TY - JOUR
AU - Mitchener, Kris James
AB - Abstract Fiscal deficits, elevated debt-to-GDP ratios, and high inflation rates suggest hyperinflation could have potentially emerged in many European countries after World War I. We demonstrate that economic policy uncertainty was a key driver pushing a subset of European countries into hyperinflation shortly after the end of the war. Germany, Austria, Poland and Hungary (GAPH) suffered from frequent uncertainty shocks—and correspondingly high levels of uncertainty—caused by protracted political negotiations over reparations payments, the apportionment of the Austro-Hungarian debt and border disputes. In contrast, other European countries exhibited lower levels of measured uncertainty between 1919 and 1925, allowing them more capacity with which to implement credible commitments to their fiscal and monetary policies. Impulse response functions show that increased uncertainty caused a rise in inflation contemporaneously and for a few months afterwards in GAPH, but this effect was absent or much more limited for other European countries. Why do hyperinflations begin? In a mechanical sense, economists have known the answer to this question at least since the monetarist revolution: money is printed in response to unsustainable fiscal policy. But ex ante, how does one identify factors that trigger hyperinflation in one country but not another, when macroeconomic indicators look broadly similar across them? For example, as a consequence of World War I (WWI), many European economies abandoned their commitments to fixed exchange rates and ran up large public debts, predisposing them to high inflation, if not hyperinflation. Belgium, Great Britain, France and Italy had their debt-to-GDP ratios drift upwards in excess of 100% by 1920 and saw their price levels double from 1913 (Table 1). Per capita direct costs associated with the war were high for both Allied and Central Power countries, and the destruction of human capital was broadly similar in countries such as France and Germany.1 All of these conditions suggest that a grim fiscal and monetary situation prevailed throughout Europe after World War I, and unfavourable macroeconomic preconditions existed as much for victors as the vanquished. Table 1. Debt and Inflation around WWI (%). . Debt-to-GDP ratio . Annual inflation . . 1913 . 1920 . (1914–18) . Austria 72.6 69.3 84.6 Belgium 49.5 102.7 59.0 France 66.3 185.5 35.9 Germany 63.0 288.5 21.2 Netherlands 61.6 59.6 41.4 Hungary 62.3 70.4 71.1 Italy 69.2 142.3 42.2 Great Britain 33.0 130.7 23.1 . Debt-to-GDP ratio . Annual inflation . . 1913 . 1920 . (1914–18) . Austria 72.6 69.3 84.6 Belgium 49.5 102.7 59.0 France 66.3 185.5 35.9 Germany 63.0 288.5 21.2 Netherlands 61.6 59.6 41.4 Hungary 62.3 70.4 71.1 Italy 69.2 142.3 42.2 Great Britain 33.0 130.7 23.1 Notes: Annual inflation rates are authors’ calculations based on wholesale price data from Statistisches Handbuch der Weltwirtschaft, except for Belgium, which are retail prices from Revue de Travail as published in Monthly Labor Review (1919, p. 121). Debt-to-GDP ratios for Belgium, France, the Netherlands, Italy and the UK are from Reinhart and Rogoff (2009) and Ritschl (2012). German debt-to-GDP figures are calculated using data from Balderston (1989) and Ritschl (2012; 2013). Apportionment of existing Austro-Hungarian debt had not been settled by 1920, so reported post-war figures are for 1918, and are based on Schulze (2005). Open in new tab Table 1. Debt and Inflation around WWI (%). . Debt-to-GDP ratio . Annual inflation . . 1913 . 1920 . (1914–18) . Austria 72.6 69.3 84.6 Belgium 49.5 102.7 59.0 France 66.3 185.5 35.9 Germany 63.0 288.5 21.2 Netherlands 61.6 59.6 41.4 Hungary 62.3 70.4 71.1 Italy 69.2 142.3 42.2 Great Britain 33.0 130.7 23.1 . Debt-to-GDP ratio . Annual inflation . . 1913 . 1920 . (1914–18) . Austria 72.6 69.3 84.6 Belgium 49.5 102.7 59.0 France 66.3 185.5 35.9 Germany 63.0 288.5 21.2 Netherlands 61.6 59.6 41.4 Hungary 62.3 70.4 71.1 Italy 69.2 142.3 42.2 Great Britain 33.0 130.7 23.1 Notes: Annual inflation rates are authors’ calculations based on wholesale price data from Statistisches Handbuch der Weltwirtschaft, except for Belgium, which are retail prices from Revue de Travail as published in Monthly Labor Review (1919, p. 121). Debt-to-GDP ratios for Belgium, France, the Netherlands, Italy and the UK are from Reinhart and Rogoff (2009) and Ritschl (2012). German debt-to-GDP figures are calculated using data from Balderston (1989) and Ritschl (2012; 2013). Apportionment of existing Austro-Hungarian debt had not been settled by 1920, so reported post-war figures are for 1918, and are based on Schulze (2005). Open in new tab We examine inflation dynamics in European countries immediately after the end of WWI. In particular, we analyse how pronounced economic policy uncertainty in a subset of countries contributed to their descent into hyperinflation shortly after the end of the war. Economic policy uncertainty has been shown to be an important driver of macroeconomic conditions (Bloom 2009; 2014), such as investment and GDP, with Carrière-Swallow and Céspedes (2013) finding substantial cross-country heterogeneity in response to uncertainty shocks. However, analysing how uncertainty shocks influence inflation has only received attention more recently (Vavra, 2014), with many aspects yet to be explored; in particular, how uncertainty shocks might influence inflation dynamics prior to hyperinflations.2 Yet as recent examples of high inflation in Venezuela and Argentina suggest, high degrees of economic uncertainty around capital controls, foreign indebtedness, national elections and government stability appear to contribute to actual and near hyperinflationary episodes.3 The recent literature on the effects of uncertainty on macroeconomic dynamics has examined several uncertainty measures, although a consensus on methods has not yet been established, even if the underlying premise of such a relationship is clear.4 For our study, we use a new cross-country data set to generate market-based measures of economic uncertainty. We then use a new empirical methodology to analyse how measured uncertainty in the interwar period affected inflation dynamics and the incidence of hyperinflation across the ten countries in our sample. Our approach is specific since it is responsive to country-specific events and institutional structures as well as general in that we use a common modelling framework across countries. Specifically, we construct a measure of uncertainty using a new interwar data set of daily spot exchange rates for ten European countries. From these high-frequency data, we develop country-specific measures of policy uncertainty based on the realised volatility (RV) of these market rates at the monthly frequency. Our approach encompasses country-specific news and events that influenced policy to generate a unique time series for each country, allowing us to draw attention to differences in the level and variability of uncertainty across our sample of countries. We show that these RV measures were elevated across Europe after WWI and particularly so in Germany, Austria, Poland and Hungary (GAPH) prior to their hyperinflations. The RV measures for the GAPH countries were many times larger than that of the Netherlands, which was neutral during the war and which we use as a benchmark for comparison. Uncertainty was also more pronounced in GAPH than in intermediate cases, such as France (Dornbusch, 1987; Prati, 1991). A detailed examination of contemporary news sources suggests high RV months are associated with events that contributed to policy uncertainty, and GAPH experienced such events with greater frequency and magnitude. We show that the high degree of measured uncertainty in GAPH was associated with weak fiscal positions, protracted political negotiations over war reparations, and unresolved national borders. The other European countries also experienced certain spikes in RV that can be associated with events leading to greater economic policy uncertainty, but they were considerably fewer in number and smaller in magnitude. We find that our RV measure of uncertainty appears related to the inability of policymakers in GAPH to formulate and commit to credible fiscal policies, suggesting a causal link between uncertainty and inflation dynamics—a point alluded to, but not formally tested, in Sargent (1982).5 To test for causality more formally, we embed our measure of uncertainty within a reduced-form macroeconomic model that includes changes in prices, industrial production and notes outstanding at a monthly frequency. Our main estimation technique is smoothed local projections (SLP), a recent innovation by Barnichon and Brownlees (2019) that permits inference on the effects of shocks in systems of equations with small samples. Specifically, we focus our analysis on monthly impulse response functions (IRFs) for RV shocks on inflation at the country level prior to the start of hyperinflation for each GAPH country and over longer periods for the six other countries that did not tip into hyperinflation. For GAPH, the SLP results show that increased uncertainty contributed to a contemporaneous rise in inflation as well as for a few months immediately following the shock. For example, in Germany, the results suggest that a one-standard-deviation increase in our RV measure leads to a contemporaneous increase in inflation of about eight percentage points and another three percentage points in the subsequent month. We find similar patterns and larger magnitudes for the three other hyperinflation countries. Moreover, in the months just prior to when hyperinflation broke out in GAPH, monthly shocks to RV were some of the largest observed across the whole sample period. By contrast, for the other European countries with lower RV measures, the effect of increased uncertainty on inflation is absent or near zero in magnitude. For example, our results for France show that a one-standard-deviation increase in uncertainty had almost no effect on its inflation rate. These findings demonstrate the utility of our methodology as it permits comparisons across countries—such as Germany and France, for example—that were on opposite sides of the reparations imbroglio as a payer and recipient, respectively. The empirical results help to elucidate how greater policy uncertainty contributed to accelerated inflation in Germany, but less so in France. Further, our methodology allows us to encompass many of the country-specific policy narratives emphasised by historians of post-WWI Europe. For example, we can examine more directly how uncertainty around issues such as the size and settlement of reparations in Germany could have contributed to driving inflation expectations further into a negative spiral (Webb, 1986). Our paper also contributes to several strands of the literature in economics. First, it complements recent research initiated by Bloom (2009) that examines the relationship between uncertainty and macroeconomic outcomes. Our research extends the work by Baker et al. (2016) and subsequent related papers by focusing on uncertainty’s effects on inflation, a macroeconomic outcome that has received comparatively less attention than investment and output.6 Our conjecture—that policy uncertainty is critical for understanding interwar European inflation dynamics—also builds on research describing how the hyperinflations of the early 1920s resulted from unbalanced fiscal and monetary policy (Cagan, 1956; Dornbusch, 1982; 1985; Sargent, 1982). Our analysis focuses on the period prior to the start of hyperinflation in order to better understand the role of measured uncertainty as a driver of inflation dynamics. We extend these earlier treatments by providing a quantitative modelling framework that can account for the differential inflation dynamics across Europe and that helps understand why the GAPH countries experienced hyperinflation. Finally, our analysis relates to recent studies on fiscal policy shocks. One strand seeks to quantify how uncertainty around the timing, magnitude and composition of fiscal policies fundamentally influence medium-term macroeconomic projections and accompanying policy actions (Paredes et al., 2015). Another vein shows that unexpected changes in fiscal volatility shocks had a sizable adverse effect on US economic activity in the late twentieth century (Fernández-Villaverde et al., 2015). The interwar period provides a clear parallel to questions of how fiscal policy uncertainty influences the broader macroeconomy as examined in these papers. In many parts of post-WWI Europe, uncertainty over reparation payments, border disputes and legacy debt obligations conspired with limited tax bases and political events to weaken policymakers’ abilities to provide direction over the future path of tax and spending policies. The greater incidence and magnitude of these uncertainty shocks for the GAPH countries are shown to contribute to their challenging inflationary environments. 