## Abstract

I identify significant effects of devaluation risk on interest rates and output using US silver coinage policy news between 1878 and 1900 as clean shocks to exchange rate expectations. The Free Silver movement heightened fears the United States would abandon the gold standard and depreciate the dollar. Because Congress, rather than a central bank, set silver coinage policy, silver policy news was likely uncorrelated with economic shocks. Corporate bonds exposed to dollar devaluation returned an additional 1 percent relative to safer bonds when silver risk decreased. Additionally, increased silver coinage risk is associated with an economically significant fall in industrial production.

## I. Introduction

LARGE fluctuations in exchange rates, especially devaluations, have been shown to have real consequences for output (Gupta, Mishra, & Sahay, 2007). Additionally, many fixed exchange rate regimes are viewed as imperfectly credible, with markets pricing in the possibility of devaluations (Mitchener & Weidenmier, 2015; Schmukler & Serven, 2002). Does devaluation risk have real costs as well? Assessing the real effects of this devaluation risk in a modern setting is an empirical, challenge for a number of reasons. First, fluctuations in devaluation risk are often caused by shocks to other economic variables, such as output or asset prices, making it difficult to identify devaluation risk effects. Second, many changes in devaluation risk are quickly followed by actual exchange rate devaluations, again raising identification challenges.

I overcome these identification issues by exploiting the unique historical and institutional features of the US monetary system at the end of the nineteenth century to estimate the effects of devaluation risk on economic activity. Between 1878 and 1900, the United States was on a gold standard (i.e., the dollar was convertible to a fixed amount of gold at the Treasury), but a political coalition of farmers and miners pressed for the additional convertibility of dollars to a fixed amount of silver. The preferred policy of this Free Silver movement would have resulted in a 50% depreciation of the dollar against gold.1

This paper consists of two distinct but complementary analyses. I first identify silver coinage news shocks and explore their effects on corporate bond credit risk, a key component of private borrowing costs, using a high-frequency event study approach. These shocks are derived from the historical financial press, while the event study uses daily corporate bond holding period returns. I then aggregate my daily credit risk premium changes from silver policy shocks to the monthly level to examine how they affected monthly industrial production and the dollar-gold interest rate spread.2

This period is an excellent setting for measuring the real effects of currency risk. Political factors, rather than economic factors, drove devaluation risk, reducing the endogeneity of the shocks. I also use the narrative record to verify that no economic news occurred on silver policy news days to further alleviate endogeneity concerns. Additionally, the United States never abandoned the gold standard between 1878 and 1900, despite the persistent threat posed by Free Silver, so I do not have to separate the effects of currency risk from the effects of an actual currency crisis. Finally, many companies were exposed to exchange rate risk on their balance sheets because devaluation would have raised their real debt burdens. Seventy percent of corporate debt was payable in “gold coin” rather than dollars and was primarily issued by companies in the nontradable sector, creating a case of currency mismatch where firms' liabilities and income streams are in different currencies.

Silver coinage news differentially affected the price of corporate bonds according to their exposure to dollar devaluation. The effects were significantly larger after the panic of 1893, when observers questioned whether the Treasury's reserves could withstand a run on gold. I obtain these results using daily corporate bond price data from over 100 firms that I hand-collected. I calculated devaluation exposures using information from earnings reports and balance sheets that were available to investors at the time. Bonds with interest coverage ratios so low that they would have defaulted solely due to a change in the dollar-gold exchange rate under Free Silver returned an additional 1% relative to bonds with the highest interest coverage ratios in response to reduced silver coinage risk. For silver news occurring after the panic of 1893, this return differential rises to 1.5%. I use the cross section of returns on event days to show that these return differentials actually reflect the gold debt burden of these bonds rather than some other correlated factor.

In the second part of the paper, I find that greater silver coinage risk leads to a substantial increase in the interest rate differential between dollar and gold-denominated assets–which I call the currency risk premium, as it reflects expected changes in the dollar-gold exchange rate and a risk premium–relative to its mean, and this effect persists for several months. Industrial production also falls by a statistically significant amount due to higher silver coinage risk, reaching a trough at twelve months after the shock, according to estimates of monthly impulse response functions to silver news. Crucially for identification, changes in industrial production and the price level do not predict the timing of silver news shocks, so these effects may be interpreted as causal. Devaluation risk had real effects because it raised expected default costs, which reduced the supply of credit going to these firms; I show suggestive evidence for this mechanism.

My work addresses several issues in macroeconomics related to currency mismatch, exchange rate regimes, and the real effects of policy uncertainty. I show that not only currency crises but also devaluation risks (without any change in policies) produce contractions in economic activity (previous cross-country studies on the output effects of currency crises include Gupta et al., 2007; Cerra & Saxena, 2008; and Bordo, Cavallo, & Meissner, 2010.) In this regard, I expand on other work studying the effect of broader political and economic uncertainty on aggregate output and firm outcomes, such as Baker, Bloom, and Davis (2016) and Ludvigson, Ma, and Ng (forthcoming); however, they discuss general policy or economic uncertainty, whereas I study a specific form of risk. Additionally, I provide evidence that exchange rate expectations matter for corporate borrowing costs due to currency mismatch on firm balance sheets: to the best of my knowledge, no previous work has argued that the balance sheet channel transmits nominal devaluation risk to corporate investment decisions.3

Identifying the balance sheet channel faces numerous challenges, and evidence in modern settings has proven mixed.4 Foreign currency assets, including derivatives, as well as “natural hedges” like revenues from exports, are hard to observe, making it difficult to control for the asset side of the balance sheet. Additionally, the firms studied are often the largest and the safest, as these are the ones that typically issue foreign currency bonds or syndicated loans. These are not problems in the time period I study. Rarely did anyone other than trade merchants and bankers hedge currency risk at this time. Further, the main companies issuing bonds during this time period were railroads that did not have explicitly gold-denominated revenues. Finally, while the railroads were relatively large companies, they were not always safe investments: 21% of the railroads in my sample defaulted during the panic of 1893.

My results also have implications for historical work on silver coinage in the United States and the broader impact of gold standard expectations on economic activity. A recent paper by Fulford and Schwartzman (2020; hereafter FS) also builds on previous work by economic historians (Friedman & Schwartz, 1963; Calomiris, 1993; Hallwood, Macdonald, & Marsh, 2000), arguing that silver coinage created expectations that the United States would leave the gold standard to show that this indeed harmed the real economy through adjustments in financial intermediation. Although this paper arrives at the same conclusions, it importantly emphasizes an additional transmission channel not discussed in FS (2020): currency mismatch on firm balance sheets.5 Further, whereas FS (2020) primarily study the cross-sectional response to a single event–the election of 1896–I study the economic response to multiple events where the probability of devaluation changed.6 Previous work linking expectations about the gold standard to output has focused on the Great Depression, but due to the bevy of policy changes during this time period, it is difficult to systematically study how gold standard uncertainty contributed to output fluctuations in the United States.7

The rest of the paper is organized as follows. Section II reviews the monetary institutions in the United States during the latter half of the nineteenth century; section III describes the role of currency mismatch in linking devaluation risk, interest rates, and output; section IV discusses the methodology and results of the daily-level empirical analysis; section V does the same for the monthly impulse response functions; section VI describes narrative evidence on silver coinage and industrial production; section VII summarizes and outlines policy implications.

## II. US Monetary and Financial Institutions, 1878–1900

### A. The Gold Standard and Silver Coinage in the United States

Prior to the Civil War (1861–1865), the United States operated under a bimetallic system where paper currency could be exchanged for a fixed amount of either gold or silver at the US Treasury. Both metals were treated as money so long as the mint convertibility ratio approximated the market convertibility ratio. When these two ratios were not equal, the metal undervalued at the mint ceased to circulate as money and was used only for private purposes. After the suspension of metallic convertibility during the Civil War, the Coinage Act of 1873 restored the fixed dollar-gold exchange rate at its historical level of $20.67 per ounce of gold, and the Resumption Act of 1875 set the date at which convertibility would resume at January 1, 1879. The Coinage Act omitted mention of silver coinage, essentially demonetizing silver and pushing the United States to a pure gold standard. Silver regained some of its previous monetary status through two legislative acts that allowed a limited amount of currency to be convertible to silver. First, the Bland-Allison Act of 1878 required the Treasury to purchase between$2 and $4 million worth of silver bullion each month and convert it to currency. Second, the Sherman Silver Purchase Act of 1890 set a fixed weight (4.5 million ounces) of silver to be purchased at the market price and coined each month. At the time of its passage, the Bland-Allison Act's minimum monthly requirement would have added roughly 1.2% annually to the total money stock in 1879, ceteris paribus. Silver purchases under the Sherman Act equaled approximately$5 million a month, which would have increased the 1890 money stock by 1.44%, ceteris paribus.

