Denouncing Odious Debts

Stephanie Collet & Kim Oosterlinck

Journal of Business Ethics

ISSN 0167-4544 Volume 160 Number 1

J Bus Ethics (2019) 160:205-223 DOI 10.1007/s10551-018-3865-7

1 23 Your article is protected by copyright and all rights are held exclusively by Springer Science+Business Media B.V., part of Springer Nature. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23 Author's personal copy

Journal of Business Ethics (2019) 160:205–223 https://doi.org/10.1007/s10551-018-3865-7

ORIGINAL PAPER

Denouncing Odious Debts

Stephanie Collet1 · Kim Oosterlinck2

Received: 25 October 2016 / Accepted: 26 March 2018 / Published online: 2 April 2018 © Springer Science+Business Media B.V., part of Springer Nature 2018

Abstract Economists have suggested it was optimal to signal the odious character of bonds when they were issued. However, since the odious debt doctrine has not been recognized by any court, one could argue that denouncing odious debts is useless. Exploiting a unique historical episode, this paper quantifes the impact of protests on odious debts. In 1906, the Russian government foated a bond in Paris to cover the costs of its war against Japan but also to raise money to crush the political movements wishing to reform Russia’s political system. Issued without parliamentary consent, this loan met with ferce opposition. Press campaigns in Great Britain, France and Germany denounced its odious character. If failure or success is determined solely by the ability to prevent a loan from being issued, then the campaign failed. We argue, however, that failure or success should be determined in light of the fnancial costs imposed on the issuer for future loans and even the ability to force a postponement of these loans. We show that these campaigns increased the yields of all Russian bonds traded in Paris, and thus Russia’s future borrowing cost. Yields on the 1906 loan, which was specially targeted by the campaigns, rose even more. However, once the press campaigns stopped, yields experienced a declining trend, highlighting the important role the press may have on odious debts.

Keywords Ethics · Odious debt · Repudiation · Financial history · Sovereign debt · Russia

JEL Classifcation F34 · G12 · G15 · N23

Introduction proposed by Sack1 (1927, p. 26) prevailed.2 In this work, Sack highlighted three criteria for defning odious debts. As a consequence of the Spanish–American War, the USA For a government to declare a debt odious, it would have to annexed Cuba. However, America refused to take over the prove that: (1) the debts had been issued without the popu- debts issued by Spain in the name of Cuba because the pro- lation’s consent, (2) the population would have received no ceeds had been used to crush the Cuban independence move- beneft from the debt issue, and (3) lenders knew that the ment. This precedent prompted legal scholars to establish a proceeds would not be used for the population and that the general framework for analyzing cases in which it was unfair population was opposed to the debt. to repay sovereign debts, or, to use the current terminol- In an ideal world, odious debts would be easy to identify. ogy, when debts could be considered as odious (Jèze 1922; In practice, Sack’s defnition is hard to rely upon. At the Sack 1927; Feilchenfeld 1931). Eventually, the defnition time the book was written (1927), the projected use of the funds raised by a loan was mentioned in the prospectus. It was therefore possible to determine whether the population Electronic supplementary material The online version of this article (https​://doi.org/10.1007/s1055​1-018-3865-7) contains would beneft from those funds (assuming they would be supplementary material, which is available to authorized users. used for the stated purposes). Nowadays the exact purpose of the borrowing is hardly ever mentioned and public funds * Kim Oosterlinck [email protected] 1 On Alexander Nahum Sack, see Ludington and Gulati (2008). 2 See for example Khalfan et al. (2003). This does not mean that 1 Research Center SAFE, Goethe University, Frankfurt, these criteria have not been criticized. For example, King (2016, p. Germany 57) considers these three criteria as too limited to refect the diversity 2 Université Libre de Bruxelles, SBS‑EM, 50 Av. Roosevelt, of odious debts which he classifes into four categories (war debts, CP 114/03, 1050 Brussels, Belgium illegal occupation debts, corruption debts and subjugation debts).

Vol.:(0123456789)1 3 Author's personal copy

206 S. Collet, K. Oosterlinck are fungible, making the defnition difcult to apply (Bolton Although the literature suggests odious debts should be and Skeel 2007).3 In some instances, furthermore, only part denounced before being issued, no paper has yet analyzed of the proceeds of the loan may have been squandered by how markets react when debts are denounced as unfair to the country’s ruler, raising the question of the treatment of repay. Empirical evidence is also missing on the efective- these partially odious debts (Ben-Shahar and Gulati 2007). ness of press campaigns and organized protests against odi- Dictatorships or governments who hurt their population may ous debts. This paper relies on a historical episode to address also borrow indirectly, making odious debts even harder to these questions. In 1906, the Russian government foated a identify. According to Hausmann,4 investors buying the JP bond in Paris to cover the costs of its war against Japan, but Morgan Emerging Bond Market Index (EMBI +) are in fact also to raise money to crush the political movements wish- gaining high returns mostly because the fund invests in Ven- ing to reform Russia’s political system (Girault 1973, pp. ezuelan debt. Indeed, even if the bonds issued by Venezuela 394, 401 and 413). Issued without parliamentary consent, represent only 5% of the index, they account for 20% of its the loan met with ferce opposition. Several Russian political yield. Hausmann has therefore suggested that investing in parties pledged they would repudiate it should they come to EMBI + is similar to rooting “for Venezuelan debt, which power.7 In parallel, press campaigns in Britain, France and means wishing for really bad things to happen to Venezue- Germany stressed the loan’s odious character, which led to la’s people”. In view of the existence of partially odious attempts to prevent its issue.8 Organized protests against the debts, Shafter (2007) suggests lenders should exercise due 1906 loan were such that for many years Russian loans were diligence. Determining the consent of the population is even a topic of debate in French politics. In France, the most vocal harder, since this leads to a judgment on the government newspaper against this loan was L’Humanité. This was due that may be more political than legal. In the absence of a to its far-left political leaning but also to the fact that almost referendum on each loan, it is in fact impossible to prove all other French newspapers had been bribed to silence by that the population agrees to an issue. As a result, the debate the Russian government. has gradually shifted from odious debts to odious regimes Did L’Humanité and the organized protests infuence mar- (Bolton and Skeel 2007; Buchheit et al. 2006). But proving kets by stressing the unethical character of the loan? Accord- that a regime is odious is only marginally easier than prov- ing to Long (1972b, p. 39), one of the frst scholars to focus ing that a debt is. on the 1906 loan, “the anti-loan campaign failed” but “it The literature has also attempted to assess whether debts revealed the continued close cooperation between Russian should be labelled as odious before they are issued. Econo- revolutionary émigrés and French socialists, united in their mists consider it optimal to signal the odious character ex eforts to expose the fnancial and political bankruptcy of ante (Jayachandran and Kremer 2006). Doing so would cre- Tsarist autocracy”. We argue that the usefulness of denounc- ate an incentive for potential lenders not to extend loans ing debts as odious should not be judged only on the ability to odious regimes, decreasing both the likelihood that such to prevent bonds from being issued or on their repudiation regimes emerge and the survival times of those that do.5 but also on the fnancial costs imposed on the issuer. More Demand could also decline because ethical agents would precisely, one should assess whether the yields of Russian refrain from lending to dictators. Empirical evidence shows that ethics play a role in investment decisions, suggesting that denouncing odious debts would infuence markets. In 7 this case, an odious debt label would send a clear signal to In this respect, the 1906 loan ofers a striking parallel with the case 6 of bonds issued recently in Venezuela. On 30 July 2017, a new con- the market. stituent assembly was elected in Venezuela. This election was con- vened following a decree issued by President Nicolas Maduro. The decision to hold these elections was heavily criticized both by the 3 King (2016, p. 190) qualifes this point by suggesting that when international community and by the opposition parties, which viewed large sums are involved concealing the odious character of the debts the decision as a move to replace a National Assembly that refused may not be so easy. Gentile (2010) suggests including covenants in to follow the president blindly. See Buchheit (2017). Following this the agreements governing the bond. These covenants would specify election, the legality of the new assembly, and its ability to legally the use of the proceeds and limit the dictators’ ability to use the funds propose a debt swap, has been questioned (see Gulati 2017). As for for other purposes since any breach on the covenant could lead inves- opposition parties, they have announced they would repudiate the tors to ask for immediate reimbursement. debts issued by the state oil company PDVSA should they come to 4 See Hausmann (2017). power (Porzecanski et al. 2017). Press campaigns against the loans 5 Despite this, and as pointed out by Ludington et al. (2010), whereas issued by the Maduro regime, now commonly referred to as “hunger economists tend to consider that it is optimal to label regimes as des- bonds”, have gained momentum (Jenkins and Zerpa 2017). potic ex ante, history has shown that the analysis of whether a debt is 8 Protests did not stop the loan issue. They were, however, successful odious or not is usually conducted ex-post. in forcing the Russian regime to postpone any future loans by several 6 As evidenced by the growing number of socially responsible (SR) years. The French government imposed a 2-year moratorium for any funds (Peillex and Ureche-Rangau 2016) and by the limited costs new loan in France (Girault 1973, p. 445) and the next Russian loan associated with the choice of SR sovereign bonds (Drut 2010). would only be issued in 1909.

1 3 Author's personal copy

207 Denouncing Odious Debts bonds were afected by the press campaigns, or whether the In Paris, the reputation of Russian bonds was so good that yields to holders of the odious 1906 loan difered substan- even years after repudiation investors still believed they tially from those on other Russian bonds. To address these would be reimbursed (Oosterlinck and Landon-Lane 2006; questions, our paper relies on a panel of Russian bonds Oosterlinck 2016). traded in Paris between 1906 and 1913. Taking into account Corruption is one of the best ways of explaining the suc- the bonds’ characteristics, we test if their yields were higher cess of Russian bonds in France. Russian agents in Paris during the press campaigns and if the 1906 loan was penal- exploited the corrupt character of the French press to get ized because of its odious character. Our results indicate good coverage in exchange for bribes (Bignon and Flan- that protests increased the required yields of all Paris-traded dreau 2011). The campaigns were such that in 1898 Rus- Russian bonds and hence Russia’s future borrowing cost. sian bonds traded at almost the same value as their French Yields on the 1906 loan, which was specially targeted by counterparts (Lysis9 1914, p. 94). According to Lysis (1914), the campaigns, rose even higher than the yields of the other the fact that a country with accumulated defcits and on the Russian bonds. Once the press campaigns stopped, however, brink of insolvency, a country without freedom, parliament yields trended downwards, highlighting the important role or industry could be perceived as creditworthy was entirely the press may play in odious debts. due to the banks’ propaganda. The French government was To develop our argument, we have organized the paper perfectly aware of the corruption. It actually encouraged as follows. The next section discusses the political context the situation, for example, when it suggested the Russian of the 1906 issue, as well as the protests staged when it was government pay the press to reduce the magnitude of the foated. “Data and Methodology” section presents the data negative news concerning the Russo–Japanese war and the and methodology. “Results and discussion” section discusses 1905 revolution (Long 1972a, b). The amounts involved the results; “Conclusion” section concludes. were substantial. According to Long (1972a, b), in just 2 years, from 1904 to 1906, FF 2.5 million was spent in press “subsidies”. For this, Russian agents bought space in various The 1906 Loan… and Organized Protests newspapers to publish “fnancial information” related to the country’s bonds (for example, drawing numbers for lottery The frst important Russian loan foated in France was issued bonds). In this way, they were actually buying the newspa- in 1888 (Girault 1973). The presence of Russian loans on the pers’ collaboration. Bignon and Flandreau (2011) estimate Paris bourse was no coincidence. At the eve of the Russian that on average, for the critical years 1905 and 1906, two revolution, in 1917, an amount close to 11,720,000 French out of sixteen pages in Le Rentier were devoted to Russian francs (FF) had been issued in Paris (Girault 1973, p. 24). bonds. The press was not alone in benefting from the “gen- The success of the Russian bonds was due to elements erosity” of the Russian empire. The head of the Paris bourse related to politics, the economy and the corrupt nature of as well as the French fnance minister also received their the country’s Third Republic. After 1890, France and Rus- share of bribes. sia moved gradually closer together. The Franco-Russian Even though Russia was used to placing international Alliance and the alliance between France and the UK gave loans, it faced extraordinary opposition when it tried to birth in August 1907 to the Triple Entente. At the turn of the foat a new issue in 1906 (Long 1972a, b). The opposition century, fearing an increasingly powerful Germany, France was such that the lead underwriter, Crédit Lyonnais, held a used its stock exchange to secure or reinforce alliances, most secret meeting with the representatives of the Tsar to decide notably with Russia. As a result, French companies were whether to go ahead with the issue (Bayle 1997). Caution encouraged to invest in Russian bonds in order to sustain was understandable. A year earlier, the Russian empire had that country’s industrialization (Oosterlinck 2016). Eco- made an unsuccessful attempt to launch a loan, despite hav- nomic motives only partially explain the presence of Russian ing paid hefty bribes to the press (Rafalovitch 1931, p. 121). loans in French investors’ portfolios. Russia had vast natural Part of the Russian opposition threatened to repudiate the resources. Furthermore, given its geographical size, Russia’s loan should they come to power (Crisp 1961). The strong huge potential market was often mentioned as a guarantee of hostility to this loan can easily be explained by the Russian good fnancial prospects. Russian debt also enjoyed a good political context at the time. reputation because the proceeds of the loans were often used to boost the economy by developing infrastructure and local industry, both generators of income (Flandreau 2003). Banks also had an interest in creating a large market for Russian 9 Lysis was the pseudonym of the French journalist Eugène Letail- loans. They were paid not only a commission for issuing leur, a staunch critic of the banking sector who denounced press cor- ruption (Bignon 2011). Lysis’s articles led, from 1907 onwards, to a them but also a fee to manage the liquidity provided by the debate with Testis (the pseudonym of the French banker Alexis Ros- Russian Treasury to pay the coupons (Girault 1973, p. 34). tand), see Ronzon-Bélot (2003).

