WORKING DRAFT. NOT FOR ATTRIBUTION.

Revealing Climate Change Opinions through Investment Behavior: Evidence from

Fukushima

Zhen Lei*

Anastasia Shcherbakova†

October 2014

Abstract: In this study we present a novel research approach to obtaining behavior-based evidence of regional climate change attitudes, using the 2011 Fukushima nuclear plant incident as a natural experiment. Our approach allows us to produce the first non-survey-based empirical evidence of a trans-Atlantic divide in public opinion on the environment and climate change, and an estimate of the relative monetary value that investors assign to fossil-based and renewable energy. This value is based on the perceived potential of these fuel types to substitute for nuclear generation in the aftermath of the Fukushima crisis. We carry out an event study to examine differences in abnormal returns of coal and renewable energy companies on European and American stock exchanges. We find that investors trading on U.S. markets exhibit a significantly more favorable perception of the profitability of coal stocks (generating a $4.2 billion increase in cumulative coal market capitalization), while investors trading on European exchanges display a more favorable perception about the profitability of renewable energy stocks (adding over $3 billion to cumulative renewable market capitalization).

Key words: Climate Change, Public Opinion, Event Study, Financial Markets, Stock Returns, Fukushima

* Leone Family Department of Energy and Mineral Engineering, Penn State Univeristy† Naveen Jindal School of Management, University of Texas at Dallas

1

1. Introduction

There has been an evident trans-Atlantic divide on climate policies in the United

States and Europe. Europe has set an ambitious emissions reduction target under the

Kyoto Protocol, implemented multiple policies and measures including cap-and-trade to

achieve that target, and played an important role in post-Kyoto international negotiations.

U.S. legislators, by contrast, failed to support the Kyoto Protocol, often questioned the

basic science of climate change, and failed in 2010 to enact a climate bill to launch a cap-

and-trade scheme to curtail carbon emissions (Carlarne, 2006; Schmidt & Haifly 2012;

Skjærseth, Bang and Schreurs, 2013).

Such a trans-Atlantic divide on climate policy has been, at least in part,

attributable to a reported difference in public opinion on climate change between the

United States and Europe. For example, surveys conducted in 2009 by the Pew Research

Center and the World Bank show that 44 percent of U.S. respondents believed at that

time that global warming was a very serious problem. A much higher percentage of

respondents in France (68 percent) and Germany (60 percent) agreed with this statement.

Only 46 percent of U.S. respondents thought climate change would harm people in their

country now or within 10 years, whereas 67 percent of respondents in France believed so

(Brechin and Bhandari, 2011).‡

Surveys and opinion polls presently constitute the primary method of obtaining

individuals’ views on climate change and other environmental concerns (Brechin and

‡ The Brechin and Bhandari (2011) study only looks at relative differences in surveyed opinions that the authors compile from multiple polls. Although the authors convey that substantial differences in responses exist, they do not carry out any statistical tests to evaluate these differences.

2

Bhandari, 2011; Lachapelle et al., 2012; Lorenzoni and Pidgeon, 2006). However, the

political and emotional charge that fuels the debate on climate change makes obtaining

accurate and honest views on such subjects rather challenging. While useful, public

opinion polls give us only a relatively simple—and often general—sense of how people

feel about a certain issue. Moreover, self-reported opinions have been found to

misrepresent individuals’ true beliefs and corresponding actions (Podsakoff et al., 2003).

For instance, survey-based studies appear to lead to a paradox that “Americans seem

concerned about global warming, yet view it as less important than nearly all other

national or environmental issues” (Leiserowitz, 2006).

This paper takes a novel approach to understanding individuals’ true attitudes

toward climate change by directly observing actions that people take when presented with

a potential energy and environmental crisis—actions that reveal their perceptions of how

grave environmental and climate change concerns are. Being able to observe what

people do in response to a challenge, rather than what they say, provides more

compelling evidence of one’s understanding of said challenge, especially when observed

actions are tied to monetary outcomes.