1. Economic Policy Uncertainty in Europe after WWI WWI has been described as a watershed moment in European economic history. The war destroyed property and killed large numbers of soldiers and civilians. Old imperial powers were dissolved, and new nation states were formed. Ethnic groups were involuntarily separated across new borders, and armed conflicts and border disputes continued even after treaties were drawn. The Great War left a legacy of fractured states grappling with high unemployment, industrial dislocation and high national debts. Existing trade flows were disrupted and then reorganised, often in patterns that did not resemble the old ones. As a result, policymakers across the continent struggled to find their footing, often attempting to replicate institutions of the past, such as the gold standard, even though fundamental political and economic change had occurred. Table 1 shows the fiscal burden that resulted from the war. Debt-to-GDP ratios for countries that existed pre- and post-WWI rose significantly, suggesting that, in the absence of rapid economic growth, new taxes and controls on spending would be necessary in the 1920s to return to prewar ratios. Because much of the spending on the war was monetised (Ferguson, 1975; Sargent, 1982; Eichengreen, 1992), prices across Europe rose considerably, as shown in the last column of Table 1. Given such widespread economic dislocation in the wake of the Great War, it is not obvious ex ante which countries in Europe might subsequently experience hyperinflation. For example, in addition to the situation shown in the table, France ran budget deficits and financed them through money creation and increased national debt in the early 1920s (Fraser and Taylor, 1990). How might one identify the likelihood of a particular country experiencing hyperinflation given macroeconomic indicators at war's end that look broadly similar across Europe? Being on the winning or losing side of a war appears insufficient for determining which countries subsequently experienced hyperinflation, especially in light of the fact that many countries on both sides of the conflict were saddled with inter-allied debts (Eichengreen, 1992). The Central Powers of Bulgaria and Turkey, the successor state to the Ottoman Empire, experienced very substantial increases in prices between 1914 and 1925 (more than 3,000%), but prices never increased by more than double in any given year. Czechoslovakia—a newly formed state carved out of a losing Central Power—saw its prices rise by roughly 1,600% between 1914 and 1921, but prices stabilised thereafter. Sargent (1982) emphasised that this new country, in contrast to another one like Poland, succeeded in reducing uncertainty over its budget prospects and did not experience hyperinflation.7 On the other hand, Russia, initially part of the victorious Triple Entente (until it withdrew from the war following the Bolshevik Russian revolution), did experience hyperinflation.8 We hypothesise that pronounced macroeconomic policy uncertainty in particular European countries was an important driver tipping countries into hyperinflation. We consider policy uncertainty to include direct policy actions, such as budget expenditures as well as indirect activities, such as negotiations over war reparations and border disputes. Focusing on war reparations, the intractable debates surrounding them cast a pall over all of post-war Europe and proved counterproductive for generating credible fiscal policy. As laid out in the Treaty of Versailles (arts. 231–2), the Allies paradoxically acknowledged that Germany lacked the resources to pay, but also insisted on payment: The Allied and Associated Governments recognise that the resources of Germany are not adequate … to make complete reparation for all such loss and damage. The Allied and Associated Governments, however, require, and Germany undertakes, that she will make compensation for all damage done to the civilian population of the Allied and Associated Powers. The individual treaties struck with other belligerents—namely, the treaties of Saint-Germain-en-Laye (1919) and of Trianon (1920) that established Austria and Hungary, respectively, as independent states—similarly specified that reparations were to be paid. Notably, all three treaties did not specify the total amount or number of years over which payments would have to be made. Instead of providing precise terms for payment of damages, a political entity, known as the Reparations Commission (RC), was charged with working out the details. The RC met repeatedly after the treaties were concluded, but the gap between what the Allies sought and what Germany and others were willing to pay remained large. In January 1921, France demanded 226 billion gold marks ($94 billion). Germany's counter proposal was just 50 billion gold marks ($21 billion) (Boemke et al., 1998, p. 410). Meanwhile, the US Congress failed to ratify the Versailles Treaty and was reduced to ‘observer status’ as the negotiations dragged on, leaving the conversation among only Europeans who often found themselves at loggerheads. On 5 May 1921, four days after the deadline initially imposed on the RC for reaching an agreement and nearly two years after the Versailles Treaty was signed, a deal was finally struck with Germany whereby it would pay 132 billion gold marks ($51.25 billion). The London Ultimatum, as it became known, required Germany to pay two billion gold marks in annuities as well as annual payments of roughly one billion gold marks (Feldman, 1993). The amount would ultimately become a source of intense political debate within Germany (as discussed below), but, after several years of uncertainty over the impact of reparations on the German budget, the immediate effect of the announcement was to temporarily calm financial markets and reduce the volatility of the German mark. In Germany, reparations began taking their toll well before a final agreement was reached.9 The Versailles Treaty specified that 20 billion gold marks ($4.76 billion) be paid immediately to provision occupying armies as well as of deliveries by Germany of ships, financial securities, natural resources (e.g., coal and timber), livestock and agricultural products. Between 1920 and 1922, the young Weimar Republic ran a budget deficit as reparations payments replaced armaments as the single biggest item in the budget—absorbing 48% of it between April 1920 and March 1923 (Young, 1925). Sargent (1982, p. 73) argues that, if reparations payments were excluded, Germany's budget deficit was not unsustainable until 1923. The reparations agreement of May 1921 proved quite costly from a political standpoint. Adolf Hitler denounced the London payments schedule and began to attract large crowds unhappy with the treaty's terms. Walter Rathenau—founder of the German Democratic Party, foreign minister of the republic, and advocate of fulfilling Germany's obligations as specified in the treaty—was assassinated on 24 June 1922 by right-wing extremists. Shortly afterwards, the RC declared Germany to be in default for its failure to deliver a shipment of timber. The issues were similar in Hungary and Austria. The amount of Austrian reparations was not fixed in the Treaty of Saint-Germain, and the country's financial situation remained murky until the RC determined its obligations. Austria's constitutional courts contended that that the country was not bound to pay obligations that arose out of the war, claiming it was a new country as of 1918. However, the signing of the international treaty suggested this legal argument would not satisfy foreign expectations (Dumberry, 2007, p. 102). Food and raw material shortages plagued the economy after the war and placed pressing financial burdens on the new government. It responded by providing both food relief and unemployment benefits, and accordingly its expenditures outpaced revenue growth. The budget deficit expanded rapidly between 1919 and 1921, and unable to resolve its budget problems, Austria declared bankruptcy in January 1921 before reparation payments began. Hungary's total reparations payment was also not fixed in the Treaty of Trianon, but was finally set in 1924 at 200 million gold crowns. In the intervening period, budget gridlock ensued. In 1920, the minister of finance, Lorant Hegedus, drafted a financial programme that reduced budget expenditures, sought tax reform, and strove for deflation, but the programme met with opposition because of internal disagreement regarding payments to the Allies. As Sargent (1982, p. 57) noted, ‘This circumstance alone created serious obstacles in terms of achieving a stable value for Hungary's currency and other debts, since the unclear reparations obligations made uncertain the nature of the resources to back those debts.’ The budget situation was further complicated by the fact that the successor states to Hungary were supposed to compensate it for its former property. However, this amount was never fixed, and the issue remained unsettled throughout the 1920s. The victors of WWI faced significant fiscal uncertainty as well from the failure to resolve the reparations question (Young, 1925, p. 8). Physical damage from the war was concentrated in north-eastern France and Belgium, and reconstruction of the war-ravaged areas required revenue. At best, Allied soldiers who were fortunate enough to return from the battlefield faced diminished employment opportunities and, at worst, destroyed communities. Adding to these domestic problems were the large international debts that Great Britain and France had accumulated during the war. Both countries had liquidated foreign assets to finance part of the war and then had turned to external finance to cover their growing needs as the war dragged on. Making post-WWI repayment more complicated was France's insistence that it would only service debts to the USA (its largest creditor) if they were fully securitised by future German reparations. This position made it difficult for France to obtain new credit from US sources (Ritschl and Straumann, 2010). Creditors, such as those in the USA, had no incentive to forgive the debts, which only intensified pressure on the French to seek fiscal solutions by demanding large reparations from the former Central Powers. The designation of post-war borders and conflicts arising from them was a second important issue increasing uncertainty over policy decision-making in post-WWI Europe, making it difficult to forecast expenditures (e.g., securing and policing borders) and derive revenues from disputed areas. Border disputes also contributed to continued high military expenditure in many European countries even after the war ended, adding significantly to treasury obligations and budget deficits (Young, 1925, p. 8) and creating uncertain future fiscal obligations so long as disputes remained unresolved. For example, the German Republic accepted its new western borders, but still faced uncertainty over territorial integrity in this direction since France and Belgium were using the threat of invasion as a ‘stick’ in negotiations over reparations payments. Germany's fears were eventually realised in 1923, when the two countries jointly occupied the Ruhr. To the east, Germany failed to fully accept the borders demarcating the newly formed states of Poland and Czechoslovakia. Territorial uncertainty was exacerbated by France's attempt to impose a cordon sanitaire to hem in the ‘diseased nation’ of Germany through its mutual-assistance alliances with these two new nations. Austria suffered immediate financial troubles as a result of having no natural borders defined by ethnicity or production, and the dismembering of the Austro-Hungarian Empire led to serious economic dislocation as new boundaries dissected railroad lines as well as production (e.g., separation of iron and coal mines from steel mills) (Prochnik, 1922). The newly formed nation of Poland was weighed down by four border disputes (two with Czechoslovakia, and one each with Germany and Lithuania) and a war with Soviet Russia, creating uncertainty as to whether the country would even last. In order to secure its borders, the Polish government committed large sums to the military; defence expenditure accounted for 53% of the budget deficit and one-third of the nation's budget in the early 1920s (Durand, 1922). The region of Upper Silesia was only forcibly removed from Germany and incorporated into Poland in June 1922 by the League of Nations, after a plebiscite mandated by the Versailles Treaty failed. Poland's border disputes and war with the Soviet Russia were slowly and fitfully resolved, dragging on from 1919 to 1921. The League of Nations’ attempt to resolve a dispute with Lithuania, simply by imposing a border, failed to resolve the dispute just before Poland tipped into hyperinflation. Uncertainty over economic policy resulted from many unresolved issues regarding the formation of new states out of the former Austro-Hungarian Empire (Austria, Hungary, Czechoslovakia, Poland and the Kingdom of Slovenes, Croats and Serbs). Internal political developments—most prominently, the rapid rise of communism in post-war Europe—further threatened the durability of these policies and governments. These successor states struggled with diminished or non-existent state capacity. Hungary was reduced to one-third of its former size and was left with few raw materials for inputs into industrial production, a reduced internal market, and no access to the sea, which diminished its capacity for international trade (Jonas, 2016). Upon dissolution, political chaos ensued. Béla Kun's communist regime in Hungary and the Romanian occupation of Budapest (August–November 1919) delayed state formation and the ability to collect revenue. Further, the assignment of the existing debt of the Austrian monarchy needed to be settled (Marks, 1978). Interest payments were suspended in 1919, and it was not until the Innsbruck Protocol of 1925 was signed that the existing unsecured debt was formally apportioned. The rise of communist movements further contributed to uncertainty in many European countries after WWI. For example, although socialists were instrumental in forcing Kaiser Wilhelm II to abdicate and establishing the Weimar Republic in Germany, more radical left-wing factions were committed to organised violence in an attempt to establish a Soviet-style communist government. The fledgling German government was forced to commit troops and resources to suppress these violent rebellions in regions such as Bavaria, even after the Republic had been declared. Industrial councils in both Germany and Austria were introduced in 1919 and 1920 in order to appease organised labour and reduce the threat of communist uprisings. Austria went one step further in using social programmes to diffuse the growth of communism, introducing unemployment insurance, but, as noted, the generous provision of social programmes led to massive expenditures without the needed revenue increases. In Poland, the threat of communism came directly from the East, via the border war with Soviet Russia. 