The deflation required to return to prewar gold parity, as well as continued deflation under the gold standard, led a coalition of farmers and miners to push for a return to bimetallism at the antebellum mint ratio of 16 ounces of silver to 1 ounce of gold. They hoped the additional money created would raise the overall price level, easing their debt burden and boosting their exports by depreciating the dollar. The two silver purchase acts described above were compromise capitulations to the Free Silver movement. The controversial aspect of the bimetallism advocated by the Free Silver movement was the 16:1 mint ratio. Relative to the market price ratio of silver to gold between 1880 and 1896, this mint ratio would have overvalued silver; by the end of this time period, the market ratio was closer to 32:1. Gold would thus have ceased to circulate as money and the dollar would have been devalued by up to 50% relative to gold.

Even these limited amounts of silver coinage created doubts about US commitment to the gold standard, leading to gold outflows that negated the inflationary effect of the silver money injections. These fears of a gold standard exit were at their highest in the aftermath of the panic of 1893, as the gold drain had pushed the Treasury's gold reserves to historic lows. The business community largely blamed the Sherman Act for the devastating panic of 1893, where annual industrial production fell 15.3% between 1892 and 1894, and President Grover Cleveland signed a Sherman Act repeal into law in November 1893 (Friedman & Schwartz, 1963; Jalil, 2015). Although the election of 1896 is widely seen as the unofficial end of the silver threat, this era of limited silver convertibility ended for good with the passage of the Gold Standard Act of 1900 (Timberlake, 1978). This law established gold as the only metal for which dollars could be exchanged at the Treasury.

### B. Bond Markets and Financial Institutions

By the end of the nineteenth century, the United States had a burgeoning market for long-term corporate debt. As mentioned in section I, railroads were the main sector issuing traded bonds, but utility and industrial companies made significant inroads during the 1890s, when bond debt of the railroads averaged 40% of US gross national product (GNP).8 For comparison, total nonfinancial corporate bond debt was around 27% of US GNP in 2017. Most railroad bonds were mortgages against the companies' property, particularly the lines of track, and were denominated in gold rather than dollars. These bonds varied significantly in their liens on the property, ranging from first or second liens on the main line of the company to those that were junior to all other claims (often numerous) on the entire property. Additionally, other junior debt was unsecured or backed only by other issues of stocks and bonds. Most industrial and utilities debt was also not mortgaged against any property, making the safety of their bonds much more dependent on their earning capacity.

Financial institutions were both directly and indirectly involved in the trading of corporate securities. Their direct role as investors of bonds and stocks was small relative to the size of the market (they only owned about 3.5% of all corporate securities), in part due to federal and state regulations on what securities they could own. Rather, the greater importance of the financial sector was in financing the purchase of stocks and bonds on credit, particularly for Wall Street traders. By 1910, one-third of all national and state bank loans was issued against stock and bond collateral (Pratt, 1912). If banks called in these loans and the borrower was unable to pay, the banks became owners of the securities, free to buy and sell these assets on the stock exchanges.

## III. Devaluation Risk and the Real Economy: The Role of Currency Mismatch

Nominal devaluation risk can affect the real economy through currency mismatch on borrowers' balance sheets (e.g., Cespedes, Chang, & Velasco, 2004). With liabilities denominated in a foreign currency but assets and income in the domestic currency, borrowers see increased real debt burdens after devaluations, raising the likelihood of default. Since asset prices incorporate expectations about all future values of the exchange rate, an increase in the probability of devaluation in the future should immediately increase borrowing costs for firms borrowing in foreign currency. Firms producing nontradable goods and services suffer the most because they do not receive the main benefit of devaluation: increased export competitiveness. As firms go out of business or lower their investment due to higher interest rates, this lowers production.

As suggested previously, currency mismatch was a common problem during the late nineteenth century. A substantial fraction of corporate debt was payable in gold rather than dollars, with over 90% issued by companies producing nontraded goods and services (particularly railroads and utilities). Depending on the year, 65% to 70% of the corporate debt in my bond data set had interest or principal (or both) denominated in gold. These “gold clauses” in bond covenants were often necessary for bonds to be traded on the London Stock Exchange. Most companies did not hold many gold-denominated assets at the time.

## IV. Daily-Level Event Study

The previous section argues that devaluation risk should affect firm borrowing costs. My empirical work first uses a daily-frequency event study to analyze the corporate bond market reaction to silver news. My primary regression compares holding period returns (HPR) between groups of bonds with different exposure to the effects of silver coinage. I later use these daily events as plausibly exogenous shocks to estimate monthly impulse response functions for the currency risk premium and industrial production. I describe the methodology used to estimate these impulse response functions at the beginning of section V.

To construct a series of events related to silver coinage policy, I use the narrative approach popularized by Romer and Romer (2004) to study modern US monetary policy. (For a detailed description about the event selection procedure see the online appendix.) I use only information available and known at the time when finding these events. To that end, I look for mentions of silver coinage policy news in the “Financial Affairs” section of the New York Times and the “Bankers Gazette” in the Commercial and Financial Chronicle between 1878 and 1900. Importantly, I drop events where there is other economic news released on the same day to avoid biasing my quantitative estimates.

The empirical analysis uses a set of 21 news shocks related to silver coinage. These events occur across 29 days, since the narrative record indicates that some news events affected financial markets across multiple days. Table 1 contains a brief description of each event. The symbols, (+) or (−), indicate whether the news appeared to increase (+) or decrease (−) expected future silver coinage.9 Events essentially fall into one of two categories: legislative action, such as the introduction of a bill to repeal the Sherman Act in December 1892, or executive branch positioning, like the election of the pro-gold William McKinley in November 1896. One recurring pattern in the events is for the Senate–where Free Silver supporters from less populous states held a greater portion of seats–to pass a Free Silver bill and for the House–where gold standard supporters from the more populous regions held the advantage–to then kill that measure in some way.10 There are several news items that are not included in this final set of events because the newspapers of the time explicitly indicate they did not alter silver coinage expectations. More details are provided in the online appendix, but this typically occurred when people discounted the possibility that the policy proposal would become law or when the president's veto was wholly expected.

Table 1.
Silver Policy News
 Date Description March 4, 1884 Juilliard v. Greenman legal tender case decision (⁠$+$⁠) February 27, 1885 Repeal of Bland-Allison voted down (⁠$+$⁠) December 9, 1885 President Cleveland calls for repeal of Bland-Allison in first message to Congress (⁠$-$⁠) December 22, 1885 Senator Beck delivers speech shooting down B-A repeal (⁠$+$⁠) June 9, 1890 Compromise silver purchase measure passes House (⁠$-$⁠) June 18, 1890 Senate passes free silver measure (⁠$+$⁠) July 8, 1890 New silver bill agreed on by Republican conferees of House and Senate (⁠$+$⁠) January 15, 1891 Free silver bill passes Senate (⁠$+$⁠) February 20, 1891 House Coinage committee votes against Senate silver bill (⁠$-$⁠) July 5, 1892 Free silver bill passes Senate (⁠$+$⁠) July 13–14, 1892 Free silver rejected in House (⁠$-$⁠) December 7, 1892 Introduction of Sherman Act repeal (⁠$-$⁠) February 9–10, 1893 House refuses to consider act repealing Sherman Act (⁠$+$⁠) June 30–July 1, 1893 President Cleveland orders emergency session of Congress to repeal Sherman Act in August (⁠$-$⁠) August 26, 28–29, 1893 House repeals Sherman Act by 2–1 majority (⁠$-$⁠) September 28, 1893 President Cleveland writes letter stating he will only accept unconditional repeal of Sherman Act (⁠$-$⁠) October 24, 1893 Compromise repeal fails to pass Senate (⁠$-$⁠) June 13, 15–16, 1896 Republicans announce campaign platform for gold standard (⁠$-$⁠) July 1, 1896 Free silver Democrats to control presidential nomination (⁠$+$⁠) August 13, 1896 William Jennings Bryan speech on Wall Str disappoints (⁠$-$⁠) November 2 & November 4, 1896 Election of Republican candidate William McKinley (⁠$-$⁠)
 Date Description March 4, 1884 Juilliard v. Greenman legal tender case decision (⁠$+$⁠) February 27, 1885 Repeal of Bland-Allison voted down (⁠$+$⁠) December 9, 1885 President Cleveland calls for repeal of Bland-Allison in first message to Congress (⁠$-$⁠) December 22, 1885 Senator Beck delivers speech shooting down B-A repeal (⁠$+$⁠) June 9, 1890 Compromise silver purchase measure passes House (⁠$-$⁠) June 18, 1890 Senate passes free silver measure (⁠$+$⁠) July 8, 1890 New silver bill agreed on by Republican conferees of House and Senate (⁠$+$⁠) January 15, 1891 Free silver bill passes Senate (⁠$+$⁠) February 20, 1891 House Coinage committee votes against Senate silver bill (⁠$-$⁠) July 5, 1892 Free silver bill passes Senate (⁠$+$⁠) July 13–14, 1892 Free silver rejected in House (⁠$-$⁠) December 7, 1892 Introduction of Sherman Act repeal (⁠$-$⁠) February 9–10, 1893 House refuses to consider act repealing Sherman Act (⁠$+$⁠) June 30–July 1, 1893 President Cleveland orders emergency session of Congress to repeal Sherman Act in August (⁠$-$⁠) August 26, 28–29, 1893 House repeals Sherman Act by 2–1 majority (⁠$-$⁠) September 28, 1893 President Cleveland writes letter stating he will only accept unconditional repeal of Sherman Act (⁠$-$⁠) October 24, 1893 Compromise repeal fails to pass Senate (⁠$-$⁠) June 13, 15–16, 1896 Republicans announce campaign platform for gold standard (⁠$-$⁠) July 1, 1896 Free silver Democrats to control presidential nomination (⁠$+$⁠) August 13, 1896 William Jennings Bryan speech on Wall Str disappoints (⁠$-$⁠) November 2 & November 4, 1896 Election of Republican candidate William McKinley (⁠$-$⁠)