More specifically, we examine behavior of investors in U.S. and European

financial markets during the unfolding of the Fukushima crisis that reflects their

perceptions about future profitability of various energy assets (i.e. stocks of fossil fuels

and renewable energy). Our approach exploits an intuitive link between people’s attitudes

toward the concept of climate change and their relative perceptions for the profitability

prospects of different energy sources. Survey-based studies have found that people more

3

concerned about climate change tend to have a more optimistic and positive evaluation of

renewable energy than of fossil fuels.§ This implies that an individual’s level of

environmental and climate concern likely frames his assessment of the relative future

profitability of various energy types. We draw on this link to gauge potential regional

differences in climate change attitudes through actual investing behavior of financial

market participants in the United States and Europe. These participants are mostly local

investors, as established by the home country bias literature.**

We also address an important challenge of differentiating investor perceptions of

the global profit potential of specific energy types on one hand from investor

expectations about local policy response to the Fukushima crisis on the other. The

former would be based on global profit expectations and thus provide evidence of

investor’s environmental biases. The latter may have nothing to do with one’s

environmental attitudes, as it would impact investment decisions through only local profit

expectations.

We use Japan’s Fukushima nuclear power crisis of March 2011 as a natural

experiment, and base our study on stock market data from the time period encompassing

this event. We apply event study methodology to these data to investigate how investors

§ For example, analysis of a large British public attitude survey by Spence et al. (2010) finds that both concerns about climate change and general environmental concerns are linked with a positive evaluation of renewables.** The home country bias, a result that investors’ equity portfolios tend to remain heavily concentrated in their domestic stock markets, has been studied for decades and remains one of the most important puzzles in international finance. The vast majority of investors does not hold foreign securities or invests only a small portion of total assets in foreign stocks, despite a relatively low correlation between stock returns of various countries and the potential benefit from international diversification. See, for example, French and Poterba (1991), Abreu, Mendes and Santos (2011), and Grinblatt and Keloharju (1991).

4

in U.S. and European financial markets, with potentially different attitudes toward

climate change and the environment, envisioned the worldwide response to the nuclear

crisis. ††

As emergency workers struggled to contain the scale of the Fukushima crisis in its

initial days, public support for nuclear power weakened noticeably and the prospects of

nuclear power worldwide dimmed.‡‡ A decline in nuclear generation without an offsetting

reduction in energy demand implies a need for compensating generation from other

energy sources.

The anticipated shift in the global generation portfolio led investors to rebalance

their energy asset holdings. We argue that the rebalancing was done in a manner

consistent with investor perceptions of how the world would respond to the nuclear crisis

(that is, their expectations about which energy type would become the most likely

substitute for nuclear generation). The increase in demand for the products and services

of these energy companies would make them more profitable and therefore more

†† Event study methodology has been widely used in finance and economics literature (Fama, 1965). There are studies that apply the event study methodology and other empirical approaches to evaluate the effect of the Fukushima crisis on financial returns of energy firms in a single region. They find that nuclear utilities (particularly Japanese utilities in the affected area and those operating nuclear power plants) suffered a severe decline in returns and a significant increase in systematic risk (Kawashima & Takeda 2012), while European nuclear and renewable sectors experienced short downward and upward returns adjustments, respectively (Ferstl et al. 2011).‡‡ Most notably, Germany committed to phase nuclear capacity out of its energy portfolio completely by the year 2022, and China halted nuclear projects due to safety uncertainty. Many other nations followed suit with safety checks and delays of new nuclear project approvals (Joskow & Parsons, 2012). In retrospect, the negative impact of the Fukushima crisis on nuclear power may have been exaggerated. Duffy (2011) notes that a global nuclear renaissance was unlikely prior to the Fukushima incident and Joskow & Parsons (2012) argue that the actual effect of Fukushima is unlikely to result in as drastic a reduction on nuclear capabilities as has been widely conceived.

5

attractive to investors.§§ Because the actions we observe have direct financial outcomes,

our approach provides evidence that reflects the relative monetary value that investors

assign to fossil-based and renewable energy in the immediate aftermath of the Fukushima

crisis.

Our results indicate that investment behavior does indeed reflect investors’

environmental perceptions. We observe positive financial effects accruing to a portfolio

of global coal companies traded on U.S. markets, on the order of 8.6 percent in excess

returns or $4.2 billion in market capitalization, but not to those listed on European

exchanges. By contrast, European exchange data reveal larger positive outcomes for a

portfolio of global renewable energy companies than was observed on U.S. exchanges.

