The Cash Nexus: Money and Politics in Modern History, 1700-2000

by Niall Ferguson

  Source: Eichengreen and Wyplosz, ‘Stability Pact’, p. 103.

  EXPECTATIONS: THE PAST AND THE PRESENT

  Whether adaptive or rational, expectations are in large measure historically based. To be sure, most financial markets are ‘weak-form-efficient’ in the sense that ‘the sequence of past prices provides no exploitable information as to the sequence of future price movements’: they follow what economists like to call a ‘random walk’.28 But even the most sophisticated econometric models – as well as the more or less informal models on which small investors make their decisions – need some past data to chew on. The significance of defaults and inflations such as those described above and in the preceding two chapters lies here.

  From the investors’ point of view, a major reason for fearing that a country might default or depreciate its currency is the simple fact that it has done so in the past. This explains why the short-term benefits of default or depreciation in reducing a government’s debt burden must be set against the longer-term costs of loss of reputation, which usually raises the cost of future borrowing. This is a crucial point if we are to understand why some countries have been able to sustain much higher absolute levels of debt than others.

  Early modern evidence confirms the link between past misdeeds and present interest rates. Most obviously, creditworthy city-states could borrow at lower interest rates than default-prone monarchs. Yields on the consolidated debt of Genoa in the second half of the fourteenth century fluctuated between 5 and 12 per cent.29 This was not unusual: in fifteenth-century Florence yields ranged between 5 and 15 per cent.30 By contrast, Habsburg aides in Antwerp yielded as much as 20 per cent in the 1520s and 1530s (though they fell to 10 per cent in 1550). The yields on short-term asientos rose steadily during the sixteenth century as the Habsburg regime’s credit-rating declined due to successive defaults: from 18 per cent in the 1820s to 49 per cent by the 1550s.31

  It was the Dutch system of public finance that proved most successful in lowering interest rates. Dutch yields fell steadily from above 8 per cent in the 1580s to 5 per cent in the 1630s, 3 per cent in the 1670s and just 2.5 per cent by the 1740s.32 Yet this was at a time when the United Provinces were steadily increasing their debt, confirming that there is no automatic correlation between the absolute size of a debt and the yield on the bonds that constitute it. The same was true of the Swiss cantonal debts and the yields on them for most of the sixteenth century.33

  By contrast, French yields in the eighteenth and nineteenth centuries reveal starkly the impact of fiscal unreliability on investor confidence. The effect of France’s institutional difficulties – above all, the ancien régime’s bad habit of defaulting every few decades – was not only to limit the absolute amount France could borrow, but to push up the cost of borrowing far above the equivalent Dutch or British figures. Because of the complex structure of the French debt, it is possible to calculate a variety of different yields. Rentes were considered biens immeubles, like land, and were not easily traded. Life annuities – the largest part of the debt by 1789 – were not bought and sold after the lives had been specified. The best available rate that can be compared with that for Britain is therefore probably that for the emprunt d’octobre (October Loan) created by the new Compagnie des Indes in 1745, though it should be remembered that this represented only a small fraction of the total French debt. Nevertheless, the figures clearly indicate the extent of the fiscal difference between France and Britain. Between 1745 and 1780 there was a significant differential between the yields on French and British government bonds of the order of between 100 and 200 basis points (one or two percentage points). As Figure 12 shows, the spread was at its widest in the early 1770s, when the yield on the French October Loan rose above 10 per cent at a time when consol yields were steady at around the 3.5 per cent mark. Other French bonds carried even higher yields.34 True, between 1780 and 1785 the spread fell below 100 basis points, and indeed all but disappeared in March 1785. But averaging out the figures for 1754 to 1789, it is clear that the cost of borrowing was significantly higher for France than for her rival across the Channel: of the order of 244 basis points, or nearly 2½ per cent.

  Figure 12. British and French bond yields, 1753–1815

  Sources: Neal, Rise of Financial Capitalism, pp. 241–57, for consol prices; French data kindly supplied by François Velde.

