Do Risk Disclosures Matter When It Counts? Evidence from the Swiss Franc Shock*

Luzi Hail University of Pennsylvania, The Wharton School

Maximilian Muhn Humboldt University of Berlin

David Oesch University of Zurich

January 2019

Abstract

We examine the relation between disclosure quality and information asymmetry among market participants following an exogenous shock to macroeconomic risk. In 2015 the Swiss National Bank (SNB) abruptly announced that it would abandon the longstanding minimum euro-Swiss franc exchange rate. We find evidence suggesting that firms with more transparent disclosures regarding their foreign exchange risk exposure ex ante exhibit significantly lower information asymmetry ex post. The gap in bid-ask spreads appears within 30 minutes of the SNB announcement and persists for two weeks. We validate the informational role of past risk disclosures with three field surveys: (1) Sell-side analysts emphasize the importance of existing (risk) disclosures in evaluating the translational and transactional effects of the currency shock. (2) Lending banks’ credit officers rely on past disclosures as the primary information source available for smaller (unlisted) firms in the immediate aftermath of the shock. (3) Investor-relations managers use existing financial filings as a key resource when communicating with external stakeholders. The results imply that risk disclosures continue to attenuate information asymmetry and the costs of adverse selection well beyond their initial publication date. JEL classification: F31, G12, G14, G15, G30, M41 Key Words: Risk disclosures, adverse selection, liquidity, information asymmetry, currency

risk, archival studies, surveys

* We appreciate the helpful comments of Ray Ball, Phil Berger, Jeremy Bertomeu, Siddharth Bhambhwani, Matthias Breuer, Donal Byard, Vedran Capkun, Ted Christensen, Holger Daske, Peter Fiechter, John Gallemore, Joachim Gassen, Christian Hofmann, Christoph Kaserer, Christian Leuz, Miao Liu, Tim Martens, Maximilian Müller, Martin Nienhaus, Natalia Reisel, Thorsten Sellhorn, and workshop participants at the 2017 American Accounting Association meeting, 2018 Swiss Winter Accounting Conference, 2018 University of Minnesota Conference, 2018 Verein für Socialpolitik Accounting Section meeting, 2018 European Accounting Association meeting, 2018 European Financial Management Association meeting, Baruch College, University of Chicago, University of Hong Kong, Hong Kong University of Science and Technology, Humboldt University of Berlin, University of Illinois at Chicago, Ludwig Maximilian University of Munich, Stockholm School of Economics, and WHU. We thank Conrad Meyer and Michèle Schnyder for their help in putting us in touch with the right people for our interviews and surveys. Maximilian Muhn gratefully acknowledges the support by the German Research Foundation through the Collaborative Research Center 649 “Economic Risk” at the Humboldt University of Berlin.

1

“The Swiss franc rocketed beyond parity with the euro on Thursday after Switzerland’s

central bank stunned markets by scrapping its long-standing cap on the strength of the

currency. The euro dropped around 30% against the franc, ending more than three years

of calm in Swiss foreign-exchange markets.” Wall Street Journal, January 15, 2015.

1. Introduction

How do market participants cope with unexpected events that affect a firm’s prospects, and

what role does the quality of existing financial statement information play in this process? In

standard economic theory, when new information arrives, it is immediately processed by economic

agents and affects their behavior. In financial markets, when a firm releases its financial reports,

previously private information about future expected cash flows enters the public domain, and it

quickly is reflected in stock prices, liquidity, and investors’ trading patterns (e.g., Beaver 1968;

Balakrishnan et al. 2014; Rogers et al. 2017). However, little is known about the role of disclosures

after they lose their initial news element. We exploit an exogenous shock to macroeconomic risk

to examine whether and how the precision (or quality) of past currency risk disclosures is related

to the information asymmetry reduction among market participants (in the sense of Verrecchia

2001). We show that the market reacts to the exogenous shock in a way consistent with more

precise past disclosures allowing investors to better process and interpret new information. We

confirm this role of past disclosures by surveying financial analysts, credit officers in lending

banks, and investor-relations managers at affected firms.

On January 15, 2015, at 10:30 in the morning, the Swiss National Bank (SNB) announced it

would abandon the minimum exchange rate between the euro (EUR) and the Swiss franc (CHF)

without further notice. The SNB introduced the exchange-rate peg in September 2011 to counter

the ongoing pressure on the Swiss franc to act as a “safe haven” amid the turmoil in global financial

markets. By increasing the supply of francs, the SNB managed to maintain the target exchange

2

rate of 1.20 EUR-CHF. The announcement to discontinue the exchange rate peg caught the market

by surprise. It had an immediate and large impact: the EUR-CHF exchange rate dropped close to

parity and Swiss stock markets lost almost 10 percent in a single day.

The SNB announcement provides an ideal setting to analyze the long-term transparency

effects of past risk disclosures. First, the unpegging of the Swiss franc represents an exogenous

shock, whose timing is random and not influenced by either firms or investors. Second, the drop

in exchange rates is economically significant and has the potential to affect the expected cash flows

of many Swiss firms because exports of goods and services make up about 70 percent of Swiss

GDP at the time. Yet, the effect is not uniform but reflects a firm’s business model and exposure

to the euro. This private information is inherently difficult to observe and requires disclosures by

the firm. Third, the initial market reaction occurs within minutes of the event, which is too short

a time span for the firm to prepare and release new disclosures, so external stakeholders must rely

on existing financial information. Fourth, while firm disclosure is typically the result of an

endogenous choice complicating our ability to measure the link between disclosure quality and

information asymmetry (e.g., Clinch and Verrecchia 2015), we mitigate this endogeneity issue by

holding disclosure quality constant and focusing on the effects of an exogenous shock to

macroeconomic risk. Thus, we believe that in our setting variation in the precision of a firm’s past

disclosures about the currency exposure of its business is a plausible path to explain variation in

the market reaction to the SNB announcement. At the same time, the informational role of past

disclosures is not obvious, because the already published information is stale, might have little

relevance for the new situation or requires special expertise to be processed. Moreover, competing

information sources could be more accurate and timely.

To test our research question, we first conduct an archival event study of changes in

information asymmetry following the Swiss franc shock conditional on the precision of past risk

3

disclosures in firms’ annual reports. For each of our 151 sample firms listed on the SIX Swiss

Exchange (equal to 98 percent of market capitalization), we construct a risk disclosure score. The

score ranges from 0 to 7 and captures the quality of a firm’s disclosure policy regarding its foreign

exchange risk exposure (not the actual risk). It reflects the extent of a firm’s disclosure regarding

items like revenues generated abroad, the currency distribution of liquid assets and liabilities, or

the sensitivity of net income to exchange rate fluctuations. A higher score indicates more precise

information for investors to interpret the subsequent shock, which should reduce the cost of

adverse selection and, in turn, lower bid-ask spreads.

We find that in the initial three days after the SNB announcement information asymmetry

among investors dramatically increases. Bid-ask spreads, on average, jump by 21 basis points or

44 percent relative to the 30-day period leading up to the event. Yet, systematic heterogeneity

exists. Firms with higher risk disclosure scores by one standard deviation experience less of an

increase in bid-ask spreads by about 7 basis points. The gap in spreads appears within 30 minutes

after the announcement by the SNB and persists for two weeks. The results are robust to the

inclusion of firm and industry-by-day fixed effects or, in an intraday analysis, the inclusion of

fixed effects for each 30-minute trading interval. They also hold when we control for firms’

general information environment and their actual currency risk exposure. In additional analyses,

we find that a firm’s currency risk exposure, measured by international sales or the historical

correlation between stock returns and EUR-CHF exchange rates, is significantly negatively related

to the stock price reaction following the Swiss franc shock. No such relation is present for our risk

disclosure score. In sum, the archival tests imply that past risk disclosures help investors

contextualize the new information from the unexpected change in risk, and that it takes firms about

two weeks to close this information gap.

4

In the second part of the study, we incorporate field data from users of financial statement

information into the analysis. We conduct surveys and structured interviews with three groups of

market participants affected by the SNB announcement: (1) sell-side analysts covering the firms

listed on SIX, (2) credit officers of lending banks, and (3) investor-relations managers at the

affected firms. Taken together, this evidence from the field supports the plausibility of the

inferences we draw from the archival tests and increases our confidence in the causality of the

established relations (Soltes 2014; Bernstein et al. 2016).

Based on 77 survey responses (9.4 percent response rate), we find that only 3 percent of

financial analysts anticipated such a surprising move by the SNB within the next three months.

Analysts expected the change in the currency regime to negatively impact the average firm, and

they reassessed the prospects of almost their entire portfolio within the first two days. In most

cases, such a reassessment led to material adjustments of the quantitative inputs in the valuation

models that analysts use to derive their earnings forecasts and stock recommendations. More to

the point, existing annual reports and financial filings served as a key information resource in the

immediate aftermath of the event, on par with private communication with management, and

ranked only behind analysts’ personal experience and industry knowledge. This role declined in

the days that followed, as ad hoc announcements gained importance. The main use of existing

financial statement information was to better contextualize and interpret the translational and

transactional currency risk exposure of the firm and to assess the effectiveness of the hedging

strategy in place. Anecdotal evidence suggests that analysts had past risk disclosures in mind when

answering these survey questions.

We next conducted a series of structured interviews with credit officers from four large Swiss

banks. Lending institutions enjoy privileged access and closer ties to their clients (e.g., Rajan

1992; Agarwal and Hauswald 2010), and given their longer horizon, they face less time pressure

5

to respond to unexpected changes in risk. Immediately after the SNB announcement (within one

week for listed firms, two to four weeks for the rest), all banks reviewed their loan portfolios and

classified firms into different risk categories depending on the materiality and perceived impact of

the shock. The main, and for non-listed firms only, information source for this triage were existing

financial disclosures including annual reports submitted during the periodic loan review. High

risk cases underwent a more detailed review over the following months and loan officers often

engaged in direct discourse with their clients to gather additional information on the impact of the

currency shock, particularly for large corporate borrowers. Two banks indicated that they

tightened their clients’ reporting requirements and frequency as a response to the event, but none

mentioned adjustments to the credit risk rating system.

Finally, based on 39 replies to a survey of investor-relations managers (26 percent response

rate), we find that firms were surprised by the currency shock, and anticipated negative effects on

future sales and earnings. Close to 50 percent of the firms started communicating with external

stakeholders on the day of the event, primarily through private channels with financial analysts

and institutional investors. The interactions were both proactive and a reaction to outside demand,

and the main goals were to reassure existing investors, reduce uncertainty, and build trust in the

firm’s reporting strategy. Consultation with key management, existing contingency plans, and

past annual reports and financial filings served as primary information sources in the preparation

of the external communication. As time went by, the communication shifted towards regular

financial filings and pre-scheduled investor events.

Our study contributes to the literature in several ways. We find that the precision of past risk

disclosures explains systematic variation in information asymmetry well beyond the initial

publication date. Put differently, disclosures not only have the ability to reduce information

asymmetry upon release (e.g., Brown et al. 2009; Shroff et al. 2013; Balakrishnan et al. 2014), but

6

can also mitigate adverse selection in the future. Our channel is different from credibly committing

to future transparent reporting (Leuz and Verrecchia 2000; Verrecchia 2001; Francis et al. 2008;

Daske et al. 2013) which leads to a persistent effect on information asymmetry. The effect in our

setting is transitory and relies on the quality of past disclosures interacting with new information.

Our study is related to Blacconiere and Patten (1994), Bonetti et al. (2018) as well as Cuny and

Dube (2018). Similar to our identification strategy, these papers use an accident in a chemical

plant in India, the Fukushima nuclear disaster, and the bankruptcy of the bond insurer Ambac,

respectively, as exogenous events. However, they focus on the short-term price reactions and the

long-term cost of capital or debt ratings effects of past disclosures, which is different from the

purely information-driven mechanism at work in our setting. Advantages of our approach are that

it allows us to disentangle the information channel from the effects of (risk) exposure and, by using

intraday data, we can get close to causal identification of the stipulated relation. As we hold the

disclosure level constant, our perspective is also different from studies investigating the change in

(voluntary) disclosures and the related market consequences upon the arrival of fundamental news

(e.g., Leuz and Schrand 2009; Balakrishnan et al. 2014).

We further contribute to the literature on how investors use financial reports. By combining

archival and survey methods (e.g., Bernstein et al. 2016), we provide a richer picture of the extent

to which investors, analysts, lenders, and employees rely on financial statement information. Our

evidence from the field corroborates the notion of a causal link between unexpected changes in

firm fundamentals and information asymmetry among investors through the means of past risk

disclosures. This result complements evidence by Drake et al. (2015; 2016) who find that investors

make more use of the SEC EDGAR database when financial statements are complex, firms report

negative or largely discretionary earnings, and there is a negative shock to firm value. In contrast,

we zoom in on one particular type of disclosures and one specific (exogenous) event, both dealing

7

with a firm’s business risk. This focus allows us to open the black box of how past disclosures

provide a channel through which they affect information asymmetry and, ultimately, firm value.

Finally, our study contributes to the literature on corporate risk disclosures. Prior studies

examine the stock market reaction to risk disclosures (e.g., Campbell et al. 2014; Heinle and Smith

2017), assess their ability to predict future risk (e.g., Jorion 2002), or investigate incentives for

strategic risk reporting (e.g., Begley et al. 2017). We test whether risk disclosures help investors

gauge the impact of a sudden realization of risk. In broader terms, our results speak to the role of

extant disclosures to interpret new information beyond just risk, for instance, if a competitor

announces a new product, new regulation is proposed, or a large merger in the industry takes place.

2. Hypothesis Development and Institutional Setting

2.1. Conceptual Underpinnings for the Informational Value of Past Risk Disclosures

We analyze the informational role of past risk disclosures. Prior archival research mostly

focuses on the information content of newly released disclosures; that is, whether the information

contained in these new disclosures changes investors’ beliefs about future returns or prices (e.g.,

Beaver 1968). The new information signal transfers previously unknown or private information

into the public domain and is expected to immediately affect stock prices, bid-ask spreads, and

investors’ trading behavior. Thus, the release of the information and the market reaction occur at

the same time. After the release, the news content and the precision of the new disclosures is

typically assumed to become stale and have no more effect on investors’ future decisions.

However, the view that financial disclosures are only useful at the time of their release might

be too short-sighted (e.g., Ball and Shivakumar 2008). They might interact with future (public or

private) information signals plus reflect on the usefulness of prior (public or private) information.

In our setting, we examine whether the precision (or quality) of past public announcements

8

complements current public information signals. Specifically, we argue that past risk disclosures

become useful again once the underlying risk suddenly changes. The arrival of unexpected news

about the state of the firm (i.e., the currency shock) increases uncertainty among investors. As

market participants scramble to assess the economic consequences of the news, past disclosures

about how the firm handles currency risk help investors contextualize and interpret the new

situation and reduce some of the uncertainty. Thus, we use the exogenous shock to examine the

relation between disclosure quality and the information asymmetry reduction among market

participants as predicted by theory (e.g., Amihud and Mendelson 1986; Diamond and Verrecchia

1991; Verrecchia 2001; Lambert et al. 2007; Goldstein and Yang 2017). Our setting is related to

Kim and Verrecchia (1994; 1997). In their models, past private information interacts with a new

public signal, thereby exacerbating the informational disadvantage of the uninformed investors.

In contrast, we study variation in the precision of past information, which is public knowledge,

while holding the distribution of private information among investors constant. We expect firms

with more precise past disclosures about risk to suffer less of an increase in information asymmetry

upon the arrival of the sudden news (i.e., leaving investors without private knowledge at less of an

informational disadvantage).1

While our study fits into the general body of work on the link between disclosure quality and

information asymmetry among market participants (e.g., Leuz and Verrecchia 2000; Daske et al.

2008; Balakrishnan et al. 2014), our information structure is distinct from other mechanisms

examined in the literature. For instance, by credibly committing to disclose information regardless

of the future outcome, a firm can reduce information asymmetry in a persistent manner (e.g.,

Diamond and Verrecchia 1991; Baiman and Verrecchia 1996). The commitment occurs in the

1 We would obtain the same prediction if sophisticated investors saw more value in acquiring additional private

information for low-quality disclosure firms upon the arrival of the sudden news (i.e., high-quality risk disclosures crowd out private information acquisition by sophisticated investors).

9

past, and because it imposes a cost (e.g., through the use of better-quality auditors), it renders

current public announcements more precise. In our setting, the initial disclosure is in the past, but

from thereon forward the information is available to all market participants and remains available

to them regardless of future outcomes. As a result, the stipulated effect of past disclosures is likely

transitory as firms can make up for them with new public announcements, and investors have

incentives to gather more information. Our analysis is also distinct from prior studies that focus

on the signaling effect of past disclosures (e.g., Blacconiere and Patten 1994; Banker et al. 1995;

Heflin and Wallace 2017). In these studies, the past public announcement proxies for the actual

risk exposure of the firms and, hence, it should be related to the price reaction upon the arrival of

the (unexpected) news that prompts a fundamental change in risk. We instead are interested in

how transparency with regard to firms’ risk exposure relates to the transaction costs that arise from

the adverse selection problem between investors with varying degrees of informedness when

interacting with each other.2

The informational role of past risk disclosures is not obvious. First, by definition, already

published disclosures are stale and potentially outdated. The information might have little

practical relevance for investors when a firm faces a new situation. Second, while past risk

disclosures are available to the public, they might require special knowledge and expertise to be

processed. For instance, only sophisticated users such as analysts or institutional investors might

have the ability to interpret and contextualize the new event against the backdrop of past

information. In this case, the availability of past disclosures would intensify the comparative

disadvantage for the uninformed investor (Indjejikian 1991; Kim and Verrecchia 1994). Third, a

2 This focus on information asymmetry is the reason why we use bid-ask spreads as the dependent variable in our tests

and not price. At the same time, we are not arguing that the overall market reaction to the Swiss franc shock was not large and unlikely to be a function of a firm’s exposure to the euro. It was, and from a price perspective, this market reaction certainly must be considered the first-order effect of the Swiss franc shock while the price reaction to the transparency effect is negligible (see also Section 3.3.4).

10

firm cannot foresee every contingency ex ante or it is too costly to provide adequate information.

Thus, past disclosures contain, at best, indirect information about the impact of a change in firm

fundamentals. Other information sources are likely more accurate and timely, and firms can

quickly react and bridge a perceived information gap with new ad hoc disclosures.

2.2. Institutional Setting of the Swiss Franc Shock

The Swiss franc has a long tradition of acting as a safe-haven currency for investors when

global financial markets are in turmoil. In 2011, fueled by the looming sovereign debt crisis in

Europe and the near-failure of the debt-ceiling negotiations in the U.S., the Swiss franc strongly

appreciated against the euro (by up to 17 percent) and the U.S. dollar (by up to 22 percent). This

currency appreciation put a strain on the many export-reliant industries in Switzerland and

increased the deflationary pressure. At this point, the SNB had already intervened multiple times

in the capital markets by broadening the monetary base and acquiring foreign currencies. In

August, the SNB declared that it would take additional steps against the strong Swiss franc (SNB

2011a). These efforts culminated in a stunning announcement on September 6, 2011, when the

SNB set a currency floor of CHF 1.20 per EUR and insisted that it would “enforce this minimum

rate with the utmost determination” as it was committed “to buy foreign currency in unlimited

quantities” (SNB 2011b). As Figure 1 illustrates, the new currency regime successfully enforced

this minimum EUR-CHF exchange rate in the years that followed. The Swiss franc was effectively

pegged against the euro, and fluctuations of the Swiss franc against other currencies were mostly

attributable to developments in the Eurozone.

Things changed abruptly in early 2015. On January 15, at exactly 10.30 am, the SNB released

an official statement declaring that it would abandon the minimum EUR-CHF exchange rate

11

without further notice (SNB 2015).3 The sudden policy change came as a surprise to market

participants. For instance, Bloomberg News (2015) had surveyed 22 economists in the first two

weeks of 2015 and none of them expected that the cap would be abandoned in 2015. A Deloitte

(2015) survey conducted in December 2014 found that only 3 out of 129 chief financial officers

of Swiss firms expected a change in the currency regime in 2015.4 Mirkov et al. (2017) conclude

that the market did not foresee the SNB decision based on an analysis of high-frequency liquidity

data in option markets and foreign exchange markets.5

The market reaction to the SNB announcement was swift. As Figure 1 indicates, the Swiss

franc immediately soared against the euro and other major currencies, as the SNB decision

effectively untied the Swiss franc from the Eurozone. At closing, the Swiss franc had risen against

the euro (U.S. dollar) by 14 percent (12 percent). The Swiss All Share index (SSIP) dropped by

almost 9 percent, marking the single largest daily decrease since its inception in 1998. The

economic consequences did not stop there. Efing et al. (2016) find that the negative wealth effects

of the Swiss franc shock were stronger for export-oriented firms with a heavy reliance on domestic

production. They find that these firms experienced larger reductions in profitability and sales over

the next year than their domestically oriented peers. Bonadio et al. (2016) analyze daily

transaction-level data from the Swiss Customs Administration and find that cross-border import

3 In the same statement, the SNB also announced to lower deposit interest rates “to ensure that the discontinuation of

the minimum exchange rate does not lead to an inappropriate tightening of monetary conditions.” Unlike the minimum exchange rate decision, analysts had expected to see lower deposit rates in the near future.

4 Our own survey evidence confirms these observations. Only 3 (20) percent of the sell-side financial analysts and 5 (36) percent of the investor-relations managers anticipated such a surprising move by the SNB within the next three months (one year). See Section 4.

5 The announcement on January 15, 2015, was also not part of the ongoing monetary policy assessments that the SNB conducts at pre-set dates every three months. The last assessment of this nature took place on December 11, 2014, with the next meeting scheduled for March 2015. The December assessment was accompanied by the headline “Swiss National Bank reaffirms minimum exchange rate,” and the press release stated that “the SNB will continue to enforce the minimum exchange rate with the utmost determination” (SNB 2014).

12

prices adjusted either instantaneously (if invoiced in EUR) or within a few working days (if

invoiced in CHF), indicating a quick exchange-rate pass-through.