### A. Preliminary Analysis

I establish whether silver coinage news is associated with greater bond price movements using two different methods. First, following Kuttner and Posen (2010), I test whether silver policy news contained additional information for bond HPR relative to days without news. Essentially, I test the null hypothesis that there was no additional variance on days with silver news. I begin by constructing bootstrap estimates of the 5th and 95th percentiles for the distribution of HPR on nonevent days. Because price volatility likely rose in the aftermath of the panic of 1893, I construct two sets of percentiles using only those trading days prior to May 1893 as one group and all nonevent days after the panic of 1893 as the other set. Nonevent days in my data set are one, five, and ten days before each event, as well as a six-month period in 1891 with no silver coinage news. I count the number of event dates that have average HPR that are either above the 95th percentile or below the 5th percentile estimated from the nonevent dates. I compare this number to a critical value from the binomial distribution. If the actual count exceeds the critical value, then I reject the null hypothesis that the variance is the same for the two groups.

My second approach uses a simple regression to test whether days with news about silver coinage correlates with greater changes in bond prices:
$Rt=α+βSilvert+xt'γ+ɛt,$
(1)
where $Rt$ is the average holding period return across corporate bonds traded on date $t$; $Silvert$ is a variable taking one of three values: 0 for days with no silver news, 1 on days where expected future silver coinage increases, and $-$1 on days when expected future silver coinage falls; $xt$ is a vector of month-year indicator variables, meant to control for the average level of bond yield volatility in a given month in a given year. Holding period returns are calculated as $(Pt+Ct)/(Pt-1)$, where $P$ denotes price and $C$ denotes coupon payment (if any). For most bond trades, there is no coupon payment, so I simply calculate HPR as the change in the natural logarithm of price. One could also use yield-to-maturity changes as a measure of the effect of silver coinage on borrowing costs, but many corporate bonds at the time had sinking funds or call provisions. This makes it difficult to accurately measure yield to maturity, so I follow Alquist and Chabot (2011) and use HPR instead.

Average HPR data are based on daily closing price data for corporate bonds traded on the New York Stock Exchange (NYSE), hand-collected from the New York Times and Wall Street Journal. For each event day, I record the closing price for each bond sold on the NYSE that day, as well as the last price at which each bond sold before the event date. I repeat this process for each of the nonevent days in my data set.

Two out of fifteen events prior to the panic of 1893 exceed the percentile bounds, while six of the fourteen events occurring after the panic surpass the bounds. The probability of two events exceeding the bounds in the pre-panic sample is 0.2699, while for the post-panic events, the probability of seeing six events outside of the percentile bounds is less than 0.00005. I therefore fail to reject that average HPR on pre-panic days with silver coinage news are more volatile than days without silver coinage news. Conversely, I easily reject this null hypothesis for the sample of events after the panic of 1893. I repeat the above process using absolute values of HPR as an additional test and again reject the null hypothesis for silver news only after the panic of 1893.

