (being 1865 the initial year). For 1850–1864 the stock would be obtained by deducting the annually minted gold and silver and the net imports of gold and silver coin from the stock in 1865.

The annual change in metallic stock provides a measure of the change in reserves, with two options available, one in which the net imports of gold and silver coin derive from the official series (option A) and another that derive from the revised estimates (option B).

See Tables 8.1, 8.2, 8.3 and 8.4.

Table 8.1 The balance of payments on current account, net capital inflow, and international indebtedness, 1850–1913 (million Pesetas): indirect ‘residual’ estimates

Table 8.2 Investment and saving, 1850–1913 (million Pesetas and % GDP): indirect estimates

Table 8.3 The balance of payments on current account, net capital inflow, and international indebtedness, 1850–1913 (million Pesetas): direct ‘residual’ estimates

Table 8.4 Investment and saving, 1850–1913 (million Pesetas and % GDP): direct estimates

Direct Estimates

See Figs. 8.17, 8.18, 8.19, 8.20, 8.21, 8.22, 8.23, 8.24 and 8.25.

Fig. 8.17

Services trade balance: net income from abroad (million Pesetas). Direct estimates. Sources: See text and Table 8.3

Fig. 8.18

Current account balance: direct estimates (million Pesetas). Sources: See text and Table 8.3

Fig. 8.19

Net capital inflow: direct estimates (million Pesetas). Sources: See text and Table 8.3

Fig. 8.20

Effective international indebtedness: direct estimates (million Pesetas). Sources: See text and Table 8.3

Fig. 8.21

International net capital flow and Spain’s net capital inflow (direct estimates) (million £). Sources: See text, Table 8.3, and Jones and Obstfeld (1997)

Fig. 8.22

Spain’s current account balance (direct estimates) (% GDP) and nominal and real effective exchange rate (1913=100), 1870–1913. Sources: See the text and Table 8.4

Fig. 8.23

Gross investment and saving (direct estimates) (% GDP). Sources: Investment, Prados de la Escosura (2017); Saving, see the text and Table 8.4

Fig. 8.24

Decomposing gross investment: private and government saving and the (negative of the) current account balance (direct estimates) (% GDP). Sources: Government Saving, Comín and Díaz Fuentes (2005); for the rest, see the text and Table 8.4

Fig. 8.25

Net capital inflow as a proportion of gross investment (direct estimates) (%). Sources: See the text and Table 8.4