Over the same time period, renewable stocks traded on European markets earned total

excess returns of 13.8 percent (or over $3 billion in market capitalization), while

renewable companies listed on U.S. exchanges gained only 3.8 percent in excess returns

(which translates to about $1.7 billion in market capitalization). These results suggest that

investors active on U.S. exchanges were more skeptical of climate change and perceived

more optimistic prospects for coal energy, relative to investors active on European

exchanges. European investors, on the other hand, were likely to be more concerned

about the effects of climate change and thus formed a more favorable profitability

expectation for renewable energy.

To the best of our knowledge, this is the first study to estimate the difference in §§ An investor’s objective is to maximize the return on his investment, given some risk tolerance. Following the Fukushima nuclear crisis, an investor would rebalance her asset portfolio toward those financial assets she believes will be more profitable in the future. For energy stocks, this implies companies associated with that energy type that is most likely to make up for a reduction in nuclear generation.

6

monetary value placed by investors on coal and renewable energy assets in the United

States and Europe, as reflected by their prospects in the aftermath of the Fukushima

crisis. Having a quantitative assessment of investors’ environmental perceptions is

important because it leads to creation of more realistic and effective environmental

strategies. For instance, no matter how much U.S. environmental regulators favor

renewable generation, they will not be able to encourage capital flows to renewable

sources if investors don’t foresee high profitability potential for renewable energy and

technology companies. Since investors’ capital allocation decisions over time feed into

firm and industry profitability, then actions taken by owners of capital implicitly

determine which government policies stand to succeed or fail.

This study is also the first to examine actual behavior, rather than opinions

reported to surveyors and polling organizations, to provide empirical evidence on the

trans-Atlantic difference in public environmental and climate change opinions. It is worth

noting that because financial investors comprise a significant but partial share of any

country’s adult population, our results reflect environmental attitudes among a subsample

of the general populace.*** However, provided that behavior-based evidence is more

*** In 2011 (the year on which our analysis focuses), households held approximately 37 percent of the total value of U.S. equities directly, while institutional investors owned 63 percent. In addition to direct holdings, households also participated in equity markets indirectly through mutual funds (44 percent of all U.S. households) and pension funds (69 percent) (see Investment Company Institute’s 2014 Investment Company Fact Book, available at http://www.ici.org/pdf/2012_factbook.pdf, and table L.213 of the Federal Reserve Statistical Release Z.1, available at http://www.federalreserve.gov/releases/z1/current/annuals/a2005-2013.pdf).Participation in the stock market is usually skewed toward higher education and income levels. In 2013, nearly 52 percent of all U.S. adults participated in equity markets either directly or indirectly. Among college graduates participation rate was 77 percent and among adults with incomes of at least $75,000 it was 80 percent (see Gallup and Pew Research Center’s numbers here: http://www.gallup.com/poll/162353/stock-ownership-stays-record-low.aspx and here:

7

compelling than evidence drawn from population surveys and opinion polls, our approach

presents a good complement to the survey-based approach. Our regional approach can

also be scaled down and used to evaluate country-level differences in public attitudes

toward climate change.†††

The paper is organized as follows: in the next section we provide a sketch of our

research framework and hypotheses, as well as the econometric specification used in our

main empirical analysis; Section 3 provides details about our data sources and

stratification methods for data samples; Section 4 presents a discussion of our empirical

results; and Section 5 concludes with some policy implications.

2. Research Framework

2.1 Global profitability prospects vs. local policy response

One challenge we face in this study is in distinguishing between investors’

perceptions of global energy profitability and their expectations about local policy

response to the Fukushima crisis. On one hand, investors form profitability expectations

based on their understanding of the global response to the crisis, which survey-based

evidence suggests is framed by their environmental attitudes (Spence et al., 2010). On the

other hand, investors’ expectations about how domestic policymakers will respond to the

Fukushima crisis will also affect investment decisions by changing the relative profit

http://www.pewresearch.org/fact-tank/2013/11/18/dow-soars-but-only-about-half-of-americans-will-benefit/).††† Country-level analysis has also been popular in the survey/opinion poll literature. Lachapelle, Borick and Rabe (2012), for example, use telephone surveys administered in the United States and Canada to compare public attitude toward climate science and climate policy between the two nations.