The effect of silver news clearly differs across time, and I argue that this time-varying effect reflects changes in the ability of the United States to maintain the gold standard; hence, silver coinage mattered for the economy insofar as it altered currency risk. At the time, the Treasury was legally required to hold a minimum of $100 million in gold reserves in order to continue issuing gold certificates that entitled holders to a fixed value of gold. Figure 1 shows that reserves fell below this threshold only after the panic of 1893, after which the magnitude of HPR on silver news days increases substantially.11 These low gold reserves induced the government to take the unprecedented step of issuing gold bonds in exchange for gold immediately in an effort to shore up gold reserves, showing the level of concern about maintaining gold convertibility after the Panic of 1893. Figure 1. Gold Reserves and Event Returns, 1884–1897 The vertical line marks the beginning of the panic of 1893. Figure 1. Gold Reserves and Event Returns, 1884–1897 The vertical line marks the beginning of the panic of 1893. Estimates of equation (1) confirm average HPR systematically changed by more on event days, but that this effect is confined to events after the panic of 1893. Column 1 of table 2 reports the coefficient on the silver event variable when the dependent variable is the average daily corporate bond HPR. This variable is negative and statistically significant at the 1% level. The estimate implies that news raising expected future silver coinage reduced corporate bond HPR 56 basis points on average. Testing for differential effects across time, column 2 of table 2, reports results when the regression includes an additional dummy that takes a value of 1 when a silver event occurs after the panic of 1893. As expected, only events after that panic produce a statistically significant effect of 92.5 basis points on bond HPR, with silver news before the panic having no effect on HPR. Table 2. Event Study: Daily Average Corporate Bond Holding Period Return  (1) (2) (3) Silver Event −0.558*** −0.088 −1.603*** (0.12) (0.08) (0.39) Post-Panic of 1893 Event −0.925*** (0.2) Treasury Gold Reserves 0.0291 (0.02) Event $×$ Gold Reserves 0.009*** (0.003) $N$ 233 233 233  (1) (2) (3) Silver Event −0.558*** −0.088 −1.603*** (0.12) (0.08) (0.39) Post-Panic of 1893 Event −0.925*** (0.2) Treasury Gold Reserves 0.0291 (0.02) Event $×$ Gold Reserves 0.009*** (0.003) $N$ 233 233 233 Results based on estimating equation (1). All specifications include month-year dummies. Heteroskedastic standard errors in parentheses. $*$$p<0.1$, $**$$p<0.05$, and $***$$p<0.01$. I then modify equation (1) to include the Treasury's gold reserves and its interaction with the silver event variable to test whether the ability to maintain gold convertibility mattered for the effect silver news on bond HPR. Column 3 of table 2 confirms this hypothesis, as the interaction term is indeed statistically significant and positive, indicating that higher gold reserves mute the impact of higher expected future silver coinage on bond HPR.12 Using the coefficient from column 3 implies that going from the gold reserves in July 1890 ($184.092 million), when the Sherman Act was passed, to the gold reserves when the Senate failed to pass a compromise repeal (ensuring a full repeal instead) of the Sherman Act in October 1893 ($84.385), increases the size of the bond market response to silver news by roughly 87 basis points. This is about 60% of the actual difference in average HPR for these two events. Additionally, I repeat these three specifications discussed above using the absolute value of HPR as the dependent variable and report results in the online appendix. I compare the magnitude of the coefficient to average HPR for two other uncertainty shocks that were the dominant news in their respective months in 1895 according to the Commercial and Financial Chronicle in order to gauge the economic magnitude of the silver news shocks. The first event was a series of financial market panics across Europe due to political unrest and the failure of a South African mining company on November 8. On this day, the average US corporate bond HPR was $-$49 basis points. The other event was a December 17 message from President Cleveland to Congress regarding a border dispute in South America between the US and the United Kingdom. Over the next two days, corporate bonds returned an average of $-$38 basis points per day. Considering the average HPR for silver news after the panic of 1893 was around 1%, I therefore interpret the effect of silver news on corporate bond yields as economically significant. ### B. Difference-in-Difference: Safe Versus Exposed Bonds To further2 establish plausibility that silver coinage news drives HPR on event days, I compare HPR between groups of bonds with different exposure to the effects of silver coinage on silver news days and nonevent days. In particular, bonds with smaller earnings cushions to absorb increases in the dollar value of their interest should have seen larger HPR in response to silver coinage news. As the expected gold debt burden increased, safe bonds with higher interest coverage ratios would see smaller increases in their probability of default relative to bonds with higher debt burdens, and thus lower HPR. The online appendix provides a simple credit risk model where this is true under plausible assumptions. Further, gold debt burdens and earnings coverage also concerned investors at the time, as the Wall Street Journal ran an article on August 3, 1896, explaining how a change in the gold debt burden due to dollar devaluation would affect one railway's profits. I calculate dollar devaluation exposure at the bond level using annual earnings and balance sheet data collected from Poor's Manual of Railroads as well as information on capital structure (i.e., seniority of different bond issues) extrapolated from the first edition of Moody's Manual (1909). Essentially, I start by collecting annual earnings data for each firm and interest payments for each bond in my sample. Then, for each firm, I determine the hierarchy of its bonds' claim to income, as well as whether the bond contained a gold clause. For each bond, I then determine whether, under a hypothetical switch to free silver, each bond would have enough available earnings to still meet its interest obligations after dollar devaluation. Available earnings are defined to be total net earnings less interest for more senior bonds. I use only the earnings data available for the reporting year prior to the event in order to best determine which bonds were most at risk of going into default at the time of the event. Using averages over several years in order to get a better sense of normal earnings for each railroad results in a set of exposed bonds that is a smaller subset of the sample used in the analysis. Exposed bonds are those that, holding everything but the dollar-gold exchange rate constant, would be pushed into default by a switch of the monetary regime to free silver.13 I then compare HPR for these exposed bonds to HPR for those that would have earned the highest credit rating using the statistics outlined by Moody (1909) when he assessed credit risk. This is a useful comparison group because these bonds had the highest interest coverage ratios and often did not contain gold clauses. (The online appendix contains further details on the calculation of bond-level exposure and the assignment of credit ratings.) Importantly, very few bonds are close to the threshold for being counted as exposed. I therefore conclude that making the exposure measure more or less restrictive fails to alter the results in any substantive way. All analysis that focuses on differential bond HPR across exposure categories uses only events in the 1890s as Poor's Manual of Railroads are unavailable for the early 1880s. Before discussing the quantitative analysis, I first present some rough evidence supporting this identification strategy. Figure 2 plots the median of the absolute value of the excess HPR of exposed corporate bonds over safe corporate bonds for ten, five, and one day before each silver event, as well as on the event days. The horizontal line at 45.75 basis points represents the median of the absolute value of the exposed bond excess HPR on all nonevent days in the sample. Excess HPR do not appear to be trending before the discovery of silver coinage news, and event-day excess HPR are the only ones above the nonevent median. Figure 2. Exposed Bond Excess Returns in Month before Silver Event Dots are the median absolute excess bond HPR of exposed bonds over safe bonds across corresponding days in the sample. The horizonal line is the median absolute value of excess HPR for exposed bonds across all nonevent days in the sample. Figure 2. Exposed Bond Excess Returns in Month before Silver Event Dots are the median absolute excess bond HPR of exposed bonds over safe bonds across corresponding days in the sample. The horizonal line is the median absolute value of excess HPR for exposed bonds across all nonevent days in the sample. My regression analysis compares HPR between two groups across time, thus following a standard difference-in-difference approach: $Ri,t=α+γExpi+β1Silvert+β2(Silvert×Expi)+xt'η1+(Speci×xt')η2+ɛit,$ (2) where the outcome variable, $Rit$, is the average holding period return for bonds in exposure group $i$ traded at date $t$.14 To construct this variable, I record the prices and potential coupon payments for all safe or exposed bonds sold on date $t$, record each bond's previous sale price, calculate the daily holding period return, and take the average across each group of bonds (exposed or safe).15 There were typically 25 to 30 safe bonds and 15 to 30 exposed bonds traded per day. The variable $Expi$ is a dummy that takes a value of 1 when the average return is for bonds most exposed to dollar devaluation based on the criteria discussed above. $Silvert$ takes one of three values: 1 on event days with news increasing expected silver coinage, $-$1 when the news lowers expected silver coinage, and 0 for nonevent days. The coefficient, $β2$, therefore captures the differential effect of silver news on exposed corporate bonds relative to safe corporate bonds. The value of $β2$ should be negative and statistically significant if silver coinage news drives HPR on event days. The variable $xt$ is a set of month-year fixed effects. Estimates of equation 2 in table 3 confirm that silver news caused exposed bond prices to change by more than safe prices on event days. Column 1 shows results for the full set of events, and the Event-Exposed interaction coefficients imply that news that increased silver coinage risk lowered exposed bond HPR by an additional 1% relative to safe bond HPR. Columns 2 and 3 then split the sample into only events before the panic of 1893 and only events after it. As shown in column 3, the differential effect on exposed bonds is greatest after the panic, averaging 1.6%, but there is still a small differential effect for events before the Panic based on column 2. Table 3. Event Study: Exposed versus Safe Corporate Bond Holding Period Return  All Events Pre-Panic Only Post-Panic Only Silver Event −0.172*** −0.078*** −0.262*** (0.05) (0.03) (0.08) Event $×$ Exposed −0.993*** −0.371*** −1.59*** (0.20) (0.12) (0.46) Exposed −0.16 −0.133 −0.142 (0.20) (0.20) (0.21) $N$ 462 434 440  All Events Pre-Panic Only Post-Panic Only Silver Event −0.172*** −0.078*** −0.262*** (0.05) (0.03) (0.08) Event $×$ Exposed −0.993*** −0.371*** −1.59*** (0.20) (0.12) (0.46) Exposed −0.16 −0.133 −0.142 (0.20) (0.20) (0.21) $N$ 462 434 440 Results based on estimating equation (2). All specifications include month-year dummies interacted with the exposed group dummy. Heteroskedastic-robust standard errors in parentheses. $*$$p<0.1$, $**$$p<0.05$, and $***$$p<0.01$. The economic significance of my estimated effects on return differentials is substantial. The 1895 nonsilver events discussed previously resulted in HPR differentials for exposed bonds of 89 and 71 basis points, similar to the average effect of silver news, though it is just half of the estimated effects of silver news after panic of 1893. My estimated effect is also sizable compared to the change in the spread between safe and speculative bonds after the United States abandoned the gold standard on April 19, 1933, an event that economic historians have highlighted as leading to rapid economic recovery (Hausman, Rhode, & Wieland, 2019; Jalil & Rua, 2016). I calculate the average HPR for ten industrial bonds, as well as the average HPR for ten high-grade railroad bonds used in the Dow Jones daily averages for these respective categories on April 19, 1933. Industrial bonds are chosen as the speculative group, reflecting the view at the time that these were riskier bonds. On April 19, 1933, these industrial bonds returned an additional 4% relative to high-grade railroad bonds. For events after the panic of 1893, the estimated average daily excess return due to silver coinage news is 1.6%, with several individual events seeing over 2.5% higher HPR for exposed bonds. Therefore, it would appear the additional return on exposed bonds to a change in the probability of a monetary regime switch is of a similar magnitude to the additional return from an actual monetary regime change. As an additional check, I perform a regression using only exposed bond HPR as the dependent variable and including safe bond HPR as a regressor: $ExpRett=α+βSilvert+γSafeRett+xt'δt+ɛt,$ (3) where $ExpRett$ is the average exposed bond HPR on date $t$; $Silvert$ is again the ${-1,0,1}$ variable corresponding to days with lower future silver coinage, no silver news, and increased future silver coinage, respectively; $SafeRett$ is the average safe bond HPR; and $xt$ is the same set of controls