2. Measuring Uncertainty during the Interwar Period Obstfeld and Rogoff (1983) show that, since households take into account future inflation when determining how much money to hold today, multiple equilibria for the path of inflation can arise. Inflation could simply depend on the growth rate of money over the growth rate of output, as in a monetarist model, or it could depend on other factors, including the intertemporal government budget constraint, as emphasised in models of the fiscal theory of the price level. For example, Woodford (1995) examines how a government's decision on how to finance its budget is crucial to the path of inflation. Further, news that causes agents to believe that expected primary surpluses will be smaller in the future can result in unanticipated inflation in the present—even before it shows up in a country's budget balance (Davig and Leeper, 2011). Our approach to understanding how economic uncertainty translates into inflation dynamics allows for the possibility that policy news—whether fiscal, monetary and even geopolitical—may influence the path of inflation, regardless of whether it veers away from the path of the money stock, as would be the case in a model of the ‘pure’ fiscal theory of the price level (McCallum and Nelson, 2005). As noted by Sargent (1982) and Leeper (1991), several European monetary authorities during the interwar period had to adjust their money stocks to meet the higher levels of government deficits caused by active fiscal policies. As discussed, across Europe, unresolved political issues, weak state capacity, uncertainty over borders, the size of national debts and other fiscal obligations, and an inability to generate tax revenues all contributed to fiscal policy uncertainty. Our empirical strategy for examining its influence on inflation dynamics needs to allow for the possibility that hyperinflation could have occurred in any European country. We thus develop country-specific measures of uncertainty to allow for within-country comparisons (i.e., to assess whether uncertainty is a driver of inflation and of possible hyperinflation) and for cross-country comparisons (i.e., to determine whether countries that experienced hyperinflation had more pronounced measures of uncertainty relative to other countries. To do so, we follow the methodology introduced in Bloom (2009) and extended by subsequent research that measures how uncertainty influences the macroeconomy. In particular, we focus on the volatility of financial market variables as our measure of economic uncertainty. Bloom (2009) created an uncertainty index based on the volatility of daily US stock returns, and Baker and Bloom (2013) show that bond yields, exchange rates and GDP-forecast volatility are additional suitable proxies for uncertainty. Similar proxies, such as the option-based VIX index of stock market volatility, have been used as uncertainty measures in more recent papers. For example, Baker et al. (2016) find a high correlation between the VIX index and other uncertainty measures. Gilchrist et al. (2014) measure uncertainty using variations in corporate bond spreads and find that uncertainty measured in this way has a strong effect on business investment. Other non-financial measures have been created; for example, Baker et al. (2016) construct a measure of US economic policy uncertainty based on newspaper coverage frequency. They found that their measure influences aggregate investment, consumption and employment. Jurado et al. (2015) generate an uncertainty index from the residuals of a factor-based macroeconomic model and show that this measure identifies fewer periods of elevated uncertainty as compared to measures relying on financial market data. Ludvigson et al. (forthcoming) similarly distinguish between financial market uncertainty and measures reflecting broader macroeconomic uncertainty. Based on our desire to maximise country coverage in the 1920s, we hand collected daily exchange rate data from issues of the Commercial and Financial Chronicle over the years 1919 to 1925. We then computed the monthly, RV based on the countries’ foreign exchange spot prices in the New York trading market.10 Exchange rates are a particularly useful asset price for measuring uncertainty since they capture both real-time economic and political ‘news’ specific to a country as well as information about a country's price dynamics, particularly during periods of economic instability.11 Such RV measures have been shown to be reasonable and consistent proxies for the true, but unobservable, variance of financial time series (Andersen et al. 2003). Accordingly, our FX-based approach provides a country-specific, high-frequency measure of economic policy uncertainty based on public data from the currency markets. An important feature of the underlying data is that they are based on exchange-rate trading in New York City, a foreign-exchange market whose institutional features and market microstructure were not affected by events on the ground in Europe. In addition, previous research on the interwar period supports the notion that foreign-exchange markets were well functioning. For example, Frankel (1977) provides empirical evidence that the exchange market for German Reichsmark operating in London was efficient.12 Einzig (1962, p. 239) argues that trading technology, such as telephone lines, ‘made it much easier to take advantage of discrepancies in rates prevailing at the same time in the different [trading] centres, and consequently those discrepancies tended to narrow down considerably’. As such, RV based on New York spot exchange rates should adequately encompass a country's economic and political news without being polluted by distortions due to country-specific marketplace characteristics. We compute a country's RV for a given month as the average of the squared daily, demeaned log first differences of its exchange rate: $$\begin{equation} {\rm{R}}{{\rm{V}}_{{\rm{kt}}}} = \frac{1}{{{{\rm{N}}_{\rm{t}}} - 1}}\sum\limits_{{\rm{i}} = 1}^{{{\rm{N}}_{\rm{t}}} - 1} {{{(\Delta {{\rm{X}}_{{\rm{ki}} + 1}} - {{\rm{\mu }}_{{\rm{kt}}}})}^2}} , \end{equation}$$(1) where k is a country index; Nt is the number of New York trading days in month t; Δxki+1 is the change in the logged spot exchange rate (in US dollars) for country k from day i to i + 1 in month t; and μkt is the average value of these changes in month t.13 Since these are measures of scale deviations, we can readily compare them in relative terms, and Table 2 shows that RV has a reasonable degree of cross-sectional and time-series variation. Notably, the GAPH countries shown in the first four columns generally have higher RV values and exhibit upswings leading into their observed hyperinflation periods. These characteristics are especially pronounced for Germany in 1923. The average RV values for the non-hyperinflation countries are clearly lower, although they do exhibit interesting spikes throughout the sample. Note that the Czech RV measure starts at a high level and drops steadily to fall in line with the other non-hyperinflation countries. Table 2. Uncertainty in Europe after WWI (Average Annual Realised Volatility). . 1919 . 1920 . 1921 . 1922 . 1923 . 1924 . 1925 . Germany 31.00 33.51 29.35 65.54 1,217.26 2.35 – Austria 56.07 51.07 55.42 46.41 – – – Poland – 53.26 63.46 16.65 162.98 10.50 12.24 Hungary – 35.07 37.80 32.53 59.78 29.77 0.38 Belgium 4.31 4.77 5.49 1.71 3.73 7.57 0.58 Great Britain 1.15 1.49 0.40 0.14 0.08 0.13 0.01 Czechoslovakia – 29.27 8.44 7.86 0.34 0.29 0.00 France 5.72 6.53 2.70 1.76 2.46 6.08 1.36 Netherland 0.58 0.89 0.41 0.10 0.07 0.08 0.00 Italy 3.43 8.17 3.64 2.40 0.96 0.45 1.42 . 1919 . 1920 . 1921 . 1922 . 1923 . 1924 . 1925 . Germany 31.00 33.51 29.35 65.54 1,217.26 2.35 – Austria 56.07 51.07 55.42 46.41 – – – Poland – 53.26 63.46 16.65 162.98 10.50 12.24 Hungary – 35.07 37.80 32.53 59.78 29.77 0.38 Belgium 4.31 4.77 5.49 1.71 3.73 7.57 0.58 Great Britain 1.15 1.49 0.40 0.14 0.08 0.13 0.01 Czechoslovakia – 29.27 8.44 7.86 0.34 0.29 0.00 France 5.72 6.53 2.70 1.76 2.46 6.08 1.36 Netherland 0.58 0.89 0.41 0.10 0.07 0.08 0.00 Italy 3.43 8.17 3.64 2.40 0.96 0.45 1.42 Notes: Monthly RV is computed using daily spot exchange rates from the New York trading market. Annual averages are simple means of that year's monthly values. Open in new tab Table 2. Uncertainty in Europe after WWI (Average Annual Realised Volatility). . 1919 . 1920 . 1921 . 1922 . 1923 . 1924 . 1925 . Germany 31.00 33.51 29.35 65.54 1,217.26 2.35 – Austria 56.07 51.07 55.42 46.41 – – – Poland – 53.26 63.46 16.65 162.98 10.50 12.24 Hungary – 35.07 37.80 32.53 59.78 29.77 0.38 Belgium 4.31 4.77 5.49 1.71 3.73 7.57 0.58 Great Britain 1.15 1.49 0.40 0.14 0.08 0.13 0.01 Czechoslovakia – 29.27 8.44 7.86 0.34 0.29 0.00 France 5.72 6.53 2.70 1.76 2.46 6.08 1.36 Netherland 0.58 0.89 0.41 0.10 0.07 0.08 0.00 Italy 3.43 8.17 3.64 2.40 0.96 0.45 1.42 . 1919 . 1920 . 1921 . 1922 . 1923 . 1924 . 1925 . Germany 31.00 33.51 29.35 65.54 1,217.26 2.35 – Austria 56.07 51.07 55.42 46.41 – – – Poland – 53.26 63.46 16.65 162.98 10.50 12.24 Hungary – 35.07 37.80 32.53 59.78 29.77 0.38 Belgium 4.31 4.77 5.49 1.71 3.73 7.57 0.58 Great Britain 1.15 1.49 0.40 0.14 0.08 0.13 0.01 Czechoslovakia – 29.27 8.44 7.86 0.34 0.29 0.00 France 5.72 6.53 2.70 1.76 2.46 6.08 1.36 Netherland 0.58 0.89 0.41 0.10 0.07 0.08 0.00 Italy 3.43 8.17 3.64 2.40 0.96 0.45 1.42 Notes: Monthly RV is computed using daily spot exchange rates from the New York trading market. Annual averages are simple means of that year's monthly values. Open in new tab RV variation can be examined in more detail using time-series plots for individual countries. Figures 1–3 display RV from July 1919 to December 1925 relative to the initial RV value for the Netherlands. This initial value was selected as a benchmark for comparison since the Netherlands was a neutral country during the war and has data available for the entire sample period. Figure 1 compares the normalised RV measures for the Netherlands and Great Britain, which are the lowest across our sample of countries. Note that Netherland's RV measure in January 1920 is much higher relative to its normalising observation in July 1919, spiking up to 3.0. Contemporary news reports suggest that this increased uncertainty can likely be attributed to the country's decision to join the League of Nations and its refusal to hand over Kaiser Wilhelm as a war criminal under the Treaty of Versailles. After that peak, the RV series declines steadily and reaches a low of 0.1 by the end of the sample in December 1925. The British measure follows a similar pattern, although it exhibits much more volatility in early 1920. These patterns suggest that uncertainty in these two economies and their political environments, while relatively elevated shortly after the end of the war with the November 1918 armistice, declined steadily over the sample period. Fig. 1. Open in new tabDownload slide Normalised RV Measures for the Netherlands and the UK. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to December 1925. Fig. 1. Open in new tabDownload slide Normalised RV Measures for the Netherlands and the UK. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to December 1925. Figure 2 shows the RV measures for four other European countries which, ex post, did not experience hyperinflation. Belgium, Czechoslovakia, France and Italy clearly exhibit much higher peaks and are more volatile, reaching levels that are ten to 20 times larger than the Dutch benchmark value. This degree of variation in the uncertainty measures speaks to the wide variation in economic conditions and fiscal policymaking environments across post-war Europe. For example, France felt justified in running large budget deficits, believing Germany would eventually pay reparations for the reconstruction of territories destroyed or damaged during WWI. The French finance minister Poincaré more or less proclaimed this in his famous quip ‘le Boche paiera tout’, or, roughly, the Germans will pay for everything. And, in order to enforce their claim to reparations and secure borders shared with Germany, France maintained significantly higher budget shares of military spending into the early 1920s. Note that the RV measures for France and Belgium spike to nearly ten before and during the occupation of Germany's Ruhr Valley in January 1923. Fig. 2. Open in new tabDownload slide Normalised RV for Non-hyperinflationary European Countries. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to December 1925. Fig. 2. Open in new tabDownload slide Normalised RV for Non-hyperinflationary European Countries. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to December 1925. Figure 3 shows the RV dynamics for the GAPH countries, those that eventually experienced hyperinflation. The time periods depicted reflect our estimation samples, which depend on data availability as well as the start of hyperinflation—as defined in the next section. (See Table 4 for further information on sources and sample periods.) Figure 3A shows the German RV series and highlights how the events of 1923—the year in which Germany tipped into hyperinflation—clearly increased uncertainty to new heights. The occupation of the Ruhr Valley and the tense continuing negotiations around reparations raised RV uncertainty by up to 550 times the benchmark. Fig. 3A. Open in new tabDownload slide Normalised RV Measure for Germany. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to June 1923; see Table 4. Fig. 3A. Open in new tabDownload slide Normalised RV Measure for Germany. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to June 1923; see Table 4. Fig. 3B. Open in new tabDownload slide Normalised RV Measure for Austria. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to September 1922; see Table 4. Fig. 3B. Open in new tabDownload slide Normalised RV Measure for Austria. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from July 1919 to September 1922; see Table 4. Figures 3B–3D show the RV measures for Austria, Poland and Hungary. These values are clearly much higher than those observed for the non-inflationary countries shown in Figures 1 and 2. Based on contemporaneous news accounts, we highlight developments that directly contributed to the countries’ exchange rate volatility and general economic uncertainty. For example, Polish uncertainty spikes around military actions and attempted peace negotiations as well as failed governments. Table 3 displays the largest ten RV values prior to the start of the hyperinflations in these countries.14 Common themes around the noted RV spikes are border disputes, controversial fiscal policies such as tax increases, and government realignments due to resignations or elections. Fig. 3C. Open in new tabDownload slide Normalised RV Measure for Poland. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from March 1920 to September 1923; see Table 4. Fig. 3C. Open in new tabDownload slide Normalised RV Measure for Poland. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from March 1920 to September 1923; see Table 4. Fig. 3D. Open in new tabDownload slide Normalised RV Measures for Hungary. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from November 1920 to June 1924; see Table 4. Fig. 3D. Open in new tabDownload slide Normalised RV Measures for Hungary. Notes: Values are relative to the RV measure for the Netherlands in July 1919. The time period displayed is from November 1920 to June 1924; see Table 4. Table 3. News in GAPH for the Ten Highest RV Months prior to Hyperinflation. Month . Rank . . Panel A. Germany Mar. 1920 10 Ruhr uprising by Communists. Resignation of Republican government under Bauer. Introduction of new income tax brackets. Kapp Putsch with call for general strike in response. Dec. 1921 6 Chancellor Wirth requests payment moratorium just six months after the first payment under the London Ultimatum Aug. 1922 4 The Allied RC unanimously decide to grant a six-month moratorium on reparations payments Oct. 1922 8 Chancellor Wirth submits a question to the cabinet council as to whether to declare bankruptcy Nov. 1922 5 Chancellor Wirth's complete moratorium efforts fails, and he resigns on 14 Nov. Dec. 1922 9 New German proposal of a five-year moratorium on reparations payments Jan. 1923 1 France occupies the Ruhr to ensure payment of war reparations Feb. 1923 3 Striking railway workers in the Ruhr begin returning to their jobs as German resistance falters Apr. 1923 7 Wilhelm Cuno's government spends more than seven times revenue received June 1923 2 Germany asks for a new reparations conference Panel B. Austria Nov. 1919 9 Emperor Karl returns to Vienna to lead a constitutional committee Dec. 1919 6 Minister of Labour Hanusch introduces the Eight-Hour Act to limit the workday Jan. 1920 3 Vienna runs out of coal, shutting down electric tramcars May 1920 2 200,000 people protest against a proposed tax on capital Jan. 1921 4 Austrian government declares itself insolvent May 1921 7 Threat of secession crisis in May and June in several Austrian provinces Sept. 1921 10 Regular Hungarian troops reoccupy Burgenland Oct. 1921 8 Bauer introduces the Social Democratic Financial Plan, increasing taxes and decreasing government expenditures June 1922 5 Finance minister Gurtler is forced to retire after protests against the new customs tariff Aug. 1922 1 Allied Powers ask the League of Nations to intervene after Austria exhausts all foreign borrowing possibilities Panel C. Poland Nov. 1920 2 Peace treaty between Poland and Soviet Russia is ratified, but negotiations continue on terms. Petlura's Ukrainian forces are driven into Polish-controlled territory by Bolshevik Russian attacks. Dec. 1920 5 Belgian diplomats attempt to resolve the Lithuanian–Polish dispute on behalf of the League of Nations Jan. 1921 9 Widespread shortages of food and other essentials reported in major cities July 1921 6 Large-scale fighting between ‘irregular’ Polish and German forces over the partition of Silesia. British troops return to Upper Silesia to help French forces occupy the area. Sep. 1921 8 The League of Nations Council holds its 14th extraordinary session in Geneva to discuss the Upper Silesia dispute between Germany and Poland Oct. 1921 7 The League settles the Upper Silesia dispute. Poland receives most of the industrial areas, but only about one-third of the contested land. Nov. 1921 1 Troubled implementation of the League settlement. In an effort to prevent a war, the League hosts a conference between Germany and Poland. July 1922 10 Fall of Prime Minister Ponikowski-led government creates political crisis. National Democrats try to set up a government under Sliwinski, but are opposed by Pilsudski. Two prime ministers are appointed in one month. June 1923 3 Chancellor Sikorski is overthrown, and Witos installed as the new prime minister. Right-wing parties control the formation of a new cabinet. July 1923 4 Witos failed to secure a $100 m. loan from Hallgarten & Co., a US firm Panel D. Hungary Nov. 1920 8 The National Assembly reluctantly ratifies the Treaty of Trianon Mar. 1921 7 Charles I, the former king, attempts a coup Nov. 1921 10 Charles I's second coup attempt Dec. 1921 4 Plebiscite to determine the fate of Burgenland; Austria receives most of it Jan. 1922 6 Prime Minister Kallay announces a budget deficit nearly double what was expected Aug. 1922 2 Cabinet meets in an extraordinary session to consider the ‘Austrian’ problem July 1923 1 Negotiations for loan from League of Nations in progress Jan. 1924 9 Terms of a League loan are ratified Mar. 1924 5 Final negotiations of the League loan, including borders and reparations schedule May 1924 3 Arrangements for the League loan are made Month . Rank . . Panel A. Germany Mar. 1920 10 Ruhr uprising by Communists. Resignation of Republican government under Bauer. Introduction of new income tax brackets. Kapp Putsch with call for general strike in response. Dec. 1921 6 Chancellor Wirth requests payment moratorium just six months after the first payment under the London Ultimatum Aug. 1922 4 The Allied RC unanimously decide to grant a six-month moratorium on reparations payments Oct. 1922 8 Chancellor Wirth submits a question to the cabinet council as to whether to declare bankruptcy Nov. 1922 5 Chancellor Wirth's complete moratorium efforts fails, and he resigns on 14 Nov. Dec. 1922 9 New German proposal of a five-year moratorium on reparations payments Jan. 1923 1 France occupies the Ruhr to ensure payment of war reparations Feb. 1923 3 Striking railway workers in the Ruhr begin returning to their jobs as German resistance falters Apr. 1923 7 Wilhelm Cuno's government spends more than seven times revenue received June 1923 2 Germany asks for a new reparations conference Panel B. Austria Nov. 1919 9 Emperor Karl returns to Vienna to lead a constitutional committee Dec. 1919 6 Minister of Labour Hanusch introduces the Eight-Hour Act to limit the workday Jan. 1920 3 Vienna runs out of coal, shutting down electric tramcars May 1920 2 200,000 people protest against a proposed tax on capital Jan. 1921 4 Austrian government declares itself insolvent May 1921 7 Threat of secession crisis in May and June in several Austrian provinces Sept. 1921 10 Regular Hungarian troops reoccupy Burgenland Oct. 1921 8 Bauer introduces the Social Democratic Financial Plan, increasing taxes and decreasing government expenditures June 1922 5 Finance minister Gurtler is forced to retire after protests against the new customs tariff Aug. 1922 1 Allied Powers ask the League of Nations to intervene after Austria exhausts all foreign borrowing possibilities Panel C. Poland Nov. 1920 2 Peace treaty between Poland and Soviet Russia is ratified, but negotiations continue on terms. Petlura's Ukrainian forces are driven into Polish-controlled territory by Bolshevik Russian attacks. Dec. 1920 5 Belgian diplomats attempt to resolve the Lithuanian–Polish dispute on behalf of the League of Nations Jan. 1921 9 Widespread shortages of food and other essentials reported in major cities July 1921 6 Large-scale fighting between ‘irregular’ Polish and German forces over the partition of Silesia. British troops return to Upper Silesia to help French forces occupy the area. Sep. 1921 8 The League of Nations Council holds its 14th extraordinary session in Geneva to discuss the Upper Silesia dispute between Germany and Poland Oct. 1921 7 The League settles the Upper Silesia dispute. Poland receives most of the industrial areas, but only about one-third of the contested land. Nov. 1921 1 Troubled implementation of the League settlement. In an effort to prevent a war, the League hosts a conference between Germany and Poland. July 1922 10 Fall of Prime Minister Ponikowski-led government creates political crisis. National Democrats try to set up a government under Sliwinski, but are opposed by Pilsudski. Two prime ministers are appointed in one month. June 1923 3 Chancellor Sikorski is overthrown, and Witos installed as the new prime minister. Right-wing parties control the formation of a new cabinet. July 1923 4 Witos failed to secure a $100 m. loan from Hallgarten & Co., a US firm Panel D. Hungary Nov. 1920 8 The National Assembly reluctantly ratifies the Treaty of Trianon Mar. 1921 7 Charles I, the former king, attempts a coup Nov. 1921 10 Charles I's second coup attempt Dec. 1921 4 Plebiscite to determine the fate of Burgenland; Austria receives most of it Jan. 1922 6 Prime Minister Kallay announces a budget deficit nearly double what was expected Aug. 1922 2 Cabinet meets in an extraordinary session to consider the ‘Austrian’ problem July 1923 1 Negotiations for loan from League of Nations in progress Jan. 1924 9 Terms of a League loan are ratified Mar. 1924 5 Final negotiations of the League loan, including borders and reparations schedule May 1924 3 Arrangements for the League loan are made Notes: Data are displayed chronologically for the ten months with the highest RV for each country. Column 2 shows the country-specific RV rank for that month. See Online Appendix A for reference sources. Open in new tab Table 3. News in GAPH for the Ten Highest RV Months prior to Hyperinflation. Month . Rank . . Panel A. Germany Mar. 1920 10 Ruhr uprising by Communists. Resignation of Republican government under Bauer. Introduction of new income tax brackets. Kapp Putsch with call for general strike in response. Dec. 1921 6 Chancellor Wirth requests payment moratorium just six months after the first payment under the London Ultimatum Aug. 1922 4 The Allied RC unanimously decide to grant a six-month moratorium on reparations payments Oct. 1922 8 Chancellor Wirth submits a question to the cabinet council as to whether to declare bankruptcy Nov. 1922 5 Chancellor Wirth's complete moratorium efforts fails, and he resigns on 14 Nov. Dec. 1922 9 New German proposal of a five-year moratorium on reparations payments Jan. 1923 1 France occupies the Ruhr to ensure payment of war reparations Feb. 1923 3 Striking railway workers in the Ruhr begin returning to their jobs as German resistance falters Apr. 1923 7 Wilhelm Cuno's government spends more than seven times revenue received June 1923 2 Germany asks for a new reparations conference Panel B. Austria Nov. 1919 9 Emperor Karl returns to Vienna to lead a constitutional committee Dec. 1919 6 Minister of Labour Hanusch introduces the Eight-Hour Act to limit the workday Jan. 1920 3 Vienna runs out of coal, shutting down electric tramcars May 1920 2 200,000 people protest against a proposed tax on capital Jan. 1921 4 Austrian government declares itself insolvent May 1921 7 Threat of secession crisis in May and June in several Austrian provinces Sept. 1921 10 Regular Hungarian troops reoccupy Burgenland Oct. 1921 8 Bauer introduces the Social Democratic Financial Plan, increasing taxes and decreasing government expenditures June 1922 5 Finance minister Gurtler is forced to retire after protests against the new customs tariff Aug. 1922 1 Allied Powers ask the League of Nations to intervene after Austria exhausts all foreign borrowing possibilities Panel C. Poland Nov. 1920 2 Peace treaty between Poland and Soviet Russia is ratified, but negotiations continue on terms. Petlura's Ukrainian forces are driven into Polish-controlled territory by Bolshevik Russian attacks. Dec. 1920 5 Belgian diplomats attempt to resolve the Lithuanian–Polish dispute on behalf of the League of Nations Jan. 1921 9 Widespread shortages of food and other essentials reported in major cities July 1921 6 Large-scale fighting between ‘irregular’ Polish and German forces over the partition of Silesia. British troops return to Upper Silesia to help French forces occupy the area. Sep. 1921 8 The League of Nations Council holds its 14th extraordinary session in Geneva to discuss the Upper Silesia dispute between Germany and Poland Oct. 1921 7 The League settles the Upper Silesia dispute. Poland receives most of the industrial areas, but only about one-third of the contested land. Nov. 1921 1 Troubled implementation of the League settlement. In an effort to prevent a war, the League hosts a conference between Germany and Poland. July 1922 10 Fall of Prime Minister Ponikowski-led government creates political crisis. National Democrats try to set up a government under Sliwinski, but are opposed by Pilsudski. Two prime ministers are appointed in one month. June 1923 3 Chancellor Sikorski is overthrown, and Witos installed as the new prime minister. Right-wing parties control the formation of a new cabinet. July 1923 4 Witos failed to secure a $100 m. loan from Hallgarten & Co., a US firm Panel D. Hungary Nov. 1920 8 The National Assembly reluctantly ratifies the Treaty of Trianon Mar. 1921 7 Charles I, the former king, attempts a coup Nov. 1921 10 Charles I's second coup attempt Dec. 1921 4 Plebiscite to determine the fate of Burgenland; Austria receives most of it Jan. 1922 6 Prime Minister Kallay announces a budget deficit nearly double what was expected Aug. 1922 2 Cabinet meets in an extraordinary session to consider the ‘Austrian’ problem July 1923 1 Negotiations for loan from League of Nations in progress Jan. 1924 9 Terms of a League loan are ratified Mar. 1924 5 Final negotiations of the League loan, including borders and reparations schedule May 1924 3 Arrangements for the League loan are made Month . Rank . . Panel A. Germany Mar. 1920 10 Ruhr uprising by Communists. Resignation of Republican government under Bauer. Introduction of new income tax brackets. Kapp Putsch with call for general strike in response. Dec. 1921 6 Chancellor Wirth requests payment moratorium just six months after the first payment under the London Ultimatum Aug. 1922 4 The Allied RC unanimously decide to grant a six-month moratorium on reparations payments Oct. 1922 8 Chancellor Wirth submits a question to the cabinet council as to whether to declare bankruptcy Nov. 1922 5 Chancellor Wirth's complete moratorium efforts fails, and he resigns on 14 Nov. Dec. 1922 9 New German proposal of a five-year moratorium on reparations payments Jan. 1923 1 France occupies the Ruhr to ensure payment of war reparations Feb. 1923 3 Striking railway workers in the Ruhr begin returning to their jobs as German resistance falters Apr. 1923 7 Wilhelm Cuno's government spends more than seven times revenue received June 1923 2 Germany asks for a new reparations conference Panel B. Austria Nov. 1919 9 Emperor Karl returns to Vienna to lead a constitutional committee Dec. 1919 6 Minister of Labour Hanusch introduces the Eight-Hour Act to limit the workday Jan. 1920 3 Vienna runs out of coal, shutting down electric tramcars May 1920 2 200,000 people protest against a proposed tax on capital Jan. 1921 4 Austrian government declares itself insolvent May 1921 7 Threat of secession crisis in May and June in several Austrian provinces Sept. 1921 10 Regular Hungarian troops reoccupy Burgenland Oct. 1921 8 Bauer introduces the Social Democratic Financial Plan, increasing taxes and decreasing government expenditures June 1922 5 Finance minister Gurtler is forced to retire after protests against the new customs tariff Aug. 1922 1 Allied Powers ask the League of Nations to intervene after Austria exhausts all foreign borrowing possibilities Panel C. Poland Nov. 1920 2 Peace treaty between Poland and Soviet Russia is ratified, but negotiations continue on terms. Petlura's Ukrainian forces are driven into Polish-controlled territory by Bolshevik Russian attacks. Dec. 1920 5 Belgian diplomats attempt to resolve the Lithuanian–Polish dispute on behalf of the League of Nations Jan. 1921 9 Widespread shortages