3. Archival Analysis of Past Risk Disclosures and Information Asymmetry

3.1. Research Design

To test the informational role of past risk disclosures around the Swiss franc shock, we

examine changes in daily bid-ask spreads around the event conditional on the quality of a firm’s

risk disclosures. Figure 2 illustrates the timeline of events. Notably, the measurement of the cross-

sectional partitioning variable occurs before the event. Based on this timeline, we estimate the

following ordinary least squares (OLS) regression model:

Log(Bid-Ask Spreadi,t) = β1 Post_SNBt + β2 Post_SNBt × FXRisk_Disci + ∑ βj Controlsi,t-1 +

∑ βk Post_SNBt × Controlsi + ∑ βl Fixed Effectsi,t + εi,t. (1)

The dependent variable, Bid-Ask Spread, is the mean value of intraday minute-by-minute

differences between bid and ask quotes (divided by the mid-point) for firm i on trading day t.

Post_SNB is a binary indicator that takes on the value of ‘1’ beginning on day t = 0 of the SNB

announcement (January 15, 2015), and ‘0’ beforehand. The sample period comprises the 33

trading days surrounding the SNB announcement, with t Î [0; +2] serving as the event window

and t Î [-30; -1] as benchmark period. Post_SNB captures the mean incremental change in bid-

ask spreads following the event. We predict that this coefficient is positive, indicating a sudden

increase in information asymmetry regarding future EUR-CHF exchange rate fluctuations.

The primary test variable, FXRisk_Disc, is a proxy for the quality of foreign exchange risk

disclosures of firm i as reported in the most recent annual report before the event (i.e., for fiscal

13

year 2013).6 We rely on annual reports because they represent by far the most comprehensive

source for currency risk information.7 We construct FXRisk_Disc by scoring seven items: (i)

revenues, (ii) assets, and (iii) costs/profits generated and held outside of Switzerland; (iv) the

currency distribution of short-term monetary assets and liabilities; (v) the exposure to and (vi) the

hedging strategy regarding foreign currency risk; and (vii) the sensitivity of net income or

shareholders’ equity to changes in foreign exchange rates. Each disclosure item receives a score

of 1 (quantitative and qualitative information), 0.5 (qualitative information), or 0 points (no or

boilerplate disclosures), and the variable FXRisk_Disc equals the sum of points, ranging from 0

(worst) to 7 (best). For ease of interpretation, we standardize the raw scores to a mean of zero and

a standard deviation of one for use in the regression analysis. In Appendix A, we provide details

on the construction of this variable together with illustrative examples.

The idea underlying the FXRisk_Disc score is to gauge the quality of the information, not the

actual risk exposure of the firm. More precise past risk disclosures should allow investors to better

assess the future cash flow implications of the currency shock, ceteris paribus, and translate into

lower information asymmetry. We expect the interaction term between Post_SNB and

FXRisk_Disc to exhibit a negative sign. Because we include firm fixed effects in the model, the

main effect of FXRisk_Disc (as well as any intercept) does not appear in Eq. (1). When we control

for daily trends in the data, the day fixed effects subsume the main effect of Post_SNB. In this

case, the interaction term is specified by the within-day, between-firm variation in bid-ask spreads.

In terms of firm-level Control Variables, we follow prior research (Chordia et al. 2000; Leuz

and Verrecchia 2000; Christensen et al. 2013), and include firm size using the Market Value of

6 Over 90 percent of the sample firms end their fiscal year on December 31, and publish the annual report within two

to four months. Most of the remaining firms have fiscal-year ends of March 31. Thus, none of the sample firms had released its 2014 annual report at the time of the SNB announcement.

7 We checked a random sample of quarterly and half-year reports as well as other financial filings occurring over the course of the fiscal year, but could not identify additional relevant risk information in those interim disclosures.

14

equity at the end of each trading day, daily Share Turnover, and Return Variability equal to the

standard deviation of half-hourly stock returns over the trading hours of the day. Depending on

the specification, we include a vector of additional firm-level control variables in the model and

interact them with Post_SNB to allow the event effects to vary across firm attributes. Because

most of the firm attributes are time-invariant, the firm fixed effects subsume the respective main

effects and only the interaction terms are defined. The model comprises firm, day, or one-digit

SIC industry-by-day Fixed Effects to account for the mean bid-ask spread along these dimensions.

We partial out the fixed effects when computing the adjusted R2 and report the more meaningful

“within” adjusted R2 values in the tables. We estimate Eq. (1) in a log-linear form using the natural

logarithm of bid-ask spreads and the time-varying control variables (lagged by one day). We draw

statistical inferences based on two-way clustered standard errors by firm and trading day.

3.2. Sample Selection and Description

The initial sample comprises 235 publicly traded firms with a primary listing in Switzerland

that are constituents of the SSIP at the end of 2014. We collect minute-by-minute intraday spread

data from Bloomberg for these firms. For each firm-minute, Bloomberg reports bid and ask quotes,

trade prices, CHF volumes, and the number of ticks. To ensure sufficient liquidity and avoid

potential market microstructure biases such as the stale quotation problem (McInish and Wood

1992), we apply a series of data filters to obtain the final sample as outlined in Table 1, Panel A.

We require firms to have, on average, at least ten trades per day and to have updated daily bid-ask

spreads on at least 141 out of the 144 trading days (≥ 97.5 percent) over the period from October

2014 to April 2015. These two criteria eliminate thinly traded firms and firms with narrow market

depth. We delete firms with missing data for the variables used in the regression analysis. The

selection procedure yields a final sample of 151 firms that make up more than 98 percent of the

capitalization of the Swiss market.

15

We further clean the data by restricting the spreads to trading hours, and delete bid and ask

quotes with zero value or volume (Ng et al. 2008). SIX trading hours are from 9.00 am to 5.30

pm, consisting of 17 half-hourly trading intervals.8 We replace missing quotes (i.e., no new bid or

ask quotes during a specific minute) by carrying the previous quotes forward within a day, but not

across days (McInish and Wood 1992). We obtain daily data by taking the mean of minute-by-

minute spreads, but do not consider negative spreads in this computation (McInish and Wood

1992; Chan et al. 1995). We use the most recent transaction prices when computing half-hourly

log returns for the return variability measure.

Panel B of Table 1 provides descriptive statistics for the variables used in the daily bid-ask

spread regressions. The mean Bid-Ask Spread is 48 basis points, which is smaller than the values

reported for broader samples (e.g., Christensen et al. 2016), and consistent with our firms being

highly traded and liquid. The mean and median values of the FXRisk_Disc score are 4 points out

of 7 possible. The interquartile range of 2 points indicates that the variable offers ample variation.

Aside from the main control variables, the panel also tabulates descriptive statistics for additional

(control and dependent) variables that we use in the sensitivity tests. Int_Sales is the percentage

of sales generated abroad as shown in the most recent annual report.9 When exact data are missing

(32 cases), we infer international sales from textual and other disclosures in the annual report.

Hist_Corr_EUR is the historical correlation between weekly stock returns and EUR-CHF

exchange rates measured in the three years prior to currency peg. In four cases with missing

historical data, we use the contemporaneous correlation with USD-CHF exchanges rates instead.

Total_Disc is a score, ranging from 1 (worst) to 6 (best), ranking the overall disclosure quality of

8 Specifically, we only consider quotes between 9.02 am and 5.20 pm. SIX randomly opens trading between 9.00 and

9.02 am (opening auction), and no trades are settled between 5.20 and 5.30 pm (closing auction). Thus, the first and last half-hourly interval are slightly shorter than 30 minutes.

9 Using Worldscope data (item no. 08731) instead of hand-collected data, and coding missing values as zero leaves our results virtually unchanged.

16

a firm’s annual report in the fiscal year 2013. This score is published annually by the Institute for

Banking and Finance of the University of Zurich. Num_Analysts is the number of analysts in

I/B/E/S that cover a firm in the week before the SNB announcement. If missing, we set the variable

to zero. Free Float is the percentage of shares available to ordinary investors at the end of the

most recent fiscal year (source: Datastream). We compute daily Stock Return as the natural

logarithm of price at the end of trading over price at the end of the previous day. In some analyses,

we also use the closing Stock Price at the end of each trading day (in CHF) as the dependent

variable. We do not winsorize or truncate any of the data.10

The median sample firm is covered by four analysts, has a free float of 69 percent, and a

market capitalization of CHF 1.26 bn. Thus, these firms are large and visible. On average, firms

generate two thirds of their sales outside of Switzerland, suggesting a heavy reliance on exports

and a potentially substantive exposure to risks from foreign currency fluctuations. However, both

measures of currency risk exposure (Int_Sales and Hist_Corr_EUR) display large variation,

suggesting that the effect of the Swiss franc shock differs across firms.

3.3. Changes in Information Asymmetry Following the Swiss Franc Shock

3.3.1. Analyses of Daily Bid-Ask Spreads

We begin the archival analyses with plotting the mean daily bid-ask spreads over the 40

trading days surrounding the event and report results in Figure 3. The solid line represents firms

with high-quality past risk disclosures (i.e., FXRisk_Disc greater than the sample median). The

dashed line stands for a synthetic control group (Abadie and Gardeazabal 2003; Abadie et al.

2010). Following Cavallo et al. (2013) and Acemoglu et al. (2016), we match each high-quality

10 When we winsorize or truncate all the continuous variables in Eq. (1) at the first and 99th percentile (not tabulated),

the results remain largely unaffected and none of the inferences change.

17

disclosure firm to a combination of all firms with an FXRisk_Disc score at or below the median

(benchmark pool) using a convex weighting matrix that minimizes the Euclidean differences in

the outcome variable on each trading day over the [-20;-1] period. By construction, the differences

in bid-ask spreads are not significant before the event. Spreads rise dramatically on the day of the

SNB announcement, remain at high levels for the initial days, and then gravitate towards pre-

announcement levels. The pattern indicates that the currency shock increases the uncertainty in

the market. More to the point, the reaction is much abated for the firms with high-quality risk

disclosures. Visually, the gap in spreads between the two groups persists until about 15 trading

days after the event.

We next conduct a bivariate analysis in a difference-in-differences framework and report

results in Table 2, Panel A. Specifically, we report the mean daily bid-ask spreads in a two-by-

two matrix for firms with high versus low FXRisk_Disc scores before and after the SNB

announcement. Before the event, firms with more transparent risk disclosures have lower bid-ask

spreads, as one would expect. However, the difference of 13.5 basis points is not significant. After

the event, bid-ask spreads for both groups significantly increase, but less so for the more

transparent firms. The gap between groups extends to 34.6 basis points and is statistically

significant. So is the difference-in-differences across the four cells, suggesting that firms with

better risk disclosures experience less of an increase in information asymmetry.

In Panel B of Table 2, we report coefficients and t-statistics from estimating variations of Eq.

(1). The model in Column (1) contains the basic liquidity controls plus firm fixed effects. The

significantly positive coefficient on Post_SNB indicates that relative to the pre-event level bid-ask

spreads, on average, increased by 21 basis points or 44 percent (e0.367 – 1) after the Swiss franc

shock. This impact is economically important. However, the effect is mitigated for firms with

more informative past risk disclosures. The significantly negative coefficient on the interaction

18

term between Post_SNB and FXRisk_Disc suggests that a one standard deviation increase in the

risk disclosure score is associated with less of a jump in bid-ask spreads by about 7 basis points or

almost a third compared to the average reaction.11 The control variables are all significant and

have the predicted signs.

In column (2), we allow the effect to differ across large and small firms and add an interaction

between Post_SNB and the natural logarithm of a firm’s mean market value over the benchmark

period (Market Value-30,-1) to the model. We further include fixed effects for each trading day,

which flexibly control for trends in the data and subsume the main effect of Post_SNB. In Column

(3), we allow for the possibility that the Swiss franc shock impacted some industries more than

others and let the day fixed effects vary by one-digit SIC industry.

In Column (4), we control for the mitigating effects of currency risk exposure and overall

disclosure quality. Specifically, we include an operations-based proxy (IntSales) and a market-

based proxy (Hist_Corr_EUR) of a firm’s actual exposure to currency risk in the model. Firms

with higher exposure to exchange rate fluctuations might face higher uncertainty immediately after

the event. We also control for the overall quality of a firm’s annual report (Total_Disc) and its

information environment (Num_Analysts). Firms with transparent annual reports likely have more

informative risk disclosures. Analysts are sophisticated users of firm information and often

possess superior information capabilities and private communication channels with management

(Brown and Rozeff 1978; Brown et al. 2015). Finally, we control for ownership structure (Free

Float). The lower the proportion of corporate insiders, the less afraid corporate outsiders should

be of trading with a better-informed party. Because all of these additional controls are time-

11 We compute the 7 basis points as (e-0.103 – 1) multiplied by the post-event mean spread (69 basis points). This effect

roughly translates into a 10 percent decrease. Relative to the pre-event mean bid-ask spread, the increase is reduced by 14 percentage points computed as the difference between the full effect (+44 percent) and the mitigated effect (+30 percent, i.e., [e0.367-0.103 – 1]). Relative to the overall mean spread (48 basis points), the mitigated effect represents a reduction of 14 percent.

19

invariant, only the interactions with Post_SNB are identified and only the Post_SNB x

Num_Analysts term is significant and negative.

In Column (5), we add concurrent daily stock returns and its interaction with Post_SNB to the

model as another way of controlling for a firm’s actual risk exposure. Firms with large negative

price reactions (and, hence, a higher exposure) to the currency shock likely face more uncertainties

about the future. We find that, on average, contemporaneous stock returns are negatively related

to bid-ask spreads. The relation is more pronounced after the event (but the difference not

statistically significant). Throughout Columns (2) to (5), the Post_SNB × FXRisk_Disc term is

largely unaffected by the inclusion of the additional controls, suggesting that a firm’s size,

exposure to currency risk, general disclosure quality, or daily trends in the data do not act as

confounding factors. In Column (6), we use Return Variability as alternative proxy of information

asymmetry and the results are very similar.

We conduct the following additional sensitivity tests (not tabulated): (i) We include two

additional proxies for a firm’s general disclosure quality, namely the length of the annual report

(measured by the natural logarithm of the number of pages) and a binary indicator marking firms

that follow international accounting standards (IFRS or U.S. GAAP). International standards

prescribe more extensive (risk) disclosures and might proxy for a more transparent information

environment. (ii) We separately estimate Panel B of Table 2 for firms following local GAAP (33

firms) or international standards (118 firms). (iii) We control for the potentially confounding

effects of new public disclosures immediately after the event and create a binary indicator for firms

that release ad-hoc statements and press briefings or participate in pre-scheduled investor events

within days [0;+2] of the SNB announcement (69 firms). (iv) We add interaction terms between

Post_SNB and the time-variant liquidity controls (Market Value, Share Turnover, and Return

20

Variability) to the model. For each of these sensitivity tests, the main coefficient of interest,

Post_SNB × FXRisk_Disc, is significantly negative and none of the inferences change.12

3.3.2. Persistence of Effect in Daily Bid-Ask Spreads

We next examine the persistence of the differential bid-ask spread reaction. Theory predicts

that the informational value of past risk disclosures is transitory after the exogenous event and that

new and updated information sources (within and outside of the firm) quickly take its place. To

get an idea about the length of the informational benefits, we estimate a variant of Eq. (1) over a

[-30;+20] window, in which we include a set of indicator variables, Post_SNBt, for each day t Î

[0;+20]. The vector of control variables is the same as in Column (2) of Table 2, Panel B. In

Figure 4, we plot the individual Post_SNBt × FXRisk_Disc interaction terms (together with the 95

percent confidence intervals). The graph indicates that the mitigating effect of past risk disclosures

on information asymmetry is largest on the day of the SNB announcement (-0.125). It diminishes

to -0.028 on day t = +4, and hovers around -0.05 until trading day t = +10, before it dissipates.

Thus, it takes about two weeks for the differences in bid-ask spreads to disappear.13

The short-lived nature of the effects is consistent with expectations and increases our

confidence in the research design. The issue then becomes one of economic significance. We

deem a reduction of the sudden increase in information asymmetry following the SNB

announcement by 7 basis points (or about 14 percent relative to the overall mean spread; see

Footnote 11) for more transparent firms economically substantial. Compared to other settings with

a persistent change in disclosure quality and/or strict enforcement of disclosure regulation, the

12 With one exception the magnitudes of the Post_SNB × FXRisk_Disc terms in the sensitivity tests are very similar

to those tabulated. When estimating the models separately for local GAAP firms and firms with international standards, the coefficients of interest become more negative for the former (-0.163 based on model 2) and less negative but still significant for the latter firms (-0.031). This result is consistent with local GAAP firms having more discretion in their risk disclosure choices as well as a less transparent information environment to begin with.

13 The two-week time frame is consistent with the survey answers of sell-side analysts (see Section 4.1 and Table 5).

21

effects that we document are smaller (as one would expect) but still large enough to matter for

market participants.14 Moreover, we find that the average daily trading volume significantly

increases by more than 85 percent over the two weeks after the SNB announcement relative to the

benchmark period (not tabulated). Thus, the informational benefits of past risk disclosures come

at a time with high uncertainty and divergence in opinions among market participants.

3.3.3. Analyses of Intraday Bid-Ask Spreads

To examine the trading pattern on the event day, we estimate a version of Eq. (1), in which

we replace Post_SNB with 17 separate indicator variables for each 30-minute trading interval of

the event day (i.e., from 9 am to 5:30 pm). For instance, the 11:30-12:00 Dummy represents the

30 minutes from 11:30 am to noon. We then interact these time indicators with FXRisk_Disc to

estimate the differential bid-ask spread reaction conditional on the quality of past risk disclosures

over the course of the event day. The SNB announcement took place at 10:30 am. Thus, we expect

markets to react from this point on but not beforehand. All variables are computed similarly as

before, but with respect to 30-minute instead of daily trading intervals. The regression contains

firm and/or time-of-day fixed effects, and we assess the statistical significance based on robust

standard errors with two-way clustering by firm and trading interval. Table 3 reports the results.

In Column (1), we combine the day of the SNB announcement with the benchmark period

(i.e., t Î [-30; 0]), and include both firm and time-of-day fixed effects. The series of interaction

terms is defined by the within-trading interval, between-firm variation in spreads. The main effects

of the time indicator variables are small and only marginally significant before 10:30 am,

14 For instance, Daske et al. (2008) report a 6.4% to 15.1% reduction in bid-ask spreads around mandatory IFRS

adoption (see their Table 4). Christensen et al. (2013) find a decrease by 17.5% to 35.1% around substantial enforcement changes of disclosure regulation in countries of the European Union (see their Table 4). Christensen et al. (2016) report a reduction of about 20% following two major changes in securities laws (see their Internet appendix).

22

suggesting no abnormal activity relative to the benchmark period. Beginning with the 10:30 to

11:00 am interval, the main effects become significantly positive. The coefficient magnitudes

suggest substantive increases in bid-ask spreads ranging from 148 percent (12.30 to 1.00 pm) to

36 percent (5.00 to 5.30 pm). The interaction terms with FXRisk_Disc are not significant before

10:30 am, but then quickly turn negative and become significant from 11:00 am onwards. They

remain significant throughout the day. Within 30 minutes of the announcement, firms with better

quality risk disclosures experience a reduced effect on information asymmetry on the order of -19

percent (12.30 to 1.00 pm) to -6 percent (3.00 to 3.30 pm).

We repeat the analysis with event-day observations (t = 0) and report results in Column (2).

The time-of-day fixed effects subsume the main effects of the time indicator variables.15 The

results are very similar. While firms with higher quality risk disclosures have no information

advantage in the trading hours before the event, a significant gap in bid-ask spreads appears within

minutes of the unexpected announcement and persists for the remainder of the day. The intraday

results provide strong support for the stipulated link between the advent of fundamental news and

information asymmetry through the channel of a firm’s past disclosures. The reaction was simply

too quick for firms and analysts to release new information at such short notice.16

3.3.4. Analyses of Daily Returns

One of the biggest concerns for our analysis is that the result of a significantly negative

relation between FXRisk_Disc and bid-ask spreads reflects a firm’s exposure to the currency risk

and not necessarily the quality of past disclosures. In our information asymmetry tests in Section

15 We do not include firm fixed effects in this specification as they would subsume one of the interaction terms with

FXRisk_Disc and render the interpretation of the remaining interactions relative to this base period. 16 When we use trade-size to infer whether trades were executed by retail or institutional investors (e.g., Lee 1992),

we find that the fraction of small trades (< CHF 7,500) sharply and significantly decreases directly after the SNB announcement (not tabulated). This drop occurs within the first 30 minutes of the announcement and applies to all firms. Both sell-side analysts and investor-relations managers confirm in their answers to our survey that they communicated mainly with institutional investors immediately after the event.

23

3.3.1, we control for this possibility by including variables such as the proportion of international

sales, the correlation between firm returns and exchange rates, or contemporaneous stock returns.

Yet, from a valuation perspective, risk exposure should be what matters most around an exogenous

shock to firm fundamentals and lead to a negative relation between a firm’s currency risk exposure

and stock prices following the SNB announcement (i.e., through an impact on firms’ growth

opportunities). Disclosure quality, if anything, has a limited effect on this relation that would work

through the cost of capital channel (e.g., Lambert et al. 2012; Goldstein and Yang 2017).

We test this line of argument in two ways. First, we examine the stock price reaction around

the SNB announcement conditional on a firm’s risk disclosure quality. Figure 5 reports the results.

The graph plots mean daily stock returns for firms with high-quality risk disclosures (i.e.,

FXRisk_Disc greater than the sample median) and a synthetic control group in the 40 days

surrounding the event. The control group is constructed in a way that there is no difference in

stock returns over the [-20;-1] benchmark period (see discussion of Figure 3 for details). As the

graph indicates, there is no differential stock price reaction between the two matched groups in the

period after the currency shock.

Second, we directly estimate the relation between stock price, disclosure quality, and currency

risk exposure. To do so, we estimate a variant of Eq. (1) with the natural logarithm of Stock Price

as the dependent variable and report the results in Table 4.17 The significantly negative coefficient

for Post_SNB in Column (1), in which we include the continuous firm-level controls plus firm

fixed effects, shows that firms, on average, experience a decline in share prices of 6.2 percent

following the SNB announcement. Our proxy for risk disclosure quality is not significant,

suggesting that changes to the adverse selection component of cost of capital are not large enough

17 The natural logarithm of Stock Price as left-hand side variable in combination with firm fixed effects is similar to a

returns specification.