as in equation (2). The inclusion of SafeRet as a control variable is inspired by the market return model sometimes used to calculate expected returns in event studies. Here, I use the average safe bond HPR as a regressor to further rule out the possibility that exposed bonds always move by some factor relative to safe bonds. As shown in table 4, silver news remains an important determinant of exposed bond HPR when I estimate equation (3). The coefficient on silver coinage news is nearly identical in columns 1 to 4 to the silver event-exposed interaction term in columns 1 to 3 in table 3. Here the silver event coefficients imply that less silver coinage risk lowered exposed bond HPR an additional 1% to 1.6%. Regardless of how one measures the “expected” exposed return on a given day, silver news has a statistically and economically significant effect on exposed bond HPR. Table 4. Event Study: Exposed-Bond Holding Period Return  All Events Pre-Panic Events Post-Panic Events Silver Event −0.983*** −0.392*** −1.613*** (0.17) (0.12) (0.28) Safe Change 1.06** 0.728 0.925** (0.45) (0.49) (0.44) $N$ 231 217 220  All Events Pre-Panic Events Post-Panic Events Silver Event −0.983*** −0.392*** −1.613*** (0.17) (0.12) (0.28) Safe Change 1.06** 0.728 0.925** (0.45) (0.49) (0.44) $N$ 231 217 220 Dependent variable is the average daily holding period return for all exposed corporate bonds traded each day. Results based on estimating equation (3). All specifications include month-year dummies. Heteroskedastic-robust standard errors in parentheses. $*$$p<0.1$, $**$$p<0.05$, and $***$$p<0.01$. Appendix tables A3 and A4 report results for equations (2) and (3) using the absolute values of the dependent variables. This is done to help mitigate mean reversion concerns, particularly when month-year fixed effects are included in the regression. Prices may have been moving sharply in one direction prior to a silver event, and any news that would tend to move prices in the opposite direction ends up moving prices sharply simply due to reversion to the mean. Though the magnitudes fall by about 50%, they are still significantly different from 0 in every case. Further, as shown in online appendix figures A2 and A3, the monthly changes in yield to maturity spreads between safe and exposed bonds often match the implied changes in yields around silver events. ### C. Silver Coinage and Default Risk Premiums: Mechanisms The results above are consistent with the currency mismatch channel of transmission for nominal devaluation risk, but the exposure variable could be correlated with other bond attributes that matter for HPR on event days. In particular, because the exposed bonds have higher credit risk more generally and have different liquidity properties than safe bonds do, there are other channels through which devaluation risk could affect bond spreads. I thus focus on the cross-sections of HPR on event days and test which bond features affect the magnitude of HPR in response to silver news. I run the following regression, $Ri,t=α+β1ZeroReturni,t+β2EarningsDepreciationi,t+β3EarningsChangei,t+Eventt'γ+ɛi,t,$ (4) where $Ri,t$ is the total holding period return of bond $i$ for each event occurring after the panic of 1893, $t$16; ZeroReturn is the fraction of days in the 1891 nonevent sample bond that $i$ had a return of 0, EarningsDepreciation is the proportional change in the available earnings for bond $i$'s interest after a hypothetical dollar devaluation against gold; EarningsChange is the change in bond $i$'s available earnings from the year prior; and $Event$ is a vector of dummies for each of the events after the panic of 1893. When the bond is sold on an event where silver coinage risk decreases, the return is multiplied by $-$1, which allows for the pooling of “good news” and “bad news” events. (A more detailed description of these variables and their sources is available in the online appendix.) ZeroReturn is used as a proxy for a bond's liquidity, with a higher number of 0-return days suggesting lower bond liquidity. Hence, more liquid bonds would be more likely to have a non-0 return on event days. As a result, more liquid bonds should see more negative returns on average on event days, given the data transformation described above. $EarningsDepreciation$ captures the role of the gold debt burden and is one part of the calculation of the exposure indicator described above. Finally, $EarningsChange$ measures the responsiveness of HPR due to changes in available earnings in the months leading up to the silver events. Default risk for bonds may have risen for reasons outside of silver coinage risk in the months prior, and insofar as silver news changed the general economic outlook, these bonds would see greater HPR on event days. I find that both ZeroReturn and EarningsDepreciation are statistically significant, and both variables have the correct sign when entered individually in the regression (columns 1 to 3 in table 5). The same variables are also statistically significant when all regressors are included. Although both a bond's liquidity and its gold debt burden affect its HPR because of silver news, the gold debt channel is quantitatively more important for the exposed bond excess HPR over safe bonds. This is because, on average, exposed bonds only had slightly fewer 0-return days when compared to safe bonds. Going from the average value of ZeroReturn for exposed bonds (0.33) to the average value for safe bonds (0.41) implies a return that is 24 basis points lower during a silver event. In comparison, going from the average value of EarningsDepreciation for exposed bonds ($-$99.5%) to that for safe bonds ($-$6.9%) implies a silver event HPR that is 1.07% lower. Table 5. Post-Panic Event HPR and Bond Characteristics  (1) (2) (3) (4) Zero-return days 4.95** 2.98** (2.06) (1.37) Earnings after depreciation 1.04*** 1.16*** (0.22) (0.10) Change in earnings 2.06 0.71 (2.02) (2.19) $N$ 639 756 504 441 $R2$ 0.139 0.1875 0.0885 0.2801  (1) (2) (3) (4) Zero-return days 4.95** 2.98** (2.06) (1.37) Earnings after depreciation 1.04*** 1.16*** (0.22) (0.10) Change in earnings 2.06 0.71 (2.02) (2.19) $N$ 639 756 504 441 $R2$ 0.139 0.1875 0.0885 0.2801 Dependent variable is holding period return (in percent) of corporate bonds traded on event days multiplied by $-$1 on ($-$) event days (see table 1). Results based on estimating equation (4). All columns include event fixed effects. Standard errors clustered at the firm level in parentheses. $*$$p<0.1$, $**$$p<0.05$, and $***$$p<0.01$. Although the gold debt burden of a bond appears to be the driving feature of a bond's HPR on days with silver news, the above regression does not rule out that changes in investors' risk-bearing capacity from silver news led to increased demand for exposed bonds. In particular, as gold hoarding ended with declines in silver risk, this should have boosted funding liquidity, as banks had more funds to loan to speculators.17 If speculator funding constraints mattered, the money market rates should have seen large changes in the opposite direction of the exposed bond excess HPR. The top panel of figure 3 plots the change in the safe-speculative spread on post-panic of 1893 event days against the change in the average call loan rate on those days.18 While there are a few days with large declines in the call rate and high exposed bond excess HPR, there is generally no relationship between money market and bond market changes. Similarly, looking at the relationship between exposed bond excess HPR and changes in the amount of loans on New York City banks' balance sheets, as is done in the bottom panel of figure 3, also shows a lack of correlation between spread changes and financial conditions in a tight window around silver coinage events. Figure 3. Event Excess Returns and Money Market Changes Graphs plot the average exposed bond excess HPR over safe corporate bonds during the silver events after the panic of 1893 against changes in other financial conditions. (Top) Daily changes in the average call loan rate on the same day. (Bottom) Average cumulative excess HPR across post-panic of 193 event days within a given week against the corresponding change in loans outstanding of New York City Clearinghouse banks during same week. Figure 3. Event Excess Returns and Money Market Changes Graphs plot the average exposed bond excess HPR over safe corporate bonds during the silver events after the panic of 1893 against changes in other financial conditions. (Top) Daily changes in the average call loan rate on the same day. (Bottom) Average cumulative excess HPR across post-panic of 193 event days within a given week against the corresponding change in loans outstanding of New York City Clearinghouse banks during same week. Finally, the narrative record indicates there was discussion about how a dollar devaluation would affect firms' abilities to repay their gold-denominated debts. For example, after the nomination of William Jennings Bryan as the Democratic presidential candidate, the Wall Street Journal ran several articles discussing railways' gold debts and the impact of a potential free silver victory. In one, they discuss that the “danger of free silver” could be met by holding gold bonds only “if it were certain that roads could meet their interest in gold” and that if the dollar devalued too much relative to gold, “it would be impossible for corporations which have only a small surplus above fixed charges to meet their interest in gold.” This is exactly the characteristic the exposure dummy and the EarningsDepreciation variables attempt to capture. While one cannot draw definite conclusions from the evidence, it strongly suggests that exposed bond HPR on event days were due more to changes in default risk than risk premiums based on investor financing conditions. ## V. Monthly Impulse Response Functions The daily bond HPR data show that financial markets responded to silver coinage risk. In this section, I estimate impulse response4 functions with monthly data on exchange rate expectations and industrial production to explore whether these responses were justified. ### A. Methodology I investigate the impact of silver coinage risk on exchange rate expectations and industrial production using the local projection technique pioneered by Jordà (2005). Impulse response functions are computed through a series of OLS regressions at different forecast horizons. This approach offers more flexibility when compared to a traditional vector autoregression (VAR) technique in estimating the effect of the shock at later horizons. In this study, I compute impulse responses from 0 to 24 months after the initial shock, where the shock is the monthly HPR differential between exposed and safe bonds due to silver coinage news. For each horizon, $h∈[0,24]$, and outcome variable, $z$, I estimate the following regression,19 $zt+h=α+βhEvent_Excesst+∑k=16[ρkDollarRiskt-k+θkln(IndProdt-k)+ϕkln(PriceLevelt-k)]+ψt+ɛt+h,$ (5) where $z$ is either industrial production or the currency risk premium; Event_Excess is the geometric average return difference between exposed and safe bonds on silver coinage news days within a month; DollarRisk is the dollar-gold sixty-day interest spread; IndProd is the Miron-Romer adjusted index of industrial production (Romer, 1994); and PriceLevel is the general index of the overall price level (NBER Macrohistory series m04051). The dollar-gold interest spread measures the currency risk premium and expands Calomiris's (1993) series for 1893 to 1896 to cover 1878 to 1900.20 Lags of Event_Excess are not included because of their strong correlation with lags of the currency risk premium, as shown below. I then use the resulting estimates for the $βh$ to calculate the impulse response functions. Simply using narrative techniques does not guarantee the silver events are truly exogenous. In particular, silver events and the differential bond HPR resulting from silver news should not be a response to past macroeconomic developments. To ensure that my silver events are not driven by shocks to other variables, I test whether industrial production, the price level, and the currency risk premium Granger-cause my Event_Excess variable. Specifically, I look at whether this is true in a VAR specification with six lags of each of the four variables. I use first differences of industrial production and the price level in the VAR because I cannot reject the null hypothesis of a unit root in either case. The null hypothesis is that industrial production, the price level, and the currency risk premium do not Granger-cause Event_Excess. ### B. Impulse Response Function Estimates I begin by showing the raw time series for the currency risk premium and the exposed bond excess HPR on event days, as these form the basis for the currency risk impulse response, which verifies whether silver news affected dollar devaluation expectations. Figure 4 shows that these shocks correlate to large changes in the sixty-day dollar-gold interest spread. It is also apparent that the currency risk premium was fairly small over this entire time period, never greater than 2%. Dollar devaluation was likely perceived to be a tail-risk event: a low-probability but very large shock. Figure 4. Silver Coinage Shocks and the Currency Risk Premium The figure plots the time series of the sixty-day dollar-gold interest rate differential (based on commercial paper rate and implied gold interest rate from sight and sixty-day bills of exchange) and the average exposed bond excess HPR over safe corporate bonds due to silver policy news during a given month. Figure 4. Silver Coinage Shocks and the Currency Risk Premium The figure plots the time series of the sixty-day dollar-gold interest rate differential (based on commercial paper rate and implied gold interest rate from sight and sixty-day bills of exchange) and the average exposed bond excess