8

potential of various energy firms locally. To make sure that we are observing activity

driven by global profit perceptions, we restrict our data sample to publicly traded

electricity fuel source companies, namely producers of coal and renewable energy

technologies, that sell their products or have operations in multiple global markets, and

calculate outcomes based on sampled stocks’ trading and operating regions.‡‡‡

In contrast to companies operating solely in the United States or European Union,

the overall profitability of energy firms with such global operations is determined by

multiple markets (including both the United States and European Union) and is thus not

fully dependent (and in some cases, not at all dependent) on the region in which the stock

shares of these firms are traded. This means that implementation of local policies

favoring a single type of energy source would have a muted effect on profit potential of

firms dealing in that specific energy type. This allows us to rule out local policy

expectations as the main cause of any differences observed in local stock market

outcomes. We hypothesize that such differences, if observed, would instead be driven by

investors’ perceptions about global profitability prospects of certain energy sources,

which reveal their intrinsic attitudes toward climate change.

‡‡‡ We focus on coal and renewable energy for two reasons. The first is that these two sources of electricity constitute a sharp contrast in terms of environmental impacts and carbon emissions. The second is that it is feasible to identify publically listed and globally operating firms that either mine coal or manufacture renewable technologies. We exclude results for natural gas—another important source of electricity generation—from the main discussion due to several ambiguities associated with it. Natural gas is cleaner than coal, but still emits carbon dioxide and other pollutants. It is also difficult to identify a sizable group of firms that focus solely on natural gas production; major gas companies also produce oil, and provide midstream services. For robustness purposes, we identify and analyze a sample of natural gas producers and midstream/service companies. We present their results in Appendix B.

9

Let us provide some support for the above hypothesis through a simple example.

China MingYang Wind Power Group and XinJiang Goldwind Science & Technology

Company are two prominent Chinese wind turbine manufacturers. MingYang Wind

Power is traded on NASDAQ in the United States, while XinJiang Goldwind is traded on

Germany’s Frankfurt Stock Exchange. Both companies sell and install wind turbines in

China, United States, Europe, and other global markets. If, in response to the Fukushima

crisis, European governments decided to boost support for renewable energy projects but

the U.S. government did not, the impact on the economic prospects of MingYang would

not necessarily be worse than on XinJiang, because both firms have sales in the U.S and

in Europe. Therefore, assuming that investors in the United States and Europe are

sophisticated, have equal access to information and make rational assessments about

firms’ future profit potential, the demand for and prices of these two companies’ stocks,

determined by their global potential for future profitability, should not vary from one

regional exchange to another. However, if investors in the United States are in general

more skeptical about climate change, they may on average have a more optimistic view

about the prospect of coal-fired electricity as a viable substitute for nuclear energy in both

the United States and Europe, relative to their European counterparts. Hence, investors’

demand for MingYang stock in U.S. markets might be lower than that for XinJiang stock

in European markets, consistent with investors’ translation of expected global response

into profitability expectations.

One might also be concerned about the presence of naïve investors. If investors do

not understand well the global nature of the sampled companies, they may simply assume

that companies whose stocks are listed on their regional exchanges mainly operate within 10

their region. Such investors would then base profitability expectations for various energy

stocks on their anticipated domestic policy response to the Fukushima crisis, and

rebalance their energy investments accordingly. However, our data sample contains

energy stocks that have global operations and are traded on multiple exchanges around

the world. The naïve assumption would then lead an investor to give equal treatment to

all stocks within the same energy category and thereby assign inappropriate profitability

expectations to stocks of companies that have financial and physical presence in several

regions around the world. To account for this possibility, we carry out separate analyses

on securities that are listed on financial markets of only one geographic region, and those

cross-listed in both regions. If naïve investors dominate and both cross-listed and region-

specific stocks are given equal treatment, we should observe no difference in returns

outcomes among these two categories of stocks.