24

to move stock prices. The same holds when we include day fixed effects and allow the stock price

reaction to differ across small and large firms (Column 2). More to the point, in Columns (3) and

(4), we separately estimate the relation between our two proxies for currency risk exposure,

Int_Sales and Hist_Corr_EUR, and share price and find significantly negative interaction terms

with Post_SNB. The results indicate that market prices reflect a firm’s actual exposure to currency

rate fluctuations, as one would expect. In terms of economic magnitude, a one-standard-deviation

increase in the Int_Sales ratio (Hist_Corr_EUR) leads to an incremental reduction in stock price

of 2.6 (1.3) percent. These effects are substantial. When we combine all variables and include the

controls for overall disclosure quality (Total_Disc and Num_Analysts) and Free Float (Column 5)

as well as allow the day fixed effects to vary by industry (Column 6), the tenor of the results

remains the same. Only risk exposure (i.e., Int_Sales) but not risk disclosure quality is

significantly related to the stock price reaction. Combined with our results in Table 2, the evidence

is consistent with disclosure quality affecting information asymmetry – as predicted by theory –

but not the price reaction around the Swiss franc shock. On the other hand, risk exposure has a

first-order effect on prices, but not on information asymmetry.

4. Field Analysis of the Use of Past Risk Disclosures by Various Stakeholders

We next discuss various stakeholders’ use of financial statement information. Specifically,

we surveyed (1) sell-side financial analysts, (2) credit officers at lending banks, and (3) investor-

relations managers working at firms affected by the Swiss franc shock. The purpose of collecting

this data in the field is to provide context for the informational role of past risk disclosures, validate

the inferences drawn from the archival tests, and ultimately increase our confidence in the causality

of the stipulated relations (Soltes 2014; Bernstein et al. 2016).

25

4.1. Reaction of Sell-Side Financial Analysts to Swiss Franc Shock

Sell-side analysts are an important information intermediary between firms and investors, and

they should be among the first to assess the impact of the currency shock on firms’ future expected

cash flows. Our survey attempts to better understand their information processing and to examine

their use of past financial statement information in the wake of this exogenous event. We develop

the survey after an extensive review of the related literature (Graham et al. 2005; Dichev et al.

2013; Nelson and Skinner 2013; Brown et al. 2015; 2016), and pre-test it with several academics

and practitioners. The survey comprises 15 questions, divided into the four sections (1) general

information about the Swiss franc shock, (2) timeliness and motivation of reaction to the event,

(3) information sources employed, and (4) methodological approach to incorporate the information

in earnings forecasts and stock recommendations. A series of demographic questions concludes.

When we administered the survey, we use a neutral tone and only briefly explained the general

purpose but did not mention that we were studying the usefulness of financial statement disclosures

for the assessment of the Swiss franc shock. These measures should help avoid several well-

known survey biases (see, e.g., Orne 1962). We are primarily concerned about the “confirmation

bias” (e.g., Nickerson 1998) which leads respondents to seek out evidence supporting their priors

or ignore data that might disprove their beliefs, the “good-subject bias” (e.g., Nichols and Maner

2008) which primes respondents to behave in a manner as the experimenter expects, and the

“Hawthorne effect” which lets individuals modify their behavior in response to being observed

(e.g., Levitt and List 2011). Appendix B provides more details on the survey together with all the

results not explicitly tabulated in the main body of the study. We cross-reference the respective

survey question (Q) when discussing the results in the text.

Our pool of suitable survey subjects comprises all the sell-side analysts covering at least one

of the 151 sample firms during the SNB announcement. We receive 77 completed surveys,

26

resulting in a response rate of 9.4 percent.18 The median analyst in our sample covers 10 to 15

firms from up to three different industries. Firms domiciled in Switzerland make up about 25

percent of the average analyst portfolio. However, a quarter of the respondents exclusively cover

Swiss firms in their research. Analysts, on average, are between 30 and 50 years old, hold a

master’s degree, and have close to 10 years of experience. They work at large brokerage houses

with more than 25 sell-side analysts. About 40 percent each work for employers headquartered in

Switzerland or the rest of Europe. See Table B1 in Appendix B for details.

The Swiss franc shock caught financial analysts by surprise. Only 3 percent of the respondents

anticipated this type of move by the SNB within the next three months and only 20 percent within

a year (Q2). This result corresponds with the archival evidence in Figure 1, and points to the

randomness of the event. The currency shock had a big impact on firm fundamentals. 85 percent

of the respondents thought that almost all of the Swiss firms in their portfolios were impacted by

the event (Q3), and in all cases the effect was expected to be negative (Q4). The Swiss franc shock

recalibrated analysts’ stance towards currency risk. Before the event, only 36 percent considered

a typical firm’s currency risk exposure important information when forming earnings forecasts or

stock recommendations (Q1). This number soared to 52 percent after the event (Q15), which is

significantly higher than before.

The Swiss franc shock triggered a swift reaction from sell-side analysts. They reassessed the

prospects of almost their entire portfolio within the first two days of the SNB announcement

(Q5.1). For about 80 percent of the firms, the reassessment led to a material adjustment of the

qualitative inputs (40 percent) and/or quantitative inputs (70 percent) in the valuation models that

analysts use to derive their earnings forecasts and stock recommendations (Q5.2). The main

18 The response rate is comparable to other surveys of executives (8.4 percent in Graham et al. 2005), chief financial

officers (5.4 percent in Dichev et al. 2013), and financial analysts (10.9 or 7.1 percent in Brown et al. 2015, 2016).

27

drivers behind the rapid response were the increased information needs and, hence, demand from

clients as well as the presumed exposure of the firm to the currency shock (Q6). This answer

already points to the potential importance of risk disclosures for analysts’ decision-making in the

immediate aftermath of the event. Analysts continued to evaluate the impact of the Swiss franc

shock after the initial response. More than 60 percent of the analysts reviewed the prospects of

their entire portfolio in more detail in the two to three weeks that followed (Q7). This time pattern

nicely corresponds with the drift we observe in the archival tests. As Figure 4 illustrates, it took

markets about two weeks to fully bridge the gap in bid-ask spreads.

We next explore the role of financial statement information in the decision process. Table 5

lists the survey answers to the questions about the importance of various information sources for

the assessment of the Swiss franc shock of a typical firm immediately after the announcement by

the SNB (Q8) and within the following two to three weeks (Q9). We list the answers in decreasing

order of importance.19 As Panel A suggests, personal experience/industry knowledge and private

communication with firm management were considered extremely important during the advent of

the fundamental news. Existing annual reports or other financial filings rank third, not statistically

different from private communication or ad hoc announcements by the firm, but more important

than market sources, media coverage, or peer information. Thus, not only did analysts consult

annual reports to gather additional insights on the economic consequences of the currency shock,

but this resource also ranked highest in terms of publicly available information. Our result

contrasts with the information sources that analysts use under “normal” circumstances. Brown et

al. (2015) find that industry knowledge and private communication with management are always

considered important (see their Table 1). Yet, annual reports or other financial filings are less

19 When we administered the survey, we presented the options in random order to the subjects to avoid any order bias

(Nelson and Skinner 2013).

28

relevant and rank substantially lower. Panel B of Table 5 provides insights into the dynamics of

the information processing by financial analysts. In the two to three weeks that follow the SNB

announcement, existing annual reports and other filings lose ground (albeit not statistically

significant), while private communication and ad hoc announcements by the firms gain

importance. These answers are consistent with newer, updated information substituting for past

disclosures, which in the short-run were used to complement the advent of news.

Asked directly what they use financial statement information for (Q10), analysts provided the

responses listed in Table 6, Panel A. The main purpose of past disclosures was to better

contextualize and interpret the translational and transactional currency risk exposure of the firm

and, to a lesser degree, to assess the effectiveness of the hedging strategy in place. These stated

goals map into the construction of our FXRisk_Disc score that we use in the archival analyses.

Thus, when searching the annual report for relevant disclosures, analysts likely considered items

included in our empirical proxy. To further corroborate such a direct link between the survey

answers and the use of risk disclosures by sell-side analysts, we closely inspect analyst reports

released within 15 days of the SNB announcement. Specifically, we search these reports for

(anecdotal) evidence that analysts were using information from past annual reports when assessing

the impact of the Swiss franc shock. We found multiple incidents in which analysts were either

directly referencing annual reports as their data source or, indirectly, were using information and

numbers contained in firms’ annual reports (or, more to the point, in the sections that we considered

in the construction of FXRisk_Disc) for their analysis and computations (see Appendix C for

details and three illustrative examples). Collectively, this evidence is consistent with analysts

having past risk disclosures in mind when referring to accounting data in the survey answers.

Panel B of Table 6 indicates that a firm’s business complexity together with variation in the

quality of past disclosures pose the main obstacles in assessing the consequences of the currency

29

shock (Q12). About half of the respondents considered assessing the impact of the event

substantively more difficult for some firms than others (Q11). In terms of how analysts

incorporated the new information into their valuation models, 75 percent responded that they have

made quantitative input adjustments, while about 40 percent reconsidered the entire situation of

the firm (including strategic responses by management) or adjusted the valuation parameters

(Q14). In sum, the results from our survey of sell-side analysts underscore the importance of past

(risk) disclosures during the advent of new information and corroborate the evidence from the

archival analyses in Section 3.

4.2. Reaction of Lending Banks to Swiss Franc Shock

Lending banks represent another group of stakeholders affected by the Swiss franc shock,

especially if their borrowers conduct business abroad. Banks enjoy privileged access and closer

ties to their clients (e.g., Rajan 1992; Agarwal and Hauswald 2010), and given their longer horizon

than equity investors, they face less time pressure to respond to unexpected changes in risk. We

conducted a series of in-depth structured interviews with senior credit officers and risk managers

of the four largest commercial banks in Switzerland.20 The interviews lasted about an hour. In the

first part, the bank representatives explained how they handled the events of January 15. In the

second part, we guided them through a questionnaire similar in structure and length to the

analyst/investor-relations manager survey. As the sample size of the interviews does not lend itself

to statistical analysis, we provide a qualitative summary of these conversations.

After the SNB announcement (within one week for listed firms, two to four weeks for the

rest), all four banks reviewed their relevant loan portfolio, and classified it into different risk

categories depending on the materiality and perceived impact of the currency shock. For this initial

20 We conducted five interviews in total, as one of the banks divides its credit office into publicly listed and privately

held clients, and we interviewed senior management at both divisions.

30

assessment, banks incorporated any information already available to them. The main, and for non-

listed firms only, information source for this triage were existing financial disclosures including

annual reports submitted during the periodic loan review (e.g., 2013 annual reports, 2014 quarterly

reports, budgets for 2015). Only in rare instances and for large (listed) clients, loan officers

established a direct communication with the borrower during the initial period of uncertainty.

Managers at one bank mentioned that they were prepared for such an event because they had

constructed a so-called heat map (i.e., risk profile) of their customer base after the initial pegging

of the Swiss franc to the euro in September 2011.

The triage typically resulted in three categories of clients: (1) borrowers heavily affected by

the Swiss franc shock, (2) borrowers without material consequences, and (3) the firms considered

somewhere in between. The first group was small (sometimes only a handful of all corporate

clients), and the consensus among our interview partners was that the Swiss franc shock might

have been the tipping point, but rarely the root cause of the financial problems for these firms.

Sometimes, these clients were handed over to the recovery division within the bank. The second

group, without material impact, was surprisingly large (up to 80 percent in some cases) and

consisted primarily of smaller, private clients with a strong focus on domestic markets. No

immediate action was taken for these firms. The third group of firms, potentially risky cases,

underwent a more detailed review over the following months, and loan officers often engaged in

direct discourse with their clients to gather additional information on the impact of the currency

shock. In some cases, this information gathering was standardized using a questionnaire. Lending

banks found that larger companies, particularly when listed on an exchange, were easier to evaluate

because more (and better) information is available and the information is more recent. At the same

time, lenders considered larger firms to be less affected by the Swiss franc shock due to natural

31

currency hedges by means of their business model and their superior ability of dealing with

exchange rate fluctuations (e.g., by derivative hedging).

With respect to the usefulness of existing financial filings, our interviews yielded the

following insights. First, while lending banks typically have closer ties with their borrowers and

less time pressure to react than investors, they still relied on this information source to evaluate the

impact of the Swiss franc shock. In fact, their timing pattern is comparable to the sell-side analysts.

In the days immediately after the SNB announcement, past disclosures were a valuable source of

information. As time went by, banks replaced “hard” information from existing financial filings

with “soft” information from personal exchanges with the client. Second, two of the four banks

indicated that they had tightened their clients’ reporting requirements and frequency as a response

to the event. This tightening could take the form of requiring half-yearly or quarterly reporting,

and suggests that existing financial disclosure practices were not considered informative enough

in the advent of sudden news. No interview partner mentioned that the bank adjusted its credit

risk rating system following the Swiss franc shock.

4.3. Reaction of Investor-Relations Managers to Swiss Franc Shock

In our third field analysis, we shift the focus away from market participants, and survey

investor-relations (IR) managers working at the affected firms. The goal is to understand the

external communication needs after the Swiss franc shock, and the role that existing financial

statements might have played in this interaction. We develop a survey comprising 15 questions

and pre-test it with several academics and practitioners. The survey is divided into four parts: (1)

general information about the event, (2) timeliness of and motivation for communication with

external stakeholders, (3) means of communication, and (4) stakeholders’ information needs. A

simple demographic question concludes. When we administered the survey, we did not mention

32

that we were studying the usefulness of financial statement disclosures for the assessment of the

Swiss franc shock. Appendix D provides more details on the survey together with all the results

not explicitly tabulated in the main body of the study. We cross-reference the respective survey

question (Q) when discussing the results in the text.

We sent the survey to IR representatives at 149 of the 151 sample firms.21 The response rate

is 26.2 percent or 39 completed surveys. 68 percent of the respondents carry either the title of

head of IR, (senior) IR manager, or head of communications; 18 percent hold the position of chief

financial officer. The respondent firms are a good representation of the sample. On average, they

do not differ significantly from non-respondent firms in terms of the firm characteristics used in

the archival tests as well as with regard to the FXRisk_Disc score. IR managers working at Swiss

firms in January 2015 were surprised by the SNB proclamation. Only 5 percent expected such a

move within the next three months, 35 percent within a year (Q2). 60 percent of the respondents

expected a negative impact on the firm’s future expected cash flows, only 8 percent foresaw a

positive effect (Q3). At many firms, information about the firm’s foreign currency exposure

always belonged to the communication with external stakeholders (Q1). While its role did not

shift for firms that considered it an integral part of their communication strategy, currency risks

received more prominent coverage after the event by firms that used to largely ignore this issue in

the past (Q15).

Firms quickly reached out to external stakeholders following the event. 46 percent of IR

departments communicated with investors, analysts, etc. on the day of the SNB announcement, 64

percent within the first week (Q4). The interactions were typically both proactive and a reaction

to outside demand (Q7). The main goals were to reassure existing investors, reduce uncertainty in

the market place, build reputation, and promote trust in the firm’s reporting strategy (Q11). Table

21 We could not identify valid email addresses of investor-relations managers at two of the sample firms.

33

7, Panel A, lists the answers to the question about the internal and external recipients of the

communication (Q5). Financial analysts were the primary target group for information about the

currency shock, followed by institutional investors, and management or other internal departments.

Communication with lending banks, retail investors, and the external audit firm rank significantly

lower. This ordering is consistent with the answers in the analyst and bank surveys, and maps into

the bid-ask spread pattern that we observe in the archival tests (see Figure 3 and Figure 4). The

uncertainty in the markets was highest shortly after the event, but sophisticated investors and

knowledgeable information intermediaries could presumably distinguish between firms with more

and less precise disclosures.

Panel B of Table 7 presents the information sources used by the IR departments in preparation

of the external communication (Q6). Consultation with key management is ranked highest, trailed

by existing internal reports, and past annual reports or other financial filings. Only 3 percent of

the respondents did not deem the latter information source important. Next, we examine the use

of different communication channels. Table 8, Panel A, reports the answers. Private

communication with analysts and investors were the preferred means of communication in the

days immediately after the SNB announcement (Q8). Ad hoc disclosures and press releases were

only rarely used. This result is consistent with the answers from the sell-side analysts (see Table

5, Panel A), and suggests that sophisticated market participants were seeking and IR departments

fulfilling the need for non-formalized ways of conveying information that goes beyond past

disclosures. In the two to three weeks that followed, communication at already pre-scheduled

events gained importance (e.g., investor days or conference calls), but private communication

retained its role as key information source (Q9).

Panel B of Table 8 indicates reasons why firms did not pursue a more proactive

communication strategy (Q10). No clear rank order emerges. 44 percent of respondents indicated

34

that upcoming regular financial filings (e.g., the imminent release of the 2014 annual report) or

pre-scheduled events were important. Fears of disclosing proprietary data or setting a disclosure

precedent did not receive much weight, even though Graham et al. (2005) list them as the two

primary reasons limiting firms’ voluntary disclosure behavior (see their Table 12).

Finally, we investigate how IR managers perceived the external information needs. Corporate

outsiders were eager to obtain more information about the currency shock. Web traffic, emails,

and phone calls were above normal levels registered during the release of the firm’s most recent

quarterly or annual earnings (Q12). Information about the firm’s transactional and translational

exposure, followed by the nature and extent of the hedging strategy, were in particularly high

demand (Q13). These perceptions nicely map into what sell-side analysts considered important

information drawn from existing financial statements (see Table 6, Panel A). Asked more directly,

26 (67) percent of IR professionals believed that information in past disclosures was extremely (at

least somewhat) important to outside stakeholders when assessing the immediate consequences of

the Swiss franc shock (Q14). This favorable impression is broadly consistent with the evidence

presented throughout our study.

5. Conclusion

We use the Swiss franc shock as an ideal experimental setting to shed light on how disclosure

quality is related to information asymmetry among market participants and to get close to causal

identification. In archival analyses, we find that firms with better risk disclosures experience a

substantially mitigated increase in information asymmetry following the currency shock. This gap

in bid-ask spreads appears within 30 minutes of the announcement by the SNB and persists for

about two weeks. Surveying sell-side analysts, lending banks, and IR managers indicates that

existing disclosures are key to evaluate the translational and transactional effects of the currency

35

shock, and sometimes (e.g., for smaller, unlisted firms) are the only resource available at short

notice. The results imply that risk disclosures continue to attenuate information asymmetry and

the costs of adverse selection well beyond their initial publication date.

One implication of our study is that being transparent has benefits beyond simply providing

useful information at the initial release date. The same set of disclosures might be relevant in the

future as it provides contextual information in case the firm’s information environment changes.

The potential benefits of reduced uncertainty in posterior trading are highest for firms with more

transparent disclosures. This insight is important as it points to reduced costs of adverse selection

that could arise from selling shares in the future. On a more fundamental level, our results

highlight that managers should care about the long-term consequences of their disclosure

decisions. Even if information appears irrelevant now, it might become useful later. Thus,

preparers, regulators and practitioners should not only consider the current decision usefulness of

a certain disclosure, but also whether this information could become relevant at some point in case

a firm’s circumstances change.

36

References

Abadie, A., A. Diamond, and J. Hainmueller. 2010. Synthetic Control Methods for Comparative Case Studies: Estimating the Effect of California’s Tobacco Control Program. Journal of the American Statistical Association 105 (490): 493-505.

Abadie, A., and J. Gardeazabal. 2003. The Economic Costs of Conflict: A Case Study of the Basque Country. American Economic Review 93 (1): 113-132.

Acemoglu, D., S. Johnson, A. Kermani, J. Kwak, and T. Mitton. 2016. The Value of Connections in Turbulent Times: Evidence from the United States. Journal of Financial Economics 121 (2): 368-391.

Agarwal, S., and R. Hauswald. 2010. Distance and Private Information in Lending. Review of Financial Studies 23 (7): 2757-2788.

Amihud, Y., and H. Mendelson. 1986. Asset Pricing and the Bid-Ask Spread. Journal of Financial Economics 17 (2): 223-249.

Baiman, S., and R.E. Verrecchia. 1996. The Relation among Capital Markets, Financial Disclosure, Production Efficiency, and Insider Trading. Journal of Accounting Research 34 (1): 1-22.

Balakrishnan, K., M.B. Billings, B. Kelly, and A. Ljungqvist. 2014. Shaping Liquidity: On the Causal Effects of Voluntary Disclosure. Journal of Finance 69 (5): 2237-2278.

Ball, R.A.Y., and L. Shivakumar. 2008. How Much New Information Is There in Earnings? Journal of Accounting Research 46 (5): 975-1016.

Banker, R.D., S. Das, and C.S. Ou. 1995. Prior Financial Ratio Information and Differential Market Response to Regulatory Changes. Journal of Accounting, Auditing & Finance 10 (4): 699-717.

Beaver, W.H. 1968. The Information Content of Annual Earnings Announcements. Journal of Accounting Research: 67-92.

Begley, T.A., A. Purnanandam, and K. Zheng. 2017. The Strategic Underreporting of Bank Risk. The Review of Financial Studies 30 (10): 3376-3415.

Bernstein, S., X. Giroud, and R.R. Townsend. 2016. The Impact of Venture Capital Monitoring. Journal of Finance 71 (4): 1591-1622.

Blacconiere, W.G., and D.M. Patten. 1994. Environmental Disclosures, Regulatory Costs, and Changes in Firm Value. Journal of Accounting and Economics 18 (3): 357-377.

Bloomberg News. 2015. SNB Unexpectedly Gives up Cap on Franc, Lowers Deposit Rate, published on January 15, 2015 [retrieved March 9, 2015]. Available from http://www.bloomberg.com/news/articles/2015-01-15/snb-unexpectedly-gives-up-cap-on-franc-lowers-deposit-rate.

Bonadio, B., A.M. Fischer, and P. Saure. 2016. The Speed of Exchange Rate Pass-Through. Working Paper, Federal Reserve Bank of Dallas.

Bonetti, P., C.H. Cho, and G. Michelon. 2018. Environmental Disclosure and the Cost of Capital: Evidence from the Fukushima Nuclear Disaster. Working Paper, University of Exeter.