HPR over safe corporate bonds due to silver policy news during a given month. The shock is always a negative 1% excess HPR of exposed corporate bonds over safe corporate bonds on silver news dates, corresponding to the estimated causal effect from equation (2). Since an increase in future silver coinage risk corresponded to a negative excess HPR, I refer to a negative realization of my shock as an increase in expected future silver coinage. The shaded areas around the impulse responses represent the 90% confidence intervals, estimated using Newey-West standard errors, following Jordà (2005). An increase in expected future silver coinage raises the dollar-gold interest spread by a statistically significant amount, as seen in the top left panel of figure 5. Within a month of an increase in silver risk, the currency risk premium rose by 11.34 basis points. Although the immediate effect on the currency risk premium is small and statistically insignificant, this estimate is driven by one event. Dropping this event from the sample results in a larger and statistically significant effect for silver news on the currency risk premium. Further, the increase in currency risk appears to be persistent, with peak effects of 15 basis points coming three and five months after the initial shock. This maximum increase is about 75% of the average currency risk premium over the entire sample period. The estimated impulse response function for this interest rate differential therefore supports the Friedman and Schwartz (1963) hypothesis that silver coinage in the United States raised expectations that the nation would abandon gold and depreciate the dollar. Figure 5. Impulse Response Functions In all panels, impulse is a $-$1% excess return of exposed bonds over safe bonds due to silver coinage news. (Top left) Based on estimating equation (5) for currency risk premium. (Top right) Estimates estimates equation (5) for industrial production. (Bottom left) Estimates equation (5) for three versions of the event series: the solid line is the original shocks series, the dashed line sets the summer of 1893 events to 0, while the dotted line drops the summer of 1893 events and the election of 1896 events in August and November of that year. (Bottom right) Regresses cumulative net gold inflows on silver news shock, six lags of net gold inflows, and month dummies added to correct for seasonality. Shaded areas are 90% confidence intervals constructed using Newey-West standard errors. Figure 5. Impulse Response Functions In all panels, impulse is a $-$1% excess return of exposed bonds over safe bonds due to silver coinage news. (Top left) Based on estimating equation (5) for currency risk premium. (Top right) Estimates estimates equation (5) for industrial production. (Bottom left) Estimates equation (5) for three versions of the event series: the solid line is the original shocks series, the dashed line sets the summer of 1893 events to 0, while the dotted line drops the summer of 1893 events and the election of 1896 events in August and November of that year. (Bottom right) Regresses cumulative net gold inflows on silver news shock, six lags of net gold inflows, and month dummies added to correct for seasonality. Shaded areas are 90% confidence intervals constructed using Newey-West standard errors. The top right panel of figure 5 displays the contractionary effect of increased expected silver coinage on industrial production. The negative effect of increasing the risk of future silver coinage on industrial production confirms the belief of the contemporary financial community about silver's impact. The impact reaches a trough at twelve months after the shock and stays negative and significantly different from 0 for several months after the trough. Further, the shape of the response mirrors that found in studies of modern monetary policy (Ramey, 2016). Industrial production drops 3.05% at the trough of twelve months and is still around 1% lower eighteen months after the silver coinage news. The trough effect is roughly 40% of the standard deviation of the twelve-month change in industrial production during this period. The estimated twelve-month effect also captures about 25% of the fall in production from January 1893 to January 1894 (the onset of the panic). Overall, these silver coinage news shocks were responsible for 25% to 30% of the variation in industrial production around the time of the trough effect, based on forecast error variance decompositions using the Gorodnichenko and Lee (2017) methodology at horizons of nine to fourteen months ahead. Additional evidence of the contractionary effects of devaluation risk on the real economy using pig iron production (often driven by changes in demand for steel rails or new locomotives) and a factory employment index can be found in the online appendix. Further, I also compute a similarly shaped impulse response function for industrial production in response to an increase in the currency risk premium using lagged values of the silver event excess HPR shock as the instrument and report the results in the online appendix. A comparison of the estimated trough effect of silver news on industrial production to the one-year response of output to twin banking and currency crises reported in Cerra and Saxena (2008) further underscores the economic cost of devaluation risk. However, given the large differences in both data and institutional context, one should be extremely cautious in drawing strong conclusions about the relative costs of devaluation risk and twin crises.21 Given that my sample includes the panic of 1893, it would be inappropriate to compare my measured effect of devaluation risk to the response solely coming from a currency crisis. Depending on the subsample of countries used in estimation, they find output (which is less volatile than industrial production) is between 4% and 6% lower one year after the onset of the twin financial crises. Although the expected devaluation during the time period was never greater than 2%, changes in silver coinage risk produced economic fluctuations of a size not far from those produced by actual devaluations. Given the economic magnitude and implications of the above findings, it is imperative to ensure they are well identified. Returning to the potential question of whether silver events are simply correlated with shocks to other economic variables, I cannot reject that changes in industrial production and the price level do not Granger-cause silver news. In other words, there is no evidence to suggest silver policy news systematically responds to shocks to production or the price level. I can reject the null hypothesis for the currency risk premium, however. Table 6 reports the $p$-values for each of these variables in the first column. If shocks to some economic variable not in the local projections regressions drive currency risk premium shocks, and these shocks in turn drive future production, then identification is still threatened. I discuss below how I work around this problem. Table 6. Granger Causality Test $p$-Values  (1) (2) (3) (4) Industrial production 0.808 0.894 0.891 0.975 Price level 0.169 0.657 0.616 0.861 Currency risk premium 0.000 0.77 0.109 0.138  (1) (2) (3) (4) Industrial production 0.808 0.894 0.891 0.975 Price level 0.169 0.657 0.616 0.861 Currency risk premium 0.000 0.77 0.109 0.138 Reported $p$-values are for null hypothesis that each variable does not Grangercause the silver event credit spread shock. Actual tests use first difference of industrial production and price level. Column 1 contains original series of excess HPR of exposed corporate bonds over safe corporate bonds during months with silver news, column 2 gives summer of 1893 events a value of 0, column 3 drops events occurring in the summer of 1893, and column 4 drops August and November 1896 events as well. The events most likely correlated with other shocks occurred after the panic of 1893 and later in the election season of 1896; I discuss the timing and outcome of all silver events in the 1890s in the online appendix. The aftermath of the panic could be contaminated for two reasons. First, the Senate broke from its typical pattern of favoring increased silver coinage and instead repealed the Sherman Act unconditionally. Senators may have switched due to the relative severity of the panic in their states or out of political favors gained in return for voting for repeal. Second, India, the other major government coining silver at the time, also suspended silver coinage at the end of June 1893, coinciding with the US panic.22 The outcome of the 1896 election may have been driven by several economic shocks, though it is difficult to determine which–if any–of these shocks affected voting behavior and whether these economic developments were responses to the silver threat (FS, 2020). I estimate paths for industrial production that are qualitatively similar to the original response function, using only those events that were unlikely to be strongly correlated with prior economic conditions as a robustness check. The bottom left panel of figure 5 plots the response of industrial production under three scenarios: estimated using the original data, estimated setting the excess HPR for the four summer of 1893 months to 0, and estimated dropping the summer of 1893 events and the August and November 1896 events, leaving eleven and nine non-0 months, respectively. In their paper on modern US monetary policy shocks, Romer and Romer (2004) follow a similar procedure for the months of nonborrowed reserve targeting in the United States between 1979 and 1981. Despite changing the values in my shocks series, the estimated paths are fairly similar. They all show silver coinage risk leading to a contraction in output, with the trough effect occurring between twelve and fourteen months in all cases. It is especially reassuring to see that dropping the events with the largest changes in excess HPR does not undo the initial finding of a negative effect of silver risk on output. As noted previously, these results say little to nothing about the effects of actual currency crises on the real economy. This is in part because most modern currency crises involve policy changes beyond just devaluing the currency (examples include the imposition of capital controls or bailouts of bankrupt firms). Additionally, countries typically abandon exchange rate pegs in order to engage in expansionary monetary policy. Further, the different mechanisms that transmit nominal exchange rates to the real economy (e.g., currency mismatch on firm balance sheets) may be more or less present in modern economies. One clear example of how devaluation risk and actual devaluation differ involves the experience of the United States when it actually left the gold standard during the Great Depression. As alluded to earlier, this devaluation was in fact expansionary: many securities increased in value, and the US economy subsequently grew. In the online appendix, I discuss potential reasons for these divergent effects of devaluation risk and devaluation. ### C. Silver Coinage Risk and Output: Mechanism Evidence This paper primarily focuses on the role of currency mismatch on firms' balance sheets in transmitting nominal devaluation risk to the real economy. Higher silver coinage risk raised default risk for firms with gold bonds, thus lowering investors' willingness to hold these debts. This in turn reduced railroad investment and, hence, industrial production. Here I present evidence consistent with this mechanism, though it is important to note that this is not the only way nominal devaluation risk can affect the real economy. FS (2020) show that fluctuations in silver risk also altered the health of the banking sector: as silver risk increased, banks' leverage decreased. Changes in gold flows in response to silver news provide some evidence of how foreign demand for American bonds changed.23 As noted by the comptroller of the currency at the time, normally US exports would have produced gold inflows; however, during the 1890s, foreigners instead sold American securities to pay for US goods (Timberlake, 1978). Indeed, one estimate of net foreign sales between 1890 and 1894 is$300 million (Friedman & Schwartz, 1963). Foreign capital was an important source of funds for railroad investment, with roughly $3 billion American securities–or 30% of total railroad debt and equity outstanding–held abroad in 1890 (Friedman & Schwartz, 1963). I estimate an impulse response for cumulative net gold inflows using monthly data on US gold flows from the Bureau of Statistics published in the Commercial and Financial Chronicle for 1880 to 1898 that I collected. I focus on the cumulative response to see if there was sustained foreign selling in total as a result of increased silver risk. I calculate the impulse response by regressing cumulative gold flows on six lags of net gold inflows, the silver coinage shock measure used in the main impulse response analysis, and a set of month dummies to account for seasonality. The results shown in the bottom right panel of figure 5 suggest that, indeed, increased silver risk led foreigners to decrease their holdings of US securities. Gold immediately flows out of the United States, reaching a trough of a$19.2 million cumulative outflow five months after the initial silver coinage news. Additionally, the timing of this trough in net gold inflows coincides with the peak of the currency risk premium in response to silver news: after five months, as the currency risk premium declines, so too do net gold outflows. The Commercial and Financial Chronicle accounts underscore the importance of foreign sales of US securities for these gold outflows: the periodical mentions European buying or selling of US assets due to silver news after nine of the thirteen events that occurred after the passage of the Sherman Act.