Finally, there is also the possibility that a global company may nevertheless do the

majority of its business in a single region, which may drive its decision to list its shares

on the local market. In this case, stock market responses to the Fukushima crisis in

different regions could still reflect investors’ expectations about domestic policies in the

aftermath of the crisis, rather than their profitability expectations for certain fuel types

that are influenced by environmental opinions. To check this, we refine our sample

further by focusing only on those companies that have major operations outside of their

listing region (see footnote 18). Focusing only on companies with major operations

outside of their listing region would lend further support to the hypothesis that investors’

responses to the crisis reflect their attitudes toward climate change and their global profit

expectations for certain energy types.11

2.2 Event study methodology

The importance of correctly identifying the event day and length of the event

window has been highlighted and addressed in event study literature.§§§ Our task of

defining the event day is simplified by the fact that the earthquake that struck northern

Japan, being a natural disaster, was in all likelihood unexpected. We confirm the

accuracy of this event day by referencing the Fukushima Nuclear Accident Update Log,

in which the International Atomic Energy Agency (IAEA) recorded all Fukushima

developments. The first IAEA Fukushima Nuclear Log update concerning problems at

the nuclear plant was issued at 21:10 GMT on Friday, March 11, 2011. This corresponds

to 16:10 Eastern Standard Time (10 minutes after the closing bell at NYSE and

NASDAQ) and 22:10 Central European Time (well after trading stopped at all European

markets).**** We define the event day (day 0) as March 11, 2011 for both U.S.- and

European-listed securities and include one trading week of pre-earthquake data in our

event window to illustrate any diversions in financial trends in the earthquake’s

aftermath.

§§§ Hillmer & Yu (1979) and Chang & Chen (1989) examine the speed with which new information is incorporated into stock prices, providing evidence that temporary arbitrage opportunities do exist in financial markets. Binder (1998) and Kothari & Warner (2006) present evidence that mis-specifying the event day or the length of the event window often leads to incorrect estimates of the effect of new information or low statistical power of significance tests. Krivin et al. (2003) suggest several empirical methods of more accurately defining the proper event window length.**** As shown in Table 1, since U.S. and European stock markets were closed at the time of the earthquake, there were no significant changes in trading activity of U.S. investors on the day of the earthquake (trading day 0, column (3)) in response to the first news of the Fukushima event. Although the coefficient estimate on nuclear firms is positive (0.97), it is not statistically significant. Investors were not able to respond to the news until the following Monday, March 14th.

12

We then proceed to define an appropriate event window—the relevant time frame

during which the Fukushima crisis affected stock market outcomes. Our data includes

three trading weeks of post-earthquake observations, since we are interested in short-term

market response to the Fukushima crisis, for which investor perceptions and instincts

play important roles.†††† We use post-earthquake data to construct two measures of stock

responses, abnormal stock turnover and abnormal stock returns, to determine the timing

and duration of investor activities following the earthquake.

Estimating event window length based on media announcements

Pinpointing an accurate length of an event window is more difficult than

specifying an event day. One classic approach is a review of prominent media

publications like the Wall Street Journal, The New York Times, and others.‡‡‡‡ Such

review generally gives a good indication of when and for how long new information was

considered to be of importance, but can in some cases prove to be insufficiently precise in

defining the event window. We attempt to improve upon this by employing two

alternative methods to define the relevant time frame. In the first variation we refine the

usual approach by narrowing our review to only those information updates that appeared

on the IAEA Fukushima Update Log. The Log suggests that the most severe concerns

revealed themselves within the first seven to 10 days after the earthquake.§§§§ This leads †††† Also note that long-horizon event studies could have some significant drawbacks, such as increased potential to misspecify the normal return generating process and reduced ability to detect abnormalities in financial performance (Kothari & Warner, 2006).‡‡‡‡ See, for example, Brown & Warner (1980), McWilliams & Siegel (1997), and Swanson (2011). Indeed, a quote on page 249 of Brown & Warner (1980) reflects how important such publications are: “…even if the researcher doing an event study has a strong comparative advantage at improving existing methods, a good use of his time is still in reading old issues of the Wall Street Journal…”§§§§ A complete list of news stories and sources is available from the authors.

13

us to define a maximum of 15 trading days for the event window length in order to

account for any second-order effects of the earthquake and residual responses by

investors.