Brown, L.D., A.C. Call, M.B. Clement, and N.Y. Sharp. 2015. Inside the “Black Box” of Sell-Side Financial Analysts. Journal of Accounting Research 53 (1): 1-47.

Brown, L.D., A.C. Call, M.B. Clement, and N.Y. Sharp. 2016. The Activities of Buy-Side Analysts and the Determinants of Their Stock Recommendations. Journal of Accounting and Economics 62 (1): 139-156.

Brown, L.D., and M.S. Rozeff. 1978. The Superiority of Analyst Forecasts as Measures of Expectations: Evidence from Earnings. Journal of Finance 33 (1): 1-16.

Brown, S., S.A. Hillegeist, and K. Lo. 2009. The Effect of Earnings Surprises on Information Asymmetry. Journal of Accounting and Economics 47 (3): 208-225.

Campbell, J.L., H. Chen, D.S. Dhaliwal, H.-m. Lu, and L.B. Steele. 2014. The Information Content of Mandatory Risk Factor Disclosures in Corporate Filings. Review of Accounting Studies 19 (1): 396-455.

Cavallo, E., S. Galiani, I. Noy, and J. Pantano. 2013. Catastrophic Natural Disasters and Economic Growth. Review of Economics and Statistics 95 (5): 1549-1561.

37

Chan, K., Y.P. Chung, and H. Johnson. 1995. The Intraday Behavior of Bid-Ask Spreads for NYSE Stocks and CBOE Options. Journal of Financial and Quantitative Analysis 30 (3): 329-346.

Chordia, T., R. Roll, and A. Subrahmanyam. 2000. Commonality in Liquidity. Journal of Financial Economics 56 (1): 3-28.

Christensen, H.B., L. Hail, and C. Leuz. 2013. Mandatory IFRS Reporting and Changes in Enforcement. Journal of Accounting and Economics 56 (2-3): 147-177.

Christensen, H.B., L. Hail, and C. Leuz. 2016. Capital-Market Effects of Securities Regulation: Prior Conditions, Implementation, and Enforcement. The Review of Financial Studies 29 (11): 2885-2924.

Clinch, G., and R.E. Verrecchia. 2015. Voluntary Disclosure and the Cost of Capital. Australian Journal of Management 40 (2): 201-223.

Cuny, C., and S. Dube. 2018. Does Reporting Quality Moderate the Relation between Economic Changes and Changes in the Cost of Debt? Working Paper, NYU.

Daske, H., L. Hail, C. Leuz, and R. Verdi. 2008. Mandatory IFRS Reporting around the World: Early Evidence on the Economic Consequences. Journal of Accounting Research 46 (5): 1085-1142.

———. 2013. Adopting a Label: Heterogeneity in the Economic Consequences around IAS/IFRS Adoptions. Journal of Accounting Research 51 (3): 495-547.

Deloitte. 2015. The Deloitte CFO Survey: A Chill Wind Blows Q4 2014 Results. Zurich, Switzerland. Diamond, D.W., and R.E. Verrecchia. 1991. Disclosure, Liquidity, and the Cost of Capital. Journal of

Finance 46 (4): 1325-1359. Dichev, I.D., J.R. Graham, C.R. Harvey, and S. Rajgopal. 2013. Earnings Quality: Evidence from the Field.

Journal of Accounting and Economics 56 (2-3): 1-33. Drake, M.S., D.T. Roulstone, and J.R. Thornock. 2015. The Determinants and Consequences of

Information Acquisition via EDGAR. Contemporary Accounting Research 32 (3): 1128-1161. Drake, M.S., D.T. Roulstone, and J.R. Thornock. 2016. The Usefulness of Historical Accounting Reports.

Journal of Accounting and Economics 61 (2-3): 448-464. Efing, M., R. Fahlenbrach, C. Herpfer, and P. Krüger. 2016. How Do Investors and Firms React to an

Unexpected Currency Appreciation Shock? Working Paper, Swiss Finance Institute. Francis, J., D. Nanda, and P.E.R. Olsson. 2008. Voluntary Disclosure, Earnings Quality, and Cost of

Capital. Journal of Accounting Research 46 (1): 53-99. Goldstein, I., and L. Yang. 2017. Information Disclosure in Financial Markets. Annual Review of Financial

Economics 9 (1): 101-125. Graham, J.R., C.R. Harvey, and S. Rajgopal. 2005. The Economic Implications of Corporate Financial

Reporting. Journal of Accounting and Economics 40 (1-3): 3-73. Heflin, F., and D. Wallace. 2017. The Bp Oil Spill: Shareholder Wealth Effects and Environmental

Disclosures. Journal of Business Finance & Accounting 44 (3-4): 337-374. Heinle, M.S., and K.C. Smith. 2017. A Theory of Risk Disclosure. Review of Accounting Studies 22 (4):

1459-1491. Indjejikian, R.J. 1991. The Impact of Costly Information Interpretation on Firm Disclosure Decisions.

Journal of Accounting Research 29 (2): 277-301. Jorion, P. 2002. How Informative Are Value-at-Risk Disclosures? The Accounting Review 77 (4): 911-931. Kim, O., and R.E. Verrecchia. 1994. Market Liquidity and Volume around Earnings Announcements.

Journal of Accounting and Economics 17 (1-2): 41-67. Kim, O., and R.E. Verrecchia. 1997. Pre-Announcement and Event-Period Private Information. Journal of

Accounting and Economics 24 (3): 395-419. Lambert, R., C. Leuz, and R.E. Verrecchia. 2007. Accounting Information, Disclosure, and the Cost of

Capital. Journal of Accounting Research 45 (2): 385-420. Lambert, R.A., C. Leuz, and R.E. Verrecchia. 2012. Information Asymmetry, Information Precision, and

the Cost of Capital. Review of Finance 16 (1): 1-29. Lee, C.M. 1992. Earnings News and Small Traders: An Intraday Analysis. Journal of Accounting and

Economics 15 (2): 265-302. Leuz, C., and C. Schrand. 2009. Disclosure and the Cost of Capital: Evidence from Firms' Responses to the

Enron Shock. NBER Working Paper.

38

Leuz, C., and R.E. Verrecchia. 2000. The Economic Consequences of Increased Disclosure. Journal of Accounting Research 38 (Supplement): 91-124.

Levitt, S.D., and J.A. List. 2011. Was There Really a Hawthorne Effect at the Hawthorne Plant? An Analysis of the Original Illumination Experiments. American Economic Journal: Applied Economics 3 (1): 224-238.

McInish, T.H., and R.A. Wood. 1992. An Analysis of Intraday Patterns in Bid Ask Spreads for NYSE Stocks. Journal of Finance 47 (2): 753-764.

Mirkov, N., I. Pozdeev, and P. Söderlind. 2017. Toward Removal of the Swiss Franc Cap: Market Expectations and Verbal Interventions. Working Paper, University of St. Gallen.

Nelson, M.W., and D.J. Skinner. 2013. How Should We Think About Earnings Quality? A Discussion of “Earnings Quality: Evidence from the Field”. Journal of Accounting and Economics 56 (2-3): 34-41.

Ng, J., T.O. Rusticus, and R.S. Verdi. 2008. Implications of Transaction Costs for the Post–Earnings Announcement Drift. Journal of Accounting Research 46 (3): 661-696.

Nichols, A.L., and J.K. Maner. 2008. The Good-Subject Effect: Investigating Participant Demand Characteristics. The Journal of General Psychology 135 (2): 151-166.

Nickerson, R.S. 1998. Confirmation Bias: A Ubiquitous Phenomenon in Many Guises. Review of General Psychology 2 (2): 175-220.

Orne, M.T. 1962. On the Social Psychology of the Psychological Experiment: With Particular Reference to Demand Characteristics and Their Implications. American Psychologist 17 (11): 776-783.

Rajan, R.G. 1992. Insiders and Outsiders: The Choice between Informed and Arm's-Length Debt. Journal of Finance 47 (4): 1367-1400.

Rogers, J.L., D.J. Skinner, and S.L. Zechman. 2017. Run EDGAR Run: SEC Dissemination in a High-Frequency World. Journal of Accounting Research 55 (2): 459-505.

Shroff, N., A.X. Sun, H.D. White, and W. Zhang. 2013. Voluntary Disclosure and Information Asymmetry: Evidence from the 2005 Securities Offering Reform. Journal of Accounting Research 51 (5): 1299-1345.

SNB. 2011a. Swiss National Bank Takes Measures against Strong Swiss Franc, published on August 3, 2011 [retrieved September 9, 2015]. Available from http://www.snb.ch/en/mmr/reference/pre_20110803/source/pre_20110803.en.pdf.

———. 2011b. Swiss National Bank Sets Minimum Exchange Rate at CHF 1.20 Per Euro, published on September 6, 2011 [retrieved September 9, 2015]. Available from http://www.snb.ch/en/mmr/reference/pre_20110906/source/pre_20110906.en.pdf.

———. 2014. Swiss National Bank Introduces Negative Interest Rates, published on September 15, 2015 [retrieved September 9, 2015]. Available from http://www.snb.ch/en/mmr/reference/pre_20150115/source/pre_20150115.en.pdf.

———. 2015. Swiss National Bank Discontinues Minimum Exchange Rate and Lowers Interest Rate to -0.75%, published on September 15, 2015 [retrieved September 9, 2015]. Available from http://www.snb.ch/en/mmr/reference/pre_20150115/source/pre_20150115.en.pdf.

Soltes, E. 2014. Incorporating Field Data into Archival Research. Journal of Accounting Research 52 (2): 521-540.

Verrecchia, R.E. 2001. Essays on Disclosure. Journal of Accounting and Economics 32 (1): 97-180.

Figure 1: Euro-Swiss Franc Exchange Rates and Swiss All Share Index Over 2011 to 2015 Period

The figure plots the euro to Swiss franc (EUR-CHF) exchange rates (left-hand scale) and the Swiss All Share Index (symbol: SSIP; right-hand scale) over the years 2011 to 2015. The SSIP includes all shares of companies listed on the Swiss exchange SIX. Exchange rates and index values are from Datastream. The graph also indicates the dates when the Swiss National Bank (SNB) established the minimum exchange rate of CHF 1.20 per euro on September 6, 2011, and subsequently abandoned this lower threshold on January 15, 2015.

September 6, 2011 January 15, 2015

4,00

06,

000

8,00

010

,000

SS

IP S

wis

s Al

l Sha

re In

dex

1.0

1.1

1.2

1.3

EUR

-CH

F Ex

chan

ge R

ate

2011 2012 2013 2014 2015

EUR-CHF Exchange Rate SSIP Swiss All Share Index

Figure 2: Timeline of Events and Overview of Research Design for Archival Analysis

The figure illustrates our research design for the archival part of the analysis. In the regression analyses, we compare bid-ask spreads during the event window to the benchmark period using the Post_SNB indicator variable. We then test for cross-sectional differences in bid-ask spreads between firms with high and low foreign exchange risk disclosure scores (FXRisk_Disc). We construct FXRisk_Disc by scoring seven disclosure items that capture a firm’s exposure to foreign exchange risk as reported in the most recent annual report before the event (see Appendix A for details).

Time

Benchmark window[days: -30 to -1]

Event window[days: 0 to +2]

(variable: Post_SNB )

Measurement of partitioning variable

Transparency of foreign exchange risk disclosuresbased on annual reports (variable: FXRisk_Disc )

Event date: January 15, 2015,

10:30 amSNB discontinues

minimum EUR-CHF exchange rate

February, March, and April 2015

Publication of annual report for FYE 2014

December 31, 2013

December 31, 2014

Fiscal year end (FYE) 2014

Fiscal year end (FYE) 2013

February, March, and April 2014

Publication of annual report for FYE 2013

!!!

Figure 3: Information Asymmetry around Swiss Franc Shock (Conditional on Risk Disclosure Quality)

The figure maps out changes in information asymmetry as measured by bid-ask spreads in the 40 trading days (i.e., days -20 to +20) surrounding the SNB announcement to discontinue the minimum EUR-CHF exchange rate on January 15, 2015 (day 0). The sample comprises 151 firms listed on SIX (see Table 1). We plot the daily mean bid-ask spreads conditional on the quality of foreign exchange risk disclosures. That is, we compare the sample firms with an above median value of the foreign exchange risk disclosure score (FXRisk_Disc) to a (synthetic) control group. We construct the control group by matching each firm with an above median value of FXRisk_Disc to a combination of all firms with a score at or below the median (benchmark pool) using a weighting matrix that minimizes the difference in the outcome variable during the pre-event period. By construction, the difference in bid-ask spreads between the two groups is not significant before the event.

-1.4

-1.2

-1-.8

-.6

Aver

age

Log(

Bid-

Ask

Spre

ad)

-20 -10 0 10 20

Trading Day

High FXRisk_Disc FirmsControl Firms

Figure 4: Persistence of Information Asymmetry Effects for High Quality Exchange Risk Disclosure Firms

The figure maps out the persistence of the information asymmetry advantages for the firms with high quality foreign exchange risk disclosure scores (FXRisk_Disc) following the SNB announcement to discontinue the minimum EUR-CHF exchange rate. We estimate a bid-ask spread model similar to the base specification (see column 2 in Table 2, Panel B) over a [-30;+20] window and plot the coefficient estimates of the interaction terms between Post_SNBt and FXRisk_Disc in the figure. Post_SNBt is a set of indicator variables separately marking each day of the post period [0;+20]. The graph also shows 95% confidence intervals comparing the coefficients to zero (based on robust standard errors clustered by firm and day).

-.15

-.1-.0

50

.05

.1

Even

t Day

x F

XRis

c_D

isc

Coe

ffici

ents

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Event Day

Figure 5: Market Performance around Swiss Franc Shock (Conditional on Risk Disclosure Quality)

The figure maps out changes in market performance as measured by stock returns in the 40 trading days (i.e., days -20 to +20) surrounding the SNB announcement to discontinue the minimum EUR-CHF exchange rate (day 0). The sample comprises 151 firms listed on SIX (see Table 1). We plot the daily mean stock returns conditional on the quality of foreign exchange risk disclosures. That is, we compare the sample firms with an above median value of the foreign exchange risk disclosure score (FXRisk_Disc) to a (synthetic) control group (see also Figure 3). By construction, the difference in stock returns between the two groups is not significant before the event.

-8-6

-4-2

02

Aver

age

Ret

urn

(%)

-20 -10 0 10 20

Trading Day

High FXRisk_Disc FirmsControl Firms

Table 1: Sample Selection and Descriptive Statistics for Archival Analysis Panel A: Sample Selection Procedure

Data Requirements Number of Firms Members of the Swiss All Share Index as published by the Swiss exchange SIX on 31.12.2014 (symbol: SSIP) 235

– thinly traded firms (i.e., firms with, on average, less than ten trades per day over the period October 2014 to April 2015) -62

– firms with narrow market depth (i.e., firms whose bid-ask spreads were not updated on more than 97.5% of trading days over the period October 2014 to April 2015)

-14

– firms with missing data for variables used in the regression analysis -8 Final sample 151

Panel B: Descriptive Statistics for Variables Used in Daily Regression Analyses [Days: -30 to +2]

(N = 4,949) Mean Std. Dev. P1 P25 Median P75 P99 Dependent Variables: Bid-Ask Spreadt 0.482 1.107 0.050 0.144 0.264 0.471 3.774 Return Variabilityt 0.528 0.660 0.052 0.235 0.351 0.549 3.300 Stock Pricet 373 750 0.27 37 107 333 4,485 Independent Variables: Post_SNBt 0.091 0.288 0 0 0 0 1 FXRisk_DiscAR 4.027 1.425 1 3 4 5 7 Market Valuet-1 8,870 32,488 26 444 1,263 4,569 232,508 Share Turnovert-1 0.230 0.772 0.002 0.042 0.114 0.242 1.967 Int_SalesAR 0.654 0.386 0.000 0.285 0.876 0.980 1.000 Hist_Corr_EUR 0.224 0.110 -0.119 0.159 0.230 0.299 0.443 Total_DiscAR 3.665 0.747 2.037 3.180 3.632 4.173 5.102 Num_Analysts 7.744 8.534 0 2 4 11 32 Free FloatAR 0.663 0.249 0.110 0.450 0.690 0.890 1.000 Stock Returnt -0.256 2.737 -11.083 -0.894 0.000 0.831 5.500

(Continued)

Table 1 (continued) The table presents details on the sample selection (Panel A) and descriptive statistics for the variables used in the daily regression analyses (Panel B). The final sample of 151 firms gives rise to 4,949 firm-day observations over the [-30;+2] day period surrounding the SNB announcement to discontinue the minimum EUR-CHF exchange rate. We measure daily Bid-Ask Spreads as the mean of minute-by-minute intraday differences between bid and ask quotes (divided by the mid-point). We compute Return Variability as the standard deviation of stock returns over half-hourly intervals on a given trading day. Stock Price is the closing price at the end of each trading day (in CHF). Post_SNB is a binary indicator that takes on the value of ‘1’ beginning on day t = 0 of the SNB announcement (January 15, 2015). To measure the quality of a firm’s foreign exchange risk disclosures, we construct FXRisk_Disc by scoring the following seven items as reported in the most recent annual report before the event (as indicated by the subscript AR): (i) revenues, (ii) assets, and (iii) costs and profits generated and held outside of Switzerland; (iv) the currency distribution of short-term monetary assets and liabilities; (v) the exposure to and (vi) the hedging strategy with regard to foreign currency risk; (vii) the sensitivity of net income or equity to changes in foreign exchange rates. We standardize the raw scores, ranging from 0 (worst) to 7 (best), for use in the regression analysis. See Appendix A for details. Market Value is the number of shares outstanding times the stock price at the end of each trading day (in CHF million). We also compute the mean market value over the [-30;-1] day period and mark the variable with a subscript (Market Value-30,-1; not tabulated). Share Turnover is the daily CHF trading volume divided by the market value at the end of the previous trading day. Int_Sales is the percentage of sales generated outside of Switzerland as shown in the most recent annual report. When exact data were missing (32 cases), we infer international sales from textual and other disclosures in the annual report. Hist_Corr_EUR is the Pearson correlation between a firm’s weekly stock returns and the EUR-CHF exchange rates over the three years leading up to the day when the minimum exchange rate of CHF 1.20 per euro was established (September 6, 2011). When return data were missing over that period (4 cases), we use the correlation between weekly stock returns and USD-CHF exchange rates over the three years leading up to the event day (t = 0). Total_Disc is a score, ranging from 1 (worst) to 6 (best), ranking the overall quality of a firm’s annual report taken from the 2014 rating as published by the Institute for Banking and Finance of the University of Zurich. Num_Analysts is the number of analysts in I/B/E/S that cover the firm (i.e., provide a one-year-ahead earnings per share forecast) before the SNB announcement. If analyst data are missing, we set the variable to zero. Free Float is the percentage of shares available to ordinary investors at the end of the most recent fiscal year (source: Datastream). We compute daily Stock Return as the natural logarithm of price at the end of trading over price at the end of the previous day. Stock price, spread, and volume data are from Bloomberg. For expositional purposes, we multiply bid-ask spreads, share turnover, return variability, and stock returns by 100, expressing them in basis points. We do not winsorize or truncate the data.

Table 2: Analysis of Daily Changes in Information Asymmetry around Swiss Franc Shock Panel A: Difference-in-Differences Analysis of Daily Bid-Ask Spreads

(N = 4,949) [Days: -30 to +2]

Before SNB Announcement

[-30;-1] (a)

After SNB Announcement

[0;+2] (b)

Difference (b) – (a)

Low FXRisk_Disc (scores ≤ median) (i) 0.528 0.867 0.339***

High FXRisk_Disc (scores > median) (ii) 0.393 0.521 0.128***

Difference (ii) – (i) -0.135 -0.346** -0.211***

(Continued)

Table 2 (continued) Panel B: OLS Regression Analysis of Daily Bid-Ask Spreads

(N = 4,949) [Days: -30 to +2]

(1) Log(Bid-Ask

Spread)

(2) Log(Bid-Ask

Spread)

(3) Log(Bid-Ask

Spread)

(4) Log(Bid-Ask

Spread)

(5) Log(Bid-Ask

Spread)

(6) Return

Variability Test Variables: Post_SNB 0.367*** – – – – – (3.30) Post_SNB × FXRisk_Disc -0.103*** -0.097*** -0.105*** -0.103*** -0.104*** -0.130*** (-5.36) (-4.18) (-4.76) (-4.34) (-4.21) (-3.01) Control Variables: Log(Market Valuet-1) -0.243*** -0.302*** -0.286*** -0.289*** -0.300*** -0.190 (-3.94) (-4.56) (-4.89) (-4.81) (-4.31) (-1.43) Log(Share Turnovert-1) -0.028*** -0.022** -0.023** -0.023** -0.023** 0.018** (-3.00) (-2.52) (-2.59) (-2.58) (-2.54) (2.45) Log(Return Variabilityt-1) 0.061** 0.091*** 0.089*** 0.089*** 0.089*** 0.038** (2.49) (7.77) (7.48) (7.30) (7.47) (2.07) Post_SNB × Log(Market Value-30,-1) – -0.024 -0.024 0.005 0.003 0.018 (-1.35) (-1.35) (0.22) (0.15) (0.29) Post_SNB × Int_Sales – – – 0.019 -0.002 – (0.26) (-0.03) Post_SNB × Hist_Corr_EUR – – – -0.170 -0.181 – (-0.71) (-0.73) Post_SNB × Total_Disc – – – 0.007 0.007 – (0.17) (0.19) Post_SNB × Num_Analysts – – – -0.005** -0.005** – (-2.06) (-2.04) Post_SNB × Free Float – – – -0.149 -0.152 – (-1.27) (-1.42) Stock Returnt – – – – -0.005** – (-2.54) Post_SNB × Stock Returnt – – – – -0.004 – (-1.30) Fixed Effects Firm Firm &

Day Firm &

Industry-Day Firm &

Industry-Day Firm &

Industry-Day Firm &

Day Adjusted R2 (within) 0.210 0.051 0.050 0.053 0.055 0.013

The sample comprises 151 firms that give rise to 4,949 firm-day observations over the [-30;+2] day period surrounding the SNB announcement to discontinue the minimum EUR-CHF exchange rate. In Panel A, we report mean daily Bid-Ask Spreads for firms with high versus low values of the foreign exchange risk disclosure score (FXRisk_Disc) before and after the SNB announcement. We assess the statistical significance of differences across cells with t-tests based on robust standard errors clustered by firm. In Panel B, we report OLS coefficient estimates and (in parentheses) t-statistics based on robust standard errors clustered by firm and day. We use the daily Bid-Ask Spread (Return Variability in column 6) as the dependent variable. For variable details see Table 1. We include firm, day, or one-digit SIC industry-day fixed effects in the regressions, but do not report the coefficients. If indicated, we use the natural log of the raw values and lag the variables by one day. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels (two-tailed).