1 3 Author's personal copy

208 S. Collet, K. Oosterlinck

The outbreak of the war with Japan in January 1904 came placing of the loan.10 At the same time, however, part of the in a period of extreme internal tensions. In January 1905, a fnancial press was minimizing the importance of the revolu- large number of workers assembled to deliver an Address tionary movements.11 In any case, by July 1906, when Pyotr to the Tsar (Figes 1997, p. 173). Troops which had been Stolypin became prime minister, political instability was no posted in the city killed close to 200 protesters, but rumours longer viewed as a major obstacle. The Russian 5% loan of soon circulated that thousands had been murdered. The event 1906, intended to cover the general needs of the Treasury, became known as “Bloody Sunday” and marked a turning saved Russia from bankruptcy. According to Siegel (2014, point in Russian history (Werth 2004, p. 42). At the same p. 86), the loan’s successful conclusion was “an indication time the Russo–Japanese war was turning to Japan’s advan- that Russia’s domestic turmoil was sufciently under control tage. The loss of Mukden in March 1905 was followed by the that the international fnancial community deemed the tsar’s maritime defeat at Tsushima in May 1905. The war ended empire a credit-worthy risk”. But that success was hard-won. when the Treaty of Portsmouth was signed on 5 Septem- As pointed out by Long (1972b), the 1906 loan was ber 1905. In parallel, the escalation of strikes within the exceptional in that it was the only Russian foreign loan country forced the Tsar to act. On 17 October 1905, the tsar which led to organized protests and even to retaliatory reluctantly signed a manifesto granting civil liberties, con- threats and intimidation. The movements opposed to the stitutional order, cabinet government and a democratically loan were very diferent in origin and nature. They included, elected parliament, the State Duma (Figes 1997, p. 191). The amongst other, the French Socialist Party, the Russian creation of the Duma was the most important concession Kadets as well as several Russian revolutionaries and the (Pipes 1990). Political activity gained ground. Constitutional Société des amis du peuple russe et des peuples annexés Democrats (Kadets) pushed for political reforms, whereas (Long 1972b). The unethical character of the 1906 loan led the Bolsheviks were preparing for armed insurrection. The other groups to protest as well. For example, the Ligue des Bolshevik insurrection began in December, following the Droits de l’Homme12 organized a protest campaign in Paris arrest of Soviet leaders. Early successes for the insurgents (Siegel 2014). The opposition started well before the issue were soon followed by crushing repression. took place. The real turning point in the campaign came The war, together with the repression of various insurgen- when Maxim Gorky launched his appeal against the loan cies, had exacted their toll on the Russian public fnances. a few days before it was issued.13 The text was published Increasing taxes was impossible in view of the prevailing in many German newspapers, in the London Times14 but in mood of agitation. Relying on local funds was also not an France only in L’Humanité. The bribes paid by the Russian option as state savings banks and commercial banks were government had dissuaded other French newspapers from experiencing mass withdrawals (Crisp 1961). A foreign loan printing Gorky’s appeal (Long 1972b). The intensity of the was thus the most logical way to deal with the state’s fnan- campaign was such that in July 1906 the French ambassa- cial imbalance. To insure against the risk that parliamen- dor to Russia, Maurice Bompard, attempted once more to tarians might oppose the loan, the government decided to convince the Russian ally to be more open, but to no avail hurry up and issue it (Sack 1927). Sergei Witte, the Russian (Poidevin 1970, p. 70). Prime Minister at the time, was convinced the Duma would Representatives of the French socialist party, such as Jean refuse the loan (Siegel 2014, p. 80). Indeed, Social Demo- Jaurès, also attacked the loan, stressing its immorality. The crats and the Socialist Revolutionaries were clearly opposed loan even became an issue debated in the context of the 1906 to any new borrowing. The loan was therefore foated before French presidential elections, with Socialist candidates put- the Duma could convene its frst meeting. Raymond Poin- ting up anti-loan posters. Paradoxically, the head of Banque caré, the French fnance minister, hesitated to give the green de Paris et des Pays-Bas considered that any sign of open- light to accept this loan on the Paris stock exchange, fear- ness from the French government regarding the Duma would ing it might be illegal without the Duma’s consent (Girault only aggravate the problem (Poidevin 1970, p. 71). During 1961). Anticipating the protests in France, Poincaré asked the third National Conference of the French socialist party, his counterpart from the Ministry of Foreign Afairs if there was any way of infuencing Russian policy and convincing the Russian government to adopt a more conciliatory stance 10 ACL, DAF 164/1. towards the Duma (Poidevin 1970, p. 69). Their eforts did 11 See for example La Revue Financière, 19 April 1906. Of course not bear fruit. Eventually, the French government reluctantly their position may well be due to bribes. 12 approved the loan on two conditions: the constitutional reor- Human Rights League. 13 ganization of the country and Russian support for its French Gorky M. Pas un sou au gouvernement russe. L’Humanité, 9 April 1906. ally at the Algeciras Conference (Girault 1961; Poidevin 14 Crédit Lyonnais The loan was also discussed in the Inter Parliamentary Congress in 1970). According to the archives, political July 1906. See for example Journal des Débats Politiques et Littérai- instability at the beginning of 1906 had an impact on the res, 20 July 1906.

1 3 Author's personal copy

209 Denouncing Odious Debts

Ilya Rubanovitch representing the Socialist Revolutionar- The archives from Crédit Lyonnais indicate that the bond ies stressed that the Russian nation would not recognize the was oversubscribed when issued.20 The syndicate had taken debts which helped the Tsar crush the revolution15 (Parti a frm commitment for the loan, with Crédit Lyonnais and Socialiste 1906, p. 17). The parties on the left of the Kadets, Banque de Paris et des Pays-Bas each taking 25% of the most notably the Social Democrats and the Socialist Revolu- issue, Hottinguer, Comptoir National d’Escompte, Société tionaries, forewarned French investors that the loan would be Générale and Crédit Industriel et Commercial being, respec- repudiated after the fall of the Tsarist regime (Long 1972b). tively, involved for 18.75, 12.5, 12.5 and 6.25% of the issue. The Socialist Revolutionaries considered illegal any loans The underwriters tried to increase the price, rationing the issued between 10 November 1905 and the convening of the number of securities as much as possible by creating fcti- constituent assembly in May 1906. Some liberal Russian tious subscriptions. Their hope was that the unmet demand newspapers shared this view. They put the blame partially on created by rationing would translate into sharp price rises the French, who agreed to foat the loan, and they suggested (Lysis 1914, p. 35). Press campaigns were also conducted to boycotting French goods as punishment (Siegel 2014, p. 82). present the Russian loan positively. However, once the bond The position of the Kadets was more ambiguous.16 Follow- had been listed, the underwriters realized that the public ing the dissolution of the Duma in July 1906, 200 deputies, was less enthusiastic than expected. Sell orders increased including 120 Kadets, pleaded for passive resistance to the and prices nosedived. To limit price decreases, the syndi- regime. In a statement made in Vyborg, the repudiation of cate began to buy back the bonds. By 23 April 1906, the debts issued without the Duma’s consent became acceptable syndicate had spent close to FF 50 million buying back the for the Kadets17 before being criticized again in September 1906 loan, the limit set by the syndicate for its interven- 1906. tions. In view of market developments, the limit had to be In France, debates about the 1906 loan continued raging increased to FF 100 million.21 The Russian government also well after it had been foated, and did not stop until after began intervening in the market.22 At some point, the inter- 1907.18 In December 1906, Le Temps19 reported that French ventions were suspended for a while because the Russian socialists would question the government about loans to Rus- government felt it was impossible to support the prices in sia. On 8 February 1907, the question of the Russian bonds view of the large numbers of sell orders.23 By the end of was debated extensively in the French parliament (Renouvin July 1906, buying resumed24 and stopped again during the 1959). At the same time, a dramatic struggle between the fall.25 The purchases made by the syndicate and the Russian Second Duma and the Tsar’s government was occurring in government became public knowledge when Paul Philippe, a Russia. French socialists fercely attacked lending to foreign French journalist,26 (1907, pp. 8–9) publicly accused Crédit nations. The socialist Gustave Rouanet systematically took Lyonnais of artifcially supporting the prices of the 1906 the example of the 1906 Russian loan to denounce the huge loan. According to Philippe (1907, p. 16), this support had transfers of French funds to Russia which, according to him, represented between one-seventh and one-eighth of total French investments in securities. Adrien Meslier considered 20 Minutes of the Syndicate Meeting held on 28 April 1906, ACL, it was the government’s responsibility to ensure that capital DAF 164/1. This success was also present in Russia, as shown by let- invested in Russia should be reimbursed. For Albert Willem, ters from the same archive. investors were bound to lose every cent invested in Russia, 21 Minutes of the Syndicate Meeting held on 23 April 1906, ACL, and loans issued without the Duma’s consent were illegal. DAF 164/1. 22 ACL, DAF 164/1. Letter dated 27 June 1906 from Banque de Paris et des Pays-Bas to the Russian fnance minister. 23 As pointed out by an anonymous referee, whom we wish to thank, 15 This was not just the perception of the socialists. For William the main focus of this paper, markets’ reactions following the denun- Lamsdorf, the Russian foreign minister, the funds would allow ciation of the 1906 loan as odious, is not the only aspect related to crushing the revolutionary movements (Poidevin 1970, p. 72). ethics as the underwriters and the Russian government’s attempt to 16 See the translation of an article by P. Milioukov published in 1936: manipulate prices may also be questioned on ethical grounds. Les “Libéraux” russes et l’emprunt de 1906. (Archives Crédit Lyon- 24 Crédit Lyonnais refused, however, to engage in the buybacks, nais, DAF 164/1). According to Milioukov, some members of the which, from then on, were conducted mostly by Banque de Paris et party had issued a statement of repudiation in April 1905 but were des Pays-Bas and the Russian government. Minutes from the issuing later disavowed by the Kadets. syndicate on 24 July 1906 (PV reunion Comité français du syndicat 17 Crisp (1961) reprints a memorandum written by two important de l’emprunt russe 5% 1906, 24 July 1906. ACL, DAF 164/1). members of the party, Prince Dolgorukov and V.A. Maklakov, which 25 Telegram from Russian Minister of Finance Vladimir Kokovtsov clearly shows their opposition to the loan in April 1906. dated 29 September 1906, ACL DAF 164/2.. 18 Opposition was also substantial abroad. For example, L’Humanité 26 Paul Philippe was a French journalist who, jointly with Eugène mentions protests against the listing of the bond in Vienna. La révolu- Letailleur, another journalist writing under the pseudonym of Lysis, tion russe. L’Humanité, 25 April 1906. fercely attacked Crédit Lyonnais and the Russian loans between 1906 19 Nouvelles du jour. Le Temps,9 December, 1906. and 1908 (Eveno 2003, p. 64).