of food and other essentials reported in major cities July 1921 6 Large-scale fighting between ‘irregular’ Polish and German forces over the partition of Silesia. British troops return to Upper Silesia to help French forces occupy the area. Sep. 1921 8 The League of Nations Council holds its 14th extraordinary session in Geneva to discuss the Upper Silesia dispute between Germany and Poland Oct. 1921 7 The League settles the Upper Silesia dispute. Poland receives most of the industrial areas, but only about one-third of the contested land. Nov. 1921 1 Troubled implementation of the League settlement. In an effort to prevent a war, the League hosts a conference between Germany and Poland. July 1922 10 Fall of Prime Minister Ponikowski-led government creates political crisis. National Democrats try to set up a government under Sliwinski, but are opposed by Pilsudski. Two prime ministers are appointed in one month. June 1923 3 Chancellor Sikorski is overthrown, and Witos installed as the new prime minister. Right-wing parties control the formation of a new cabinet. July 1923 4 Witos failed to secure a $100 m. loan from Hallgarten & Co., a US firm Panel D. Hungary Nov. 1920 8 The National Assembly reluctantly ratifies the Treaty of Trianon Mar. 1921 7 Charles I, the former king, attempts a coup Nov. 1921 10 Charles I's second coup attempt Dec. 1921 4 Plebiscite to determine the fate of Burgenland; Austria receives most of it Jan. 1922 6 Prime Minister Kallay announces a budget deficit nearly double what was expected Aug. 1922 2 Cabinet meets in an extraordinary session to consider the ‘Austrian’ problem July 1923 1 Negotiations for loan from League of Nations in progress Jan. 1924 9 Terms of a League loan are ratified Mar. 1924 5 Final negotiations of the League loan, including borders and reparations schedule May 1924 3 Arrangements for the League loan are made Notes: Data are displayed chronologically for the ten months with the highest RV for each country. Column 2 shows the country-specific RV rank for that month. See Online Appendix A for reference sources. Open in new tab 3. The Role of Uncertainty on Macroeconomic Outcomes in Post-WWI Europe Recent scholarship suggests that high-volatility regimes are particularly relevant for capturing uncertainty's impact on macroeconomic outcomes. For example, Danielsson et al. (2018) find that large deviations from standard conditions of financial market volatility, both positive and negative, contribute to the incidence of banking and financial crises in the near future. Jurado et al. (2015), using a broader measure of macroeconomic uncertainty, find that large positive innovations due to uncertainty led to sizable declines in real activity using post-World War II US data. Given that our country-specific RV measure well encompasses political and economic uncertainty in post-WWI Europe, we use these measures to examine what effect uncertainty had on the inflation dynamics of the period. Table 4 and Online Appendix B provide details about the sources and price indices used to compute monthly inflation rates for the ten European countries in our sample. As shown in Table 4, there are several constraints that define our sample period. First, we are limited by the availability of exchange-rate data. The earliest month for which we have sufficient data to conduct our multivariate analysis is July 1919, when the daily exchange-rate series are first available for certain countries. Another constraint concerns the availability of monthly inflation figures, which are available in the early 1920s for GAPH as well as for Belgium, Czechoslovakia and the Netherlands. The last column of Table 4 lists the number of monthly observations available for each country. Table 4. Model Estimation Sample Periods. Country . Start date . Reason . End date . Reason . Number of months . Austria Jan. 1921 Monthly CPI starts Sept. 1922 Month before League of Nations loan and control 21 Germany Apr. 1920 Monthly WPI starts June 1923 Month before defined hyperinflation 39 Hungary Nov. 1922 Monthly WPI starts June 1924 New central bank commences operations; RV near zero 20 Poland Sept. 1920 Monthly WPI starts Sept. 1923 Month before defined hyperinflation 37 UK July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Belgium Feb. 1921 Monthly WPI starts Apr. 1925 Monthly notes data ends 52 Czech. July 1919 Monthly WPI starts Apr. 1925 Monthly notes data ends 51 France July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Holland Jan. 1920 Monthly WPI starts Apr. 1925 Monthly notes data ends 64 Italy July 1919 Monthly RV starts Nov. 1924 Monthly notes data ends 66 Country . Start date . Reason . End date . Reason . Number of months . Austria Jan. 1921 Monthly CPI starts Sept. 1922 Month before League of Nations loan and control 21 Germany Apr. 1920 Monthly WPI starts June 1923 Month before defined hyperinflation 39 Hungary Nov. 1922 Monthly WPI starts June 1924 New central bank commences operations; RV near zero 20 Poland Sept. 1920 Monthly WPI starts Sept. 1923 Month before defined hyperinflation 37 UK July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Belgium Feb. 1921 Monthly WPI starts Apr. 1925 Monthly notes data ends 52 Czech. July 1919 Monthly WPI starts Apr. 1925 Monthly notes data ends 51 France July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Holland Jan. 1920 Monthly WPI starts Apr. 1925 Monthly notes data ends 64 Italy July 1919 Monthly RV starts Nov. 1924 Monthly notes data ends 66 Notes: The ‘Reason’ column explains the binding date availability restrictions. WPI = Wholesale Price Index and CPI = Consumer Price Index. Open in new tab Table 4. Model Estimation Sample Periods. Country . Start date . Reason . End date . Reason . Number of months . Austria Jan. 1921 Monthly CPI starts Sept. 1922 Month before League of Nations loan and control 21 Germany Apr. 1920 Monthly WPI starts June 1923 Month before defined hyperinflation 39 Hungary Nov. 1922 Monthly WPI starts June 1924 New central bank commences operations; RV near zero 20 Poland Sept. 1920 Monthly WPI starts Sept. 1923 Month before defined hyperinflation 37 UK July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Belgium Feb. 1921 Monthly WPI starts Apr. 1925 Monthly notes data ends 52 Czech. July 1919 Monthly WPI starts Apr. 1925 Monthly notes data ends 51 France July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Holland Jan. 1920 Monthly WPI starts Apr. 1925 Monthly notes data ends 64 Italy July 1919 Monthly RV starts Nov. 1924 Monthly notes data ends 66 Country . Start date . Reason . End date . Reason . Number of months . Austria Jan. 1921 Monthly CPI starts Sept. 1922 Month before League of Nations loan and control 21 Germany Apr. 1920 Monthly WPI starts June 1923 Month before defined hyperinflation 39 Hungary Nov. 1922 Monthly WPI starts June 1924 New central bank commences operations; RV near zero 20 Poland Sept. 1920 Monthly WPI starts Sept. 1923 Month before defined hyperinflation 37 UK July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Belgium Feb. 1921 Monthly WPI starts Apr. 1925 Monthly notes data ends 52 Czech. July 1919 Monthly WPI starts Apr. 1925 Monthly notes data ends 51 France July 1919 Monthly RV starts Dec. 1924 Monthly notes data ends 66 Holland Jan. 1920 Monthly WPI starts Apr. 1925 Monthly notes data ends 64 Italy July 1919 Monthly RV starts Nov. 1924 Monthly notes data ends 66 Notes: The ‘Reason’ column explains the binding date availability restrictions. WPI = Wholesale Price Index and CPI = Consumer Price Index. Open in new tab In terms of measuring the relationship between uncertainty and inflation, one necessary modelling assumption is deciding on when hyperinflation began for the countries that experienced it. Our intuition is that when a country crosses over into a state of hyperinflation, the stability of the macroeconomic relationships that we would hope to capture with a model breakdown sufficiently as to likely provide spurious inference. Because there is no standard definition, we use a 100% increase in prices in a month to denote the onset of hyperinflation in a country (i.e., a doubling of the level of prices over a month). Thus, as shown in the top panel of Table 4, our sample period for Germany ends in June 1923, with a monthly inflation rate of about 86%, since the inflation rate in the following month was 136%.15 Similarly, our end date for Poland is September 1923, when the monthly inflation rate was 32%, since inflation reached 132% in the following month. Due to external intervention, the start of hyperinflation for Austria and Hungary are defined differently. For the Austrian sample, the monthly inflation rate reaches a maximum of only 83% in August 1922. In the following month, the League of Nations extended the country a significant loan and assumed control of certain monetary functions. As this represented a regime shift and effectively ended trading in the kronen (i.e., no reported price fluctuations in New York trading), our RV measure of uncertainty could not be calculated beyond September 1922. For Hungary, monthly inflation rates were certainly elevated, reaching 68% in July 1923 and 52% in February 1924, but never exceeded 100%.16 Accordingly, we also use regime change to mark the sample's endpoint at June 1924—when the Hungarian National Bank was formed and the krone effectively stopped trading before the forint was introduced in August 1924. Before describing our econometric approach, Table 5 presents simple summary statistics on the correlations between the countries’ monthly inflation rates (denoted as πk,t) with contemporaneous and lagged values of the changes in their RV values: $$\begin{equation} {\rm{\rho }}\left( {{{\rm{\pi }}_{{\rm{kt}}}},\Delta {\rm{R}}{{\rm{V}}_{{\rm{kt}} - {\rm{j}}}}} \right),j \in \left[ {0,6} \right]. \end{equation}$$(2) Table 5. The Correlation between Inflation and Lagged Changes in RV. . Lag j . Country . 0 . 1 . 2 . 3 . 4 . 5 . 6 . Germany 0.38* 0.24* 0.01 −0.08 0.15 0.17* 0.13 Austria 0.27* 0.11 0.23* 0.26 −0.36* −0.22 −0.17 Poland 0.06 0.25* 0.09 −0.07 −0.08 −0.20* 0.00 Hungary 0.25* 0.51* 0.09 −0.16 0.04 −0.21 −0.27 UK −0.06 0.03 0.06 −0.04 −0.08 −0.03 0.02 Belgium 0.20* −0.23* −0.05 0.11 −0.03 −0.18* 0.22* Czech. −0.01 −0.09 −0.27* −0.02 0.07 0.07 0.04 France 0.07 0.01 0.13 −0.06 0.15* −0.02 −0.08 Netherlands 0.23* −0.22* 0.21* −0.25* 0.21* 0.07 −0.22* Italy 0.23* −0.31* 0.18* −0.09 −0.01 0.03 0.13 . Lag j . Country . 0 . 1 . 2 . 3 . 4 . 5 . 6 . Germany 0.38* 0.24* 0.01 −0.08 0.15 0.17* 0.13 Austria 0.27* 0.11 0.23* 0.26 −0.36* −0.22 −0.17 Poland 0.06 0.25* 0.09 −0.07 −0.08 −0.20* 0.00 Hungary 0.25* 0.51* 0.09 −0.16 0.04 −0.21 −0.27 UK −0.06 0.03 0.06 −0.04 −0.08 −0.03 0.02 Belgium 0.20* −0.23* −0.05 0.11 −0.03 −0.18* 0.22* Czech. −0.01 −0.09 −0.27* −0.02 0.07 0.07 0.04 France 0.07 0.01 0.13 −0.06 0.15* −0.02 −0.08 Netherlands 0.23* −0.22* 0.21* −0.25* 0.21* 0.07 −0.22* Italy 0.23* −0.31* 0.18* −0.09 −0.01 0.03 0.13 Notes: Correlations are for ρ(πkt, ΔRVkt−j), jϵ[0,6] as described in the text. Asterisks denote statistical significance at the 10% level. Open in new tab Table 5. The Correlation between Inflation and Lagged Changes in RV. . Lag j . Country . 0 . 1 . 2 . 3 . 4 . 5 . 6 . Germany 0.38* 0.24* 0.01 −0.08 0.15 0.17* 0.13 Austria 0.27* 0.11 0.23* 0.26 −0.36* −0.22 −0.17 Poland 0.06 0.25* 0.09 −0.07 −0.08 −0.20* 0.00 Hungary 0.25* 0.51* 0.09 −0.16 0.04 −0.21 −0.27 UK −0.06 0.03 0.06 −0.04 −0.08 −0.03 0.02 Belgium 0.20* −0.23* −0.05 0.11 −0.03 −0.18* 0.22* Czech. −0.01 −0.09 −0.27* −0.02 0.07 0.07 0.04 France 0.07 0.01 0.13 −0.06 0.15* −0.02 −0.08 Netherlands 0.23* −0.22* 0.21* −0.25* 0.21* 0.07 −0.22* Italy 0.23* −0.31* 0.18* −0.09 −0.01 0.03 0.13 . Lag j . Country . 0 . 1 . 2 . 3 . 4 . 5 . 6 . Germany 0.38* 0.24* 0.01 −0.08 0.15 0.17* 0.13 Austria 0.27* 0.11 0.23* 0.26 −0.36* −0.22 −0.17 Poland 0.06 0.25* 0.09 −0.07 −0.08 −0.20* 0.00 Hungary 0.25* 0.51* 0.09 −0.16 0.04 −0.21 −0.27 UK −0.06 0.03 0.06 −0.04 −0.08 −0.03 0.02 Belgium 0.20* −0.23* −0.05 0.11 −0.03 −0.18* 0.22* Czech. −0.01 −0.09 −0.27* −0.02 0.07 0.07 0.04 France 0.07 0.01 0.13 −0.06 0.15* −0.02 −0.08 Netherlands 0.23* −0.22* 0.21* −0.25* 0.21* 0.07 −0.22* Italy 0.23* −0.31* 0.18* −0.09 −0.01 0.03 0.13 Notes: Correlations are for ρ(πkt, ΔRVkt−j), jϵ[0,6] as described in the text. Asterisks denote statistical significance at the 10% level. Open in new tab These correlations provide a simple view into the co-movements of these two series in our sample countries. The top panel presents the correlations for GAPH. The first point to notice is that, with the exception of Poland, there is a fairly large contemporaneous correlation between changes in GAPH uncertainty and inflation, ranging from 0.25 to 0.38. Although the correlations are calculated over the relatively short sample periods available, these values are positive, relatively large and statistically significant at the 10% level. The second point to notice is that the one-lag correlations remain positive and significant for all GAPH countries except Austria, whose second-lag correlation fits this pattern. These correlation patterns provide intuitive support for the hypothesis that economic uncertainty is an important driver of inflation in countries experiencing hyperinflation. The correlations shown in the bottom panel of Table 5 suggest a different result for the other countries in the sample. Great Britain, Czechoslovakia and France show barely any correlation at any lag between these two variables. The contemporaneous correlations for Belgium, the Netherlands and Italy are positive and significant, but their subsequent lagged correlations oscillate in a pattern that negates the initial contemporaneous correlations. Overall, these correlation patterns provide preliminary support for the hypothesis that uncertainty, as reflected in our RV measures, was not likely pronounced enough in these countries to influence their inflation dynamics and certainly not to induce hyperinflation.17 To estimate more fully the effects that uncertainty had on the economic dynamics of European economies after WWI, and inflation in particular, we frame our analysis using the following regressions at a monthly frequency: $$\begin{eqnarray} {{\rm{\pi }}_{{\rm{kt}} + {\rm{j}}}} = && {{\rm{\alpha }}_{{\rm{kj}}}} + {{\rm{\beta }}_{1{\rm{kj}}}}\Delta {\rm{R}}{{\rm{V}}_{{\rm{kt}}}} + {{\rm{\beta }}_{2{\rm{kj}}}}\Delta {\rm{R}}{{\rm{V}}_{{\rm{kt}} - 1}} + {{\rm{\beta }}_{3{\rm{kj}}}}{{\rm{\pi }}_{{\rm{kt}} - 1}} + {{\rm{\beta }}_{4{\rm{kj}}}}\Delta {\rm{I}}{{\rm{P}}_{{\rm{kt}} - 1}} \\ && +\, {{\rm{\beta }}_{2{\rm{kj}}}}\Delta {\rm{Note}}{{\rm{s}}_{{\rm{kt}} - 1}} + {{\rm{\varepsilon }}_{{\rm{kt}} + {\rm{j}}}}, \end{eqnarray}$$(3) where