Table 3: Analysis of Intraday Changes in Information Asymmetry around Swiss Franc Shock

Log(Bid-Ask Spread)

as Dependent Variable

(1)

[Days: -30 to 0] (2)

[Days: 0]

Test Variables: 09:00-09:30 Dummy -0.025* (-1.73) – 09:00-09:30 Dummy × FXRisk_Disc -0.005 (-0.23) -0.022 (-1.00) 09:30-10:00 Dummy 0.022 (1.49) – 09:30-10:00 Dummy × FXRisk_Disc -0.037 (-1.63) -0.045 (-1.52) 10:00-10:30 Dummy 0.030** (2.00) – 10:00-10:30 Dummy × FXRisk_Disc 0.006 (0.45) -0.003 (-0.14) 10:30-11:00 Dummy 0.701*** (22.71) – 10:30-11:00 Dummy × FXRisk_Disc -0.057 (-1.55) -0.079* (-1.92) 11:00-11:30 Dummy 0.902*** (27.17) – 11:00-11:30 Dummy × FXRisk_Disc -0.122*** (-3.06) -0.159*** (-3.86) 11:30-12:00 Dummy 0.753*** (26.34) – 11:30-12:00 Dummy × FXRisk_Disc -0.074*** (-2.63) -0.107** (-2.91) 12:00-12:30 Dummy 0.836*** (25.05) – 12:00-12:30 Dummy × FXRisk_Disc -0.146*** (-3.95) -0.164*** (-4.23) 12:30-13:00 Dummy 0.911*** (26.78) – 12:30-13:00 Dummy × FXRisk_Disc -0.206*** (-6.10) -0.197*** (-6.21) 13:00-13:30 Dummy 0.730*** (20.12) – 13:00-13:30 Dummy × FXRisk_Disc -0.183*** (-5.14) -0.202*** (-4.66) 13:30-14:00 Dummy 0.693*** (19.99) – 13:30-14:00 Dummy × FXRisk_Disc -0.169*** (-5.12) -0.179*** (-4.73) 14:00-14:30 Dummy 0.615*** (22.62) – 14:00-14:30 Dummy × FXRisk_Disc -0.137*** (-4.76) -0.142*** (-4.10) 14:30-15:00 Dummy 0.486*** (22.85) – 14:30-15:00 Dummy × FXRisk_Disc -0.117*** (-4.94) -0.131*** (-4.26) 15:00-15:30 Dummy 0.426*** (18.61) – 15:00-15:30 Dummy × FXRisk_Disc -0.061** (-2.15) -0.072* (-2.12) 15:30-16:00 Dummy 0.373*** (16.51) – 15:30-16:00 Dummy × FXRisk_Disc -0.072*** (-2.76) -0.080** (-2.40) 16:00-16:30 Dummy 0.328*** (13.80) – 16:00-16:30 Dummy × FXRisk_Disc -0.096*** (-4.08) -0.133*** (-3.94) 16:30-17:00 Dummy 0.386*** (15.05) – 16:30-17:00 Dummy × FXRisk_Disc -0.127*** (-4.92) -0.151*** (-4.08) 17:00-17:30 Dummy 0.309*** (14.14) – 17:00-17:30 Dummy × FXRisk_Disc -0.139*** (-6.29) -0.172*** (-5.02) Control Variables Included Included Fixed Effects Firm & Time-of-Day Time-of-Day Adjusted R2 (within) 0.101 0.626 N 64,012 2,112

The sample comprises 151 firms that give rise to 64,012 (2,112) firm-intraday observations over the [-30;0] day period leading up to (on day t = 0 of) the SNB announcement to discontinue the minimum EUR-CHF exchange rate. The announcement took place at 10:30 am on January 15, 2015. We use the intraday Bid-Ask Spread as dependent variable (log transformed), and compute it as the mean of minute-by-minute differences between bid and ask quotes (divided by the mid-point) over 30-minute intervals. For each 30-minute interval i during trading hours (i.e., from 9 am to 5:30 pm) we construct a binary indicator that takes on the value of ‘1’ during that time slot. For instance, the 11:30-12:00 Dummy represents the 30 minutes from 11:30 am to noon. Technically, the first time slot of the day only comprises 28 minutes as SIX opens trade in each share at random within the first two minutes; the last slot of the day only comprises 20 minutes before the closing auction begins. To measure the quality of a firm’s foreign exchange risk disclosures, we construct FXRisk_Disc by scoring seven disclosure items as reported in the most recent annual report before the event. We include the natural logs of Market Value, Share Turnover, and Return Variability (computed similarly as in Table 1, but at the end of each trading interval i) as control variables, each lagged by 17 trading intervals, as well as firm and time-of-day fixed effects in the regressions, but do not report the coefficients. The table reports OLS coefficient estimates and (in parentheses) t-statistics based on robust standard errors clustered by firm and trading interval. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels (two-tailed).

Table 4: Analysis of Daily Returns around Swiss Franc Shock

(N = 4,949) [Days: -30 to +2]

(1) Log(Stock

Price)

(2) Log(Stock

Price)

(3) Log(Stock

Price)

(4) Log(Stock

Price)

(5) Log(Stock

Price)

(6) Log(Stock

Price) Test Variables: Post_SNB -0.062*** – – – – – (-6.39) Post_SNB × FXRisk_Disc -0.005 -0.001 – – -0.003 -0.001 (-1.49) (-0.35) (-0.95) (-0.19) Risk Exposure Variables: Post_SNB × Int_Sales – – -0.067*** – -0.063*** -0.075*** (-5.62) (-5.17) (-4.61) Post_SNB × Hist_Corr_EUR – – – -0.115** -0.013 0.012 (-2.59) (-0.31) (0.30) Control Variables: Log(Market Valuet-1) 0.193 0.173 0.168 0.172 0.168 0.161 (1.40) (1.38) (1.38) (1.37) (1.38) (1.41) Log(Share Turnovert-1) 0.004 0.004 0.003 0.004 0.003 0.003 (1.53) (1.53) (1.39) (1.51) (1.42) (1.14) Log(Return Variabilityt-1) -0.009** -0.004* -0.003 -0.004* -0.003 -0.003 (-2.21) (-1.83) (-1.61) (-1.81) (-1.63) (-1.32) Post_SNB × Log(Market Value-30,-1) – -0.013*** -0.010*** -0.009*** -0.006 -0.007 (-4.09) (-3.38) (-2.78) (-1.24) (-1.46) Post_SNB × Total_Disc – – – – -0.004 -0.006 (-0.68) (-0.97) Post_SNB × Num_Analysts – – – – -0.001 -0.001 (-0.59) (-0.74) Post_SNB × Free Float – – – – -0.013 -0.011 (-0.80) (-0.69)

Fixed Effects Firm Firm & Day

Firm & Day

Firm & Day

Firm & Day

Firm & Industry-Day

Adjusted R2 (within) 0.357 0.193 0.226 0.199 0.227 0.226

The sample comprises 151 firms that give rise to 4,949 firm-day observations over the [-30;+2] day period surrounding the SNB announcement to discontinue the minimum EUR-CHF exchange rate. The table reports OLS coefficient estimates and (in parentheses) t-statistics based on robust standard errors clustered by firm and day. We use the daily Stock Price as the dependent variable. For variable details see Table 1. We include firm, day, or one-digit SIC industry-day fixed effects in the regressions, but do not report the coefficients. If indicated, we use the natural log of the raw values and lag the variables by one day. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels (two-tailed).

Table 5: Survey of Sell-Side Analysts – Information Sources to Assess Economic Impact of Swiss Franc Shock Panel A: Short-term Information Sources Question 8: Immediately after the SNB announcement (i.e., within the first one or two days), how important were the following information sources for your assessment of the impact of the Swiss franc shock on a typical firm you followed? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 76)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Personal experience including company/industry knowledge 5.84 *** 3-9 72.00 5.33 (2) Private communication with management or investor relations 5.61 *** 4-9 68.42 10.53 (3) Existing annual reports or other financial filings 5.35 *** 1 and 5-9 57.33 9.33 (4) Ad hoc information provided by the firm after the shock 4.91 *** 1-3 and 5-9 48.68 18.42 (5) Commercial data providers (e.g., Bloomberg) 4.00 1-4 and 7-9 24.32 24.32 (6) Stock price reactions 3.97 1-4 and 8-9 27.03 29.73 (7) Peer firms (e.g., other firms’ ad hoc information) 3.42 *** 1-6 and 8-9 9.46 35.14 (8) Media coverage (e.g., press articles, TV features) 2.76 *** 1-7 and 9 5.41 54.05 (9) Peer analysts (e.g., other analysts’ forecasts revisions) 2.14 *** 1-8 0.00 67.12

Panel B: Medium-term Information Sources Question 9: Relative to the previous question, which information sources gained or lost importance in the two to three weeks following the SNB announcement for your assessment of the impact of the Swiss franc shock on a typical firm you followed? [3-point Likert Scale: 1 to 3]

Responses (Maximum possible N = 76)

Significantly Different Than [2 = Similar Importance]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Gained Importance

[3]

Lost Importance

[1] (1) Private communication with management or investor relations 2.37 *** 3-9 44.74 7.89 (2) Ad hoc information provided by the firm after the shock 2.32 *** 3-9 36.00 4.00 (3) Personal experience including company/industry knowledge 2.13 ** 1-2 and 6-9 22.67 9.33 (4) Stock price reactions 2.03 1-2 and 9 24.00 21.33 (5) Peer firms (e.g., other firms’ ad hoc information) 1.97 1-2 and 8-9 14.86 17.57 (6) Existing annual reports or other financial filings 1.89 1-3 13.33 24.00 (7) Commercial data providers (e.g., Bloomberg) 1.86 ** 1-3 6.76 20.27 (8) Peer analysts (e.g., other analysts’ forecasts revisions) 1.80 *** 1-5 6.76 27.03 (9) Media coverage (e.g., press articles, TV features) 1.76 *** 1-5 8.11 32.43

The sample comprises up to 76 answers from respondents to a survey sent to sell-side analysts covering at least one firm listed on SIX during the SNB announcement. The table lists the average ratings (in decreasing order) of the responses to the survey question indicated in the table header as well as the percentages of responses in the two tails of the Likert scale. We also report the results from t-tests comparing the average rating of an item to the mid-point (where ***, **, and * stand for statistical significance at the 1%, 5%, and 10% levels) and to the average ratings of all the other items. For the latter test, we report the item numbers that are statistically different at the 10% level, using Bonferroni-Holm adjusted p-values to correct for multiple pair-wise comparisons.

Table 6: Survey of Sell-Side Analysts – Role of Financial Statements in Assessing Economic Impact of Swiss Franc Shock Panel A: Usefulness of Existing Annual Report Disclosures Question 10: How important were existing annual reports and other financial filings for your assessment of the following dimensions of a firm’s currency exposure immediately after the Swiss franc shock (i.e., within the first one or two days)? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 77)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Translational exposure (e.g., arising from foreign subsidiaries) 5.84 *** 3-6 73.33 6.67 (2)

Transactional exposure (e.g., currency mismatch between

revenues and costs) 5.84

***

3-6

74.03

7.79

(3) Hedging strategy 4.80 *** 1-2 and 4-6 36.84 13.16 (4) One-time gains/losses (e.g., on cash holdings) 4.36 1-2 36.84 25.00 (5) Operating and strategic responses (e.g., relocation decisions) 3.93 1-3 26.67 32.00 (6) Sensitivity to indirect effects (e.g., GDP growth) 3.89 1-3 19.74 28.95

Panel B: Firm-level Heterogeneity in the Usefulness of Existing Annual Report Disclosures Question 12: Why was it more difficult to assess the impact of the Swiss franc shock for some firms than for others? Due to differences in the … [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 69)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Complexity of firms’ operations 5.71 *** 3-4 69.57 2.90 (2) Quality and availability of financial information 5.54 *** 3-4 62.32 5.80 (3) Uncertainty about strategic responses 4.29 * 1-2 14.71 13.24 (4) Volatility of underlying business models 4.26 1-2 23.53 13.24

See the notes to Table 5.

Table 7: Survey of Investor Relations Managers – Demand for and Sources of Information on Economic Impact of Swiss Franc Shock Panel A: Recipients of Internal and External Communication Question 5: How important were the following internal and external stakeholders in your decision to start communicating about the impact of the Swiss franc shock on your firm? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 39)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Financial analysts (sell-side or buy-side) 6.13 *** 2-8 74.36 0.00 (2) Institutional investors 5.72 *** 1 and 4-8 64.10 5.13 (3) Management or other internal departments (e.g., accounting) 5.38 *** 1 and 5-8 64.10 12.82 (4) Media and financial press 5.08 *** 1-2 and 5-8 38.46 7.69 (5) Suppliers and customers 3.95 1-4 and 7-8 15.38 20.51 (6) Banks and other lending institutions 3.79 1-4 12.82 25.64 (7) Retail investors 3.38 ** 1-4 7.69 35.90 (8) External audit firm 3.23 *** 1-5 10.26 43.59

Panel B: Information Sources for External Communication Question 6: How important were the following information sources for your own preparation in advance of your communication with external stakeholders about the impact of the Swiss franc shock? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 39)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Consultation with key management (e.g., CFO) 6.21 *** 2-6 78.95 0.00 (2) Existing internal reports (other than financial statements) 5.10 *** 1 and 4-6 51.28 10.26 (3) Existing annual reports or other financial filings 4.85 *** 1 and 6 30.77 2.56 (4) Newly prepared ad hoc information and reports 4.36 1 and 6 35.90 25.64 (5) Feedback from analysts, the media, or the stock market 4.26 1-3 and 6 28.21 12.82 (6) Consultation with outside experts, auditors, suppliers, etc. 2.95 *** 1-5 5.13 48.72

The sample comprises up to 39 answers from respondents to a survey sent to investor relations managers at firms listed on SIX that were potentially affected by the SNB announcement. The table lists the average ratings (in decreasing order) of the responses to the survey question indicated in the table header as well as the percentages of responses in the two tails of the Likert scale. We also report the results from t-tests comparing the average rating of an item to the mid-point (where ***, **, and * stand for statistical significance at the 1%, 5%, and 10% levels) and to the average ratings of all the other items. For the latter test, we report the item numbers that are statistically different at the 10% level, using Bonferroni-Holm adjusted p-values to correct for multiple pair-wise comparisons.

Table 8: Survey of Investor Relations Managers – Communication with External Stakeholders about Economic Impact of Swiss Franc Shock Panel A: Communication Channels for Initial Response Question 8: Immediately after the SNB announcement (i.e., within the first one or two days), how important were the following channels for your firm’s communication with external stakeholders about the impact of the Swiss franc shock? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 39)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Private communication with analysts or investors 5.31 *** 2-4 66.67 15.38 (2)

Presentations and Q&A at already pre-scheduled events (e.g.,

investor days or conference calls) 4.41

4

43.59

30.77

(3) Communication through media and financial press 3.74 1 and 4 17.95 30.77 (4) Ad hoc announcements or press releases 2.36 *** 1-3 5.13 64.10

Panel B: Firm-level Heterogeneity in the Initial Response Question 10: What motives did prevent you from communicating more proactively with external stakeholders about the impact of the Swiss franc shock? Please indicate the importance of the following motives. [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 36)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Upcoming regular financial filings or pre-scheduled events 4.56 5-8 44.12 20.59 (2) Uncertainty about development of foreign exchange rates 4.34 5-8 20.00 17.14 (3) Limited impact of Swiss franc shock on firm 4.25 8 27.78 22.22 (4)

Little perceived information needs (e.g., existing annual

reports or other financial filings provide accurate picture) 4.15

8

20.59

14.71

(5) Uncertainty about the effect of Swiss franc shock on firm 3.57 1-2 and 8 22.86 40.00 (6) Attracting unwanted scrutiny by investors, lending banks, etc. 3.29 ** 1-2 17.65 44.12 (7)

Fear of setting a disclosure precedent that might be difficult to

maintain or is inconsistent with current policies 3.15

***

1-3

8.82

38.24

(8) Giving away company secrets or harming competitive position 2.80 *** 1-5 2.86 45.71

See the notes to Table 7.

1

Appendix A: Construction of Foreign Exchange Risk Disclosure Score (FXRisk_Disc)

A.1. Components of FXRisk_Disc

We capture the quality of a firm’s exchange risk disclosures with a manually coded score

(FXRisk_Disc), ranging from 0 to 7. Higher values represent better disclosures. FXRisk_Disc

comprises seven items. We award each item either 1, 0.5, or 0 points. We identify these items

by scouring several disclosures that were released after the 2015 SNB announcement to

discontinue the minimum EUR-CHF exchange rate (e.g., the “events after the balance sheet

date” and “outlook” sections of the 2014 annual reports, press releases, or analyst presentations)

for a randomly selected group of firms. We choose this timing because we want to determine –

in hindsight – which disclosure items likely were informative for investors to assess the effect of

the Swiss franc shock, and then check whether firms already provided these items beforehand.

Item I captures the fraction of revenues that are generated outside of Switzerland (“Revenues

Generated Abroad”). The higher a firm’s proportion of international sales, the more sensitive it

should react to currency fluctuations, assuming there is at least some degree of price elasticity.

Item II provides information about the geographical distribution of a firm’s assets and, indirectly,

about its cost structure (“Assets Held Abroad”). If these are monetary assets denominated in

foreign currencies or non-current assets bounded to a specific location (i.e., non-transferable),

then a sudden shock to the exchange rate might reduce their value in CHF. On the flip side, the

effects of currency fluctuations might be mitigated through expenses that are denominated in the

same foreign currencies. Item III reflects the geographical distribution of costs and/or profits

(“Costs Incurred/Profits Generated Abroad”). A shock to CHF exchange rates should have less

of an effect on firms with a larger fraction of their costs incurring abroad. Item IV considers

the currency denominations of short-term financial positions (“Currency Distribution of

Short-term Monetary Assets and Liabilities”). This information allows investors to assess the

impact of one-time translational gains and losses on cash and near-cash balance sheet items.

2

Item V captures the aggregate foreign currency information of the entire balance sheet in all

major currencies (“Foreign Currency Exposure”). Firms sometimes report the non-functional

currency holdings of the group companies and not the aggregate foreign currency positions per

se. Item VI provides information on a firm’s hedging strategy with respect to foreign exchange

risk (“Hedging of Foreign Currency Risk”). Currency hedges should reduce the effects of

sudden exchange rate fluctuations. Item VII provides a direct estimate of the translational

effects of a (hypothetical) percentage change in exchange rates on net income and/or

shareholders’ equity (“Sensitivity of Net Income or Equity to Foreign Exchange Rate Changes”).

Table A1 details the coding guidelines we used to construct the FXRisk_Disc score.

An important feature of the FXRisk_Disc score is that it does not capture the actual currency

risk exposure, but rather the firm’s information policy regarding this risk. For example, we

assign the same score to a firm with no hedging strategy that discloses this fact as to a firm with

very detailed quantitative disclosures on how it is hedging currency risks. Furthermore, there

likely exist interdependencies among the disclosure items. For example, investors might put

more weight on the disclosures of the geographical asset distribution if a cost split by country is

missing. Aggregating multiple items into a single score should help us to capture the overall

quality of these disclosures, and at the same time reduce measurement error.

A.2. Data Collection Process

We start the data collection process by downloading the most recent annual reports for the

sample firms that were available before the SNB announcement (i.e., for the fiscal year 2013).

We use the English language version of the reports if available. For each disclosure item, we

electronically search the respective document using keywords like “currency,” “foreign

exchange,” “Euro,” etc. In addition, we read the potentially relevant sections of the annual

reports (e.g., footnotes on segment reporting and risk disclosures). Once we have identified all

the relevant passages with regard to a firm’s foreign currency risk exposure, we apply the coding

guidelines depicted in Table A1. The coding is done through the lens of an investor who tries

3

to gauge the exposure of a firm to the Swiss franc shock. Table A2 provides illustrative

examples for the seven disclosure items. Each example represents an excerpt from the annual

report of a different firm, which has received a full score for the respective item. Notably, the

level of detail and the format vary substantially across firms. Some firms provide the

information in the form of a textual description (qualitative or quantitative), others in the form of

a table or figure.

The data collection concludes with the computation of the aggregate FXRisk_Disc score.

In Table A3, we provide descriptive statistics on the scoring of the seven items for the 151

sample firms. Disclosures on the geographical distribution of revenues and assets as well as

information on the hedging strategy are by far the most common, with more than 70 percent of

the firms receiving full scores. 59 percent also provide quantitative measures of the economic

sensitivities of currency rate fluctuations. The remaining disclosure items are less widely used

with only about 25 percent of firms receiving full scores.

4

Table A1: Coding Guidelines for Construction of FXRisk_Disc

Disclosure Item Coding Explanation

I. Revenues generated abroad

We search for tables, figures, or text explaining revenues generated in Switzerland versus abroad. We award 1 point if the firm separately lists this information among its geographical segment disclosures or we can determine it from comparable sources (e.g., the firm states that it exclusively operates in Switzerland). We award 0.5 points if the firm only provides qualitative information on foreign revenues, or does not report foreign revenues for all of its major segments. We award 0 points otherwise.

II. Assets held abroad We search for tables, figures, or text explaining the portion of non-current, operating, or total assets located in Switzerland versus abroad. We award 1 point if the firm separately lists this information among its geographical segment disclosures or we can determine it from comparable sources (e.g., the firm states that it exclusively operates in Switzerland). We award 0.5 points if the firm only provides qualitative information on foreign assets. We award 0 points otherwise.