1 3 Author's personal copy

210 S. Collet, K. Oosterlinck a dramatic impact on prices, at least until January 1907, organized by the various Russian parties, as well as the pre- when the support is supposed to have stopped. At the time he cipitation with which the loan was issued, strongly suggest claimed that the bonds would fetch a third of their value in that opposition was ferce. Since part of the proceeds were the absence of the interventions, which he denounced as ille- used to prevent further insurrections, it is also reasonable gal. Accusations were also made in the press: for example, to assume that the proceeds did not beneft the people. This in Le Temps in December 1906, the journalist deplored these point was explicit in the appeal written by Gorky. His piece interventions.27 Interventions by the Russian government did stresses that the money would allow a criminal government not stop in 1906. In his memoirs, the Russian fnance min- to pursue its crimes, leading to mass-murder.31 In an article ister Kokovtsov explicitly prides himself for having bought published on 14 April 1906, the loan is presented as a hostile back the 5% loan in the fall of 1907 when prices were around act against the Russian population.32 In view of the protests 69–70 of par to sell them back during the frst half of 1908 against the loan, French investors knew all these elements. at 87–88, making a proft of 1 million rubles for the State As a result, there is little doubt that the 1906 loan may be Bank (Girault 1973, p. 475).28 qualifed as odious. In fact, as part of the negotiations held Regarding the legality of the 1906 loan, scholars have after the repudiation in the 1920s, Russian representatives since attempted to determine whether or not the debt was would still stress the diference between pre- and post-1906 illegal or odious. For Renouvin (1959) the proceeds of the loans (Delaisi 1930, p. 10). This view was shared by many 1906 loan were used to fght the revolution.29 By contrast, in France: the 1906 debt was odious while “older” debts Crisp (1961) considers that the proceeds were meant to were not, as they had helped Russia to develop economically. cover extraordinary war expenditures. The fact that this loan did not have the consent of the Duma was also stressed (Sack 1927), even though for some authors such as Crisp Data and Methodology (1961), consent was unnecessary in view of the planned use of the funds. Since the Duma was never consulted, how- To analyze the market’s reaction to the press campaigns ever, it is easy to argue that the population was opposed to denouncing the 1906 loan as odious, we frst analyze the the debt. Considering the composition of the Duma at the yields of Russian bonds and then exploit the perceived dif- time, opposition to the loan would have been signifcant. In ference between the 1906 loan and those issued previously fact, it is because this opposition was expected that the loan by the Russian regime. To do so, we have constructed a was foated before the Duma could convene.30 The protests panel of long-term Russian bonds traded on the Paris Stock Exchange, the main foreign market for Russian bonds at the

27 time. Our panel is comprised of 13 bonds. The prices of "On commence à regretter ici qu’on n’ait pas laissé, livré à lui- même, le marché des fonds russes à un moment où la faiblesse était these bonds were collected on a weekly basis from the Cote déjà justifée". Le Temps, 5 December 1906. Translation: We are Ofcielle des Agents de Change for a period stretching from beginning to regret that we have not left the market for Russian funds May 1906 to April 1913. at a time when its weakness was already justifed. 28 The numbers of prices changes, as well as the highest and This raises an additional ethical question, whether it was inher- lowest value for any given day, were also collected. As the ently fair for the Russian government to exploit the bad reputation of the 1906 loan to buy it back cheaply. The fgures pointed out by the Russian fnance minister are fare from market prices suggesting either a private block sale or an attempt by the Minister to paint himself as a shrewd fnance minister (we thank an anonymous referee for this Footnote 30 (continued) observation, which we had failed to make). 29 the elected members of the Duma because it did so for a reason that "en janvier 1906, un gros emprunt… pour combler le défcit con- elected members of the Duma could not approve. sidérable que la guerre de Mandchourie et les événements révolution- 31 naires de 1905 ont creusé dans son budget, et pour assurer, selon les ] “Cet argent aidera seulement à accomplir des massacres. Ne don- termes employés par Lamsdorf, la "compression défnitive du mou- nez pas un sou aux bourreaux du peuple russe, bourreaux de corps et vement révolutionnaire". Translation: in January 1906, a large loan… bourreaux d’esprits! II m’est pénible de penser que l’Europe i civili- to make up for the considerable defcit that the Manchurian invasion sée, témoin de la façon dont un pouvoir barbare, par crainte de perdre and the revolutionary events of 1905 have made in its budget, and to sa position dans le pays, opprime, torture, tue des milliers d’hommes, ensure, in the words of Lamsdorf, the "defnitive compression of the que cette Europe aide précisément ce pouvoir dans son crime.” in revolutionary movement”. Gorky M. Pas un sou au gouvernement russe. L’Humanité, 9 April 30 1906. Translation: This money will only help to carry out massa- This point is clearly stressed in the Dolgorukov-Maklakov mem- cres. Do not give a penny to the executioners of the Russian people, orandum. “En contractant un emprunt (…) quelques jours avant executioners of bodies and executioners of minds! It is painful for l’ouverture de la Douma, le Gouvernement montre qu’il veut sou- me to think that a civilized Europe, which is witnessing the way in straire cette opération au contrôle des élus parce qu’il l’a fait dans un which a barbarous power, for fear of losing its position in the coun- but que les élus ne pourraient approuver”. Translation: By contract- try, oppresses, tortures, and kills thousands of men, that this Europe ing a loan (…) a few days before the opening of the Duma, the Gov- is helping precisely the political power in its crime. ernment shows that it wants to remove this deal from the control of 32 La révolution russe. L’Humanité, 14 April 1906.

1 3 Author's personal copy

211 Denouncing Odious Debts

French stock exchange did not at the time report measures of collaterals as diferences in guarantees may afect bond for liquidity, we rely on the number of daily price changes prices (Chabot and Santarosa 2017; Esteves and Tunçer as a proxy for liquidity. This approach has been applied by 2016). In our case, however, no collateral was specifcally Hautcoeur and Petit-Koñczyk (mimeo) for the Paris stock pledged for the reimbursement of any of the bonds. The loan exchange. It may be viewed as an extension of the clas- issued in 1906 was a long-term (50-year) bond paying a sic zero-return measure, which links the absence of price yearly 5% coupon due on May 1 and November 1. It was change to low liquidity (Sarig and Warga 1989; Chen et al. traded on several markets, but serial numbers were specifc 2007). Capturing the impact of liquidity is, however, tricky. to a given stock exchange. According to Freymond (1995), With an issued amount of FF 2,250,000,000, the 1906 loan 72% of the 1906 bond was traded in Paris. Bonds could was the largest loan ever issued by the Russian government be redeemed at will in Amsterdam, Berlin, Brussels, Paris, and was thus unlikely to be illiquid. New bonds, also called London or Geneva at a fxed parity stipulated on the bonds. on-the-run bonds, are more liquid than older (of-the-run) However, this option could only be exercised after 2 years.37 bonds. The market therefore usually requires a liquidity The bond was further explicitly protected from any conver- premium to hold of-the-run bonds (Goyenko et al. 2011). sion scheme for the frst 10 years. In our case one needs, however, to be more nuanced. The Table 1 provides the characteristics of the bonds used in difculties faced by the underwriter to unload the bonds our analysis. All of them are long-term instruments with probably meant that part of the issues was concentrated in maturities at issuance ranging from 50 years to infnity (in a few hands which would then have reduced liquidity. In the case of perpetuities). The coupons are mostly equal to this case, liquidity would have gradually increased over time 4%, with the 1906 loan standing out at 5% and three other with bonds being illiquid in a frst phase. Large sales by the bonds at 3%. The 1906 loan is also very diferent from the major underwriters could also have depressed prices when others as the amount borrowed dwarfs that of other loans they occurred. This is, for example, explicitly mentioned by (2250 million francs, compared with 700 million for the the Economist in June 1906 “The drop in the new Russian next largest issue). Only a few bonds were issued by the loan has caused no surprise, as it was expected that when the Rothschilds, as their investment bank decided not to lend large subscribers began to unload the premium of 6 per cent to Russia because of the pogroms taking place there. Some could not be maintained”.33 Furthermore, paying in instal- bonds could be called back after a specifed date, and most ments also led to strains on the markets when calls for the of them were traded on several exchanges. bonds were due. In order to ofer a meaningful comparison, the yields for The bonds in our panel had diferent characteristics which each issue have been computed. Finance textbooks con- we retrieve from an ofcial publication of the Russian Min- sider the yield to maturity (YTM) as the standard measure istry of Finance.34 From this source, we collected the fol- to assess the rate of return of fxed-income securities (Bodie lowing variables: the coupon rate, the maturity, the under- et al. 2011, p. 479). By construction, the YTM assumes that writer (as Flandreau and Flores 2009 have shown that at the there will be no default and that the bondholder will keep time underwriters played a certifying role35), the number of the bond until maturity. In many instances, reimbursement markets where bonds were cross-listed,36 and the existence of the bonds was partially done on basis of a sinking fund. of conversion options. We also checked for the existence One could argue that sinking funds afect the duration of the bonds as part of the issue is reimbursed before maturity. To address this issue, we assume that, if a bondholder will- ing to keep its bond up to maturity saw his bond recalled, 33 The Economist, 23 June 1906, p. 1050. The authors thank an that bondholder would immediately use the proceeds of the anonymous referee for pointing out this element. bond’s sale to buy a new bond at a similar price on the mar- 34 Manuel du Porteur de Fonds Russes. Emprunts émis ou garantis par l’état et côtés à l’étranger. 1903. Published in St. Petersburg by ket. Under this assumption, sinking funds have no impact the Russian Ministry of Finance. For the 1906 loan we rely on the on the YTM. prospectus as our main source pre-dates the issue of the loan. Furthermore, since the 1906 loan was callable, using its 35 Lawyers are often presented as another class of gatekeepers. Brad- YTM without taking into account its call option may be ley et al. (2014) show, however, that they played a minor role before misleading. Indeed since the beginning of the nineteenth WWII and were almost never mentioned in the prospectus. Our source does not mention them either. century, the ability to call a loan—or convert it, using the 36 This clause was mostly inserted for convenience. Since at the time all countries considered here adhered to the gold standard exchange rates were de facto fxed. Before the outbreak of war, prices on 37 “L’emprunt sera sérié de façon que l’une de ces places ne puisse both exchanges were extremely close. Because of the war, arbitrage vendre sa part sur les autres avant deux ans.” La révolution russe. was forbidden and, after the repudiation, prices began to diverge as L’Humanité, 16 April 1906. Translation: The loan will be issued in French bondholders were hoping to be bailed out by their government series so that one of these places will not be able to sell its share on (Bernal et al. 2010). the others before 2 years.