Δ represents the log first difference of the series. For each country k in our sample at projection horizon j, we regress changes in prices (π) on its own one-period lag as well as one-period lags of industrial production (ΔIP), notes (or money) in circulation (ΔNotes), and changes in our RV variable, both lagged and contemporaneous. Sources and definitions of these variables are described in Online Appendix B.18 The ΔNotes variable has been highlighted as a way of gauging fiscal policy regimes (Frankel, 1980; Sargent, 1982; Webb, 1986). As government expenditures fluctuated due to volatile tax revenues, reparations negotiations and military actions, different countries were, to varying degrees, financing deficits through the issuance of debt obligations or by monetising them. The latter is clearly a precursor to rising inflation and loss of currency value. Thus, the growth of notes in circulation, particularly for the GAPH countries, is an important variable in our empirical framework. Following the identification assumption used by Leduc and Liu (2016), we include contemporaneous and lagged values of the RV series, such that the contemporaneous uncertainty variable does not respond to macroeconomic shocks in the current period but can influence the other series. The macroeconomic data were hand collected from government documents and League of Nations publications as described in Online Appendix B. We estimate these regressions to generate IRFs using the smooth local projection (SLP) methodology developed by Barnichon and Brownlees (2019). Estimation based on this methodology has a number of advantages for our data set relative to the local projections (LP) methodology developed by Jordà (2005). Using the LP method, IRFs are estimated using a sequence of predictive regressions of the variable of interest on its driving variables over different horizons; that is, to generate an IRF over the period from T to T + j, one would run J regressions with the same explanatory variables and a dependent variable whose time index increments from T to T + j and reducing our sample size. The IRF is then simply the sequence of regression coefficients on the structural shock parameter. The same logic holds for the estimated confidence bands. However, as noted by Ramey (2016), this method requires a large number of estimated parameters and has been found to be erratic under certain circumstances, especially for longer IRF horizons.19 In contrast, the SLP method imposes the restriction that an IRF be a smooth function across the horizon of interest. This restriction, as well as the shrinkage estimation technique employed, reduces the number of estimated parameters, and provides clearer inference between variables across the horizon of interest. Given that our country samples contain between 20 and 66 monthly observations, the SLP method provides overwhelming benefits in terms of parameter estimation, IRF generation and meaningful inference relative to other estimators. Figure 4 displays SLP-generated IRFs for the effect of a unit (or one-standard-deviation) increase in uncertainty on the monthly inflation rate for GAPH.20 The IRFs exhibit a common downward trend from above zero at the start of the projection horizon to below zero by the sixth month. This general pattern supports the hypothesis that an increase in uncertainty corresponds to an increase in inflation in the contemporaneous month and the month immediately following, although the effect declines to zero by the third month and either remains there or reverses in subsequent months. Importantly, the SLP-generated confidence bands (plotted at the 95% level) show that these positive relationships are statistically significant in GAPH for the contemporaneous month and the first month of the projection. Thus, our empirical results support the view that uncertainty shocks lead to an immediate, if relatively short-lived, increase in inflation. Of course, for some countries in our sample, such shocks occur with high frequency, so uncertainty can cumulate to have a large effect on the inflation rate. Fig. 4. Open in new tabDownload slide The Responsiveness of Inflation to Uncertainty—Hyperinflation Sample (Impulse Response Functions). Notes: The graphs show the IRFs of a one-standard-deviation change in ΔRVkt (the uncertainty measure) on inflation πkt+j based on SLP, and as described in the text. The grey lines represent the 95% confidence intervals around the SLP regression coefficients, which are based on standard errors estimated using the Newey–West methodology. Fig. 4. Open in new tabDownload slide The Responsiveness of Inflation to Uncertainty—Hyperinflation Sample (Impulse Response Functions). Notes: The graphs show the IRFs of a one-standard-deviation change in ΔRVkt (the uncertainty measure) on inflation πkt+j based on SLP, and as described in the text. The grey lines represent the 95% confidence intervals around the SLP regression coefficients, which are based on standard errors estimated using the Newey–West methodology. The magnitudes of the effects of uncertainty on contemporaneous inflation are summarised in Table 6. As shown in the first row, the standard deviation of log changes in the monthly German RV measure is 1.21, which is equivalent to an increase of 235% in RV value. The model-implied response of the monthly inflation for Germany is 0.139, which is an increase of 0.18 standard deviations or, equivalently, a 7.9 percentage point increase in inflation. Table 7 shows the effect for the first month of the project horizon. As shown in the first row, the model-implied response of inflation to the RV shock is 0.064. When combined with the same one-standard-deviation increase in Germany's RV measure in month t, that leads to an increase in the inflation rate in month t + 1 of 0.30 standard deviations, which corresponds to an increase of 3.3 percentage points in value. Thus, a one-standard-deviation increase in uncertainty translates into a cumulative 11.2 percentage-point increase in Germany's monthly inflation rate. As shown in Figure 4, after two months, the cumulative effect was zero, suggesting that uncertainty shocks effects on inflation are short-lived. Table 6. The Contemporaneous Effect of an Uncertainty Shock on Inflation. Country . RV SD (ln delta) . RV change (%) . Coefficient . Contemporaneous effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.139 0.168 0.65 7.90 Austria 1.32 272.7 0.107 0.140 0.61 17.10 Hungary 1.72 455.8 0.096 0.165 0.91 15.28 Poland 1.21 235.3 0.031 0.037 0.18 8.04 UK 1.17 223.4 −0.001 −0.001 −0.05 −0.50 Belgium 1.25 248.6 0.007 0.008 0.13 0.85 Czech. 1.88 556.5 −0.005 −0.010 −0.25 −0.80 France 1.07 190.8 0.000 0.000 0.01 0.00 Netherlands 1.15 214.7 0.005 0.005 0.06 1.13 Italy 1.01 174.6 0.003 0.003 0.04 1.08 Country . RV SD (ln delta) . RV change (%) . Coefficient . Contemporaneous effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.139 0.168 0.65 7.90 Austria 1.32 272.7 0.107 0.140 0.61 17.10 Hungary 1.72 455.8 0.096 0.165 0.91 15.28 Poland 1.21 235.3 0.031 0.037 0.18 8.04 UK 1.17 223.4 −0.001 −0.001 −0.05 −0.50 Belgium 1.25 248.6 0.007 0.008 0.13 0.85 Czech. 1.88 556.5 −0.005 −0.010 −0.25 −0.80 France 1.07 190.8 0.000 0.000 0.01 0.00 Netherlands 1.15 214.7 0.005 0.005 0.06 1.13 Italy 1.01 174.6 0.003 0.003 0.04 1.08 Notes: Estimates are based on SLP as described in the text. The table shows the effect of an uncertainty shock in country k (i.e., an innovation to Δln(RVkt) on the contemporaneous change in its inflation rate πkt using the SLP model of Equation (2). The first column presents the magnitude of a one-standard-deviation change in Δln(RVkt) based on each country's RV series over the sample periods defined in Table 4. The second column translates that increase into a percentage increase in the level of the RVkt series. The third column reports the SLP estimation coefficients of the contemporaneous change in Δln(RVkt) on πkt. The fourth column is the product of the first and third columns, which is the numerical effect of a one-standard-deviation increase of the contemporaneous change in Δln(RVkt) on πkt. The fifth column transforms the numerical effect into standard deviations of πkt. The last column presents the transformation of the model-implied effect on πkt into a percentage point increase in the level of the πkt series. Open in new tab Table 6. The Contemporaneous Effect of an Uncertainty Shock on Inflation. Country . RV SD (ln delta) . RV change (%) . Coefficient . Contemporaneous effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.139 0.168 0.65 7.90 Austria 1.32 272.7 0.107 0.140 0.61 17.10 Hungary 1.72 455.8 0.096 0.165 0.91 15.28 Poland 1.21 235.3 0.031 0.037 0.18 8.04 UK 1.17 223.4 −0.001 −0.001 −0.05 −0.50 Belgium 1.25 248.6 0.007 0.008 0.13 0.85 Czech. 1.88 556.5 −0.005 −0.010 −0.25 −0.80 France 1.07 190.8 0.000 0.000 0.01 0.00 Netherlands 1.15 214.7 0.005 0.005 0.06 1.13 Italy 1.01 174.6 0.003 0.003 0.04 1.08 Country . RV SD (ln delta) . RV change (%) . Coefficient . Contemporaneous effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.139 0.168 0.65 7.90 Austria 1.32 272.7 0.107 0.140 0.61 17.10 Hungary 1.72 455.8 0.096 0.165 0.91 15.28 Poland 1.21 235.3 0.031 0.037 0.18 8.04 UK 1.17 223.4 −0.001 −0.001 −0.05 −0.50 Belgium 1.25 248.6 0.007 0.008 0.13 0.85 Czech. 1.88 556.5 −0.005 −0.010 −0.25 −0.80 France 1.07 190.8 0.000 0.000 0.01 0.00 Netherlands 1.15 214.7 0.005 0.005 0.06 1.13 Italy 1.01 174.6 0.003 0.003 0.04 1.08 Notes: Estimates are based on SLP as described in the text. The table shows the effect of an uncertainty shock in country k (i.e., an innovation to Δln(RVkt) on the contemporaneous change in its inflation rate πkt using the SLP model of Equation (2). The first column presents the magnitude of a one-standard-deviation change in Δln(RVkt) based on each country's RV series over the sample periods defined in Table 4. The second column translates that increase into a percentage increase in the level of the RVkt series. The third column reports the SLP estimation coefficients of the contemporaneous change in Δln(RVkt) on πkt. The fourth column is the product of the first and third columns, which is the numerical effect of a one-standard-deviation increase of the contemporaneous change in Δln(RVkt) on πkt. The fifth column transforms the numerical effect into standard deviations of πkt. The last column presents the transformation of the model-implied effect on πkt into a percentage point increase in the level of the πkt series. Open in new tab Table 7. The Effect of an Uncertainty Shock on Inflation One Month Forward. Country . RV SD (ln delta) . RV change (%) . Coefficient . One-month effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.064 0.078 0.30 3.32 Austria 1.32 272.7 0.006 0.007 0.03 9.65 Hungary 1.72 455.8 0.071 0.122 0.67 15.28 Poland 1.21 235.3 0.053 0.064 0.31 4.71 UK 1.17 223.4 0.003 0.004 0.00 0.00 Belgium 1.25 248.6 −0.006 −0.008 −0.12 −0.40 Czech. 1.88 556.5 −0.002 −0.003 −0.08 −1.10 France 1.07 190.8 0.002 0.002 0.05 0.00 Netherlands 1.15 214.7 −0.008 −0.009 −0.10 0.04 Italy 1.01 174.6 −0.020 −0.020 −0.33 −0.97 Country . RV SD (ln delta) . RV change (%) . Coefficient . One-month effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.064 0.078 0.30 3.32 Austria 1.32 272.7 0.006 0.007 0.03 9.65 Hungary 1.72 455.8 0.071 0.122 0.67 15.28 Poland 1.21 235.3 0.053 0.064 0.31 4.71 UK 1.17 223.4 0.003 0.004 0.00 0.00 Belgium 1.25 248.6 −0.006 −0.008 −0.12 −0.40 Czech. 1.88 556.5 −0.002 −0.003 −0.08 −1.10 France 1.07 190.8 0.002 0.002 0.05 0.00 Netherlands 1.15 214.7 −0.008 −0.009 −0.10 0.04 Italy 1.01 174.6 −0.020 −0.020 −0.33 −0.97 Notes: Estimates are based on SLP as described in the text. The table shows the effect of an uncertainty shock in country k (i.e., an innovation to Δln(RVkt) on the contemporaneous change in its inflation rate πkt using the SLP model of Equation (2). The first column presents the magnitude of a one-standard-deviation change in Δln(RVkt) based on each country's RV series over the sample periods defined in Table 4. The second column translates that increase into a percentage increase in the level of the RVkt series. The third column reports the SLP estimation coefficients of the contemporaneous change in Δln(RVkt) on πkt. The fourth column is the product of the first and third columns, which is the numerical effect of a one-standard-deviation increase of the contemporaneous change in Δln(RVkt) on πkt. The fifth column transforms the numerical effect into standard deviations of πkt. The last column presents the transformation of the model-implied effect on πkt into a percentage point increase in the level of the πkt series. Open in new tab Table 7. The Effect of an Uncertainty Shock on Inflation One Month Forward. Country . RV SD (ln delta) . RV change (%) . Coefficient . One-month effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.064 0.078 0.30 3.32 Austria 1.32 272.7 0.006 0.007 0.03 9.65 Hungary 1.72 455.8 0.071 0.122 0.67 15.28 Poland 1.21 235.3 0.053 0.064 0.31 4.71 UK 1.17 223.4 0.003 0.004 0.00 0.00 Belgium 1.25 248.6 −0.006 −0.008 −0.12 −0.40 Czech. 1.88 556.5 −0.002 −0.003 −0.08 −1.10 France 1.07 190.8 0.002 0.002 0.05 0.00 Netherlands 1.15 214.7 −0.008 −0.009 −0.10 0.04 Italy 1.01 174.6 −0.020 −0.020 −0.33 −0.97 Country . RV SD (ln delta) . RV change (%) . Coefficient . One-month effect (ln delta) . Number of inflation SD . Inflation change (p.p.) . Germany 1.21 235.3 0.064 0.078 0.30 3.32 Austria 1.32 272.7 0.006 0.007 0.03 9.65 Hungary 1.72 455.8 0.071 0.122 0.67 15.28 Poland 1.21 235.3 0.053 0.064 0.31 4.71 UK 1.17 223.4 0.003 0.004 0.00 0.00 Belgium 1.25 248.6 −0.006 −0.008 −0.12 −0.40 Czech. 