The economic repercussions of the defeat of the Free Silver movement in the election of 1896 also received considerable attention in the media. Railway Age sent a survey to railroad and industrial companies specifically asking how the businesses adjusted their activity in response to the pro-gold victory. Some companies reported increases in hours and employment for car shops or orders for new equipment since the results of the election became known. Often the responses note that it was the first time in years the shop had worked this many hours. A majority of companies also reported plans to further shed workers and decrease purchases of equipment if Bryan had won the election. Clearly, the risk of abandoning the gold standard mattered for economic decisions, especially railroads' investment decisions.

## VI. Conclusion

This paper studies the impact of devaluation risk on interest rates and output using the historical experience of the US with silver coinage and the gold standard. I find that increases in expected future silver coinage raised dollar devaluation expectations while lowering bond prices and industrial production. I argue that one channel through which these contractionary effects emerged is the presence of currency mismatch on firm balance sheets. Firms with gold debts most at risk of defaulting in the event of actual devaluation saw the biggest changes in their bond prices in response to silver news. These default fears led to the withdrawal of foreign capital in particular, forcing companies to reduce their investment. Since actual devaluations exert a contractionary effect through similar channels, it may be unsurprising that I find that nontrivial devaluation risk produces output costs that constitute a sizable fraction of those costs estimated for twin currency and banking crises.

My findings have implications for current policymakers. One of the largest macroeconomic sources of uncertainty in recent years has been the future of the euro. Greece, in particular, has come close to dropping out of the euro and adopting a new, depreciated currency, while all of its debt would still be payable in euros. My work implies that uncertainty about the value of the Greek currency acted as an independent drag on the Greek economy outside of the additional negative shocks hitting the Greek economy and the uncertainty surrounding other policies that would change with a Greek exit from the euro. Further, given new work highlighting the economic expansion that occurred once the United States actually devalued the dollar against gold during the Great Depression, my results imply that countries do more harm by trying to maintain their exchange rate pegs (Hausman et al., 2019; Jalil & Rua, 2016).

Finally, given my evidence for the balance sheet mechanism, this suggests policymakers should work to limit currency mismatch on nonfinancial corporate balance sheets. The need to borrow in foreign currency likely has deep structural causes, so simply limiting the amount of foreign currency that firms can borrow may not be welfare improving. Instead, countries would be better off changing the broader institutions that encourage foreign currency debts, making research into these causes imperative.

## Notes

1

The Free Silver movement advocated a mint convertibility ratio of 16 ounces of silver for 1 ounce of gold at a time when the market prices of silver and gold fluctuated between 20 and 32 ounces of silver per ounce of gold. This would have exhausted the Treasury's gold reserves and force it to suspend gold convertibility, leading to a 50% depreciation of the dollar against gold.

2

Like this paper, Gertler and Karadi (2015) use high-frequency monetary policy shocks to help identify lower-frequency effects of monetary policy on industrial production.

3

Past work has almost exclusively focused on the pricing of exchange rate expectations into government bonds (Schmukler & Serven, 2002; Powell & Sturzenegger, 2003; Bordo, Meissner, & Weidenmier, 2009; Mitchener & Weidenmier, 2015). An exception is Bailey and Bhaopichitr (2004), who argue that exchange rate risk was priced into Asian stock returns in the early twentieth century.

4

For example, Kim, Tesar, and Zhang (2015) find that foreign currency borrowing harmed small firms in Korea during the East Asian crisis, and Niepmann and Schmidt-Eisenlohr (2017) show that dollar appreciation raises the probability of default for dollar-denominated loans issued abroad. Bleakley and Cowan (2008) and Kalemli-Ozcan, Kamil, and Villegas (2016) find that foreign currency liabilities played no role in transmitting exchange rate shocks.

5

Our shared conclusion is also important, as this view is not held by all historians of the era (see Friedman & Schwartz, 1963, for a discussion).

6

FS (2020) then use this cross-sectional response to construct a time series of a “credibility index,” which they then use to study macroeconomic fluctuations and commitment to the gold standard.

7

Previous work by Temin and Wigmore (1990), Romer (1992), Edwards, Longstaff, and Marin (2015), Sumner (2015), and Jalil and Rua (2016) has focused on individual episodes during the Great Depression.

8

This is based on data published in Poor's Manual of Railroads and Romer (1989).

9

The agreement of Republicans from the House and Senate on a new silver bill in July 1890 is difficult to classify. I mark it as (+) because actual silver coinage did increase, but it is unclear if silver coinage increased by as much as people expected it to before the Sherman bill passed. Subsequent results do not depend on the classification of this event, however.

10

For more information on the politics of silver coinage and the monetary standard, see Hepburn (1903), Hollingsworth (1963), Timberlake (1978), and Unger (1964).

11

The gold reserves data are taken from the 1897 Treasurer's Report.

12

The effects may also have been larger after the panic of 1893 because the implied devaluation after adopting free silver was larger due to the widening difference between the market and mint price ratios for silver and gold. I do not test for this channel because the silver-to-gold price ratio is highly correlated with the Treasury's gold reserves.

13

This refers to any bond whose change in available earnings (if any) from a dollar devaluation would prevent that bond's (potentially altered) interest from being paid in full. Many of the bonds in my sample were already in default, particularly after the panic of 1893; however, the firm's earnings may have changed since the default date, so the calculation is still meaningful.

14

My model for HPR is a variant of the “constant-mean” model of expected returns used in event studies.

15

Several bonds were infrequently traded, and this could bias the results, since the return is calculated as if the bond was sold at its previous sale price on the day just prior. When I look at HPR on event days for a set of bonds that last traded no more than a week prior to the event, the HPR are similar to those using the entire set of bonds traded on an event day. See the online appendix for more details.

16

For multiday events, I compute the cumulative holding period return across days in the event.

17

The change in bank balance sheets from silver risk is documented in FS (2020).

18

The average call loan rate is simply the sum of the quoted high and low rates divided by two.

19

Depending on the information/selection criterion, the optimal lag length varies from two to eleven lags. I choose to use a lag length of 6, selected by the HQIC, but my findings are generally robust to the number of lags.

20

This series takes the interest rate differential between the sixty-day commercial paper rate in New York City, which was payable in dollars, and the implied gold interest rate from sixty-day and “sight” (spot) bankers' bills of exchange to capture the currency risk premium. Data are from the National Monetary Commission's Statistics for the United States, 1867–1909 and the Commercial and Financial Chronicle.

21

One difference in the studies is that Cerra and Saxena (2008) use output data, while this study focuses on industrial production. Even in modern data, industrial production is more volatile than real GDP (in the United States, industrial production is three times more volatile than output.) Further, the Miron-Romer index I use is more volatile than the modern index produced by the Federal Reserve. Davis (2004) constructs an annual industrial production index to better match the modern index, but the Davis index has annual changes of similar magnitude to the Miron-Romer index between 1884 and 1900.

22

Most of the debate on India's monetary system in the British Empire throughout the 1890s centered on the effects of US legislation on Indian economic prospects rather than the reverse. Additionally, aside from the suspension of silver coinage in 1893, the only other major Indian monetary shock during this time was the switch to a gold exchange standard in late 1898, well after the last US silver event in my data set.

23

As Friedman and Schwartz (1963) noted, sales of American securities to foreigners and gold exports are both sources of foreign currency. For a given current account balance, changes in gold flows are met by equal and opposite changes of foreign purchases of American securities.

24

The November 4, 1893, issue of the Commercial and Financial Chronicle describes how “for the last three or more years under the increasing incertitude as to the stability of our measure of values, this matter of borrowing money by railroad corporations has been growing more and more onerous.” As a result, railroads “economized the amount of of work done as far as possible,” while paying for this work with “temporary loans so as not to sacrifice their bonds, hoping all the time for a better market.”