  These differentials were based on past experience of which bonds were most likely to be defaulted upon: there was an element of ‘prepaid repudiation’.35 But there was also clearly a jump in yields in the aftermath of defaults in 1759 and 1770, suggesting that the market was not wholly clairvoyant.36 Moreover, the amount the French government had to pay on new loans issued in the period was almost always significantly higher than the yield on the October Loan. The ten-year loan issued in May 1760 yielded 9.66 per cent to investors, at a time when the October Loan yield was 6.87 per cent. The life annuities which were the most common form of new loan yielded as much as 11 per cent in 1771.37 These were far higher rates than the British government had to pay for old or new loans. Because both countries were on a fixed specie standard from 1726 until the 1790s – gold in the British case, bimetallic in the French – the differential mainly reflected the greater French default risk (though the greater liquidity of the British market doubtless played a part).38 This was what Bishop Berkeley meant when he said that credit was ‘the principal advantage which England hath over France’.39 As Isaac de Pinto put it in 1771, when French yields were soaring: ‘It is not credit that has ruined the finances of France … On the contrary, it was the failure of credit in time of need that did the mischief.’40

  The key difference between France and Britain in the eighteenth century, then, was not a matter of economic resources. France had more. Rather, it was a matter of institutions. Britain had the superior revenue-collecting system, the Excise. After the Glorious Revolution, Britain also had representative government, which not only tended to make budgets transparent, but also – more importantly – reduced the likelihood of default, since the bondholders who had invested in the National Debt were among the interests best represented in Parliament.41 The National Debt itself was largely funded (long-term) and transparently managed (especially after the advent of the consol). And the Bank of England – which again had no French analogue – also guaranteed the convertibility of the currency into gold (save in an extreme emergency), reducing if not eliminating the risk of default through inflation. It was these institutions which enabled Britain to sustain a much larger debt/GDP ratio than France because they ensured that the interest Britain paid on her debt was substantially less than France paid on hers. If one seeks a fiscal explanation for Britain’s ultimate triumph over France in their global contest, it lies here.

  But the crucial point is that financial institutions depend for their effectiveness on credibility. It is highly significant, in this context, that each time the chances of a Stuart Restoration rose – for example during the 1745 Jacobite Rising – so too did the yield on government bonds.42 To contemporaries, there was no guarantee that the regime change brought about by the Glorious Revolution would endure, and that the lineal descendants of the Hanoverians would still reign in Britain more than three hundred years after the deposition of James II. The possibility could not wholly be dismissed – even after the crushing of the ’45 at Culloden – that a combination of the French abroad and the Highland Scots at home might restore the Stuarts. But by comparison with the risks of default facing investors in French bonds, the danger was remote.

  It is at first sight surprising, in this light, that the political crisis of 1789 did not have a bigger impact on French yields. Though the yields on the new loans issued in 1782 and 1784 rose above 11 per cent, this happened a year earlier, in 1788. They oscillated around the 9 per cent mark in 1789 and 1790, but then fell to between 5 and 6 per cent in the first half of 1791. The October Loan was even less affected by the first phase of the Revolution, as Figure 12 shows, never
rising above 8 per cent, far less than the yield peak in 1771.43 What this suggests is that the market initially welcomed the advent of constitutional government, not least because the alternative was clearly another major default.44 As the 1790s went on, however, the trauma of war, the Terror and default sent French yields soaring from around 6 per cent to above 60 per cent.45 The French Revolutionary Wars could be financed only by printing money: for most of the 1790s substantial long-term government borrowing was simply out of the question.

  It was only slowly that French credit recovered from these shocks. Under Napoleon, yields declined from a peak of around 12.5 per cent in 1802 to below 6 per cent in late 1807, and held more or less steady at around 6 per cent until the winter of 1812, when defeat in Russia dealt a fatal blow to Napoleon’s ambitions. The decisive reverse at Leipzig in October 1813 saw French yields leap up to 10 per cent; and their subsequent recovery was cut short in March 1815 by the news of Napoleon’s return from Elba and the Hundred Days that culminated at Waterloo. Napoleon’s defeat and the substantial reparations imposed on France kept yields high thereafter: in 1816–17 they averaged between 8 and 9 per cent. But with the withdrawal of Allied troops and the normalization of the restored Bourbon regime’s relations with the victors of Waterloo, yields declined steadily, falling below 5 per cent in 1825 for the first time since 1753.46 French institutions gradually became more like British ones: the tax system was reformed by the Revolution and Empire; the Chamber of Deputies became more representative after 1830; the issuance of 5 and 3 per cent rentes perpetuelles became the basis of public borrowing; and the Banque de France, another legacy of Bonaparte, managed the specie currency. Nevertheless, recurrent revolutionary episodes – in 1830, 1848 and 1870–1871 – periodically revived in investors’ minds the memory of the 1790s. Not until 1901 did French yields for the first time fall before British.