1.88 556.5 −0.002 −0.003 −0.08 −1.10 France 1.07 190.8 0.002 0.002 0.05 0.00 Netherlands 1.15 214.7 −0.008 −0.009 −0.10 0.04 Italy 1.01 174.6 −0.020 −0.020 −0.33 −0.97 Notes: Estimates are based on SLP as described in the text. The table shows the effect of an uncertainty shock in country k (i.e., an innovation to Δln(RVkt) on the contemporaneous change in its inflation rate πkt using the SLP model of Equation (2). The first column presents the magnitude of a one-standard-deviation change in Δln(RVkt) based on each country's RV series over the sample periods defined in Table 4. The second column translates that increase into a percentage increase in the level of the RVkt series. The third column reports the SLP estimation coefficients of the contemporaneous change in Δln(RVkt) on πkt. The fourth column is the product of the first and third columns, which is the numerical effect of a one-standard-deviation increase of the contemporaneous change in Δln(RVkt) on πkt. The fifth column transforms the numerical effect into standard deviations of πkt. The last column presents the transformation of the model-implied effect on πkt into a percentage point increase in the level of the πkt series. Open in new tab The SLP IRFs for Austria, Poland and Hungary exhibit similar patterns of positive and significant responses early in the projection horizon, and then decline to nearly zero by the third month. As Table 6 shows, the estimated contemporaneous coefficient for Austria means that a one-standard-deviation shock to uncertainty leads to an increase in the monthly inflation rate of 17 percentage points. If we then add the one-month-lagged effect, recorded in Table 7 as a 9.7 percentage point increase, the empirical results suggest that Austrian inflation would rise by 26.7 percentage points after a one-standard-deviation increase in uncertainty. The contemporaneous effects for one-standard-deviation increases in uncertainty for Poland and Hungary result in inflation rates that, respectively, are roughly 8 and 15 percentage points higher. Incorporating the one-month-lagged response, results in monthly inflation rates that are 12.8 and 30.3 percentage points higher, respectively. These results strongly suggest that uncertainty played an important role in determining inflation dynamics during the interwar period for countries that subsequently experienced hyperinflations. In contrast, Figure 5 shows the more limited role that uncertainty played in the inflation dynamics of the other European countries; i.e., the IRFs are essentially flat at or oscillating slightly near zero. The estimated responses are not statistically significantly different from zero at the 10% level in almost all cases. The magnitudes of these estimated responses are also notably smaller, ranging between −0.01 and +0.01 at the contemporaneous and one-month horizons. As shown in Tables 6 and 7, the marginal effects of an increase of a one-standard-deviation change in the RV measure on the inflation rate are quite small for the non-hyperinflation countries, with most effects less than a 1 percentage point change in absolute value. For example, in Great Britain and the Netherlands, the cumulative two-month effect of a one-standard-deviation change in their logged RV measures on monthly inflation is −0.5 and +1.2 percentage points, respectively, which is an order of magnitude smaller than for GAPH. Fig. 5. Open in new tabDownload slide The Responsiveness of Inflation to Uncertainty—Other European Countries (Impulse Response Functions). Notes: The graphs show the IRFs of a one-standard-deviation change in ΔRVkt (the uncertainty measure) on inflation πkt+j based on SLP, as described in the text. The grey lines represent the 95% confidence intervals around the SLP regression coefficients, which are based on standard errors estimated using the Newey–West methodology. Fig. 5. Open in new tabDownload slide The Responsiveness of Inflation to Uncertainty—Other European Countries (Impulse Response Functions). Notes: The graphs show the IRFs of a one-standard-deviation change in ΔRVkt (the uncertainty measure) on inflation πkt+j based on SLP, as described in the text. The grey lines represent the 95% confidence intervals around the SLP regression coefficients, which are based on standard errors estimated using the Newey–West methodology. Czechoslovakia's SLP results are particularly interesting to compare to GAPH. As Figure 5 shows, inflation in Czechoslovakia initially responds to increased uncertainty by declining slightly. The contemporaneous and one-period-ahead projections shown in Tables 6 and 7 are negative, though not both statistically significant, and cumulate to a decline of nearly two percentage points. So, why are the effects of uncertainty for Czechoslovakia's inflation rate different in sign and size from GAPH when its RV was roughly equal to Germany's and Hungary's in 1920? After all, it too had border disputes, imperial legacy debts and new institutions—all conditions that more than likely contributed to another newly created country's (i.e., Poland's) high measured uncertainty and eventual experience with hyperinflation. There are several likely reasons. Under the leadership of finance minister Alois Rasin, the country first committed to obtaining a stable currency in 1919, shortly after it gained its independence (Sargent, 1982). Czechoslovakia then eventually adopted a fixed exchange rate by going onto the gold standard. As emphasised by Kydland and Prescott (1977), a hard peg, like Czechoslovakia adopted, tied the hands of the monetary and fiscal authorities and committed the policymakers to following rules. This decision no doubt helped to reduce uncertainty about the future path of policy. Czechoslovakia's policymaking was then consistent with its hard peg: it implemented a capital levy to move the budget towards balance, as Germany did, but it was able to do so with less uncertainty over reparations that saddled Germany. Finally, it faced border disputes, just as Poland did, but it did not engage in a prolonged war. Rather, it was able to neutralise its most serious external threat, Hungary, by entering into the Little Entente in August 1922—a date consistent with when we see the Czechoslovakia's RV began to decline. 4. Conclusion The physical and political destruction wrought by WWI traumatised Europe and had economic reverberations long after the official peace treaties were signed. Both the victors and the vanquished faced massive costs of reconstruction, uncertainty regarding political borders and tax revenues, and extended negotiations over war reparations. Europe thus faced an environment of heightened uncertainty that deeply affected economic expectations and outcomes. We develop a methodology utilising country-specific events that shows uncertainty to be an important driver of the inflation dynamics for GAPH, but much less so for other European countries. We show how uncertainty over reparations increased policy uncertainty in both France (as a receiver of reparations) and Germany (as a payer). However, the effects of uncertainty were significantly greater for Germany and contributed to the country's inflationary spiral that led to hyperinflation (Webb, 1986). For our analysis, we examine the relationship between uncertainty and inflation dynamics within an economy. Our results show that in a period of heightened general uncertainty, such as the interwar period, higher relative levels of uncertainty were significant drivers of inflation dynamics for the countries that subsequently experienced hyperinflation. In contrast, lower relative levels of uncertainty in other European countries did not influence their inflation dynamics. These joint empirical results align well with recent work by Vavra (2014), who shows how heightened economic uncertainty differentially influences the price-setting mechanism, monetary policy and output growth. The results are also in line with the finding by Danielsson et al. (2018)—that periods of unusually high and low uncertainty are more meaningful indicators of financial stability (i.e., banking crises in their work) than just the level value of uncertainty itself. Our results support the view that heightened economic uncertainty—both in relative terms within a country and across countries—contributed to the macroeconomic and price dynamics that led to rising levels of inflation in the countries that tipped into hyperinflation. Finally, our findings relate to the work of Cole and Kehoe (1996; 2000) and subsequent by othe scholars showing that financial crises that threaten a country's ability to repay its debt may quickly become self-fulfilling, and lead to a debt crisis. The parallel to fiscal policy and inflation expectations in the interwar period is clear. If economic agents believed that GAPH debts could not be repaid as long as budget discipline was non-existent, then tipping into hyperinflation becomes self-fulfilling. Further analysis of the fiscal conditions of these countries at the monthly level and their relationships with our uncertainty measure should be a fruitful direction for future work. Additional Supporting Information may be found in the online version of this article: Online Appendix Replication Package Footnotes 1 Bogart (1920) estimates per capita net direct war costs equal to $766 for Great Britain, $613 for France, $343 for Italy, $557 for Germany and $352 for Austro-Hungary. To put these figures in perspective, Great Britain's 1913 GDP per capita was $225.92. (Great Britain's 1913 nominal GDP was £2.122 billion, its population was 45.649 million, and the prewar parity was $4.86 per pound sterling). Broadberry and Harrison (2005, table 1.5) use an alternative measure of war costs (total government spending as a share of GDP) for a smaller number of belligerents, and show that these figures are roughly comparable at the end of the war for France (53.5%) and Germany (50.1%), while somewhat smaller for Great Britain (35.1%). Further, using war deaths as a percentage of the population, Broadberry and Harrison (2005, table 1.10) estimate that relative to prewar stocks, human capital declined by 7.2% in France and 6.3% in Germany. 2 Hyperinflations are relatively rare. Hanke and Krus (2013) report 56 hyperinflations over the period 1920–2008. 3 According to IMF statistics, in 2018, Venezuela and Argentina respectively had inflation rates exceeding 1,000% and 30%. Hanke and Bushnell (2017) propose Venezuela as the 57th episode of hyperinflation since 1920. 4 Jurado et al. (2015) distinguish between different types and measures of uncertainty; namely, measures reflecting financial market uncertainty—such as Bloom (2009) and our proposed measure—and measures reflecting broader macroeconomic uncertainty. Ludvigson et al. (forthcoming) also distinguish between different measures of uncertainty (financial and real) employing the methodology developed in Jurado et al. (2015). Questions of how these different measures influence macroeconomic dynamics remain a topic of debate; see Ludvigson et al. (forthcoming) and Carriero et al. (2018) for opposing empirical conclusions. 5 In particular, Sargent (1982, p. 75) states that ‘[f]rom the viewpoint that the value of a state's currency and other debt depends intimately on the fiscal policy it intends to run, the uncertainty about the reparations owed by the German government necessarily cast a long shadow over its prospects for a stable currency’. 6 For example, Leduc and Liu (2016) found that for post-war US data, an increase in uncertainty leads to an increase in unemployment and a decline in inflation; see also Caggiano et al. (2014) and Gilchrist et al. (2014). Vavra (2014) provides a model and empirical results linking uncertainty shocks to inflation and output growth. Bianchi and Melosi (2017) show that economic uncertainty linked to fiscal policy contributed to US inflation dynamics during the Great Recession. Basu and Bundick (2017) trace out the effects of uncertainty shocks on macroeconomic variables under different pricing regimes. 7 Similarly, Japan, Italy and Germany, all losers in WWII, did not experience hyperinflation. 8 Limited monthly data for Russia prevented us from including it in our analysis. 9 The constitution creating its new post-war form was signed on 11 August 1919. 10 Baillie et al. (1993) examine weekly exchange-rate volatility during this period based on data from Einzig (1937), which was sourced from the Commercial and Financial Chronicle. Peel and Spira (2015) examine the consistency of these two data sources. Concerns regarding the potential staleness of these spot prices were examined by assessing the percentage of trading days with returns of zero. These calculations are reported in Online Appendix M. While further research is needed, our results suggest that the market was functioning well enough across our sample period to process news developments and register changes in RV uncertainty. 11 In his description of interwar hyperinflations, Sargent (1982) provides examples of purchases of foreign assets and capital controls, which suggest how important exchange rates were to the economic environment of the time. For further discussion of the German hyperinflation and exchange rates, see Frankel (1977; 1980), Salemi (1980) and Webb (1986). 12 See also the follow-up discussion in a comment by Salemi (1980) and a response by Frankel (1980). 13 Please note that RV calculations based on spot exchange rates denominated in British pounds were very similar. 14 We provide a table listing the top five RV events for the non-hyperinflation countries in Online Appendix A. 15 The hyperinflation reached a peak of a monthly inflation rate of 569% in October 1923 before the reichsmark was decommissioned and replaced in December 1923; see Webb (1986). Salemi (1980, p. 766) chose the same endpoint of June 1923 for his study based on judgement, supporting our intuition by stating that ‘a regression attempting to explain [German inflation] between February 1921 and August 1923 might do well by explaining [German inflation] for only the last few months’. 16 Sargent (1982) describes Hungary as experiencing hyperinflation by noting that its price index increased by a factor of 263 between January 1922 and April 1924. 17 The results of formal Granger-causality testing of the bivariate relationship between the countries’ uncertainty and inflation series support this inference and are available upon request. 18 We conducted a number of robustness tests using additional macroeconomic series, such as labour market series, and alternative uncertainty measures as control variables. Our main empirical findings remain qualitatively similar. We report the detailed results of these robustness tests in the appendices. 19 See Plagborg-Møller and Wolf (2019) for further discussion. 