1 3 Author's personal copy

212 S. Collet, K. Oosterlinck

Table 1 Characteristics of Russian Bonds Issue year Maturity at issue Principal borrowed Coupon rate (%) Rothschild Term to call Non Trading conver- places sion

Russia 1880 1881 81 years FF 600,000,000 4.00 0 0 3 Russia 1889 1889 81 years FF 500,000,000 4.00 0 0 3 Russia 1890 em2-3 1890 80 years FF 300,000,000 4.00 0.5 0 3 Russia 1890 em4 1891 60 years FF 41,764,000 4.00 1 0 3 Russia 1893 em5 1894 81 years FF 178,037,500 4.00 0 1 2 Russia 1894 em6 1894 81 years FF 454,400,500 4.00 0 1 3 Russia Consolide em1-2 1890 81 years FF 700,000,000 4.00 1 0 4 Russia Consolide em3 1891 80 years FF 320,000,000 4.00 1 0 4 Russia 1901 1901 Perpetuity FF 424,000,000 4.00 1 14.01.1914 0 1 Russia 1891 1892 81 years FF 500,000,000 3.00 0 0 2 Russia 1896 1896 Perpetuity FF 400,000,000 3.00 1 01.01.1911 0 3 Russia 1894 1894 79 years FF 166,500,000 3.00 0 0 2 Russia 1906 1906 50 years FF 2,250,000,000 5.00 0 01.05.1916 0 3

terminology of the time—was a common feature for sover- leads to an upward bias in the analysis. As a result, “with eign bonds. France and Great Britain had converted part of embedded options, coupon yields can be a misleading meas- their sovereign debt during the nineteenth century. Conver- ure of the cost of borrowing” (Tomz and Wright 2013, p. 9). sions did not concern only major powers with sound public We address the fact that three of our loans (the 1896, the fnances. For instance, for 1886 alone, Neymarck (1902, p. 1901 and the 1906 loan) were callable in two ways. Firstly, 239) mentions planned or actual conversions for Belgium, we construct a dummy variable taking a value of one if the Denmark, Holland, Japan, Norway and Switzerland. As for bond is not callable. Secondly, we rely on the fact that a call- Russia, it had converted part of its debts at the end of the able bond is similar to an option-free bond minus the call nineteenth century (Neymarck 1902; Girault 1972; Waller option. We therefore compute the value of the call, subtract 1979). According to Siegel (2014), the consolidation of Rus- it from the bond and then compute the yield to maturity.39 sian loans had reduced the interest to be paid by 12.5 million Since the call option is an option on a bond, the method- gold rubles per year. Last but not least, in 1909 5% Russian ology developed by Black and Scholes (1973) cannot be Treasury Bills were converted into long-term bonds (Girault applied as it assumes infnite potential future values which 1973, pp. 489–490). are inconsistent with fxed-income securities. We rely there- At the beginning of the twentieth century, the conversion fore on the methodology developed by Ho and Lee (1986) mechanism was well understood. Contemporaneous authors to value options on fxed-income securities. The Ho–Lee explicitly discussed the impact of potential conversions on bond prices. For example, the price compression mechanism described in modern textbooks (Bodie et al. 2011, p. 481) was already mentioned at the time (Courtois 1902, p. 161). Footnote 38 (continued) Investors now, as then, are unwilling to pay a high price for But concern that the loan could be converted was raised by others and a bond which may be called by its issuer, as the coupon rate became seriously debated once Russian yields began to decline. 39 stated on the original bond may eventually be replaced by a Using yields to call would have been another approach but the bonds pay diferent coupons and hence were not as likely to be called, lower one. Expectations that the 1906 loan would be called an important element which the use of yields to call does not allow were already present when it was issued. L’Humanité noted addressing. Since the 1906 loan had one of the highest coupons, for example that, in view of its rate, a conversion could not bondholders would have expected the Russian government to convert be ruled out.38 When there is price compression, using YTM this issue before any other. This intuition is confrmed by the contem- poraneous press. For example, in December 1910 there were rumours on the Paris stock exchange that the 5% Russian bonds would be con- 38 “L’emprunt pourra être remboursé en 1916 à 100%. Vu le taux verted to 4.5% ones (Semaine Financière. Le Temps, 26 September à 5% cette possibilité n’est pas à exclure.” in La révolution russe. 1910). This would have left the 4% 1891 bond out of the conversion L’Humanité, 16 April 1906. Translation: The loan may be repaid in and the use of yields to call would therefore have biased our results. 1916 at 100%. Given the 5% rate, this possibility cannot be excluded. More fundamentally, and as pointed out by an anonymous referee, in Of course one could argue that, in view of its political leaning, view of its size, calling the 1906 loan would have entailed a strong L’Humanité, would attempt to discredit the new loan by all means. stress on Russian public fnances in order to fnance the conversion.

1 3 Author's personal copy

213 Denouncing Odious Debts

7

6

5

4

3

2 Average Yield of Russian Bonds Traded In Paris Min Max Russia 1906

1

0 2 3 7 8 1 1 2 2 6 6 7 7 7 8 8 8 8 9 9 9 0 0 0 0 0 0 1 1 1 1 2 2 3 2 9/5/1906 1/9/1907 9/1/1909 1/5/1910 8/3/1910 7/5/1911 1/1/1913 4/3/1907 7/8/1908 3/1/1911 5/2/1906 8/7/1907 3/4/1908 6/9/1909 6/5/1912 2/3/1909 10/9/191 1/18/191 9/27/191 3/13/191 2/20/190 2/16/191 9/14/191 8/16/191 2/12/191 7/25/190 6/26/190 5/27/190 9/30/190 4/28/190 3/30/191 5/11/191 6/22/191 5/24/191 11/8/191 4/24/191 3/26/191 5/15/190 8/19/190 4/12/191 1/31/191 6/13/190 9/18/190 4/15/190 7/21/190 7/17/191 3/17/190 1/22/190 12/7/191 8/28/191 10/17/1906 11/28/1906 10/13/1909 11/24/1909 11/20/1912 12/20/1911 10/30/1907 12/11/1907 11/11/1908 12/23/1908 10/26/1910

Fig. 1 Yield to maturity of the 1906 loan, average yield to maturity for the Russian bonds and high and low yields to maturity methodology derives the term structure of interest rates, by liquidity. This may not have been the case during the period imposing arbitrage-free conditions and using martingale considered in this analysis. By comparing the yields of the probabilities. The derived term structure of interest rates is 1906 loan with the yields of other Russian bonds, it is pos- a stochastic object where the initial term structure coincides sible to determine the extent to which bondholders were with the empirically observed one. This is done by calibrat- penalizing odious bonds. If bonds perceived as odious or ing the empirical market data with the model. The model issued without parliamentary consent traded at lower prices, requires estimating the volatility of interest rates, for which the government might have an incentive to win parliamen- we took the average volatility of Russian interest rates on tary consent for future issues. the Paris market. On basis of the call value derived by the Figure 1 provides the value of the YTM for the 1906 loan, Ho–Lee model, we computed a “conversion-free” yield to the average YTM for all Russian bonds as well as the highest maturity for the bonds with a conversion option. and lowest YTM. Over the whole period, yields experience a The yields provide insights into the costs of borrowing. declining trend. Amongst the Russian bonds, the 1906 loan At issue, the yield on the 1906 loan was the highest the Rus- stands out. Indeed for almost the whole period, it is the bond sian government had had to pay in France. Whereas Russia with the highest yield, rendering its fgure almost undis- could borrow at 3% in 1896, the yield at issue on the 1906 tinguishable from the one reporting the highest yield. The loan was 6.4% (Girault 1973, pp. 435–436). Since the yields yield of the 1906 loan is even more striking since both the on traded securities serve as benchmark for future issues, the Russian government and the underwriting banks intervened changing pattern of yields to maturity makes it possible to on the market to limit its price decline (Poidevin 1970, p. gauge an issuer’s future borrowing costs. Increases in yields 71). Without these interventions, the yields would have been thus entail a long-term cost. In the short run, however, the even higher and the diference should therefore be viewed as increase may beneft the borrower, which may be able to a lower bound. Quite logically, the average yield is substan- buy back the bonds on the market at a cheaper price.40 This tially lower than the yield of the 1906 loan. Figure 1 further was certainly the case for Russia. But to beneft from the shows the existence of substantial diferences in yields for low prices, the Russian government had to have sufcient the various bonds. This is striking as the diference in yields between bonds issued by the same sovereign is usually close to zero. Indeed, when two bonds are issued by the same 40 The authors thank an anonymous referee for pointing this out. country, macroeconomic variables afect them in a similar

1 3 Author's personal copy

214 S. Collet, K. Oosterlinck way. More generally, news related to the issuer should have expect the impact of expectations of legal action to be minor. a similar efect on all its bonds. Since the odious debt doctrine has not been recognized by Since our aim is to analyze yields, we use two macro- any court, the literature suggests that legal action on basis economic control variables capturing the overall economic of the odious debt doctrine is unlikely. Indeed, despite its situation of Russia over the years: Debt on population and moral appeal, “few countries have invoked odious debt as Export on debt.41 The data for these series come from the grounds for repudiation; none has succeeded45” (Gelpern Global Finance Database [see Flandreau and Zumer (2004) 2007, p. 81). Buchheit and Gulati (2008) also highlight how and the update in Accominotti et al. (2011)]. We also add the limited the invocation of this doctrine has been, noting that yields on French long-term bonds, the Rente,42 to control for it consists mostly of the Cuban case and of “the writings broad interest-rate movements on the French market. of an obscure law professor of the early twentieth century, In general, the difference in bonds’ yields might be and a single arbitral award of the same vintage”. Precedents explained by the various bonds’ characteristics. To control were even more limited in 1906 since the arbitral case men- for the diferences between the loans, we rely on the follow- tioned by Buchheit and Gulati (2008), the Tinoco case, was ing variables43: only settled in 1923. Investors were thus unlikely to believe that legal action would be taken against this loan.46 Even • Liquidity: number of daily price changes. nowadays legal actions following defaults remain uncom- • MaturityRemaining: Number of years remaining before mon (Panizza et al. 2009). One should stress, however, the bond is reimbursed. that exceptions may exist. The case of bonds issued by the • NonConvertibility: is a dummy variable taking a value of Maduro regime in Venezuela has attracted the interest of 1 if the bond cannot be converted. legal scholars, who have suggested that the legality of these • Odious: is a dummy variable taking a value of 1 for the bonds might be challenged in court and even suggested ways 1906 loan. to restructure these debts (Buccheit and Gulati 2017). • Rothschild is a dummy variable taking a value of 1 if If investor were unlikely to expect successful legal action, Rothschild was involved in the bond issue. one could conclude that denouncing debts as odious is in fact • TradingPlaces: captures the number of trading places a useless. At a time when reputation was of prime importance given bond was traded at. (Flandreau and Flores 2009; Tomz 2012), investors may, • Volatility: we proxy securities’ volatility by using the however, have anticipated that repudiating a tainted loan range of prices (high and low) for a given day. would have entailed less costs in terms of reputations than • Principal Amount: principal amount at the date of issue. repudiating other debts. Or in the same vein, governments might be more reluctant to help their citizens holding these In addition to explaining the reaction of the yields of all bonds and therefore “be less willing, to apply informal sanc- Russian bonds to the protests and the press campaigns, we tions, (such as sending gunboats) when the time came to do are interested in the specifc evolution of the yields of the so” (Ben-Shahar and Gulati 2007). This raises the question 1906 loan. For the 1906 loan specifcally, the diference in of an idiosyncratic risk of repudiation. yields could be attributed to several (non-mutually exclu- sive) elements: fear of legal action against the 1906 loan, idiosyncratic risk of repudiation of the 1906 loan or protests Footnote 44 (continued) against this loan. following the Elliott versus Peru ruling. It is in general hard to assess It is unfortunately impossible to measure the impact of whether markets price properly specifc legal features of sovereign expectations of legal action or of the idiosyncratic risk of bonds. For example, the impact of Collective Actions Clause on the repudiation.44 Even if it is impossible to rule this out, we price of sovereign debts is ambiguous (Bradley and Gulati 2013). 45 One notable recent exception is Ecuador, but this of course is well after the case analyzed here (King 2016). Markets reacted, however, 41 Tomz and Wright (2007, p. 352) fnd “a negative but surprisingly when the odious character of sovereign debts was debated. This was weak relationship between economic output in the borrowing coun- the case, for example when the USA refused after the Spanish-Amer- try and default on loans from private foreign creditor”. It is, however, ican war to recognize any responsibility for the debts issued in the standard to control for these macroeconomic variables. name of Cuba (see Collet 2013). 46 42 The rates of the Rentes are retrieved from Vaslin (1999). Note, however, that according to King (2016, pp. 96–100), even 43 The descriptive statistics of the diferent variables may be found in if the expression “odious debt” has seldom been used in the past, Appendix 1. subjugation debts (debts hostile or contrary to the major interest of 44 the population) are fagrantly breaching international law and should A large literature in law and fnance assesses the interplay between therefore not be enforceable. Their perception would have been legal conditions and bond prices and lending. This literature is, how- strengthened by the fact that, at the time, sovereigns were still subject ever, seldom dealing with sovereign debts (Chamon et al. 2015). to the rule of absolute immunity and not the restrictive rule which has Recent papers on this topic include Bradley et al. (2010) who analyze nowadays become more and more accepted (Weidemaier 2014; Ver- pari pasu the impact of the provisions clause on the pricing of bonds dier and Voeten 2015; Weidemaier and Gulati 2018).