Estimating event window length based on stock turnover

In our second approach of identifying the proper length of the even window, we

allow data on trading activity to suggest the appropriate cutoff, rather than imposing any

arbitrary limits (a method employed in Krivin et al., 2003). More specifically, we use

daily turnover information (the portion of a company’s total existing shares traded on a

given day) on the New York Stock Exchange to determine statistically significant

abnormalities in trading patterns. A significant change in a stock’s trading activity is a

signal that investors’ perception of that stock’s future profitability has changed. We

expect that abnormal trading volumes post-Fukushima will persist until all significant

new information is absorbed by financial markets, after which stock turnover should

return to its normal levels. The details of this approach are described below.

To account for intra-day correlation of outcomes among firms specializing in the

same energy source, we aggregate firms in each of the three fuel source categories

(nuclear, coal and renewables) into portfolios. We then construct a daily portfolio

turnover variable. This is done by calculating the ratio between each portfolio’s trading

volume on any given day (the sum of volumes across all firms within a portfolio) and the

number of its shares outstanding on that day (Eq. 1).

14

, (1)

where refers to the turnover of portfolio p on day t; is the traded volume of

stock i (assigned to portfolio p) on day t; is the total number of shares outstanding

of stock i on day t; and is the total number of firms included in portfolio p. We calculate

market-level turnover similarly to portfolio-level turnover, substituting NYSE-wide trading

volumes and shares outstanding for firm-level data.

We then estimate the market model for “normal” daily turnover at the portfolio level

(normal turnoverp,t) using 2010 data, conditioning on daily market-level turnover ( )

as in Eq. 2. The estimated residuals, εp,t , of Eq. 2 serve as the reference (or normal) pre-

earthquake daily abnormalities in turnover, and their standard deviation serves as the population

standard deviation under the null hypothesis that there is no effect of the Fukushima crisis on

stock turnover.

(2)

We apply estimated parameter coefficients of α0 and α1 of Eq. 2 to generate

theoretical “normal” daily portfolio turnover in the days following the earthquake, and

subtract this from actual observed turnover to generate the resulting abnormalities

(abnormal turnoverp,t) for the days following the earthquake (Eq. 3).

(3)15

We evaluate statistical significance of each daily post-earthquake abnormality in

stock turnover using z-statistics and their corresponding critical values, constructed from

abnormal turnover of the year 2010 (Eq. 4).

z = (abnormal turnoverp,t≥0) / σ(abnormal turnoverp,t∈2010) (4)

The results of turnover analysis are shown in Figure 1 and Table 1, both

presenting the daily turnover abnormalities for all available post-earthquake data for each

of the three different portfolios of energy firms. Daily abnormal turnover is statistically

significant (see Table 1) during the first three, four, and seven trading days after the

earthquake for coal, renewable and nuclear portfolios, respectively, which is consistent

with the event timeline suggested by the IAEA Fukushima Log. These results suggest

that the majority of new event-related information was incorporated into financial

markets during the first calendar week of the incident, though residual trading

abnormalities persisted into the second week.

Based on this turnover evidence, we expect that the majority of abnormalities in

stock returns, to which we turn next, will occur within three to seven trading days

following the earthquake. For robustness purposes, we calculate and present results for

the full 14-day post-event data sample.

Estimating the effect of Fukushima on energy stock returns

Our main outcome of interest is the return abnormalities accruing to securities of

each type of energy firm in the U.S. and European markets. Our returns analysis follows

the same approach as the examination of turnover described above. To improve the

16

efficiency of our estimates, we carry out all estimation at the portfolio level, using value-

weighted portfolio returns, , to account for firms of varying size, as in Eq. 5:

, (5)

where is the daily return of security i on day t, and is the security-specific weight, set

equal to the market capitalization of the company on the previous day (calculated as the product

of the previous day’s stock price, pricei,t-1, and number of stock shares outstanding, outstandingi,t-

1).

We estimate the normal value-weighted returns trend for each industry portfolio p

in each region, the United States and Europe, via the market model, utilizing year 2010

data (Eq. 6).

(6)

As before, we subtract estimated normal portfolio returns from observed pre- and

post-earthquake portfolio returns, generating daily abnormalities. We then calculate the

population standard deviation of abnormal returns for each region and portfolio using

abnormal returns of the year 2010, and construct a set of daily post-event z-statistics to

determine the statistical significance of daily abnormal returns of each energy portfolio

after the onset of the Fukushima crisis, in each of the two regions examined.