III. Costs incurred/profits generated abroad

We search for tables, figures, or text explaining the portion of costs incurred and/or profits generated in Switzerland versus abroad (in CHF versus other currencies). We award 1 point if the firm separately lists this information among its geographical segment disclosures or we can determine it from comparable sources. We also award 1 point if the firm provides information on the sensitivity of its profit margin to currency fluctuations. We award 0.5 points if the firm only provides partial information on the cost and profit distribution by currency (e.g., purchases or sourcing by country). We award 0 points otherwise.

IV. Currency distribution of short-term monetary assets and liabilities

We search for tables, figures, or text explaining the portion of short-term monetary assets and liabilities held in Switzerland versus abroad (in CHF versus other currencies). We award 1 point (0.5 points) if the firm lists this information for at least two (one) of the following items: cash and cash-equivalents, accounts receivables, or accounts payables. We award 0 points otherwise.

V. Foreign currency exposure

We search for tables, figures, or text explaining the (net) currency exposure of the firm, either on the aggregate group level or on the level of the individual group companies if non-functional currency holdings are shown. We award 1 point if the firm details its currency exposure by (major) balance sheet item. We award 0.5 points if the firm provides aggregate information on its currency exposure. We award 0 points otherwise.

VI. Hedging of foreign currency risk

We search for tables, figures, or text explaining the firm’s strategy of hedging foreign currency risk. We award 1 point if the firm provides quantitative information on its hedging strategy (e.g., a table showing currency forwards). The firm also receives 1 point if it indicates that it does not hedge its foreign currency exposure. We award 0.5 points if the firm provides qualitative (or incomplete) information on its hedging strategy. We award 0 points for missing or boilerplate disclosures.

VII. Sensitivity of net income or equity to foreign exchange rate changes

We search for tables, figures, or text explaining the sensitivity of net income and/or shareholders’ equity to changes in CHF exchange rates. We award 1 point if the firm provides information on the (hypothetical) translational effect of a percentage change in exchange rates. We award 0.5 points if the firm only provides qualitative information on exchange rate sensitivity or does not refer to specific percentage changes (e.g., provides the Value-at-Risk of its currency exposure). We award 0 points for missing or boilerplate disclosures (or if the disclosures exclusively relate to non-CHF exchange rates).

The table presents details on the scoring instructions for the seven items that make up the variable FXRisk_Disc, which we use to measure the quality of a firm’s foreign exchange risk disclosures. Each disclosure item receives a score of 1, 0.5, or 0 points. FXRisk_Disc is the sum of points assigned to the seven items.

5

Table A2: Illustrative Examples of Company Disclosures for Components of FXRisk_Disc Panel A: Item I – Revenues generated abroad

Source: Bell, Annual Report 2013, p. 33

Panel B: Item II – Assets held abroad

Source: Nobel Biocare, Annual Report 2013, p. 97

Panel C: Item III – Costs incurred/profits generated abroad

Source: Clariant, Annual Report 2013, p. 80

(Continued)

33

12 Financial Reporting Bell Group 44 Financial Reporting Bell LtdAppendix to Consolidated Balance SheetAppendix to Consolidated Income Statement

Appendix to Consolidated Income Statement

in CHF thousand 2013 Difference 2012

16. Operating income

Product groups

Fresh meat 864 620 4.3 % 828 637

Charcuterie own production 379 653 10.0 % 345 085

Charcuterie purchased 81 055 –4.7 % 85 082

Poultry 364 312 4.3 % 349 301

Meat specialities (game, rabbit and others) 17 715 8.8 % 16 288

Seafood 127 087 6.1 % 119 784

Other sales 10 083 –16.4 % 12 066

Product groups Switzerland 1 844 525 5.0 % 1 756 243

Charcuterie 708 913 0.9 % 702 643

Other sales 67 052 –0.9 % 67 634

Product groups international 775 965 0.7 % 770 277

Sales by product group 2 620 490 3.7 % 2 526 520

Distribution channels

Sales to Coop Group 1 381 494 5.7 % 1 307 100

Sales to other affiliated companies 17 683 1.5 % 17 416

Sales to wholesale 438 097 3.2 % 424 557

Sales to end consumers 7 251 1.1 % 7 170

Distribution channels Switzerland 1 844 525 5.0 % 1 756 243

Sales to Coop Group 18 685 40.8 % 13 269

Sales to wholesale 690 228 0.1 % 689 375

Sales to end consumers 67 052 –0.9 % 67 634

Distribution channels international 775 965 0.7 % 770 278

Sales by distribution channel 2 620 490 3.7 % 2 526 520

Sales by country

Switzerland 1 844 525 1 756 243

Germany 447 224 449 021

France 120 169 110 553

Spain, Benelux 74 207 77 990

Eastern Europe 134 366 132 713

Sales by country 2 620 490 3.7 % 2 526 520

Additional proceeds from Coop Group 2 416 1.1 % 2 389

Additional proceeds from affiliated companies 21 736 32.1 % 16 455

Additional third-party proceeds 31 505 –9.6 % 34 848

Other operating proceeds Switzerland 55 657 3.7 % 53 692

Other operating proceeds international 5 329 –29.3 % 7 535

Other operating proceeds 60 986 –0.4 % 61 227

Reductions in proceeds with Coop Group 26 091 52.1 % 17 151

Other reductions in proceeds 5 186 80.5 % 2 873

Reductions in proceeds Switzerland 31 276 56.2 % 20 024

Reductions in proceeds international 52 411 –11.7 % 59 344

Reductions in proceeds 83 688 5.4 % 79 369

A 10-year contract (with a commitment to supply and purchase) with Coop came into effect as of January 01, 2001. This contract has been extend-ed for additional five years. The supply of products to Coop is carried out under market conditions in consideration of Coopʼs purchase volume. Sales reductions include a bonus agreement on volume and sales figures which is stipulated in advance on a yearly basis by means of a business plan.

NobelBiocareAnnualReport2013–Financialreporting–Notestotheconsolidatedfinancialstatements 97

The allocation of assets is based on their geographical location and is as follows:

Non-current assets

in EUR thousands  2013 2012

  restated

United States  78,421 85,630

Israel  68,289 66,357

Sweden  34,159 39,159

Switzerland  38,169 37,782

Other countries  35,025 47,696

Total  254,063 276,624

Non-currentfinancialassets  3,282 3,940

Deferred tax assets  13,870 17,729

Total non-current assets  271,215 298,293

5 Personnel expenses

Personnel expenses are split as follows:

in EUR thousands  2013 2012

  restated

Wagesandsalaries  178,285 169,667

Socialsecuritycosts  30,266 30,241

Definedcontributionplanexpenses  6,577 6,518

Definedbenefitplanexpenses  5,373 5,033

Costforshare-basedpaymenttransactions  4,939 6,117

Total personnel expenses  225,440 217,576

Personnelexpensesarerecognizedinthefollowinglineitemsintheincomestatement:

in EUR thousands  2013 2012

  restated

Cost of goods sold  42,373 38,732

Research and development expenses  33,547 26,869

Selling and marketing expenses  114,703 121,673

General and administrative expenses  34,527 29,937

Finance cost  290 365

Total personnel expenses  225,440 217,576

Attheendof2013,NobelBiocarehad2,487employeesworldwide(2012:2,496).Headcountdecreasedby9whiletheGroupallocatedmoreresourcestoR&Dandrealizedefficiencygainsinnon-customerfacingfunctions.

Theincreaseinpersonnelexpensesfrom2012ismainlyduetohigherexpensesrelatedtobonusaccrualsandexpensesforrestructuringinconjunctionwiththefunctionalrealignmentprogram,whichprimarilyaffectscorporatefunctionsinGoth-enburg, Sweden. In 2013, the Group built restructuring provisions of EUR 4,320k for personnel expenses in relation to this program. The functional split is as follows: EUR 896k within the cost of goods sold, EUR 2,327k within research and devel-opment, EUR 184k within sales and marketing, and EUR 913k within general and administrative expenses.

Improved gross and EBITDA marginsClariant’s gross margin for continuing operations remained stable at 28.7!% (2012: 28.9!%). The adverse currency developments offset the positive effect from a good product mix, higher volume, and the result from the successfully implemented Clariant Excel-lence efficiency improvement program. Raw materials prices remained fairly stable.

As a result of lower costs, Group earnings before interest, tax, depreciation, and amortization (EBITDA) before exceptional items rose 5!% from CHF 817 million to CHF 858 million. EBITDA in-creased 9 % in local currencies. Consequently, the EBITDA margin rose from 13.5!% to 14.1!%.

SALES BY REGION – CONTINUING OPERATIONS CHF m

2013

2012

Change in % Change in LC2 in %

Europe 2%321 2!228 4 2MEA1 452 537 –!16 –!14North America 996 956 4 6Latin America 931 903 3 16

Asia!/ Pacific 1%376 1!414 –!3 3

1!Middle East & Africa 2!LC = local currencies

SALES STRUCTURE BY CURRENCY 2013 in %

Euro 42

US dollar 37

Japanese yen 3

Swiss franc 0

Emerging markets 18

COST STRUCTURE BY CURRENCY 2013 in %

Euro 49

US dollar 28

Japanese yen 3

Swiss franc 5

Emerging markets 15

EBITDA BEFORE EXCEPTIONAL ITEMS – FIVE-YEAR OVERVIEW CHF m

2013 8582012 8171

2011 975!2

2010 9012009 495 0 200 400 600 800 1!000

1!restated for IAS 19 2!as reported

EBITDA MARGIN BEFORE EXCEPTIONAL ITEMS – FIVE-YEAR OVERVIEW in %

2013 14.12012 13.5!1

2011 13.2!2

2010 12.72009 7.5 0 2 4 6 8 10 12 14 16

1!restated for IAS 19 2!as reported

80

6

Table A2 (continued) Panel D: Item IV – Currency distribution of short-term monetary assets and liabilities

Source: Calida, Annual Report 2013, p. 34 and p. 39

Panel E: Item V – Foreign currency exposure

Source: Belimo, Annual Report 2013, p. 72

Panel F: Item VI – Hedging of foreign currency risk

Source: Cembra Money Bank, Annual Report 2013, p. 53

(Continued)

T.

�66<[�:#'M ,<L SL�&' �##<V:SM L'#'3Z��6' �L' 8�&' ��M'& <: 3:&3Z3&V�6 �MM'MM8':S �:& L'#':S '\@'L3':#'D �'#'3Z��6'M @�MS &V'�] 8<L' S1�: ?U_ &�]M �L' VMV�66] [L3SS': <,, �'#�VM' '\@'L3':#' 1�M M1<[: S1�S S1'] �L' 0':'L�66] :<S L'#<Z'L��6'D

��! !"�C �//3H�1�"> EL4D EL4E��6�:#' �M �S ? �:V�L] AUJ-P?B ATJ;_OB�&&3S3<:M A?J__TB A?JU?OB�S363M�S3<: -?P *T_�'Z'LM�6 P._ ?JO?_�\#1�:0' L�S' &3,,'L':#'M ATUB UT��/�1�" �> �C D4 �"�"0�"= 6E;((L7 6EJ'@47

�L�&' �##<V:SM L'#'3Z��6' �L'�5 &<[: �] #VLL':#] �M ,<66<[M$

�F=="1�,"> 3& ="/"G�1�" &3= C=�!" ���3F1C> ="�",G��/" EL4D EL4E��� ?JP*; ?JO??��� TTJ;-; ?TJ_?*��� ?J?.O TUT��� .J?TT .-.�S1'L PJ_.P ?J_T;�3C�/ 1"C (@;2?( 4@;'('

D9 �C*"= �F=="1C ="�",G��/"> EL4D EL4E

�'#'3Z��6'M ,L<8 0<Z'L:8':S �VS1<L3S3'M *J-*T ?J;-*�'#'3Z��6'M ,L<8 @':M3<: ,V:&M T T?�L'@�]8':SM S< MV@@63'LM ?J-?* ?J?.U�VLL':S S�\ #L'&3SM .JU;* +�V:&L] L'#'3Z��6'M .JTP- *__�3C�/ 1"C 4$;?@? D;2D4

(9 1G"1C3=,"> EL4D EL4E

��[ 8�S'L3�6M *JOT. -J??_�'832,3:3M1'& 0<<&M OJ.;. .J??U�3:3M1'& 0<<&M P.J?T* U.JTO?�3C�/ $L;D@@ DD;'2D

�1' 3:&3Z3&V�6 3S'8M <, 3:Z':S<L3'M &3M#6<M'& ��<Z' �L' �6L'�&] L'&V#'& �] �66<[�:#'M <, ��� ?UJ;T_5 AU_?U$ ��� ?UJP.T5BD

�<S'M S< S1' #<:M<63&�S'& ,3:�:#3�6 MS�S'8':SM 3: ��� ?J___

�<:M<63&�S'& ,3:�:#3�6 MS�S'8':SM U_?T����� �L<V@

T;

29 �=�!" ���3F1C> 5�J��/" EL4D EL4E

�L�&' �##<V:SM @�]��6' T.J-?* OJ_;?�3C�/ D(;'4$ ?;L24

�L�&' �##<V:SM @�]��6' �L'�5 &<[: �] #VLL':#] �M ,<66<[M �M �S S1' L'@<LS3:0 &�S'$

�F=="1�,"> 3& ="/"G�1�" &3= C=�!" ���3F1C> 5�J��/" EL4D EL4E��� ?J?-_ UJP.?��� UPJ-P* ?JO-?�S1'L PJ*__ UJP;;�3C�/ D(;'4$ ?;L24

�L�&' �##<V:SM @�]��6' &< :<S �'�L 3:S'L'MS �:& �L' VMV�66] @�]��6' [3S13: T_ &�]MD

4L9 �C*"= �F=="1C /,��,/,C,"> EL4D EL4E

�3��363S3'M S< 0<Z'L:8':S �VS1<L3S3'M U_JUO_ UJPPT�3��363S3'M S< #<88'L#3�6 �0':SM .*P PTO�'L3Z�S3Z' ,3:�:#3�6 3:MSLV8':SM O?T +�V:&L] 63��363S3'M ?UJ?T; ?J.TT�3C�/ DD;@L$ (;?DD

�VLL':S 63��363S3'M &< :<S �'�L 3:S'L'MS �:& 1�Z' �: �Z'L�0' @�]8':S S'L8 <, T 8<:S1MD

449 ���=F"! "I5"1>"> �1! !"&"=="! ,1�30" EL4D EL4E

:Z<3#'M :<S ]'S L'#'3Z'& OJ.;T -J?.P�##LV'& @'LM<::'6 '\@':M'M .JPO. TJ?O?�S1'L &','LL'& 3:#<8' -J;*O -JOU.�3C�/ 4$;4'( 4(;L(4

�<:M<63&�S'& ,3:�:#3�6 MS�S'8':SM U_?T����� �L<V@

�<S'M S< S1' #<:M<63&�S'& ,3:�:#3�6 MS�S'8':SM 3: ��� ?J___

72 Belimo Annual Report 2013

The Group’s bank debts are not subject to any foreign exchange risks as these loans were all taken out by the Swiss companies in their functional currency, the CHF.

In order to limit the risks from foreign exchange fluctuations in merchandise transactions, Belimo aims to employ natural hedging as the primary strategy, seeking to keep cash inflows and outflows in a specific currency in balance as far as possible. Invoices between Group companies are mainly issued in the currency of the company receiving the invoice. Foreign Group companies pro-cure almost all their goods from the Swiss central production and distribution company BELIMO Automation AG and issue their sales invoices to third parties mainly in local currency. Exchange rate risks thus affect the Swiss company almost exclusively, with the result that the risks can be managed more easily.

In order to hedge the remaining net positions, Group Treasury selectively enters into forward foreign currency hedging transactions. Mainly the EUR and USD are hedged. These currencies regularly have a surplus of incoming payments.

The following table shows the foreign exchange risks from financial instruments whose currency differs from the functional currency of the Group company holding them.

in CHF 1 000 AUD CAD CHF EUR GBP HKD PLN USD

At December 31, 2012

Cash and cash equivalents 1 697 6 327 3 5 412 2 864 142 19 7 887

Trade accounts receivable 639 3 208 310 9 905 1 525 2 192 3 097 14 135

Other receivables 57 20 370 8 252

Financial assets 33

Trade accounts payable – 3 144 – 3 876 – 2 243

Other payables – 1 057 – 254 – 2 451

Other non-current financial liabilities – 1 201

Currency exposure 2 335 9 535 – 3 830 10 006 4 759 2 342 3 116 17 611

At December 31, 2013

Cash and cash equivalents 3 381 10 688 343 15 238 1 268 94 744 6 137

Trade accounts receivable 570 2 829 441 13 134 1 969 2 023 2 491 14 160

Other receivables 9 6 130

Financial assets 35

Trade accounts payable – 5 766 – 2 283 – 1 349

Other payables – 111 – 847 – 245 – 772

Current financial liabilities – 1 264

Currency exposure 3 951 13 406 – 5 829 24 588 3 237 2 123 3 235 18 340

Notes to the consolidated financial statements

72 Belimo Annual Report 2013

The Group’s bank debts are not subject to any foreign exchange risks as these loans were all taken out by the Swiss companies in their functional currency, the CHF.

In order to limit the risks from foreign exchange fluctuations in merchandise transactions, Belimo aims to employ natural hedging as the primary strategy, seeking to keep cash inflows and outflows in a specific currency in balance as far as possible. Invoices between Group companies are mainly issued in the currency of the company receiving the invoice. Foreign Group companies pro-cure almost all their goods from the Swiss central production and distribution company BELIMO Automation AG and issue their sales invoices to third parties mainly in local currency. Exchange rate risks thus affect the Swiss company almost exclusively, with the result that the risks can be managed more easily.

In order to hedge the remaining net positions, Group Treasury selectively enters into forward foreign currency hedging transactions. Mainly the EUR and USD are hedged. These currencies regularly have a surplus of incoming payments.

The following table shows the foreign exchange risks from financial instruments whose currency differs from the functional currency of the Group company holding them.

in CHF 1 000 AUD CAD CHF EUR GBP HKD PLN USD

At December 31, 2012

Cash and cash equivalents 1 697 6 327 3 5 412 2 864 142 19 7 887

Trade accounts receivable 639 3 208 310 9 905 1 525 2 192 3 097 14 135

Other receivables 57 20 370 8 252

Financial assets 33

Trade accounts payable – 3 144 – 3 876 – 2 243

Other payables – 1 057 – 254 – 2 451

Other non-current financial liabilities – 1 201

Currency exposure 2 335 9 535 – 3 830 10 006 4 759 2 342 3 116 17 611

At December 31, 2013

Cash and cash equivalents 3 381 10 688 343 15 238 1 268 94 744 6 137

Trade accounts receivable 570 2 829 441 13 134 1 969 2 023 2 491 14 160

Other receivables 9 6 130

Financial assets 35

Trade accounts payable – 5 766 – 2 283 – 1 349

Other payables – 111 – 847 – 245 – 772

Current financial liabilities – 1 264

Currency exposure 3 951 13 406 – 5 829 24 588 3 237 2 123 3 235 18 340

Notes to the consolidated financial statements

on a monthly basis to the Swiss regulator. As of 31 December 2013, the Group’s LCR was 985 %, well above the regulatory requirement of 100 %.

In January 2014, BCBS revised its guidance on the Net Stable Funding Ratio (“NSFR”). Under the new rules the Group’s NSFR as of 31 December 2013 was above the rec-ommended limit of 100 %.

Market RisksThe exposure to market risk factors is very limited. The Group has very limited foreign currency (“FX”) risk since it predominantly operates in the Swiss Consumer Lending market, and underwrites exclusively in Swiss francs. The main source of market risk is interest rate risk (“IRR”).

The Group predominantly has fixed inter-est rate assets and liabilities; therefore it does not face risk from basis mismatches. The primary source of IRR stems from timing mismatches between the expected time of repricing of assets and liabilities respectively. As of 31 December 2013,

the Group does not use any hedging in-struments to manage its IRR.

The Group monitors IRR by conducting a repricing gap analysis, assuming his-torically observed repayment behaviour for assets and contractual due dates for liabilities. The interest rate risk shock applied to the repricing gaps is one of a parallel shift in prevailing interest rates of +/– 100 basis points and +/– 200 basis points, taking into account relevant caps and floors for the “shocked” curve.

The Group reports the forecasted values of economic value of equity (lifetime) and earnings at risk (next 12 months) on a weekly basis, as per FINMA circular requirement, and monitors actual IRR performance against the internally de-fined triggers.

The following table shows the interest rate repricing gaps by expected time to repricing for its interest rate-sensitive financial instruments as of 31 Decem - ber 2013:

Interest Rate Repricing Gaps

At 31 December 2013 (CHF in millions)

Non- interest bearing

On Sight

0– 3 months

4– 6 months

7– 9 months

10– 12 months

13– 18 months

19– 24 months

2– 3 years

3– 4 years

4– 5 years

5+ years

Rate-sensitive Assets

780 63 497 455 399 345 554 434 555 252 48 197

Rate-sensitive Liabilities

944 0 544 329 230 348 322 541 860 338 104 16

Net gap – 165 63 – 47 125 169 – 4 231 – 107 – 305 – 86 – 56 181

Cumulative gap 63 15 141 310 306 537 431 126 40 – 16 165

The Group borrows and lends exclusively in Swiss francs. Its exposure to FX fluc-tuations is limited to that derived from supplier invoices denominated in foreign currencies. The Group monitors such FX

exposure closely, and takes immediate action if it exceeds internally set trig-gers. As at 31 December 2013, the Group does not use hedging instruments to manage its FX risk.

53

Annual Report 2013 Risk Management

7

Table A2 (continued) Panel G: Item VII – Sensitivity of net income or equity to foreign exchange rate changes

Source: Cicor, Annual Report 2013, p. 73

The table presents examples taken from firms’ annual reports for each of the seven items that make up the variable FXRisk_Disc. Each illustrative disclosure item shown in the panels received a score of 1 point (out of 1, 0.5, or 0 points). FXRisk_Disc is the sum of points assigned to the seven items.