1 3 Author's personal copy

215 Denouncing Odious Debts

One could argue that the 1906 bond faced a unique risk need a new loan soon.48 Since the most important part of the of repudiation or default. In this case, Russia would have press campaigns happened between April 1906 and Febru- defaulted only on this specifc loan. The literature on sov- ary 1907, it is impossible to disentangle an idiosyncratic ereign debt shows that selective defaults exist. Sovereigns default efect from a press campaign. The press campaigns may distinguish between domestic and foreign debts (Erce waged against the 1906 loan could in theory have afected and Diaz-Cassou 2011; Acharya and Rajan 2013), between investors’ willingness to buy Russian bonds in several ways. classes of lenders (Aizenman et al. 2013) or between types First, the attacks waged against the 1906 loan when it was of debt instruments (King 2016). In general, however, within issued may have led investors to revise their view regarding a given group, sovereigns usually attempt to minimize the Russian loans in general. Protests may have led investors to variations of haircuts across bondholders (Zettelmeyer et al. develop fundamental concerns with the creditworthiness of 2013). The government of Ecuador recently decided to sin- Russia and with the potential for price manipulation. This gle out its 2012 and 2030 Global Bonds by alluding to their would have led to an increase in the yields of all Russian odious character, amongst other elements, to justify its deci- bonds. But by targeting a specifc loan as being odious, the sion (Gentile 2010; King 2016, pp. 90–91). Historical prec- campaigns may also have led to a diferent treatment for edents could have infuenced bondholders’ perception. A law this loan in particular. This could have happened in many passed in Mexico in 1883 repudiated some of the Mexican ways. First, the press campaigns, by highlighting the odious debts because they had been issued “by the governments character of this specifc loan, would have reduced demand that pretended to exist in Mexico from 17 December 1857 to and hence prices. Second, by denouncing the loan, the press 24 December 1860 and from 1 June 1863 to 21 June 1867” could have convinced investors that investment banks would (Sack 1927, 158; King 2016, 72). Bondholders could have be unwilling to support the loan should it face difculties on therefore expected that one of the political parties opposed to the market. the bond would take power and then decide to repudiate only To assess the importance of protests, we rely on two vari- the 1906 loan. Some political parties could have decided to ables, for which the descriptive statistics can be found in follow such a path. But most parties would probably not have Appendix 1: decided to do so. The extreme left was expected to repudi- ate all the bonds, viewing the Tsarist regime as odious and • BadPublicityIndex: is a continuous variable counting thus all its debts as such, and this proved to be the case in the number of words with a negative connotation associ- 1917 (Oosterlinck 2016). When rumours about the dissolu- ated to the Russian loans in the newspaper L’Humanité. tion of the Second Duma started to spread, even the press L’Humanité was chosen because of its political orienta- hostile to Crédit Lyonnais viewed the risk of default for all tion and because it was one of the few newspapers not Russian bonds in a similar way. For example, in its 1 March infuenced by Russian bribes (Rafalovitch 1931). Up till 1907 issue, Bourse Libre, directed by Paul Philippe, recom- 1908, mentions of Russian loans appeared on a regular mended its readers to sell all their Russian securities. The basis in L’Humanité. Broadly speaking the articles were fact that a clear opponent to Crédit Lyonnais recommended either discussing Russian loans in general or focusing a similar course of action when the Russian left was gaining their attention on the 1906 loan specifcally. momentum strongly suggests that market participants were • Protest: is a dummy variable which takes a value of 1 for not expecting a selective default. The Kadets’ position was at the period during which most of the protests were held frst ambiguous, but very quickly most of its leaders pledged (April 1906–February 1907) to recognize the loan. The probability of seeing only the 1906 loan repudiated would then rest on the Social Demo- To determine the impact of protests on the yields, we crats taking power and then repudiating the loan. But was performed a panel regression to explain the yields of the 13 this likely? Certainly not by all historical accounts. Russian Bonds from our panel. More specifcally, we run Another possibility could have been to see the loan fail the following model: because some investors would have been reluctant to pay the = + +…+ + yit x1it 1it x2it 2it xkit kit it instalments when due. Indeed, according to the Financial (1) i = 1, … , N t = 1, … , T Times in July 1906, unwillingness to pay these instalments For and 47 y i t x could directly impact the markets. This impression would In this model, it is the yield of bond on day , i = 1, … , N certainly have been exacerbated by the fact that the fnan- ( ) are vectors of covariates, are the corre- cial press at the time reported rumours that Russia would sponding vector coefcients, and is the error term. Panel

48 See for example The Crisis in Russia. Financial Times, 21 July 47 The Stock Market. Financial Times, 21 July 1906. 1906.

1 3 Author's personal copy

216 S. Collet, K. Oosterlinck regression may require the use of fxed or random efects. In a second phase, we rely on the Bai and Perron (1998, To formally test between fxed efects and random efects, 2003) test to determine the breaks and their magnitude. In a Hausman test is performed (Hausman 1978). The Haus- an initial phase, the tests proposed by Bai and Perron (2003) man test assesses the existence of systematic diference in (the UDmax and WDmax tests) can determine if at least one coefcient between fxed and random efects panel regres- break is present. If the frst step indicates the presence of at sions. When the null hypothesis of no systematic diference least one break, then the number of breaks can be assessed is rejected, fxed efect panel regression should be used. on basis of a sequential examination of the sup FT statistic. In addition to our panel analysis, we rely on a structural break methodology to determine the main turning point in the spread between the yield of the 1906 loan and the aver- Results and Discussion age yield of the Russian bonds used in the panel regres- sion. By using the spreads instead of the yields of the 1906 Figure 1 shows that the yields of the 1906 loan were higher loan only, we correct for events afecting all Russian bonds than the yields for other bonds. To explain this diference, in a similar way and therefore focus on events driving the we analyze the yields of the Russian loans on basis of a yields of the 1906 loan specifcally. The structural break panel regression. Table 2 provides the results of the panel approach is commonly used in economic history.49 Since analysis under diferent specifcations. We frst regress only breaks are determined endogenously, this method captures the yields on our main variables of interest: the variables the perceptions at the time of the events, and thus guarantees linked to the protests and the interaction terms with the odi- that the analysis will not be afected by any ex-post bias. The ous nature of the 1906 loan, using robust standard errors analysis rests on the break test developed by Bai and Perron (regression 1 and 2). Once control variables are included, (1998, 2003), which makes it possible to detect multiple the Hausman tests reject the null hypothesis of no systematic structural breaks. We apply the methodology on the entire diference between fxed and random efects; we therefore period to identify when investors reassessed the diference report only the fxed efects panel regression (regression 3 between the two bonds (measured here by their spread). In and 4). practice, a break is said to occur when the intercept of the Identification of variables which are invariant within autoregressive process suddenly changes. Those shifts are bonds, such as the Odious Debt variable, is not possible due identifed in two steps. to the inclusion of fxed efects. An alternative way of investi- In the frst phase, the analysis determines the exact num- gating the link between yields and these variables is to run an ber of breaks occurring during the whole period. More for- OLS regression and to cluster standard errors within bonds. mally, we consider the following multiple linear regression In this setting, errors are independent across clusters but cor- with m breaks and thus m + 1 regimes related within clusters. This allows time-invariant variables

� � such as Odious Debt, Rothschild and Trading Places to be = +  + , = − 1 + 1, … yt xt zt j ut t Tj Tj (2) modelled. We therefore regress the yields on our main vari- ables of interest and all the control variables, including the j = 1, … , m + 1 For . In this model, yt is the spread between variables which are time-invariant using clustered standard the yield of the 1906 loan and the average yield of the Rus- errors. The cluster used in regression 5 and 6 is Odious Debt, sian bonds on day t, xt and zt are vectors of covariates, and while Trading Places is used in regression 7 and 8.51 j = 1, … , m + 1 and δj ( ) are the corresponding vector coef- The control variables have the expected sign: higher fcients, and ut is the error term. This frst step allows us to Exports on Debts lead to lower yields, and, in most specifca- separate the data into m + 1 regimes. To do so, the procedure tions, higher Debt on Population increases the yields. Higher requires setting a maximum for the number of breaks. In our Volatility is linked to higher yields, an observation consistent case, this maximum was set at eight, the results being robust with expectations as higher risk should be compensated by 50 to a higher number of breaks. Equation (2) is then esti- a higher required risk premium. Our proxy for liquidity is mated for the spread over the whole period (2 May 1906–16 either negative, which would be expected since liquid bonds April 1913). are less risky, or is not statistically diferent from zero. This observation may refect the fact that our proxy for liquidity fails to fully capture liquidity efects. Bonds foated by the Rothschilds trade at a lower yield, an observation consist- 49 See for example Willard et al. (1996), Weidenmier (2002), Brown ent with the results from Flandreau and Flores (2009). Bonds and Burdekin (2002), Zussman et al. (2008) or Flandreau and Ooster- traded on several exchanges have a statistically signifcantly linck (2012). 50 The procedure also requires setting a value for a trimming param- eter h which depends on the maximum number of breaks. In our case the trimming parameter h is equal to 0.1. 51 Note that using other clusters provides similar results.