We repeat the estimation procedure on abnormal returns for subsamples of cross-

listed and region-specific firms, and for region-specific firms with major operations in

17

and outside of the listing region (see last paragraph of section 3.2 for more details).

3. Data

3.1 Analytic sample

Our data sample covers the time period between January 4, 2010 and March 31,

2011, and comes from three primary sources. The first is the Center for Research in

Security Prices (CRSP) database, from which we extract daily observations on nine

companies with nuclear operations, four global nuclear stock indices, 15 coal firms, and

42 renewable energy (solar and wind) companies listed on three U.S. Stock

Exchanges.*****,††††† These data include each company’s identifying characteristics, listing

status, industrial classification code, primary exchange, number of trades, trade volumes,

closing price, shares outstanding, and returns, as well as the value-weighted return of the

exchange.

We supplement these data with daily aggregate observations on New York Stock

Exchange’s (NYSE) shares, number of trades, and dollar volume of trades for the same

time period. ‡‡‡‡‡

Our third data source, the Bloomberg database, provides European stock market

data on seven coal and 40 renewable energy (solar and wind) companies listed on ten

***** The global nuclear stock indices are exchange-traded funds (ETF) composed primarily of electrical utilities with nuclear generating assets, nuclear generation equipment manufacturers, and uranium miners.††††† The New York Stock Exchange (NYSE), NYSE Amex, and NASDAQ are included. The full sample of companies was extracted using relevant SIC codes, then narrowed down to producers with international sales and operations by reviewing each company’s annual reports.‡‡‡‡‡ These data can be downloaded from www.nyxdata.com .

18

exchanges in six countries.§§§§§ Among the variables included in this dataset are

companies’ identifying characteristics, industry classification, primary listing exchange,

number of trades, trade volumes, closing price, and level and percent returns. For a

complete list of all firms included in this study, please refer to Tables A1 through A3 in

Appendix A.

3.2 Sample stratification

Cross-listed and region-specific firms

The first refinement of our data entails dividing our sample of firms into two

groups. The first is those stocks that were traded on both U.S. and European exchanges at

the time of the crisis (cross-listed firms). The second group is those stocks whose listing

was unique to only one region (region-specific firms). The main hypothesis behind this

sub-analysis is that firms that are traded in both regions are likely subject to arbitrage.

Consider investors tracking energy stock prices in both U.S. and European markets.

Suppose these investors notice that the price of a cross-listed coal stock increases on U.S.

exchanges but not on European exchanges. This would make the stock traded on

European exchanges appear to be underpriced relative to its U.S. counterpart, and

encourage traders to buy shares in European markets, in turn reducing the regional price

gap. Firms traded in only one region will be less likely to be subject to such arbitrage

activities and will thus reflect perceptions of local investors more accurately. If there

exist significant differences in investors’ attitudes toward climate change and profit

§§§§§ European exchanges included in analysis are Euro Comp and the stock markets of Germany (Berlin, Frankfurt, Munich, and Stuttgart), Belgium (Brussels), the United Kingdom (London), Spain (Madrid), Italy (Milan), and France (Paris).

19

expectations toward certain energy sources, we would be more likely to see a difference

in returns outcomes among region-specific energy firms than among cross-listed

securities.

Moreover, this refined methodology allows us to test the extent to which naïve

investors dominate financial markets. As discussed earlier, naïve investors assume that

companies whose stocks are listed on their regional exchanges mainly operate within

their region. This leads naïve investors to treat all stocks within the same energy category

equally, and suggests that any observed changes in returns of a category of energy stocks

should be similar for every stock within a category, without regard for each company’s

actual location of operational and financial presence. This test of the dominance of naïve

investors is important because such investors make it difficult to distinguish between

market outcomes driven by expectations about local policies and expectations about

global profitability, rooted in sophisticated investors’ attitudes toward the environment

and climate change.

Major operations relative to listing region

In our second refinement, we divide region-specific firms into those whose main

operations occur within the listing region and those that have major operations outside of

the region in which their stock shares are traded.****** This affords us a more robust test

****** In order to determine whether or not a company has major operations outside of its listing region, we examined the websites and financial reports of all region-specific companies in our data sample. Those companies that prominently mentioned regions of operation that lay outside of their listing region were categorized as having major operations outside the listing region. For example, Peabody Energy and Arch Coal are among the top coal companies in the U.S. Both trade on the New York Stock Exchange, but paint a different picture about their global presence in their operating statements. Peabody Energy lists its operations in “US mining, Australia mining, Peabody in China, Peabody in India, Peabody in Indonesia and Peabody in Mongolia.”