Cicor | Financial Report 2013 | Consolidated Financial Statements

73

December 2013 in CHF 1 000 CHF EUR USD SGD Other

Cash and cash equivalents 37 42 2 209 32 49Trade receivables net 283 11 281 5 516 585 13Other receivables 1 2 12 114 –Trade payables –118 –5 312 –4 267 –429 –43Short-term financial liabilities – –3 196 –3 027 – –Other payables and accruals – –42 –91 –1 079 –288Total 203 2 775 352 –777 –269

December 2012 in CHF 1 000 CHF EUR USD SGD Other

Cash and cash equivalents 42 –4 828 –1 776 46 37Trade receivables net 1 509 11 749 4 759 28 13Other receivables – 11 208 – –Trade payables –85 –6 306 –3 782 –223 –38Short-term financial liabilities – –889 – – –Other payables and accruals – –13 – –514 –165Total 1 466 –276 –591 –663 –153

A 5 % strengthening of a certain currency would have increased the currency gain or loss on the basis of the foreign currency balances as of 31 December in the listed way. This analysis assumes that all other variables remain unchanged:

2013 in CHF 1 000 currency loss currency gain

5 % revaluation of EUR vs. CHF – 1855 % revaluation of EUR vs. RON 37 –5 % revaluation of RON vs. CHF 3 –5 % revaluation of USD vs. CHF 135 –5 % revaluation of USD vs. RON 72 –5 % revaluation of USD vs. SGD – 248

2012 in CHF 1 000 currency loss currency gain

5 % revaluation of EUR vs. CHF – 305 % revaluation of EUR vs. RON 40 –5 % revaluation of RON vs. CHF 1 –5 % revaluation of USD vs. CHF 132 –5 % revaluation of USD vs. RON 84 –5 % revaluation of USD vs. SGD – 194

Interest rate riskThe interest rate risk is the risk that there is a change in market value or future cash flow of a financial instrument if there is a change in interest rate.

The Group’s exposure to market risk for changes in interest rates relates primarily to the Group’s interest bearing financial debts. The Group’s policy is to manage its interest cost using a mix of fixed and variable debt. For an amount of CHF 37 million the inter-est rate was reduced in 2013 from an average of 2.4 % to an average of 2.3 %. At the reporting date the interest rate profile of the Group’s interest-bearing financial instruments is presented in note 12.

Had the interest rates in general been 100 basis points lower/higher at 31 December 2013, the pre-tax profit would have been TCHF 425 (2012: TCHF 312) higher/lower on an annual basis mainly as a result of interest expense on variable borrowings. The Group does not account for any fixed rate financial assets or liabilities at fair value with consequent profit and loss effects and does not contract derivatives as interest hedging instruments.

8

Table A3: Descriptive Statistics of Components of FXRisk_Disc

Disclosure Item

(N = 151)

Scoring Totals (in %) Mean

1 point 0.5 points 0 points

I. Revenues generated abroad 78.81 % 12.58 % 8.61 % 0.85 II. Assets held abroad 73.51 % 4.64 % 21.85 % 0.76 III. Costs incurred/profits generated abroad 15.89 % 8.60 % 75.50 % 0.20 IV. Currency distribution of short-term

monetary assets and liabilities 25.17 % 22.52 % 52.32 % 0.36

V. Foreign currency exposure 24.50 % 9.93 % 65.56 % 0.29 VI. Hedging of foreign currency risk 84.11 % 7.95 % 7.95 % 0.88 VII. Sensitivity of net income or equity to

foreign exchange rate changes 58.94 % 15.89 % 25.17 % 0.67

The table presents descriptive statistics on the scoring of the seven items that make up the variable FXRisk_Disc. Each disclosure item receives a score of 1, 0.5, or 0 points. FXRisk_Disc is the sum of points assigned to the seven items. The sample comprises 151 firms listed on the Swiss exchange (SIX) during the SNB announcement to discontinue the minimum EUR-CHF exchange rate on January 15, 2015.

9

Appendix B: Sell-Side Analysts Survey

We developed our initial draft of the sell-side analyst survey after an extensive review of the

related literature (e.g., Brown et al. 2015, 2016; Dichev et al. 2013; Nelson and Skinner 2013;

Graham et al. 2005). We circulated the draft among several academics and practitioners to gain

feedback, and implemented the revised version of the survey in Qualtrics, an online survey tool.

The survey comprises 15 questions, divided into four topical sections, and is followed by a

number of demographical questions. We display each survey section en bloc, but respondents

have the possibility to skip certain questions and/or answers and can exit the survey at any point

in time.1 Wherever possible, we randomize the order of the answering choices. However, we

do not randomize the order of the questions to not interrupt the logical flow of the survey.2

To identify a pool of suitable survey subjects we collect the names and contact information

of all sell-side analysts covering at least one sample firm during the event period. We search

the following four information sources: (1) companies’ investor relations websites; (2) analyst

coverage information as provided by S&P Capital IQ; (3) analyst coverage information as

provided on the website of the Swiss Exchange (SIX); and (4) author information from actual

analyst reports collected from Investext. The first two sources do not provide historical data,

and hence yield coverage information as of February and March 2016 (i.e., the time of data

collection). The latter two sources allow us to identify analyst coverage as of a specific point in

time. We use the SIX website to extract the names of all analysts who issued an opinion on a

specific firm in the 60 days following the event date (January 15, 2015). In addition, we scour

the analyst reports on Investext that were issued over the same time frame for information on

authorship (including analyst team members). We then update analysts’ contact information

based on the more recent data from S&P Capital IQ and company websites, and discard analysts

who left the profession in the interim or vanished from the database. We further exclude 1 In addition, we split the third section of the survey into two screens (i.e., Questions 8 and 9 are displayed together and separately from Question 10) to avoid any focus on “existing annual reports,” which is the topic of Question 10. 2 The survey instrument is available upon request from the authors.

10

analysts working for one of the three major banks in Switzerland, because we approach them

using a slightly different method (described below). This procedure allowed us to identify 773

sell-side analysts who covered a sample firm during the event, of which 743 turned out to have a

valid e-mail address. We use the latter number to calculate the response rate for this group.

We sent out a personalized e-mail with a survey link to the subjects on April 12, 2016. The

e-mail explained why we address them and the general purpose of the survey, however, we did

not mention that we were examining the usefulness of financial statement information for the

assessment of the Swiss franc shock. The e-mail also indicated the length of the survey and the

confidentiality of the responses. We sent out a reminder to non-responders after one week and

again after four weeks, but only if they were working for brokerage houses whose policy was not

to disallow survey participation or if they were covering multiple Swiss firms. We received a

total of 66 responses until the end of May 2016, giving rise to a response rate of 8.9%. This

number is rather conservative as our pool of subjects contains many analysts who were only

tangentially affected by the Swiss franc shock (e.g., analysts from U.S. brokerage houses that

only covered one single large-cap Swiss firm such as Nestlé).

For the analysts at the three major Swiss banks, we relied on personal contacts (i.e., senior

analysts or analyst team leaders) to forward the original e-mail to the people on our subjects’ list.

These e-mails were sent in April and May 2016, and were followed up with two reminders. We

received 11 responses out of 72 possible, yielding a response rate of 15.3% for this subset. In

what follows, we report the demographics of all the 77 respondents to our sell-side analyst

survey (Table B1) together with the responses to the survey questions not tabulated in the main

text (Tables B2 to B4).

11

Table B1: Demographics of the Respondents to the Sell-Side Analysts Survey

(Maximum possible N = 77) Sample % Sample %

Primary Industries Covered Age of Analyst

Retail/Wholesale 10.34 <30 12.99 Construction 6.03 30 – 39 33.77 Chemicals 4.31 40 – 49 37.66 Software/Technology 6.03 50 – 59 14.29 Health care/Pharmaceuticals/

Biotechnology 6.03 60+ 1.30

Telecommunications/Media 4.31 Insurance 7.76 Education of Analyst Real estate 3.45 Bachelor’s degree 15.58 Banks and other Finance 8.62 Master’s degree 54.55 Manufacturing consumer goods 11.21 CPA, CFA, or AZEK degree* 23.38 Manufacturing industrials 17.24 Ph.D./doctoral degree 2.60 Consulting/Business services 3.45 Other 3.90 Transportation/Energy/Utilities 8.62 Other 2.59 Years as Sell-Side Analysts

<4 years 19.48 Number of Industries Covered 4 – 9 years 31.17

1 40.26 10+ years 49.35 2 – 3 49.35 4+ 10.39 Size of Current Employer

One sell-side analyst 0.00 Number of Firms Covered 2 – 4 sell-side analysts 6.49

1 2.60 5 – 10 sell-side analysts 7.79 2 – 4 7.79 11 – 25 sell-side analysts 29.87 5 – 9 23.38 26 – 50 sell-side analysts 12.99 10 – 15 49.35 More than 50 sell-side analysts 42.86 16 – 25 12.99 25+ 3.90 Headquarter of Current Employer

Switzerland 40.26 % of Swiss Firms Covered Rest of Europe 42.86

0 % 0.00 USA 11.69 1 – 24 % 46.75 Rest of World 5.19 25 – 49 % 14.29 50 – 74 % 6.49 75 – 99 % 7.70 100 % 24.68

* CPA = Certified Public Accountant; CFA = Chartered Financial Analyst; AZEK = Degree of the Swiss Training Center for Investment Professionals. The table lists the demographic information of the participants in the survey sent to sell-side analysts covering at least one firm listed on SIX during the SNB announcement (77 answers).

12

Table B2: General Information About Analysts’ Perception of the Swiss Franc Shock

Panel A: Importance of Foreign Currency Exposure Before Swiss Franc Shock Question 1: Before the Swiss franc shock, how important was the foreign currency exposure of a typical Swiss firm you followed for your earnings forecasts and stock recommendations? [7-point Likert Scale: 1 to 7]

Responses (N = 77)

Average Rating

Significantly Different Than [4 = Neutral]

Not at all Important [1 or 2]

Neutral [3 to 5]

Extremely Important [6 or 7]

% of respondents who answered 4.88 *** 10.39 53.24 36.36

Panel B: Expectations about Occurrence of Swiss Franc Shock in the Future Question 2: As of the beginning of January 2015 (i.e., directly before the Swiss franc shock), when did you expect that the Swiss National Bank would most likely abandon the minimum Euro exchange rate, if at all? [Multiple Choice]

Responses (N = 77)

Within the Next Not in the Foreseeable

Future 3 Months 3 to 6 Months 6 to 12 Months 12 to 24 Months

% of respondents who answered 2.60 2.60 14.29 22.08 58.44

Panel C: Percentage of Covered Firms Affected by Swiss Franc Shock Question 3: What percentage of Swiss firms you followed was at least somewhat affected by the SNB announcement (positively or negatively)? [0 – 100 %] Responses (N = 77) 0 – 24 % 25 – 49 % 50 – 74 % 75 – 99 % 100 %

% of respondents who answered 7.80 2.60 3.90 15.59 70.13

Panel D: Impact of Swiss Franc Shock on Typical Firm Question 4: How has the Swiss franc shock affected the typical Swiss firm in your portfolio? [5-point Likert Scale: 1 to 5]

Responses (N = 77)

Average Rating

Significantly Different Than [3 = Neutral]

Negatively [1 or 2]

Neutral [3]

Positively [4 or 5]

% of respondents who answered 1.94 *** 84.42 15.58 0.00

(Continued)

13

Table B2 (continued)

Panel E: Importance of Foreign Currency Exposure After Swiss Franc Shock Question 15: After the Swiss franc shock, how important is the foreign currency exposure of a typical Swiss firm you follow for your earnings forecasts and stock recommendations? [7-point Likert Scale: 1 to 7]

Responses (N = 77)

Average Rating

Significantly Different Than [4 = Neutral]

Not at all Important [1 or 2]

Neutral [3 to 5]

Extremely Important [6 or 7]

% of respondents who answered 5.52 *** 1.30 46.76 51.95

The sample comprises up to 77 answers from respondents to a survey sent to sell-side analysts covering at least one firm listed on SIX during the SNB announcement. The panels list the average ratings of the responses to the survey question indicated in the panel header as well as the percentages of responses in various bins of the Likert scale (or any other scale mentioned in the question). If applicable, we report the results from t-tests comparing the average rating of an item to the mid-point (where ***, **, and * stand for statistical significance at the 1%, 5%, and 10% levels). If multiple selections are available, we also compare the average rating of an item to the average ratings of all the other items, using Bonferroni-Holm adjusted p-values at the 10% level for the comparison.

14

Table B3: Timeliness of and Motivation for Analysts’ Reaction to the Swiss Franc Shock

Panel A: Initial Response to Swiss Franc Shock for Firms Covered Question 5.1: What is the percentage of Swiss firms you followed for which you individually assessed the impact of the SNB announcement immediately thereafter (i.e., within the first one or two days)? (Note: Alternatives must add up to 100 %) [0 – 100 %]

Responses (N = 77)

Firms Assessed Firms Not Assessed No Material Effect Material Effect

% of respondents who answered 17.68 77.26 5.06

Question 5.2: Of those firms with a material effect (as indicated in question 5.1), how often did your assessment result in one of the following adjustments? (Note: Multiple adjustments are possible per firm) [0 – 100 %]

Responses (N = 77)

Qualitative Adjustment

Quantitative Adjustment

Change in Earnings Forecast/Stock Recommendation

% of respondents who answered 40.38 69.51 63.22

Panel B: Motivation for Initial Response to Swiss Franc Shock for Firms Covered Question 6: What was the motivation to provide a quantitative or qualitative assessment of the effect of the SNB announcement on firms you followed immediately after the Swiss franc shock (i.e., within the first one or two days)? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 76)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Demand from clients (e.g., institutional or retail investors) 6.13 *** 3-9 77.63 5.26 (2) Likely exposure of the firm to the currency shock 5.89 *** 3-9 74.32 6.76 (3) Demand from internal departments (e.g., trading desk, risk

management) 4.93 *** 1-2 and 5-9 57.53 21.92

(4) Relative importance of the firm (e.g., size or liquidity) 4.67 *** 1-2 and 7-9 38.36 17.81 (5) Availability of firm-specific information 4.18 1-3 and 9 26.39 20.83 (6) Personal job description or performance evaluation 4.08 1-2 34.25 31.51 (7) Peer pressure (from analysts inside or outside your firm) 3.82 1-4 23.29 30.14 (8) Reputation with management of the companies you follow 3.66 1-4 23.29 34.25 (9) Complexity of a firm’s operations 3.56 * 1-5 15.49 33.80

(Continued)

15

Table B3 (continued)

Panel C: Medium-term Response to Swiss Franc Shock for Firms Covered Question 7: What is the percentage of Swiss firms you followed for which you individually (re-)assessed the impact of the SNB announcement in the two to three weeks after the Swiss franc shock? Responses (N = 77) 0 – 24 % 25 – 49 % 50 – 74 % 75 – 99 % 100 %

% of respondents who answered 20.78 2.60 9.10 5.20 62.34

See the notes to Table B2.

16

Table B4: Methodology of Assessing Impact of the Swiss Franc Shock for Firms Covered

Panel A: Heterogeneity in Valuation Approach Question 11: Assessing the impact of the Swiss franc shock (quantitatively and/or qualitatively) was … [Multiple choice] Responses (N = 77)

Very Similar for All Firms

Relatively Similar for Most Firms

Considerably More Difficult for Some Firms Than Others

% of respondents who answered 7.79 45.45 46.75

Panel B: (Re-)Assessment of Economic Outlook for Firms Covered Question 14: Which of the following techniques did you apply to incorporate the impact of the Swiss franc shock on a typical firm you followed in your earnings forecasts or stock recommendations? [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 75)

Significantly Different Than

[4 = Sometimes]

% of Respondents Who Answered Average Rating

Significantly Different Than

Always [6 or 7]

Never [1 or 2]

(1) Use of existing models with adjustment of quantitative inputs for individual firms (e.g., sales and margins figures)

5.93 *** 2-5 74.76 5.33

(2)

Re-assessment of entire situation of the individual firm (including strategic responses like relocation decisions)

4.29 1 and 5 38.67 25.33

(3) Adjustment of existing models (e.g., adjustment of weights or introduction of a new currency risk category)

4.27 1 and 5 43.24 32.43

(4) Use of existing models with adjustment of qualitative inputs for individual firms (e.g., market risk assessment)

3.75 1 and 5 22.67 37.33

(5) Switching to new valuation models 2.05 *** 1-4 6.76 72.97

See the notes to Table B2. We do not report the answers to Question 13 “please indicate the ticker symbol of the three firms you followed for which the assessment of the impact of the Swiss franc shock was most and least difficult,” because of its potentially confidential nature.

17

Appendix C: Excerpts from Analyst Reports Issued After the SNB Announcement

We searched for all analyst reports that were released for our sample firms within 15 days of

the SNB announcement using Thomson Reuters’ Investext. We were able to identify 528

reports covering 87 individual firms that were released within this timespan. We then read the

reports and searched for (anecdotal) evidence of whether and how analysts were using

information in the risk disclosures of firms’ annual reports to evaluate the impact of the Swiss

franc shock on individual firms. We could identify several instances in which analysts were

directly referring to annual report disclosures or, indirectly, were using information from firms’

annual reports. Table C1 provides three select examples.

The first example (Panel A) is an analyst report on two firms, Straumann and Sonova, issued

by JP Morgan on January 20. Since both firms are discussed in the same report, we can directly

compare which information sources were used for each firm by the same analyst and trace it

back to the individual firm’s annual report. Interestingly, the analyst report shows a detailed

revenue and cost split by currency for Straumann which is taken from the annual report (i.e., the

analyst report states “Annual Report 2013” as the source). For Sonova, the revenue split by

geographical region (but not currency) is provided, but the cost split is missing. The source is

again the most recent annual report before the Swiss franc shock (i.e., “Annual Report FY14;”

fiscal year 2014 ended in March of 2014). The difference in informativeness of the annual

reports for the two firms is consistent with our reading and analysis of the two documents and is

reflected in the respective risk disclosure scores. Straumann has a FXRisk_Disc score of 6

compared to only 4.5 for Sonova.

The second example (Panel B) is an analyst report on ABB issued by Morgan Stanley on

January 15 (i.e., the day of the event). The report provides an estimate of ABB’s Swiss franc

exposure (revenues and cost). More to the point, it mentions that the information for this

assessment is from the 2013 Form 20-F. Further down, it also refers to a 17% asset base in

18

Swiss francs, which is mentioned in the firm’s annual report in a section that we used when

coding up ABB’s risk disclosure score.

The third example (Panel C) is an analyst report on Holcim issued by Jefferies on January

22. It provides an estimate of the impact of the sudden drop in Swiss franc exchange rates.

The report explicitly lists that the analysts have used “the currency sensitivity table on page 160

of Holcim’s 2013 annual report” for their estimates. This information was also included when

we coded up Holcim’s risk disclosure score.

19

Table C1: Illustrative Examples of Risk Disclosure Discussions in Analyst Reports Panel A: Analyst Report on Straumann and Sonova (issued by JP Morgan on January 20, 2015)

Source: JP Morgan, “Straumann/Sonova - Impact of CHF moves”, p. 2 and 4.

(Continued)

2

Europe Equity Research20 January 2015

Anasuya Sarma(44-20) 7742-6264anasuya.sarma@jpmorgan.com

StraumannFX ImpactStraumann’s revenue exposure is EUR 40%, CHF 12%, USD/CAD/AUD 28%, while cost exposure is CHF 45%, EUR 21%, USD/CAD/AUD 22%. The company has short-term cash flow hedges in place for both the EUR and USD, with the USD hedged over a 3-month time frame.

Figure 1: Straumann revenue FX exposure Figure 2: Straumann cost FX exposure

Source: Annual Report 2013. Source: Annual Report 2013.

Table 1: Straumann FX sensitivity (+/- 10%)Revenue EBIT

EURCHF +/- CHF 25m +/- CHF 15mUSDCHF +/- CHF 17m +/- CHF 7mJPYCHF +/- CHF 4m +/- CHF 2mSource: Company data.

A 15.7% fall in the EURCHF is estimated to have a negative impact of 16.8% (CHF 24m) on Straumann’s EBIT, the 14.2% fall in the USDCHF will have a further negative impact of 7.1% (CHF 10m), and the ~15% fall in the JPYCHF will have a negative impact of 2.1% (CHF 3m). This implies a total negative EBIT impact of 26.0% (CHF 36m).

Straumann management have estimated the impact to be roughly CHF 40m with up to CHF 10m salvaged through immediate steps. It has commented to the press that immediate measures such as a headcount freeze and abandoning non-business travel will make up some of this. Further, the company will likely see a currency benefit in FY15E from the lower Brazilian Real on the 2nd tranche of the Neodent acquisition.

EUR, 41%

USD/CAD/AUD, 27%

CHF, 11%

Other, 21%

USD/CAD/AUD, 19%

EUR, 21%

CHF, 47%

Other, 13%

4

Europe Equity Research20 January 2015

Anasuya Sarma(44-20) 7742-6264anasuya.sarma@jpmorgan.com

SonovaFX ImpactSonova derives only 1% of revenues from CHF, while it is the largest cost currency.

Figure 5: Sonova Sales exposure

Source: Annual Report FY14.

Table 2: Sonova FX sensitivity (+/- 5%)Sales EBITA

EURCHF +/- CHF 26m +/- CHF 14mUSDCHF +/- CHF 38m +/- CHF 12mSource: Company data.

A 15.7% fall in the EURCHF is estimated to have a negative impact of 9.3% (CHF 44m) on Sonova’s EBITA and a 14.2% fall in the USDCHF will have a furthernegative impact of 7.2% (CHF 34m) in FY15E. This implies a total negative EBITA impact of 16.8% (CHF 78m).

Updated forecasts and price targetWe adjust our Sonova forecasts to account for the CHF moves. This results in EPS downgrades of 10% in FY15E (year-end March) followed by 14-15% downgrades thereafter. Our earnings downgrades have driven the reduction in our Sep-15 price target to CHF121 (CHF141).

Switzerland, 1%

EMEA (ex Switzerland),

39%

US, 37%

Americas (ex US), 11%

Asia Pacific, 11%

20

Table C1 (continued) Panel B: Analyst Report on ABB Ltd. (issued by Morgan Stanley on January 15, 2015)

Source: Morgan Stanley, Analyst Report “ABB-Quantifying the Swiss Franc Impact”, p. 1.

Panel C: Analyst Report on Holcim (issued by Jefferies on January 22, 2015)

Source: Jefferies, “Holcim (HOLN VX) Reducing Price Target and Estimates”, p. 1.