1 3 Author's personal copy

217 Denouncing Odious Debts

Table 2 Panel regression, dependent variable yields to maturity Specifcations (1) (2) (3) (4) (5) (6) (7) (8) (9)

OdiousBond – – – – 0.681*** 0.676*** 0.681** 0.676** Protest 0.637*** 0.208*** 0.317*** 0.317*** OdiousPro- 0.672*** 0.468*** 0.468*** 0.468*** test BadPublici- 0.998*** 0.374*** 0.551*** 0.551*** tyIndex OdiousXIn- 0.948*** 0.662*** 0.659*** 0.659*** dex Liquidity − 0.0140* − 0.0140* − 0.00468 − 0.00463 − 0.00468 − 0.00463 − 0.0251** Volatility 0.219*** 0.217*** 0.134*** 0.127*** 0.134*** 0.127** 0.319*** MaturityRe- 0.000378*** 0.000367*** 0.000232*** 0.000237*** 0.000232* 0.000237* 1.56e−05 maining DebtPopula- 9.711 10.18 63.32*** 62.39*** 63.32** 62.39** 137.4*** tion ExportDebt − 4.630** − 4.901*** − 4.898*** − 4.942*** − 4.898*** − 4.942*** − 5.996** Trading- 1.328*** 1.355*** 1.328* 1.355* 0.303* Places Rothschild − 1.988*** − 2.033*** − 1.988* − 2.033* − 0.273* NonConv − 0.185** − 0.198* − 0.185 − 0.198 0.264 FrenchRate 1.618*** 1.704*** 1.618*** 1.704*** 0.988*** Principal − 0.000233 Constant 4.411*** 4.375*** − 5.199** − 4.941*** − 12.02*** − 12.45*** − 12.02*** − 12.45*** − 6.578*** Observations 3.011 3.011 3.010 3.010 3.010 3.010 3.010 3.010 2.646 R2 0.290 0.280 0.733 0.738 Fixed efects 1 1 1 1 0 0 0 0 0 Clusters 0 0 0 0 Odious debt Odious debt Trading Trading Trading places places places Number of 13 13 13 13 13 13 13 13 12 (1906 bonds omitted)

***p < 0.01, **p < 0.05, *p < 0.1 higher yield, an observation for which we have no explana- Our main variables of interest, protest and bad public- tion. Yields are positively related to the yields on the French ity index, are in all specifcations positive and statistically government bond, the Rente. The non-convertibility coef- signifcant. Protesting against the issue of the 1906 loan cient is either negative or not statistically diferent from zero. had thus an impact on the yields of all Russian bonds. This This is in line with priors: bonds which could be converted impact is large in all specifcations. If one follows the odi- were less attractive to their holders. The remaining maturity is ous regime approach, then one could argue that all the debts associated with a higher yield, an observation consistent with issued by the Tsarist regime were odious and were therefore an upward sloping yield curve. One could argue that the yields afected by the protests. The price movements may also have are afected by the amounts borrowed. As shown in Table 1, refected a change in the nature of the investors. There is no the amounts borrowed for the 1906 loan are so high compared precise data regarding bondholders, but archival evidence with those of other loans that collinearity prevents us from indicates that individual investors were an important seg- running a regression with both terms. To address this issue, ment of the Russian bond market.53 One could argue that we test whether the amounts borrowed increase the yields in individuals were more sensitive than institutional investors a subsample composed of all bonds except the 1906 one. The results (specifcation 9) show that this is not the case.52 53 The archives from the French bondholders association (Associa- tion Nationale des Porteurs Français de Valeurs Mobilières) as well 52 One could of course argue that because the amount of the 1906 as pamphlets from the British Union of Russian Bondholders show loan was so high, the results for the other bonds might not be rep- bondholders’ diversity in terms of social classes. Oosterlinck and resentative because of non-linearity. Unfortunately, not much can be Landon-Lane (2006) mention that, in 1919, 1.6 million of French done to address this concern. investors fled claims related to Russian securities.

1 3 Author's personal copy

218 S. Collet, K. Oosterlinck to the odious character of the bond. Bearing this in mind, A note detailing the expenses made to foat the bond men- one way to read the price movements could be the following. tion payments for press “advertising55” amounting to almost In an initial phase, institutional investors made a frm com- FF 1.6 million. mitment to the loan. Subscriptions began before the protests. In order to determine the main specific drivers of the By the time the loan was issued, protests and press cam- yield to maturity of the 1906 loan, we look for dates when paigns against it had gathered momentum. Some individual the spread between the yields of the 1906 loan and the investors then began to realize the odious nature of the bond average yield of Russian loans experienced structural and decided to sell their holdings. Institutional investors, less breaks. We focus only on the most statistically signifi- concerned by ethics, would then have gradually arbitraged cant breaks and therefore report and analyze break points between this loan and the other Russian loans. These trans- identified at the 95% confidence level. The breakpoint actions would have taken time because banks already had a methodology singles out three break dates: 9 January signifcant amount of Russian loans in their portfolios but 1907; 17 November 1909 and 26 June 1912. The first and this could explain why the spread narrowed. third breaks have a negative impact on the spread, while The Odious Debt dummy variable (or 1906 loan dummy) the second one increases it. As is standard in economic is statistically diferent from zero and positive in all regres- history, the breaks are then linked to potential explana- sions. Being labelled as odious thus increases the yield of tions found in the contemporaneous press or in secondary the bond. The interaction term between the odious dummy sources.56 As will be detailed below, these breaks may be variable and both protests variables (protest and bad public- linked to bondholders’ reassessment of the odious nature ity index) is not only statistically signifcant but also impor- of the Russian regime. Actions which reinforced the tant in magnitude. Both results would tend to indicate that as perception that the Russian regime would become more long as press campaigns denounced the 1906 loan as odious democratic (mainly smooth elections at the Duma) were its yields were strongly afected and this in addition to the perceived positively, whereas actions limiting the demo- general impact on Russian bonds described previously. How- cratic process were penalized. Most breaks are thus linked ever, once the press campaigns stopped, the odious nature to political events taking place in Russia. The role played of this specifc bond became less relevant even though it by the Duma on markets may seem extremely important. remained observable. Investors’ attitude to ethics may thus It is noteworthy to recall that at the time “it was con- have changed over time. The most ethical bondholders prob- fidently assumed that Russia’s dependence on Western ably refused to buy any Russian bond as this would have finance, renewed in 1906 with the biggest foreign loan in meant indirect support for an autocratic regime. Others may its history, would force him [the Tsar] to retain the liberal have decided to exclude only the 1906 loan, either because structure of the state” (Figes 1997, p. 216). The press they were convinced that it was “unfair” or because they also made direct links between the position towards the felt it was socially unacceptable to hold this loan. And of Duma and bond prices. For example in its 23 June 1906 course, some bondholders may have been indiferent to the issue, The Economist57 noted “The heavy fall in Russian ethical issues. Both extremes (the most and least ethical securities that has been recorded this week is the natural bondholders) were unlikely to change their position. But outcome of the renewed anarchy that has arisen in Russia, the middle-ground may have been afected by the environ- and the bad faith shown by the authorities in dealing with ment. One could argue that as long as the 1906 loan was the Duma”. Political actions which altered this perception mentioned as being odious, these investors refused to buy were therefore highly likely to affect markets. it. However, once the press gave less space to this question, the social pressure to refuse the bond eased. In addition, the 55 attention that investors paid to this specifc bond declined. At the time, buying advertising space was a very common way to bribe newspapers. These factors may have prompted some investors to engage 56 The methodology is not exempt from criticism. For example, if in arbitrage, buying the “cheap” 1906 bond and selling the two important events (one increasing the spread, the other decreas- other, more expensive Russian loans. In this case, the press ing it) occur on the same date, the methodology will in all likeli- would have played an important role. By stressing the odi- hood fail to capture them as they will cancel each other out [see for ous character of the 1906 loan, the media would have had a example Frey and Waldenström (2007) and Oosterlinck and Ureche- Rangau 2012]. Furthermore, when looking at the press, the researcher direct impact on the bond price. At the time, bankers high- may only consider events which ft with his theory. In our case, the lighted the major part played by the press. An internal memo events likely to move the spread are well circumscribed and this bias for the president of Crédit Lyonnais, the lead underwriter, is therefore less likely to occur. suggested bribing the press to prevent negative coverage.54 57 The fall in Russian securities. The Economist, 23 June 1906 p. 1047. See also Foreign government securities. The Economist, 23 February 1907 p. 347 and The stock market. The Economist, 23 54 Crédit Lyonnais archives, DAF 162, Undated internal memo but in March 1907 p. 507 on the link between the Second Duma election all likelihood issued in 1906, Note pour Monsieur le Président. and the prices of Russian bonds.

1 3 Author's personal copy

219 Denouncing Odious Debts

Break 1: 9 January 1907; Magnitude: − 32.09 Basis from the Thousand and One Nights.61 The public at large Points probably ignored why Steeg decided not to question the government on Russia, but his reversal may have been The frst break observed in January 1907 is very likely interpreted by some as a sign that the Russian govern- linked to the pressure exerted by the French government to ment was not as bad as usually painted. Furthermore, if present the political situation in Russia in a positive way. the socialists’ accusations were somehow perceived as The break date corresponds to the opening of the electoral unfounded, then buying the 1906 loan would have been campaign for the election of the Second Duma. Part of the less unethical than previously thought. French press estimated at the time that the Kadets were in The break may, however, also refect a general improve- the best position to win the elections as the political parties ment in the Russian economy if one assumes that the yield from the left were divided.58 The fact that elections were of the largest loan issued (the 1906) would be more afected taking place in a constitutional manner may explain the by news than the other Russian bonds. This explanation can observed break. By signalling its respect of institutions, be found in the Financial Times and the French newspaper the Russian government’s image probably improved and Le Matin, which both reported that the fnancial situation of may have led investors for whom ethics mattered to revise Russia as more satisfactory than expected and that no new their views of the regime. Eventually, the left emerged loans would therefore be needed.62 victorious from the elections, taking 222 seats compared Alternatively one could argue that the break is due to with 98 for the Kadets, 60 for the Centrists and 54 for the elements related to the way the 1906 loan had been foated. Octobrists59 (Figes 1997, pp. 224–225). The Second Duma In France, the loan could either be paid up-front (in which convened in February 1907 and remained into place for case the bondholders received a certifcate stating the loan only a short period of time.60 had been paid in full) or in instalments. The subscription The break occurs also close to a time when the French price was 88% of par, an amount future bondholders would parliament was locked in heated debate. The loan was have to pay as follows: A frst instalment of 10% had to be attacked by members of the Socialist Party. Jean Jaurès, paid up front, a second (18%) between 2 and 5 May, 1906; a for example, highlighted that the Russian loans were third (20%) on 1 August 1906; a fourth (20%) on 2 Novem- issued by a government that had little concern for its ber and the last instalment on 1 February 1907. In addi- population and only managed to float this loan because tion, accrued interest had to be added at a rate of two per it was being assisted by bankers interested only in short- cent from 1 August 1906 onwards.63 Since the identifed term profits (Girault 1973, p. 468). By contrast, the breakpoint occurs close to 1 February 1907, one could argue French government, as well as French bankers, did not that the break refects the successful completion of the loan, really mind the repression in Russia. Their main objec- an element which would have driven both the yields of the tive was to keep good relations with an important ally 1906 loan and the spread downwards. In Paris, the market and a good client. As a result, the French government was quoted both type of certifcates: those for full payment and eager to limit protests and questions in parliament. To do so, it undertook to “convince” part of the opposition 61 to curb their attacks. It managed, for example, to reach "D’ailleurs, que n’a-t-on pas dit au cours de cette discussion, qui an agreement with Théodore Steeg, a radical socialist, rappelle par bien des côtés un conte des Mille et une Nuits? On a révélé que M. Noetzlin, le négociateur du dernier emprunt russe, avait who had shown his intention to question the government touché 1 0/0, puis 20/0 sur le montant de cet emprunt, ce qui ferait, on Russia (Girault 1973, pp. 446–447). Press campaigns dans l’un ou l’autre cas, 12 ou 24 millions. C’est beaucoup pour un were also waged to portray the socialist intervention as seul homme. Que c’est beau l’imagination!" Translation: Besides, ludicrous. For example, in its 11 February 1907 issue, Le what has not been said during this discussion, which recalls in many ways a tale of the Thousand and One Nights? It has been revealed Temps made fun of socialists’ action, stressing that the that M. Noetzlin, the negotiator of the last Russian loan, had received accusations against the underwriters resembled the Tales 1 per cent, and then 2 per cent, on the amount of this loan, which would, in one or the other case, be 12 or 24 million. That’s a lot for one man. How beautiful imagination is!. 62 We thank an anonymous referee for this comment. The Financial Times (4 January 1907, p.5, Dalziels Telegrams.), citing an article in Le Matin of the previous day, states that ’an investigation made on the spot by a high and competent French personage shows that the 58 Bulletin de l’étranger. Le Temps, 11 January 1907. Russian fnancial year will close under the best of conditions, and that 59 The Octobrist Party was a liberal group favoring moderate politi- the fnancial situation generally is much more satisfactory than could cal reform (Figes 1997, pp. 194–195). have been expected…There is consequently no danger of a defcit, 60 In June 1907, Stolypin, who was known mostly for the repressions neither will there be any need to have recourse to a new loan’. conducted under his supervision, dissolved the Duma and changed 63 Document from Crédit Lyonnais dated 19 April 1906 (Crédit lyon- the electoral laws to get a parliament more to his taste. nais archives ACL DAF 164/1.