20 Note that the parameter standard errors are estimated using the Newey–West methodology. The estimated, country-level SLP coefficients are available upon request. Notes The data and codes for this paper are available on the Journal website. They were checked for their ability to reproduce the results presented in the paper. The views expressed here are those of the authors and not necessarily those of the Federal Reserve Bank of San Francisco or the Board of Governors of the Federal Reserve System. We thank Regis Barnichon, Steven Broadberry, Vasco Cúrdia, Thea Don-Siemion, Eric Fischer, Mark Harrison, Kilian Rieder, Kirsten Wandschneider, Òscar Jordà, Sylvain Leduc, Zheng Liu, Max-Stephan Schulze and conference participants at the 2018 SITE conference on the “Macroeconomics of Uncertainty and Volatility”, 2017 CEPR Economic History Symposium, the Eighth World Congress of Cliometrics, the Federal Reserve Bank of Richmond, and the University of California, Riverside for helpful comments and suggestions. We also thank the editor and referees for very helpful suggestions. We especially thank Regis Barnichon for sharing his smoothed local projection code with us, and Thilo Albers and David Chambers for sharing data. Tesia Chuderewicz, Neil Gerstein, William Hedberg, Erin Klein, Kevin Pearson and Javier Quintero provided invaluable research assistance. References Andersen T. , Bollerslev T., Diebold F.X., Labys P. ( 2003 ). ‘ Modeling and forecasting realized volatility ’, Econometrica , vol. 71 , pp. 579 – 625 . Google Scholar Crossref Search ADS WorldCat Baker S.R. , Bloom N., ( 2013 ). ‘ Does uncertainty reduce growth? Using disasters as natural experiments ’, NBER Working Paper 19475 . OpenURL Placeholder Text WorldCat Baker S.R. , Bloom N., Davis S.J. ( 2016 ). ‘ Measuring economic policy uncertainty ’, Quarterly Journal of Economics , vol. 131 , pp. 1593 – 636 . Google Scholar Crossref Search ADS WorldCat Balderston T. ( 1989 ). ‘ War finance and inflation in Britain and Germany, 1914–1918 ’, Economic History Review , vol. 42 ( 2 ), pp. 222 – 44 . Google Scholar Crossref Search ADS WorldCat Baillie R. , Bollerslev T., Redfrearn M.R. ( 1993 ). ‘ Bear squeezes, volatility spillovers and speculative attacks in the hyperinflation 1920s foreign exchange ’, Journal of International Money and Finance , vol. 12 , pp. 511 – 21 . Google Scholar Crossref Search ADS WorldCat Barnichon R. , Brownlees C. ( 2019 ). ‘ Impulse response estimation by smooth local projections ’, Review of Economic and Statistics , vol. 101 , pp. 522 – 30 . Google Scholar Crossref Search ADS WorldCat Basu S. , Bundick B. ( 2017 ). ‘ Uncertainty shocks in a model of effective demand ’, Econometrica , vol. 85 ( 3 ), pp. 937 – 58 . Google Scholar Crossref Search ADS WorldCat Bianchi F. , Melosi L. ( 2017 ). ‘ Escaping the great recession ’, American Economic Review , vol. 107 , pp. 1030 – 58 . Google Scholar Crossref Search ADS WorldCat Bloom N. ( 2009 ). ‘ The impact of uncertainty shocks ’, Econometrica , vol. 77 ( 3 ), pp. 623 – 85 . Google Scholar Crossref Search ADS WorldCat Bloom N. ( 2014 ). ‘ Fluctuations in uncertainty ’, Journal of Economic Perspectives , vol. 28 , pp. 153 – 76 . Google Scholar Crossref Search ADS WorldCat Boemeke M.F. , Feldman G.D., Glaser E. ( 1998 ). Versailles: A Reassessment after 75 Years , Cambridge : Cambridge University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Bogart E.L. ( 1920 ). Direct and Indirect Costs of the Great World War , New York : Oxford University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Broadberry S. , Harrison M. ( 2005 ). ‘ The economics of World War I: an overview ’, in ( Broadberry S., Harrison M., eds.), The Economics of World War I , Cambridge : Cambridge University Press , pp. 3 – 40 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Cagan P. ( 1956 ). ‘ The monetary dynamics of hyperinflation ’, in ( Friedman M., ed.), Studies in the Quantity Theory of Money , pp. 25 – 117 ., Chicago : University of Chicago Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Caggiano G. , Castelnuovo E., Groshenny N. ( 2014 ). ‘ Uncertainty shocks and unemployment dynamics in U.S. recessions ’, Journal of Monetary Economics , vol. 67 , pp. 78 – 92 . Google Scholar Crossref Search ADS WorldCat Carrière-Swallow Y. , Céspedes L.F. ( 2013 ). ‘ The impact of uncertainty shocks in emerging countries ’, Journal of International Economics , vol. 90 , pp. 316 – 25 . Google Scholar Crossref Search ADS WorldCat Carriero A. , Clark T.E., Marcellino M. ( 2018 ). ‘ Endogenous uncertainty ’, Federal Reserve Bank of Cleveland Working Paper 2018-05 . OpenURL Placeholder Text WorldCat Cole H.L. , Kehoe T.J. ( 1996 ). ‘ A self-fulfilling model of Mexico's 1994–1995 debt crisis ’, Journal of International Economics , vol. 41 , pp. 309 – 30 . Google Scholar Crossref Search ADS WorldCat Cole H.L. , Kehoe T.J. ( 2000 ). ‘ Self-fulfilling debt crises ’, The Review of Economic Studies , vol. 67 ( 1 ), pp. 91 – 116 . Google Scholar Crossref Search ADS WorldCat Commercial and Financial Chronicle . ( Various years ). various issues , vol. 109–121 . Danielsson J. , Valenzuela M., Zer I. ( 2018 ). ‘ Learning from history: volatility and financial crises ’, Review of Financial Studies , vol. 31 , pp. 2774 – 805 . Google Scholar Crossref Search ADS WorldCat Davig T. , Leeper E.M. ( 2011 ). ‘ Monetary-fiscal policy interactions and fiscal stimulus ’, European Economic Review , vol. 55 , pp. 211 – 27 . Google Scholar Crossref Search ADS WorldCat Dornbusch R. ( 1982 ). ‘ Monetary problems of post-communism: lessons from the end of the Austro-Hungarian Empire ’, Weltwirtschaftliches Archiv , vol. 128 , pp. 391 – 424 . Google Scholar Crossref Search ADS WorldCat Dornbusch R. ( 1985 ). ‘ Stopping hyperinflation: lessons from the German inflation experience of the 1920s ’, NBER Working Paper 1675 . OpenURL Placeholder Text WorldCat Dornbusch R. ( 1987 ). ‘ Lessons from the German inflation experience of the 1920s ’, in ( Dornbusch R. et al. , eds.), Macroeconomics and Finance: Essays in Honor of Franco Modigliani , Cambridge : Cambridge University Press , pp. 337 – 66 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Dumberry P. ( 2007 ). State Succession to International Responsibility , Netherlands : Brill Publishing . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Durand E D. ( 1922 ). ‘ Public finance of Poland ’, United States Department of Commerce, Supplement to Commerce Reports, Trade Information Bulletin No. 32, Washington, DC . OpenURL Placeholder Text WorldCat Eichengreen B. ( 1992 ). Golden Fetters: The Gold Standard and the Great Depression , New York : Oxford University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Einzig P. ( 1937 ). The Theory of Forward Exchange , London : Macmillan . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Einzig P. ( 1962 ). The History of Foreign Exchange , London : Macmillan . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Feldman G. ( 1993 ). The Great Disorder: Politics, Economics, and Society in the German Inflation, 1914–1924 , New York : Oxford University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ferguson A. ( 1975 ). When Money Dies: The Nightmare of the Weimar Collapse , London : William Kimber . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Fernández-Villaverde J. , Guerrón-Quintana P., Kuester K., Rubio-Ramírez J., ( 2015 ). ‘ Fiscal volatility shocks and economic activity ’, American Economic Review , vol. 105 , pp. 3352 – 84 . Google Scholar Crossref Search ADS WorldCat Frankel J.A. , ( 1977 ). ‘ The forward exchange rate, expectations, and the demand for money: the German hyperinflation ’, American Economic Review , vol. 67 , pp. 653 – 70 . Google Scholar OpenURL Placeholder Text WorldCat Frankel J.A. ( 1980 ). ‘ The forward exchange rate, expectations, and the demand for money: the German hyperinflation: reply ’, American Economic Review , vol. 70 , pp. 771 – 5 . Google Scholar OpenURL Placeholder Text WorldCat Fraser P. , Taylor M.P. ( 1990 ). ‘ Modeling risk in the interwar foreign exchange market ’, Scottish Journal of Political Economy , vol. 37 , pp. 241 – 58 . Google Scholar Crossref Search ADS WorldCat Gilchrist S. , Sim J.W., Zakrajšek E. ( 2014 ). ‘ Uncertainty, financial frictions, and investment dynamics ’, NBER Working Paper 20038 . OpenURL Placeholder Text WorldCat Hanke S.H. , Bushnell C. ( 2017 ). ‘ On measuring hyperinflation: Venezuela's episode ’, World Economics , vol. 18 ( 3 ), pp. 1 – 18 . Google Scholar OpenURL Placeholder Text WorldCat Hanke S.H. , Krus N. ( 2013 ). ‘ World hyperinflations ’, in ( Whaples R., Parker R., eds.), Routledge Handbook of Major Events in Economic History , pp. 367 – 77 ., New York : Routledge . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Jordà Ò. ( 2005 ). ‘ Estimation and inference of impulse responses by local projections ’, American Economic Review , vol. 95 ( 1 ), pp. 161 – 82 . Google Scholar Crossref Search ADS WorldCat Jonas P. ( 2016 ). ‘ The economic consequences of Trianon ’, Hungarianhistory.com , http://www.hungarianhistory.com/lib/tria/tria40.htm . Google Scholar OpenURL Placeholder Text WorldCat Jurado K. , Ludvigson S.C., Ng S. ( 2015 ). ‘ Measuring uncertainty ’, American Economic Review , vol. 105 , pp. 1177 – 216 . Google Scholar Crossref Search ADS WorldCat Kydland F.E. , Prescott E.C. ( 1977 ). ‘ Rules rather than discretion: the inconsistency of optimal plans ’, Journal of Political Economy , vol. 85 , pp. 473 – 92 . Google Scholar Crossref Search ADS WorldCat League of Nations . ( 1937 ). Monetary Review, Money and Banking 1936/37 , vol. 1 , Geneva : League of Nations . Google Scholar League of Nations . ( Various years ). Monthly Bulletin of Statistics , Geneva : League of Nations . Google Scholar League of Nations . ( 1938 ). Report on Exchange Control , Geneva : League of Nations . Google Scholar League of Nations . ( Various years ). Yearbook , Geneva : League of Nations . Google Scholar Leduc S. , Liu Z. ( 2016 ). ‘ Uncertainty shocks are aggregate demand shocks ’, Journal of Monetary Economics , vol. 55 , pp. 1113 – 28 . Google Scholar OpenURL Placeholder Text WorldCat Leeper E.M. ( 1991 ). ‘ Equilibria under ‘Active’ and ‘Passive’ monetary and fiscal policies ’, Journal of Monetary Economics , vol. 27 , pp. 129 – 47 . Google Scholar Crossref Search ADS WorldCat Ludvigson S. , Ma S., Ng S. ( forthcoming ). ‘ Uncertainty and business cycles: exogenous impulse or endogenous response? ’, American Economic Journal: Macroeconomics . Google Scholar OpenURL Placeholder Text WorldCat McCallum B.T. , Nelson E. ( 2005 ). ‘ Monetary and fiscal theories of the price level: the irreconcilable differences ’, Oxford Review of Economic Policy , vol. 21 ( 4 ), pp. 565 – 83 . Google Scholar Crossref Search ADS WorldCat Marks S. ( 1978 ). ‘ The myths of reparations ’, Central European History , vol. 11 ( 3 ), pp. 231 – 55 . Google Scholar Crossref Search ADS WorldCat Monthly Labor Review . ( 1919 ). ‘ Prices and cost of living ’, Monthly Labor Review , vol. 9 ( 2 ), pp. 87 – 126 . OpenURL Placeholder Text WorldCat Obstfeld M. , Rogoff K. ( 1983 ). ‘ Speculative hyperinflations in maximizing models: can we rule them out? ’, Journal of Political Economy , vol. 91 ( 4 ), pp. 675 – 87 . Google Scholar Crossref Search ADS WorldCat Paredes J. , Pérez J.J., Perez-Quiros G. ( 2015 ). ‘ Fiscal targets: a guide to forecasters? ’, European Central Bank Working Paper 1834 . OpenURL Placeholder Text WorldCat Peel D.A. , Spiru A.M. ( 2015 ). ‘ Consistency of two major data sources for exchange rates in the interwar period and evidence on the behavior of exchange rates during hyperinflation: the case of Germany ’, Working Paper, Lancaster University Management School . OpenURL Placeholder Text WorldCat Plagborg-Møller M. , Wolf C.K. ( 2019 ). ‘ Local projections and VARs estimate the same impulse responses ’, Working Paper, Princeton University . OpenURL Placeholder Text WorldCat Prati A. ( 1991 ). ‘ Poincaré's stabilization: stopping a run on government debt ’, Journal of Monetary Economics , vol. 27 (2), pp. 213 – 39 . Google Scholar Crossref Search ADS WorldCat Prochnik E.L.G. ( 1922 ). ‘ Possibilities of the economic and financial reconstruction of Austria ’, Annals of the American Academy of Political and Social Science , vol. 102 ( July ), pp. 20 – 6 . Google Scholar Crossref Search ADS WorldCat Ramey V. ( 2016 ). ‘ Macroeconomic shocks and their propagation ’, Handbook of Macroeconomics , vol. 2(A) , pp. 71 – 162 . Google Scholar Crossref Search ADS WorldCat Reinhart C.M. , Rogoff K. ( 2009 ). This Time Is Different , Princeton : Princeton University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ritschl A. , Straumann T. ( 2010 ). ‘ Business cycles and economic policy, 1914–1945 ’, in ( Broadberry S., O'Rourke K.H., eds.), The Cambridge Economic History of Modern Europe, Volume 2, 1870 to the Present , pp. 156 – 80 ., Cambridge University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Ritschl A. ( 2012 ). ‘ The German transfer problem, 1920–33: a sovereign debt perspective ’, European Review of History , vol. 19 ( 6 ), pp. 943 – 64 . Google Scholar Crossref Search ADS WorldCat Ritschl A. ( 2013 ). ‘ Reparations, deficits, and debt default: the Great Depression in Germany ’, in ( Crafts N., Fearon P., eds.), The Great Depression of the 1930s: Lessons for Today , Oxford : Oxford University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Salemi M.K. ( 1980 ). ‘ The forward exchange rate, expectations, and the demand for money: the German hyperinflation: comment ’, American Economic Review , vol. 70 , pp. 763 – 70 . Google Scholar OpenURL Placeholder Text WorldCat Sargent T. ( 1982 ). ‘ The ends of four big inflations ’, in ( Hall R.E., ed.), Inflation: Causes and Effects , Chicago : NBER , pp. 41 – 98 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Schulze M.-S. ( 2005 ). ‘ Austria–Hungary's economy in World War I ’, in ( Broadberry S., Harrison M., eds.), The Economics of World War I , Cambridge : Cambridge University Press , pp. 77 – 111 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Statistisches Handbuch der Weltwirtschaft . ( Various years ). Verlag für Sozialpolitik, Wirtschaft und Statistik Paul Schmidt, Berlin . Google Scholar Vavra J. ( 2014 ). ‘ Inflation dynamics and time-varying volatility: new evidence and an S-s interpretation ’, Quarterly Journal of Economics , vol. 129 , pp. 215 – 58 . Google Scholar Crossref Search ADS WorldCat Webb S.B. ( 1986 ). ‘ Fiscal news and inflationary expectations in Germany after World War I ’, Journal of Economic History , vol. 46 , pp. 769 – 94 . Google Scholar Crossref Search ADS WorldCat Woodford M. ( 1995 ). ‘ Price level determinacy without control of a monetary aggregate ’, Carnegie-Rochester Conference Series on Public Policy , vol. 43 , pp. 1 – 46 . Google Scholar Crossref Search ADS WorldCat Young J.P. ( 1925 ). European Currency and Finance: Commission of the Gold and Silver Inquiry, United States Senate , vols. 1 and 2 , Washington, DC : Government Printing Office . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © 2020 Royal Economic Society. Published by Oxford University Press. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Uncertainty and Hyperinflation: European Inflation Dynamics after World War I
JF - The Economic Journal
DO - 10.1093/ej/ueaa067
DA - 2021-02-10
UR - https://www.deepdyve.com/lp/oxford-university-press/uncertainty-and-hyperinflation-european-inflation-dynamics-after-world-i8LgCUiVyL
SP - 450
EP - 475
VL - 131
IS - 633
DP - DeepDyve
ER -