  Figure 13. Major bond yields since 1700 (annual averages)

  Sources: France 1746–93: Velde and Weir, ‘Financial Market and Government Debt Policy in France’. France, Germany, Italy to 1959: Homer and Sylla, Interest Rates; from 1960: OECD. Britain 1700–1800: Global Financial Data; 1800–1850, 1914–59: Mitchell and Deane, Abstract of British Historical Statistics; 1850–1914: Klovland, ‘Pitfalls’, p. 185; from 1960: ONS, US: Global Financial Data.

  Notes: UK: Consols (corrected yield); France: 1797–1824: 5 per cent rentes; 1825–1949: 3 per cent rentes; 1950–1959: 5 per cent rentes; Germany: to 1869: Prussian 4 per cents, 3.5 per cents; 1870–1908: Reich 3 per cents; 1909–26: high-grade corporate bonds; 1927–44: government loans; 1948–53: high-grade bonds; 1956–9: government loans; Italy: 1924–49: 3.5 per cents; 1950–1969: 5 per cents. All countries except Britain 1960–1999: long-term bonds (OECD standardized measure).

  The French experience of past default and depreciation as a cause of higher bond yields is far from unique. To provide a long-term perspective, Figure 13 shows yields since 1700. Another obvious case when a major default led to a sustained risk-premium on a country’s bonds can be seen from the 1920s until the 1950s. Like the French experience of the 1790s, the German hyperinflation of 1919–23 left scars on investors’ memories that were legible in bond yields for years afterwards. And the high yields on post-1923 German bonds had profound effects. For example, it was the tightness of the bond market in the later 1920s that choked off local government investment in housing, a key harbinger of the approaching Slump.47 Moreover, a ‘Keynesian’ response to the Slump at the Reich level was more or less ruled out by the fear that deficit finance would reignite persistent public fears of a second great inflation.48 Only by covertly issuing the so-called ‘Mefo-bills’ – in reality, short-term government bills – through the bogus ‘Metallurgical Research Office’ was Schacht able to finance the first phase of Nazi rearmament.49 Yet it was not only hyperinflation that traumatized bondholders and thereby circumscribed future fiscal policy. As a de facto default, the Italian ‘forced conversions’ of 1926 and 1934 had a comparable effect, pushing up the cost of any subsequent borrowing by the fascist regime and necessitating illegal devices such as secret loans from the cities of Milan and Rome.50 The French experience of high but not hyperinflation in the 1920s might seem like the optimal post-war policy, but the experience was one reason the French government adhered grimly to the gold standard in the 1930s, while Britain was able to reap the benefits of devaluation.51 We shall revert to this point in Chapter 11.

  ‘EVENTS, DEAR BOY’

  Yet it would be misleading to suggest that past behaviour is the sole determinant of yield differentials. For investors in bonds are as much interested in any current indications of a government’s future fiscal and monetary policy as they are interested in the policies of the past. This presents an awkward problem for economic theory, in that investors do not rely purely on economic data when forming their expectations of future policy. They are as much interested in political events.

  To illustrate this point in the context of a large and liquid domestic debt market, I have calculated the annual percentage increase in the yield on consols since 1754.52 Such a measure differs from the more usual measure of absolute increases or decreases in yields expressed in terms of basis points. The reason for using the percentage change is simply that an increase of 100 basis points pushes up the cost of borrowing by more in relative terms when yields rise from 2 to 3 than when they rise from 7 to 8. By this measure the twentieth century has been only slightly more volatile than the eighteenth, with the nineteenth an interlude of comparative stability. The troubled year 1974 occupies a special place at the top of the league table of rises in British bond yields, with the average annual long-bond yield going up by some 38 per cent. Second only to 1974 in the list of yield hikes was 1797, the year of the suspension of gold payments by the Bank of England: Pitt’s ‘Political Ravishment’ of ‘The Old Lady of Threadneedle-Street’.53 The year 1998, however, witnessed the biggest ever annual decline in yields, one of that year’s least commented upon historic firsts and in large part a response to the successful transition to Bank of England operational independence.54

  Figure 14. The yield on consols (end-of-month figures), 1754–1998

  Sources: 1753–1823: Neal, Rise of Financial Capitalism, pp. 241–57; 1824–42: Spectator (closing prices on last Saturday of each month); 1843–9: The Economist; 1850–1914: Klovland, ‘Pitfalls’, pp. 184 f.; 1914–62: Capie and Webber, Monetary History, pp. 514–27; 1963–98: ONS, Financial Statistics.