## REFERENCES

Alquist
,
Ron
, and
Benjamin
Chabot
, “
Did Gold-Standard Adherence Reduce Sovereign Capital Costs?
Journal of Monetary Economics
58
(
2011
),
262
272
.
Bailey
,
Warren B.
, and
Kirid
Bhaopichitr
, “
How Important Was Silver? Some Evidence on Exchange Rate Fluctuations and Stock HPR in Colonial-Era Asia,
77
(
2004
),
137
174
.
Baker
,
Scott R.
,
Nicholas
Bloom
, and
Steven J.
Davis
, “
Measuring Economic Policy Uncertainty,
Quarterly Journal of Economics
131
(
2016
),
1593
1636
.
Bleakley
,
C. Hoyt
, and
Kevin
Cowan
, “
” this review 90
(2008)
,
612
626
.
Bordo
,
Michael D.
,
Alberto
Cavallo
, and
Christopher M.
Meissner
, “
Sudden Stops: Determinants and Output Effects in the First Era of Globalization, 1880–1913,
Journal of Development Economics
91
(
2010
),
227
241
.
Bordo
,
Michael D.
,
Christopher M.
Meissner
, and
Marc D.
Weidenmier
, “
Identifying the Effects of an Exchange Rate Depreciation on Country Risk: Evidence from a Natural Experiment,
Journal of International Money and Finance
28
(
2009
),
1022
1044
.
Calomiris
,
Charles W.
, “
Greenback Resumption and Silver Risk: The Economics and Politics of Monetary Regime Change in the United States, 1862–1900
,” in
Michael D.
Bordo
and
Forrest
Capie
, eds.,
Monetary Regimes in Transition
(
Cambridge
:
Cambridge University Press
,
1993
).
Cerra
,
Valerie
, and
Sweta Chaman
Saxena
, “
Growth Dynamics: The Myth of Economic Recovery,
American Economic Review
98
(
2008
),
439
457
.
Cespedes
,
Luis Felipe
,
Roberto
Chang
, and
Andres
Velasco
, “
Balance Sheets and Exchange Rate Policy,
American Economic Review
94
(
2004
),
1183
1193
.
Commercial and Financial Chronicle. Multiple volumes.
Davis
,
Joseph H.
, “
An Annual Index of U.S. Industrial Production, 1790–1915,
Quarterly Journal of Economics
119
(
2004
),
1177
1215
.
Edwards
,
Sebastian
,
Francis A.
Longstaff
, and
Alvaro Garcia
Marin
, “
The U.S. Debt Restructuring of 1933: Consequences and Lessons
,”
NBER working paper
21694
(
2015
).
Friedman
,
Milton
, and
Anna J.
Schwartz
,
A Monetary History of the United States
,
1867
1960
(
Princeton, NJ
:
Princeton University Press
,
1963
).
Fulford
,
Scott L.
, and
Felipe
Schwartzman
, “
The Benefits of Commitment to a Currency Peg: Aggregate Lessons from the Regional Effects of the 1896 U.S. Presidential Election,
” this review
102
(
2020
),
600
616
.
Gertler
,
Mark L.
, and
Peter
, “
Monetary Policy Surprise, Credit Costs, and Economic Activity,
American Economic Journal: Macroeconomics
7
(
2015
),
44
76
.
Gorodnichenko
,
Yuriy
, and
Byoungchan
Lee
, “
A Note on Variance Decomposition with Local Projections
,”
NBER working paper
23998
(
2017
).
Gupta
,
Poonam
,
Deepak
Mishra
, and
Ratna
Sahay
, “
Behavior of Output during Currency Crises,
Journal of International Economics
72
(
2007
),
428
450
.
Hallwood
,
C. Paul
,
Ronald
Macdonald
, and
Ian W.
Marsh
, “
Realignment Expectations and the U.S. Dollar 1890–1897: Was There a ‘Peso Problem’?
Journal of Monetary Economics
46
(
2000
),
605
620
.
Hausman
,
Joshua K.
,
Paul W.
Rhode
, and
Johannes F.
Wieland
, “
Recovery from the Great Depression: The Farm Channel in Spring 1933,
American Economic Review
109
(
2019
),
42472
.
Hepburn
,
A. Barton
,
History of Coinage and Currency in the United States and the Perennial Contest for Sound Money
(
Norwood, MA
:
Norwood Press
,
1903
).
Hollingsworth
,
J. Rogers
,
The Whirligig of Politics: The Democracy of Cleveland and Bryan
(
Chicago
:
University of Chicago Press
,
1963
).
Jalil
,
Andrew J.
, “
A New History of Banking Panics in the United States, 1825–1929: Construction and Implications,
American Economic Journal: Macroeconomics
7
(
2015
),
295
330
.
Jalil
,
Andrew J.
, and
Gisela
Rua
, “
Inflation Expectations and Recovery in Spring 1933,
Explorations in Economic History
62
(
2016
),
26
50
.
Jordà
,
Òscar
, “
Estimation and Inference of Impulse Responses by Local Projections,
American Economic Review
95
(
2005
),
161
182
.
Kalemli-Ozcan
,
Sebnem
,
Herman
Kamil
, and
Carolina
Villegas
, “
What Hinders Investment in the Aftermath of Financial Crises? Insolvent Firms or Illiquid Banks?
” this review
98
(
2016
),
756
769
.
Kim
,
Yun Jung
,
Linda L.
Tesar
, and
Jing
Zhang
, “
The Impact of Foreign Liabilities on Small Firms: Firm-Level Evidence from the Korean Crisis
,”
Journal of International Economics
97
(
2015
)
209
230
.
Kuttner
,
Kenneth N.
, and
Posen
, “
Do Markets Care Who Chairs the Central Bank?
Journal of Money, Credit and Banking
42
(
2010
),
347
371
.
Ludvigson
,
Sydney C.
,
Sai
Ma
, and
Serena
Ng
, “
Uncertainty and Business Cycles: Exogenous Impulse or Endogenous Response?
American Economic Journal—Macroeconomics
(
forthcoming
).
Mitchener
,
Kris J.
, and
Marc D.
Weidenmier
, “
Was the Classical Gold Standard Credible on the Periphery? Evidence from Currency Risk,
Journal of Economic History
75
(
2015
),
479
511
.
Moody
,
John
,
(
New York
:
Analyses Publishing Company
,
1909
).
Niepmann
,
Friederike
, and
Tim
Schmidt-Eisenlohr
, “
Foreign Currency Loans and Credit Risk: Evidence from U.S. Banks
,”
CESifo working paper series
6700
(
2017
).
Poor
,
Henry V.
,
Poor's Manual of the Railroads of the United States
, multiple volumes.
Powell
,
Andrew
, and
Federico
Sturzenegger
, “
Dollarization: The Link between Devaluation and Default Risk
,” in
Eduardo
Levy Yeyati
and
Frederico
Sturzenegger
, eds.,
Dollarization: Debates and Policy Alternatives
(
Cambridge, MA
:
MIT Press
,
2003
).
Pratt
,
Sereno S.
,
The Work of Wall Street
(
New York
:
Appleton
,
1912
).
Ramey
,
Valerie A.
, “
Macroeconomic Shocks and Their Propagation
,”
NBER working paper
21978
(
2016
).
Romer
,
Christina D.
, “
The Prewar Business Cycle Reconsidered: New Estimates of Gross National Product, 1869–1908
,”
Journal of Political Economy
97
(
1989
)
1
37
.
Romer
,
Christina D.
What Ended the Great Depression?
Journal of Economic History
52
(
1992
),
757
784
.
Romer
,
Christina D.
Journal of Economic History
54
(
1994
),
573
609
.
Romer
,
Christina D.
, and
David H.
Romer
, “
A New Measure of Monetary Shocks: Derivation and Implications
,”
American Economic Review
94
(
2004
).
1055
1084
.
Schmukler
,
Sergio
, and
Luis
Serven
, “
Pricing Currency Risk under Currency Boards,
Journal of Development Economics
69
(
2002
),
367
391
.
Sumner
,
Scott
,
The Midas Paradox: Financial Markets, Government Policy Shocks, and the Great Depression
(
Oakland, CA
:
Independent Institute
,
2015
).
Temin
,
Peter
, and
Barrie
Wigmore
, “
The End of One Big Deflation,
Explorations in Economic History
27
(
1990
),
483
502
.
Timberlake
,
Richard H.
,
The Origins of Central Banking in the United States
(
Cambridge, MA
:
Harvard University Press
,
1978
).
Unger
,
Irwin
,
The Greenback Era: A Social and Political History of American Finance, 1865–1879
(
Princeton, NJ
:
Princeton University Press
,
1964
).

## Author notes

I am deeply grateful to Dora Costa for her unyielding support and guidance. I also thank Leah Boustan for detailed feedback on this paper. Andy Atkeson, Daniel Beltran, Gillian Brunet, François Geerolf, Eric Hilt, Francis Longstaff, Tim Schmidt-Eisenlohr, and Felipe Schwartzman also provided comments that greatly improved this paper. For helpful comments at an earlier stage of this project I thank Michael Bordo, Gary Gorton, Alan Taylor, François Velde, and Marc Weidenmier. Yuqing Wu and Peter Fuzhe Zhang provided outstanding research assistance. The views in this paper are solely my responsibility and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or of any other person associated with the Federal Reserve system. All errors are my own.

A supplemental appendix is available online at http://www.mitpressjournals.org/doi/suppl/10.1162/rest_a_00841.