The table presents excerpts taken from analyst reports illustrating how financial analysts used risk disclosures from firms’ annual reports to evaluate the impact of the Swiss franc shock in the days immediately following the SNB announcement. We highlighted the relevant portions of the text in yellow.

Ben.Uglow@morganstanley.com

Lucie.Carrier@morganstanley.com

Robert.J.Davies@morganstanley.com

Andrew.Caldwell@morganstanley.com

MORGAN STANLEY & CO. INTERNATIONAL PLC+

Ben Uglow

+44 20 7425-8750

Lucie A Carrier

+44 20 7677-0884

Robert J Davies, Ph.D.

+44 20 7425-2057

Andrew Caldwell

+44 20 7425-2177

ABB ( ABBN.VX , ABBN VX )

Capital Goods / Switzerland

Stock RatingStock Rating Equal-weightEqual-weight

Industry ViewIndustry View In-LineIn-Line

Price targetPrice target SFr 24.00SFr 24.00

Shr price, close (Jan 14, 2015) SFr 19.70

52-Week Range SFr 24.80- 18.56

Mkt cap, curr (mn) US$44,050

Net debt (12/14e) (mn)* US$231

*EV (12/14e) (mn) US$50,374

* = GAAP or approxim ated based on GAAP

Fiscal Year Ending 12/13 12/14e 12/15e 12/16e

Revenue (US$ mn)** 41,848 40,036 41,654 44,114

EBITDA (US$ mn)** 5,705 5,489 6,354 7,122

EBIT (US$ mn)** 4,387 4,129 4,939 5,624

EPS (US$)** 1.29 1.23 1.51 1.79

Prior EPS (US$)** - - - -

Consensus EPS (US$)§Consensus EPS (US$)§ 1.391.39 1.211.21 1.411.41 1.601.60

P/E** 20.4 17.3 12.8 10.8

*EV/Revenue 1.5 1.3 1.1 1.0

*EV/EBITDA 11.2 9.2 7.2 6.2

*EV/EBIT 14.6 12.2 9.2 7.8

Div yld (%) 2.9 3.7 4.3 4.5

FCF yld ratio (%)** 4.2 6.1 7.2 8.7

Net debt (US$ mn)* 1,538 231 977 767

Net debt/EBITDA** 0.3 0.0 0.2 0.1

ModelWare EPS (US$) 1.23 1.15 1.47 1.75

Unless otherwise noted, all m etrics are based on Morgan Stanley ModelWare fram ework

§ = Consensus data is provided by Thomson Reuters Estim ates

* = GAAP or approxim ated based on GAAP

** = Based on consensus m ethodology

e = Morgan Stanley Research estim ates

Industry View

In-Line

Stock Rating

Equal-weight

Price Target

SFr 24.00

January 15, 2015

ABBABB

Quantifying the Swiss Franc Impact

Although ABB reports in US dollars, we need to remember that its

SFr costs exceed revenues. On this basis, there will be some negative

impacts – a 'worst-case' EPS effect would be 7%. We expect the

eventual impact to be much lower.

S wiss revenue exposure is 2% – but cost base is higher. Although ABB

reports in US$ (a translation positive), there is an imbalance between its Swiss

Franc revenue and cost base. According to information in the 2013 Form 20-F,

ABB generated ~98% of its revenues from customers outside of Switzerland.

While not necessarily exactly the same as 'sales by destination' it implies a 2%

revenue exposure only. Assessing the cost base is much less clear, and our

assumptions are not confirmed by the company. The key items are: (i) 5% of

the 150,000 headcount is based in Switzerland; but (ii) 17% of its long-lived

assets are in Switzerland (PPE).

EBIT margin falls ~90bps and EPS by 7% – but note this is a 'worst-case'

scenario: In Exhibit Exhibit overleaf, we have made two assumptions: (i) that 10% of

ABB's manufacturing cost base is still in Switzerland – this seems intuitively

too high, but we are trying to reflect the ~17% 'asset base' data; (ii) that 5% of

ABB's SG&A is in Switzerland. If these assumptions are correct, at US$/SFr0.88,

then gross margin would be ~80bps lower, reported EBIT margin falls from

11.9% to 11.0% – and FY15 EPS is reduced 7.3% to $1.40. We believe that this

is a 'worst-case' assumption, and expect that once clarified by the company,

the final EPS impact should be significantly lower.

What is the transaction impact? We see this as fairly limited. With the

exception of a handful of products (eg. certain large power transformers, parts

of the automation portfolio) ABB is not a major net exporter; additionally, from

a transaction standpoint, we would expect that any headwind from Swiss Franc

appreciation should be more than offset by Euro weakness versus US dollar.

The main conclusion is that the main effect within the company is around

inflation of Swiss-related costs.

Investment view – EW, but there could be a 'tactical opportunity': given

the 11% share price move versus our worst-case 7% EPS impact, this feels like

an overreaction. At $17.6 and on reduced earnings, ABB now trades at ~12.5x

FY15 EPS, ~9.5x EBIT and the shares are at their lowest level since November

2012. In a near-term sense, we would expect to see some kind of trading

bounce. However, from a fundamental standpoint we would remain cautious

ABB ahead of 4Q results February 5th, given end-market exposures (Utility, Oil

& Gas, Petchem and Mining) which are challenged.

Morgan Stanley does and seeks to do business withcompanies covered in Morgan Stanley Research. As a result,investors should be aware that the firm may have a conflictof interest that could affect the objectivity of MorganStanley Research. Investors should consider MorganStanley Research as only a single factor in making theirinvestment decision.

For analyst certification and other important disclosures,For analyst certification and other important disclosures,refer to the Disclosure Section, located at the end of thisrefer to the Disclosure Section, located at the end of thisreport.report.+= An alysts employed by n on -U .S. aff iliates are n o t reg istered w ith F INRA, may

n ot be associated person s o f th e member an d may n ot be su b ject to NASD/NYSE

restriction s on commu n ication s w ith a su b ject compan y, pu b lic appearan ces an d

trad in g secu rities h eld by a research an alyst accou n t.

| January 15, 2015ABB

MORGAN STANLEY RESEARCH

1

CHF Prev. 2013A Prev. 2014E Prev. 2015E Prev. 2016E

Rev. (MM) -- 19,719.0 18,996.0 19,254.0 21,547.0 18,915.0 24,030.0 20,844.0

EBITDA (MM) -- 3,896.0 3,687.0 3,691.0 4,274.0 3,479.0 5,116.0 4,021.0

EPS -- 3.91 3.92 3.87 5.01 4.22 6.67 5.09

Div. Yield 2.15% 2.15% 2.48% 2.90%

Dividend

FY Dec -- 1.30 1.50 1.30 1.75 1.50 2.10 1.75

FY P/E 15.4x 15.6x 14.3x 11.9x

Price Performance

JAN-14 MAY-14 SEP-14 JAN-15

90

80

70

60

50

COMPANY NOTE

Target | Estimate Change

Switzerland | Industrials | Building Materials 22 January 2015

Holcim (HOLN VX)Reducing Price Target and Estimates.Retaining Buy Rating

EQU

ITY RESEARC

H EU

ROPE

BUYPrice target CHF80.00

(from CHF94.00)Price CHF60.40

Financial SummaryNet Debt (MM): CHF9,461.0

Market Data52 Week Range: CHF84.66 - CHF56.50Total Entprs. Value (MM): CHF29,127.2Market Cap. (MM): CHF19,666.2Insider Ownership: 18.3%Institutional Ownership: 28.5%Shares Out. (MM): 325.6Float (MM): 224.1Avg. Daily Vol.: 1,392,208

Mike Betts *Equity Analyst

44 (0) 20 7029 8692 mbetts@jefferies.com * Jefferies International Limited

Key Takeaway

Although we are also adjusting our estimates and price target for therecently completed asset swap with Cemex and for our lower cement volumeexpectations for 2015 and 2016, their impact is dwarfed by the impact ofthe stronger Swiss Franc. The latter accounts for CHF600m of the CHF795mreduction in our 2015 EBITDA estimate. It is also the main factor reducing our12-month price target from CHF94 to CHF80.

Stronger Swiss Franc. Our estimates are based on current exchange rates. Previously wehad used the rate of 1.21 Swiss Francs to the Euro, whereas we are now using a rate of 1.02.The Swiss Franc has also strengthened significantly against most other countries. The impacton EBITDA is automatically calculated in our model. The adjustments lower down the P&Lare more difficult to estimate, but we have used the currency sensitivity table on page 160of Holcim’s 2013 annual report which shows what would have been the impact in 2013 ofa 5% depreciation in other currencies versus the Swiss Franc on sales, EBITDA, net income,cash flow and net debt.

Reduction in EBITDA estimates. Our 2014 operating EBITDA estimate is similar topreviously, increasing by just CHF4m to CHF3,691m. Our 2015 estimate declines by 18%from CHF4,274m to CHF3,479m and our 2016 estimate falls by 22% from CHF5,116m toCHF4,021m.

Valuation/RisksLafarge (LG FP, €59.59, Buy) and Holcim announced in April 2014 that they plannedto merge at a one-for-one share exchange ratio. We therefore value the two businessescombined. Our 12-month price target of CHF80 is based on our estimate of LafargeHolcim’smid-cycle EBITDA excluding associates but including synergies, of CHF6,151m multiplied byour estimate of LafargeHolcim’s future long-term EV/EBITDA ratio of 8.0x. This is similar toHolcim’s average rating over the last 10 years of 8.2, but slightly lower than Lafarge’s of 9.2x.The main risk to our estimates and rating would be if the Swiss Franc or Euro strengthened,or if cement prices increase by less than we have assumed. Also, if trading is worse than weexpect in some of the key markets, such as India, Canada, Mexico and Algeria.

Jefferies does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that Jefferies may have a conflictof interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision.Please see analyst certifications, important disclosure information, and information regarding the status of non-US analysts on pages 8 to 12 of this report.

21

Appendix D: Investor Relations Survey

We developed the survey of firms’ investor-relations reaction to the Swiss franc shock

similarly to the sell-side analyst survey (see Appendix B). After receiving feedback from

academics and practitioners on the initial draft, we implemented the survey in Qualtrics. The

survey comprises 15 questions, divided into four topical sections, and is followed by a few

demographical questions. We display each survey section en bloc, but respondents have the

possibility to skip certain questions and/or answers and can exit the survey at any point in time.

Wherever possible, we randomize the order of the answering choices. However, we do not

randomize the order of the questions to not interrupt the logical flow of the survey.3

Our subject pool comprises the 151 sample firms. For each firm, we identify the name and

contact information of the person responsible for investor relations. Our primary data sources

are the company’s website and firm information provided on the website of the Swiss Exchange

(SIX). Because not all firms have a dedicated investor relations department or a designated

head of investor relations, we search the company functions using the following algorithm: (1)

head of investor relations, (2) (senior) investor relations manager, (3) head of communications,

(4) assistant to chief financial officer, (5) investor relations team (e.g., e-mail addresses that start

with info@, ir@, or investors@), (6) chief financial officer, and (7) chief executive officer.

This way we were able to identify 149 valid e-mail addresses, which we use as the denominator

in the calculation of the response rate.4 Table D1, Panel B, shows the job titles of the survey

respondents, and indicates that about 70% of the participants were either head of investor

relations, head of communications, or senior investor relations managers.

To maximize turnout, we proceeded like follows: first, we contacted the Swiss Society for

Investor Relations (an association of leading investor relations professionals), and asked them to

send out the survey to their member firms on our behalf. 49 sample companies are represented

3 The survey instrument is available upon request from the authors. 4 We sent out our survey to 148 of the 149 people identified in the search because one firm volunteered to pre-test the survey at an earlier stage and provided us with feedback on the questions.

22

among their members. Second, we directly contacted the remaining subjects through e-mail.

In both cases, the initial survey request was administered at the end of May 2016. We sent out

a personalized e-mail with a survey link to the subjects. The e-mail explained the general

purpose of the survey, however, we did not mention that we were examining the usefulness of

financial statement information for the assessment of the Swiss franc shock. The e-mail also

indicated the length of the survey and the confidentiality of the responses. We sent out a

reminder to non-responders after one week and again after four weeks, and made sure that the

Swiss Society for Investor Relations reminded its members once. We received a total of 39

responses until the end of July 2016, giving rise to a response rate of 26.2%. This number

seems high, but it is part of investor relations managers’ job to reply to external information

requests.

In what follows, we report the demographics of the 39 respondents to our investor relations

survey (Table D1) together with the responses to the survey questions not tabulated in the main

text (Tables D2 to D4).

23

Table D1: Demographics of the Respondents to the Investor Relations Survey Panel A: Characteristics of Respondent Firms

Full Sample (N = 151) Respondent Firms (N = 37) Mean Std. Dev. Mean Std. Dev. Bid-Ask Spreadt 0.512 1.247 0.383 0.400 FXRisk_DiscAR 4.020 1.434 4.054 1.229 Market Valuet-1 8,814 32,438 5,045 8,718 Share Turnovert-1 0.229 0.333 0.200 0.185 Return Variabilityt-1 0.531 0.555 0.465 0.177 Int_SalesAR 0.655 0.387 0.606 0.391 Hist_Corr_EUR 0.223 0.111 0.230 0.111 Total_DiscAR 3.662 0.750 3.835 0.717 Num_Analysts 7.715 8.553 8.649 8.169 Free FloatAR 0.663 0.250 0.646 0.252 Stock Returnt -0.254 0.325 -0.319 0.277

Panel B: Job Titles of Survey Participants

Job Title (N = 38) N % Head of Investor Relations 12 31.58 % (Senior) Investor Relations Manager 9 23.68 % Chief Financial Officer 7 18.42 % Head of Communications 5 13.16 % Treasury or Accounting 3 7.89 % Other 2 5.26 %

Panel A lists descriptive statistics of several firm attributes for the full sample and the 37 firms that self-identified and responded to the survey sent to investor relations managers at firms listed on SIX during the SNB announcement (two firms did not provide a ticker symbol). The panel reports means and standard deviations for the variables used in the daily information asymmetry regressions (see Table 1 for details). For each firm, we take the mean variable value over the [-30;+2] day period surrounding the SNB announcement to compute the sample distribution. None of the differences in means across samples is significant at conventional levels using t-tests. In Panel B, we report the job titles of the survey respondents (one respondent did not provide a job title).

24

Table D2: General Information About Investor Relations Managers’ Perception of the Swiss Franc Shock

Panel A: Importance of Foreign Currency Exposure Before Swiss Franc Shock Question 1: Before the Swiss franc shock, how prevalent was your firm’s foreign currency exposure in the communication with external stakeholders (e.g., investors, financial analysts)? [7-point Likert Scale: 1 to 7]

Responses (N = 39)

Average Rating

Significantly Different Than [4 = Neutral]

Not at all Important [1 or 2]

Neutral [3 to 5]

Extremely Important [6 or 7]

% of respondents who answered 4.44 12.82 56.41 30.77

Panel B: Expectations about Occurrence of Swiss Franc Shock in the Future Question 2: As of the beginning of January 2015 (i.e., directly before the Swiss franc shock), when did you expect that the Swiss National Bank would most likely abandon the minimum Euro exchange rate, if at all? [Multiple Choice]

Responses (N = 39)

Within the Next Not in the Foreseeable Future 3 Months 3 to 6 Months 6 to 12 Months 12 to 24 Months

% of respondents who answered 5.13 10.26 20.51 10.26 53.85

Panel C: Impact of Swiss Franc Shock on Typical Firm Question 3: How has the Swiss franc shock affected your firm’s business? [5-point Likert Scale: 1 to 5]

Responses (N = 39)

Average Rating

Significantly Different Than [3 = Neutral]

Negatively [1 or 2]

Neutral [3]

Positively [4 or 5]

% of respondents who answered 2.26 *** 64.10 28.21 7.69

Panel D: Importance of Foreign Currency Exposure After Swiss Franc Shock Question 15: After the Swiss franc shock, how prevalent is your firm’s foreign currency exposure in the communication with external stakeholders (e.g., investors, financial analysts)? [7-point Likert Scale: 1 to 7]

Responses (N = 39)

Average Rating

Significantly Different Than [4 = Neutral]

Not at all Important [1 or 2]

Neutral [3 to 5]

Extremely Important [6 or 7]

% of respondents who answered 4.95 ** 5.12 66.67 28.21

(Continued)

25

Table D2 (continued)

The sample comprises up to 39 answers from respondents to a survey sent to investor relations managers at firms listed on SIX during the SNB announcement.

The panels list the average ratings of the responses to the survey question indicated in the panel header as well as the percentages of responses in various bins

of the Likert scale (or any other scale mentioned in the question). If applicable, we report the results from t-tests comparing the average rating of an item to the

mid-point (where ***, **, and * stand for statistical significance at the 1%, 5%, and 10% levels). If multiple selections are available, we also compare the

average rating of an item to the average ratings of all the other items, using Bonferroni-Holm adjusted p-values at the 10% level for the comparison.

26

Table D3: Timeliness and Means of External Communication About the Swiss Franc Shock by Investor Relations Managers

Panel A: Timeliness of Initial Response Question 4: At what point in time did you start communicating with external stakeholders (e.g., investors, financial analysts, etc.) about the qualitative and/or quantitative impact of the Swiss franc shock on your firm? [Multiple Choice]

Responses (N = 39)

Within ... of the SNB Announcement More Than Three Months 1 Day 1 Week 1 Month 3 Months

% of respondents who answered 46.15 17.95 25.64 5.13 5.13

Panel B: Nature of Initial Response Question 7: Immediately after the SNB announcement (i.e., within the first one or two days), would you describe your firm’s communication with external stakeholders about the impact of the Swiss franc shock as rather reactive or proactive? [7-point Likert Scale: 1 to 7]

Responses (N = 39)

Average Rating

Significantly Different Than

[4 = Both]

Mostly Reactive [1 or 2]

Both Reactive and Proactive

[3 to 5]

Mostly Proactive [6 or 7]

% of respondents who answered 4.15 23.08 53.84 23.07

Panel C: Medium-term Communication Channels Question 9: Relative to Question 8, which information channels gained or lost importance in the two to three weeks following the SNB announcement? [3-point Likert Scale: 1 to 3]

Responses (Maximum possible N = 39)

Significantly Different Than [2 = Similar Importance]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Gained Importance

[3]

Lost Importance

[1] (1) Presentations and Q&A at already pre-scheduled events (e.g.,

investor days or conference calls)

2.41 *** 3-4 43.59 2.56

(2) Private communication with analysts or investors 2.23 ** 4 28.21 5.13

(3) Communication through media and financial press 2.13 1 and 4 25.64 12.82

(4) Ad hoc announcements or press releases 1.92 1-3 7.89 15.79

(Continued)

27

Table D3 (continued)

Panel D: Goals of External Communication Question 11: What were the most important goals of your communication with external stakeholders about the impact of the Swiss franc shock? Please indicate the importance of the following goals. [7-point Likert Scale: 1 to 7] Responses (Maximum possible N = 38)

Significantly Different Than [4 = Neutral]

% of Respondents Who Answered Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Reassure existing investor base 5.82 *** 3-9 60.53 0.00

(2) Reduce uncertainty and the “information risk” that investors

place on your company

5.65 *** 6-9 59.46 2.70

(3) Promote reputation for transparent reporting 5.26 *** 1 and 6-9 50.00 5.26

(4) Build confidence in firm’s reporting strategy 5.18 *** 1 and 6-9 42.11 7.89

(5) Build confidence in firm’s operating strategy 5.18 *** 6-9 47.37 7.89

(6) Correct/prevent undervaluation of the stock price 4.11 1-5 and 8-9 21.05 15.79

(7) Level the playing field between different types of investors 3.95 1-5 and 8-9 21.62 21.62

(8) Increase stock liquidity and/or reduce stock volatility 3.26 *** 1-7 7.89 34.21

(9) Attract new investors and increase general awareness of firm 3.26 *** 1-7 7.89 36.84

See the notes to Table D2.

28

Table D4: Information Needs of External Stakeholders After the Swiss Franc Shock

Panel A: Communication Channels for Information Requests Question 12: Compared to your firm’s last earnings announcement, how many more or less information requests through the various channels did you receive from external stakeholders after the SNB announcement? [5-point Likert Scale: 1 to 5]

Responses (Maximum possible N = 39)

Significantly Different Than

[3 = Same]

% of Respondents Who Answered Average Rating

Significantly Different Than

Much More [5]

Much Less [1]

(1) Phone calls 3.87 *** 2-4 28.95 0.00

(2) E-mails 3.72 *** 1 and 3-4 25.64 2.56

(3) Website hits 3.26 ** 1-2 3.23 0.00

(4) Downloads of financial reports or other company documents 3.25 *** 1-2 0.00 0.00

Panel B: Nature of Information Requests Question 13: Based on your impression, what type of information was particularly relevant to external stakeholders in their assessment of the Swiss franc shock for your firm? Firm-specific information about … [7-point Likert Scale: 1 to 7]

Responses (Maximum possible N = 38)

Significantly

Different Than [4 = Neutral]

% of Respondents Who Answered

Average Rating

Significantly Different Than

Extremely Important [6 or 7]

Not at all Important [1 or 2]

(1) Transactional exposure (e.g., currency mismatch between

revenues and costs)

5.13 *** 5-6 57.89 18.42

(2) Translational exposure (e.g., arising from foreign subsidiaries) 5.05 *** 52.63 13.16

(3) Hedging strategy 4.97 *** 50.00 13.16

(4) Sensitivity to indirect effects (e.g., GDP growth) 4.53 ** 21.05 7.89

(5) One-time gains/losses (e.g., on cash holdings) 4.43 1 29.73 13.51

(6) Operating and strategic responses (e.g., relocation decisions) 4.14 1-3 21.62 21.62

Panel C: Usefulness of Existing Annual Report Disclosures Question 14: Based on your impression, how relevant was the information provided in the existing annual report and financial filings to external stakeholders in their assessment of the Swiss franc shock immediately after the SNB announcement (i.e., within the first one or two days)? [7-point Likert Scale: 1 to 7]

Responses (N = 39)

Average Rating

Significantly Different Than [4 = Neutral]

Not at all Important [1 or 2]

Neutral [3 to 5]

Extremely Important [6 or 7]

% of respondents who answered 4.74 *** 7.69 66.67 25.64

See the notes to Table D2.