1 3 Author's personal copy

220 S. Collet, K. Oosterlinck those for partial payments. Up till 1 August 1906, prices 96.00 are provided for a certifcate for a 28% payment (5% 1906, listed under the terminology “Certifcats prov. Francais, 92.00

28% payés”). From 1 August to 2 November that certifcate 68% paid was replaced with one for a 48% payment, which itself was 88.00 eventually replaced as from February 1 by a certifcate for e 48% paid a 68% payment. Pric 28% paid Paying the instalments required very large amounts of 84.00 cash. This may have prompted investors to sell prior to the call. There is no data regarding supply and demand for the 80.00 1906 loan. However, the number of daily price changes, as well as the highest and lowest prices for a given day, is avail- 76.00 able. We rely on this data to assess whether the market was 18-4-1906 7-6-1906 27-7-1906 15-9-1906 4-11-1906 24-12-1906 12-2-1907 Date more active than usual prior to the dates when the instal- Fully Paid(average) Partially paid (average) ments were due. More precisely, we compare the number of diferent prices (as proxy for liquidity) and the price range Fig. 2 for the certifcates prevailing before the instalments and 1906 Certifcate Prices (Fully Paid and Partially Paid) those that succeeded them. The last data point for the certif- cate representing a 28% payment has two daily changes, with Break 2: 17 November 1909; Magnitude: 17.62 Basis a low and high price of 79.40 and 79.45. On the frst trading Points date of the certifcate with a 48% payment, prices changed 16 times during the day, with a low of 78.50 and a high of The break date coincides with the date of the Finish crisis of 80.00. Similarly, on the last date, the certifcate with a 48% 1909. In October 1909, the Russian government requested payment was traded, prices changed 11 times, with a low from Finland a contribution to fnance its military expen- of 84.65 and a high of 85.15, respectively. The certifcate ditures.65 On 17 November 1909, the Finnish parliament replacing the 48% (the 68%) experienced 21 changes on its (the Finnish Diet) refused to approve these expenditures.66 frst trading day with a low and a high of 84.15 and 85.05, This position led to the dissolution of the Diet by Impe- respectively. If investors had rushed to sell their certifcates rial Ukase.67 In France, l’Humanité presented the dissolu- before the call, then the number of daily changes and the tion of the Diet as a direct attack on the fundamental laws price range would have been higher for the last trading day of Finland.68 This position was not confned to the far-left of a given certifcate than for the frst trading day of the press. The Economist would for example write “the Rus- new certifcate. This is not what is observed.64 This tends to sian government’s latest eforts to override the rights of the prove that investors refrained from selling prior to the call Grand Duchy of Finland is likely to prove detrimental to the to avoid the cost of obtaining sufcient liquidity to meet it. creditworthiness of the Russian Empire”.69 Another way to deal with this issue is to look at the price Far from being anecdotal, the position adopted by the diference between the various certifcates. Indeed, if mar- Russian government, and more precisely by Stolypin, ket participants feared that a large number of subscribers showed that neither were prepared to respect the new con- would refuse to buy the bond, and thus pay the last instal- stitutional situation which had emerged in Finland after 1905 ment, then the price diference between the various certif- (Seton-Watson 1967, p. 668). Stolypin’s personal opposi- cates should refect this concern. Figure 2 shows the price tion to Finland was also fuelled by the important number diference between the fully and partially paid 1906 bond. of Russian revolutionaries who had used Finland as a safe The diference between the highs and lows of each certif- refuge. Investors for whom ethics mattered may therefore cate averages 0.27 for the frst period, 0.62 for the second have refrained from buying the 1906 loan or may even period and 0.67 for the last period, the two last fgures being strongly infuenced by outliers on 17 and 31 October 1906, and 21 November 1906. The fact that the fully paid certif- cate traded at a higher price than the provisional ones can 65 La question fnlandaise. Le Temps, 18 November 1909. hardly be viewed as a sign that market participants were 66 E. Roubanovitch, La lutte contre la réaction russe, L’Humanité, 1 seriously concerned about the outcome of the loan issue. December 1909. 67 La question fnlandaise. Le Temps, 20 November 1909. 68 E. Roubanovitch, La lutte contre la réaction russe. L’Humanité, 1 64 This fnding holds if, instead of looking at the last trading day December 1909. only, one considers the whole month before the instalment. 69 Finland and Russia. The Economist, 30 October 1909, p. 859.

1 3 Author's personal copy

221 Denouncing Odious Debts have decided to sell it in view of the repression exercised debts, it may be wondered whether markets consider odious in Finland. debts any diferently from other debts and whether it pays to denounce these debts as odious. Break 3: 26 June 1912; Magnitude: − 16.29 Basis This paper exploits a unique case, the 1906 loan issued by Points Russia, to determine markets’ perception of debts presented as unfair to repay and to establish whether protests have For the third break, two potential explanations may be sug- an impact on the issue of such loans. This case is unique, gested. The break date is indeed close to the transition from for several reasons. First, only one loan was singled out by the Third to the Fourth Duma and to the visit of Raymond the press as being odious. Even opponents to the autocratic Poincaré to Russia. In November 1907, the Third Duma Tsarist regime made a distinction between this specifc loan started working under a broad majority of conservative par- and other debts issued by Russia. This loan was questioned ties, with the pro-government parties controlling 287 out of because its proceeds were used to crush revolutionary move- 443 seats (Figes 1997, p. 225). It remained in power until ments in Russia but also because it was foated without par- 1912. Even though the importance of the Duma declined liamentary consent. Second, the fotation led to massive over time (Figes 1997, p. 244), markets may have perceived protests and press campaigns in the countries where it was a smooth political transition in a positive way. to be listed. In fact, this loan was so controversial that it even The break might, however, also be linked to the planned became a point of discussion during the French presidential visit of Raymond Poincaré to Russia. Between 1908 and elections. 1912 the relationship between France and Russia had come To assess the impact of press campaigns denouncing odi- under pressure. The British and French position during the ous debts and regime, we rely on a panel of Russian bonds Bosnian crisis had left Russia isolated and, as a result, sym- traded in Paris. Our results frst indicate that the press cam- pathy for France had plummeted in Russia (Tomaszewski paigns had a negative impact on all Russian loans, suggest- 1999). The growing tensions with Germany meant France ing that, following the press campaigns, many investors con- had to guarantee that Russia would come to its aid should sidered the regime as odious. The results further show that a war break out. In this setting, Poincaré’s visit marked “a market participants were reluctant to hold the 1906 loan in signifcant turning point in the evolution of the Franco- particular and still preferred other loans issued by Russia. Russian alliance” (Tomaszewski 1999, p. 373). The press This diference of treatment disappeared once the intensity reported the planned visit,70 and one may conjecture that of the press campaigns diminished. Once protests and nega- investors viewed it as recognition of the importance of the tive press reports subsided, investors began to buy “cheap” Russian ally by the French government. If France felt Russia 1906 bonds. Our results thus highlight the importance of deserved to be treated as an equal, bondholders may have protests and press campaigns. While some investors would reconsidered their position on the country’s debt. However, never have bought the 1906 loan, others were infuenced by since the link is not mentioned in the press, this proposi- the negative press coverage. The press has thus an important tion should be viewed as a suggestion rather than a frm role to play in terms of ethics because it can remind investors statement. of the unethical character of “odious” debts and regimes.

Compliance with Ethical Standards Conclusion Conflict of interest The authors declare that they have no conficts of Dictators often rely on the issuance of foreign sovereign interest. bonds to fund their regime. From an ethical point of view, Human and Animals Rights This article does not contain any studies lending to dictators is tantamount to an admission of sup- with human participants or animals performed by any of the authors port for their regime. Journalists, for example, have argued that foreign debts are “the most visible remaining symbol of 71 international complicity” with a “dictator’s bloody reign”. References Legal scholars have attempted to limit the willingness to lend to such regimes by developing the doctrine of odious debt. But since there is no international law defning such Archival Sources

Crédit Lyonnais archives, Paris, France: ACL DAF 162, ACL DAF 70 See for example l’Aurore 15 June 1912 or Le Temps 16 June 1912. 163, ACL DAF 164/1, ACL DAF 164/2, ACL DAF 00104/1, ACL 21 AH 57, ACL 98 AH 86. 71 Jim Hoagland Iraq’s Bigger Picture. Washington Post, 10 October 2004.