  Note: The possibility existed that if consols reached 100 they could be redeemed. This creates difficulties for the calculation of correct yield figures in the late nineteenth century, which have been addressed by Klovland.

  Inferences about political causation can be made with slightly more confidence when the same calculation is made using monthly figures (see figure 14).55 Here the months that stand out are November 1792, March 1778 and March 1803, in each of which the yield on consols rose by more than 14 per cent. In fourth place comes June 1974 (an increase of just under 13 per cent), followed by March 1814 (12 per cent). It is possible that all these dramatic fluctuations were due to monetary factors: in all but one case Britain was off gold. On the other hand, it is at least suggestive that each jump in yields coincided with a major international or domestic political crisis. On 6 November 1792 French forces defeated the Austrian army at Jemappes and overran the Austrian Netherlands (modern Belgium); on the 19th, the French National Convention offered its support to all peoples wishing to overthrow their governments. Similarly, the collapse in bond prices in March 1778 came shortly after the United States had signed two treaties with France, leading Britain to declare war on France. In March 1803 Napoleon’s annexations of Italian territory and interference in the affairs of Switzerland were causing grave concern in London: war broke out again with France in May. The rise in yields that occurred in March 1814 is the exception, since it coincided with Napoleon’s defeats at Laon and Arcis-sur-Aube, and the fall of Boulogne (12 March) and
Paris (31 March). However, the 13 per cent rise in June 1974 came after a catalogue of political reverses for the newly installed Labour government: the collapse of the Sunningdale agreement in Northern Ireland (28 May) and the explosion of a bomb outside Westminster Hall (17 June) signalled a serious deterioration in the Ulster crisis; while the government suffered a succession of parliamentary defeats, leading ultimately to a second dissolution on 20 September, a mere seven months since the previous election.

  The apparent link between political events and the bond market is even closer if one considers the experience of inter-war France (see Figure 15). The biggest increase in the yield on rentes at any time in the history of the Third Republic came in August 1925, when the rate rose by more than 10 per cent.56 This might seem surprising. In June a preliminary agreement had been signed with Germany confirming the existing West European borders, a deal finalized that October at Locarno. On 13 July French troops had evacuated the Rhineland. However, events elsewhere were less peaceful. A revolt in Morocco against Spanish and French rule had broken out in May 1925; the government resolved to crush it, sending General Pétain to lead a substantial force against the rebels. The war went on until May 1926.57 The other major crises on the French bond market are easier to fathom. On four occasions between 1933 and 1939, yields rose by between 8.5 and 10 per cent: March 1933, March 1935, April 1937 and January 1939. It seems at least plausible that the deterioration of Franco-German relations and the possibility of another great war were behind these rises. On 5 March 1933 the Nazis consolidated their power in Germany with a sweeping election victory; eleven days later the new government sank the latest British disarmament plan at Geneva by insisting that the brown-shirted Nazi stormtroops (the SA) should not be counted as part of Germany’s armed forces. And the Enabling Law of 23 March gave Hitler dictatorial powers. On 1 March 1935 the Saarland was restored to Germany; two weeks later Hitler repudiated the disarmament clauses of the Versailles Treaty and reintroduced conscription. April 1937 saw Belgium released from her obligations under the Locarno Treaty and Guernica bombed flat by German aircraft. The crisis of January 1939 differed only because it related to Italy rather than Germany (indeed, on 6 December 1938 France and Germany had signed a pact confirming the inviolability of their existing frontiers). But on 17 December Italy had denounced its 1935 agreement with France regarding Corsica and Tunisia, prompting the French prime minister Édouard Daladier to make a defiant visit to both places. The ground was cut from under Daladier on 10 January when Chamberlain and Halifax visited Rome for talks with Mussolini.