1 3 Author's personal copy

222 S. Collet, K. Oosterlinck

Published References Verein für Socialpolitik/German economic association session asset and bond markets (B09-V3), Muenster. Chen, L., Lesmond, D. A., & Wei, J. (2007). Corporate yield spreads Accominotti, O., Flandreau, M., & Rezzik, R. (2011). The spread of and bond liquidity. Journal of Finance, 62(1), 119–149. empire: Clio and the measurement of colonial borrowing costs. Collet, S. (2013). How big is the fnancial penalty for unfair debt? Economic History Review, 64(2), 385–407. the case of cuban bonds at the time of independence. European Acharya, V. V., & Rajan, R. G. (2013). Sovereign debt, government Review of Economic History, 17(3), 364–387. myopia, and the fnancial sector. Review of Financial Studies, Courtois, A. (1902). Traité des opérations de bourse et de change. 26(6), 1526–1560. Paris: Garnier Frères. Aizenman, J., Hutchison, M., & Jinjarak, Y. (2013). What is the risk Crisp, O. (1961). The Russian liberals and the 1906 Anglo-French of European sovereign debt defaults? Fiscal space, CDS spreads loan to Russia. The Slavonic and East European Review, 39(93), and market pricing of risk. Journal of International Money and 497–511. Finance, 34, 37–59. Delaisi, F. (1930). Les Soviets et la dette russe en France. Paris: Pub- Bai, J., & Perron, P. (1998). Estimating and testing linear models with lications de la conciliation Internationale. multiple structural changes. Econometrica, 66(1), 47–78. Drut, B. (2010). Sovereign bonds and socially responsible investment. Bai, J., & Perron, P. (2003). Computation and analysis of multiple Journal of Business Ethics, 92(1), 131–145. structural change models. Journal of applied econometrics, 18(1), Erce, A., & Diaz-Cassou, J. (2011). Selective sovereign defaults. Glo- 1–22. balization and Monetary Policy Institute Working Paper 127, Fed- Bayle, F. (1997). Emprunt russe. Indemnisation mode d’emploi. Paris: eral Reserve Bank of Dallas. Balland. Esteves, R. P., & Tunçer, A. C. (2016). Feeling the blues Moral hazard Ben-Shahar, O., & Gulati, M. (2007). Partially odious debts? Law and and debt dilution in Eurobonds before 1914. Journal of Interna- Contemporary Problems, 70(4), 47–81. tional Money and Finance, 65, 46–68. Bernal, O., Oosterlinck, K., & Szafarz, A. (2010). Observing bailout Eveno, P. (2003). L’argent de la presse française. Paris: Editions du expectations during a total eclipse of the sun. Journal of Interna- CHTS. tional Money and Finance, 29(7), 1193–1205. Feilchenfeld, E. (1931). Public debts and state succession. New York: Bignon, V. (2011). Conseiller et infuencer. La presse fnancière en Macmillan. France à la Belle Époque, 1881–1914. Traverse Revue d’histoire, Figes, O. (1997). A people’s tragedy. The Russian Revolution 1891– 18(3), 41–68. 1924. New York: Penguin Books. Bignon, V., & Flandreau, M. (2011). The economics of badmouth- Flandreau, M. (2003). Caveat emptor: Coping with sovereign risk ing: Libel Law and the Underworld of the Financial Press in under the international gold standard, 1871–1913. In M. Flan- France before World War I. Journal of Economic History, 71(3), dreau, C.-L. Holtfreich, & H. James (Eds.), International fnancial 616–653. history in the twentieth century, system and anarchy (pp. 17–50). Black, F., & Scholes, M. (1973). The pricing of options and corporate Cambridge: Cambridge University Press. liabilities. Journal of Political Economy, 81(3), 637–654. Flandreau, M., & Flores, J. H. (2009). Bonds and brands: foundations Bodie, Z., Kane, A., & Marcus, A. J. (2011). Investments and portfolio of sovereign debt markets, 1820–1830. Journal of Economic His- management. New York: McGraw-Hill. tory, 69(3), 646–684. Bolton, P., & Skeel, D. (2007). Odious debts or odious regimes? Law Flandreau, M., & Oosterlinck, K. (2012). Was the emergence of the and Contemporary Problems, 70(4), 83–107. international gold standard expected? Evidence from Indian Bradley, M., Cox, J. D., & Gulati, M. (2010). The market reaction to Government Securities. Journal of Monetary Economics, 59(7), legal shocks and their antidotes: Lessons from the sovereign debt 649–669. market. Journal of Legal Studies, 39(1), 289–324. Flandreau, M., & Zumer, F. (2004). The making of global fnance. Bradley, M., & Gulati, M. (2013). Collective action clauses for the OECD, Development Center Studies Paris. Eurozone. Review of Finance, 18(6), 2045–2102. Frey, B., & Waldenström, D. (2007). Using fnancial markets to analyze Bradley, M., Salvatierra, I. D. L., & Gulati, M. (2014). Lawyers: Gate- history: The case of the Second World War. Historical Social keepers of the sovereign debt market? International Review of Research, 32(4), 330–350. Law and Economics, 38, 150–168. Freymond, J. (1995). Les emprunts russes de la ruine au rembourse- Brown, W. O., & Burdekin, R. C. K. (2002). German debt traded in ment. Histoire de la plus grande spoliation du siècle. Paris: Edi- London during World War II: A British perspective on Hitler. tion du Journal des Finances. Economica, 69(276), 655–669. Gelpern, A. (2007). Odious, not debt. Law and Contemporary Prob- Buchheit, L. (2017). Enforcing new Venezuelan public debt instruments lems, 70(3), 81–114. (Maduro bonds, part 2). The Financial Times, 18 August 2017. Gentile, C. M. (2010). The market for odious debt. Law and Contem- Buchheit, L. C., & Gulati, G. M. (2008). Odious debts and nation- porary Problems, 73(4), 151–174. building: When the incubus departs. Mane Law Review, 60, Girault, R. (1961). Sur quelques aspects fnanciers de l’alliance franco- 477–485. russe. Revue d’histoire moderne et contemporaine, 8, 67–76. Buccheit, L., & Gulati M. (2017). How to restructure venezuelan debt. Girault, R. (1972). Conjoncture et investissement international: Les Duke Law School Public Law & Legal Theory Series 2017-52. placements français en Russie. Un exemple à la fn du XIXe siècle. Buchheit, L. C., Gulati, G. M., & Thompson, R. B. (2006). The Revue Economique, 23(5), 889–918. dilemma of odious debts. Duke Law Journal, 56, 1201–1262. Girault, R. (1973). Emprunts russes et investissements français en Russie. Chabot, B., & Santarosa, V. (2017). Don’t cry for Argentina (or other 1887–1914. Paris: Armand Colin. Sovereign borrowers): Lessons from a previous era of sovereign Goyenko, R., Subrahmanyan, A., & Ukhov, A. D. (2011). The term debt contract enforcement. Capital Markets Law Journal, 12(1), structure of bond market liquidity and its implications for expected 9–37. bond returns. Journal of Financial and Quantitative Analysis, 46(1), Chamon, M., Schumacher, J., & Trebesch, C. (2015). Foreign law Annual 111–139. bonds: Can they reduce sovereign borrowing costs?. In Maduro bonds conference 2015: Economic development-theory and policy. Gulati, M. (2017). . The Financial Times, 8 August 2017. Hausman, J. A. (1978). Specifcation tests in econometrics. Economet- rica, 46(6), 1251–1271.

1 3 Author's personal copy

223 Denouncing Odious Debts

Hausmann, R. (2017). The hunger bonds. Project Syndicate, May 26, Porzecanski, K., Jenkins, C., & Soto, N. (2017). Wall Street is starting to 2017 available at: https​://www.project-syndi​ ​cate.org/comme​ntary​ get nervous about Venezuela’s ‘Hunger Bonds. Bloomberg 9 June /madur​o-venez​uela-hunge​r-bonds​-by-ricar​do-hausm​ann-2017- 2017, available at; https://www.bloom​ berg.com/news/artic​ les/2017-​ 05?barri​er=acces​sreg. 06-09/venezuela-bond-repud​ iatio​ n-jitte​ rs-grow-after​ -goldm​ an-sachs​ ​ Ho, T. S., & Lee, S. B. (1986). Term structure movements and pric- -deal. ing interest rate contingent claims. Journal of Finance, 41(5), Rafalovitch, A. (1931). L’abominable vénalité de la presse… d’après 1011–1029. les documents des archives russes (1897–1917). Paris: Librairie du Jayachandran, S., & Kremer, M. (2006). Odious debt. The American Eco- Travail. nomic Review, 96(1), 82–92. Renouvin, P. (1959). Les relations franco-russes à la fn du XIXe siècle Jenkins, C., & Zerpa, F. (2017). Goldman Sachs Just Ignited the Ven- et au début du XXe siècle. Cahiers du monde russe et soviétique, ezuela ‘Hunger Bonds’ Movement. Bloomberg, 31 May 2017, https​ 1(1), 128–147. ://www.bloom​berg.com/news/artic​les/2017-05-31/goldm​an-sachs​ Ronzon-Bélot, C. (2003). Banquiers de la Belle Époque. Actes de la -gave-big-hand-to-venez​uela-hunge​r-bonds​-movem​ent. recherche en sciences sociales, 146–147, 8–20. Jèze, G. (1922). Cours de science des fnances et de législation fnancière Sack, A.-N. (1927). Les efets de la transformation des états sur leurs française. Théorie du crédit public. Giard, Paris: M. Giard. dettes publiques et autres obligations fnancières. Paris: Recueil Khalfan, A., King, J., & Tomas, B. (2003). Advancing the odious debt Sirey. doctrine. Centre for International Sustainable Development Law Sarig, O., & Warga, A. (1989). Bond price data and bond market liquidity. (CISDL) working paper 11. Journal of Financial and Quantitative Analysis, 24(3), 367–378. King, J. (2016). The doctrine of odious debt in international law. A Seton-Watson, H. (1967). The Russian Empire 1801–1917. Oxford: restatement. Cambridge: Cambridge University Press. Oxford University Press. Long, J.-W. (1972a). Russian manipulation of the French Press, 1904– Shafter, J. (2007). The due diligence model: A new approach to the prob- 1906. Slavic Review, 31(2), 343–354. lem of odious debts. Ethics & international Afairs, 21(1), 49–67. Long, J.-W. (1972b). Organized protest against the Russian 1906 loan. Siegel, J. (2014). For peace and money. French and British Finance in Cahiers du Monde Russe et Soviétique, 13(1), 24–39. the Service of the Tsars and Commissars. Oxford: Oxford University Ludington, S., & Gulati, M. (2008). A convenient un-truth: Fact and Press. fantasy in the doctrine of odious debts. Virginia Journal of Interna- Tomaszewski, F. (1999). Pomp, circumstance, and realpolitik: The evo- tional Law, 48(3), 595–639. lution of the Triple Entente of Russia, Great Britain, and France. Ludington, S., Gulati, M., & Brophy, A. L. (2010). Applied legal history: Jahrbücher für Geschichte Osteuropas, 3, 362–380. Demystifying the doctrine of odious debts. Theoretical Inquiries in Tomz, M. (2012). Reputation and international cooperation: sovereign Law, 11(1), 247–281. debt across three centuries. Princeton: Princeton University Press. Lysis. (1914). Contre l’Oligarchie Financière en France, La réponse de Tomz, M., & Wright, M. L. (2007). Do countries default in “bad times”? Lysis aux établissements de Crédit. Paris: Albin Michel. Journal of the European Economic Association, 5(2–3), 352–360. Manuel du Porteur de Fonds Russes. (1903). Emprunts émis ou garantis Tomz, M., & Wright, M. L. (2013). Empirical research on sovereign debt par l’état et côtés à l’étranger. St. Petersburg: Russian Ministry of and default. NBER working paper 18855. Finance. Vaslin, J. M. (1999). Les rentes françaises au XIXe siècle-Le marché Neymarck, A. (1902). Finances contemporaines. Paris: Guillaumin. des fonds d’État était-il efcient. Unpublished doctoral dissertation, Oosterlinck, K. (2016). Hope springs eternal. French bondholders and Université d’Orléans. the Repudiation of Russian Sovereign Debt. New Haven: Yale Uni- Verdier, P. H., & Voeten, E. (2015). How does customary international versity Press. law change? The case of state immunity. International Studies Quar- Oosterlinck, K., & Landon-Lane, J. (2006). Hope springs eternal: French terly, 59(2), 209–222. bondholders and the Soviet Repudiation (1915–1919). Review of Waller, J. (1979). La naissance du rouble-or. Cahiers du monde russe et Finance, 10(4), 507–535. soviétique, 20(3–4), 285–304. Oosterlinck, K., & Ureche-Rangau, L. (2012). Interwar Romanian sov- Weidenmier, M. (2002). Turning points in the U.S. Civil War: Views from ereign bonds: The impact of diplomacy, politics and the economy. the Grayback market. Southern Economic Journal, 68(4), 875–890. Financial History Review, 19(2), 219–244. Weidemaier, W. M. C. (2014). Sovereign immunity and sovereign debt. Panizza, U., Sturzenegger, F., & Zettelmeyer, J. (2009). The economics University of Illinois Law Review, 1, 67–114. and law of sovereign debt and default. Journal of Economic Litera- Weidemaier, W. M. C., & Gulati, M. (2018). Difering perceptions? Mar- ture, 47(3), 1–47. ket practice and the evolution of sovereign immunity law. Journal of Parti socialiste. (1906). 3e Congrès National tenu à Limoges les 1er, Law & Social Inquiry. (forthcoming). 2, 3 et 4 Novembre 1906, Compte Rendu Analytique. Paris: Parti Werth, N. (2004). Histoire de l’Union Soviétique. Paris: Presses Univer- Socialiste. sitaires de France. Peillex, J., & Ureche-Rangau, L. (2016). Identifying the determinants Willard, K., Guinnane, T., & Rosen, H. (1996). Turning points in the of the decision to create socially responsible funds: An empirical Civil War: Views from the Greenback market. American Economic investigation. Journal of Business Ethics, 136(1), 101–117. Review, 86(4), 1001–1018. Philippe, P. (1907). Les Fonds Russes et les Manœuvres du Crédit Lyon- Zettelmeyer, J., Trebesch, C., & Gulati, M. (2013). The Greek debt nais. Paris: La Bourse libre. restructuring: an autopsy. Economic Policy, 28(75), 513–563. Pipes, R. (1990). La révolution russe. Paris: Presses Universitaires de Zussman, A., Zussman, N., & Nielsen, M. O. (2008). Asset market per- France. spectives on the Israeli–Palestinian confict. Economica, 75(297), Poidevin, R. (1970). Finances et Relations Internationales 1887–1914. 84–115. Paris: Armand Colin.

1 3