20

of the differences between investors’ global profit expectations (which are framed by

their environmental opinions) and their local policy expectations. For example, if

renewable energy companies traded on European exchanges that have major operations

in the United States see abnormally high returns, it would suggest that European investors

act more on their preferences for renewable energy than on their expectation of European

policy response to the Fukushima crisis. Such a result would thus provide evidence in

favor of our principal hypothesis of global profit-driven investment behavior, which, we

argue, is informed by investors’ own environmental attitudes.

4. Results

Our main results are contained in Figures 2 and 3 and Tables 2 through 4. In

order to give the reader a more intuitive sense of the full effect of the Fukushima crisis on

returns outcomes, Figures 2 and 3 display a cumulative sum of abnormal returns over

time for coal and renewable portfolios, respectively. To supplement the discussion that

follows, Tables 2 and 3 provide daily abnormal returns figures and their respective z-

statistics, from which the Figures were derived.

Our results indicate that investment behavior does reflect investors’

environmental attitudes. We observe positive financial effects accruing to a portfolio of

global coal companies traded on U.S. markets, on the order of $4.2 billion, but not to

those listed on European exchanges. Relatively inexpensive coal-fired power plants

cause more pollution and carbon emissions per kilowatt-hour (kWh) of electricity than

Arch Coal writes that, along with its subsidiaries, it has “a leading position in every major U.S. coal basin, with mining complexes in Wyoming, Colorado, Illinois, West Virginia, Kentucky, Virginia and Maryland.”

21

nuclear power, natural gas, and renewables. A stronger positive response toward coal

generation by investors active on U.S. exchanges suggests that, relative to investors

active on European exchanges, those active on U.S. exchanges were more skeptical of

climate change and had a more optimistic assessment of the potential of coal-based

power to replace nuclear generation in the aftermath of the Fukushima crisis.

Figure 2 exhibits cumulative abnormal returns of coal securities traded on U.S.

and European financial exchanges.†††††† Following the earthquake at trading day 0, we

can see that the paths of the two regions’ abnormal returns diverge, with U.S.-listed coal

firms accruing excess gains and European coal companies incurring slight earnings

losses. This diversion in trends suggests that U.S. and European investors’ perception of

coal companies’ post-Fukushima profit potential differed significantly. Columns (1) and

(2) of Table 2 show the daily abnormal returns for these coal stocks. Excess gains of

U.S.-listed coal companies are higher than excess gains of their European counterparts.

Excess returns in the United States were not statistically significant after the third

day of the event window, suggesting that financial markets absorbed Fukushima’s impact

on the U.S. coal industry during a very short time period. If we sum up only those daily

return abnormalities that are statistically significant at the five percent level or better

during 14 trading days following the earthquake, we get a cumulative excess gain of 8.6

percent for U.S. markets. To put it another way, U.S. investors were willing to pay 8.6

percent more for a share of coal stock on March 31, 2011 (14 trading days after the

earthquake that triggered the Fukushima crisis) than on March 11, 2011. This gain

translates to a combined increase in market capitalization of U.S.-listed coal companies †††††† Due to data limitations, there are no cross-listed coal firms in our sample.

22

of $4.2 billion over the course of just three days.‡‡‡‡‡‡

Restricting the sample further to only those coal companies that have major

operations outside of their listing region, we see that the results remain qualitatively

similar, and somewhat more pronounced in magnitude (see panel (b) of Figure 2 and

columns (3) and (4) of Table 2). This supports the notion that local policy expectations

were not the driving force behind investors’ response to the Fukushima crisis.

By contrast, European exchange data reveal larger positive outcomes for a

portfolio of global renewable energy companies, when compared to the stock price boost

realized by U.S.-listed renewable firms (see Figure 3 and Table 3). Adding up

statistically significant abnormalities over 14 post-earthquake trading days, we find that

renewable firms traded only on European exchanges earned cumulative excess returns of

13.8 percent. Their U.S.-traded peers realized a cumulative excess gain of only 3.8

percent over the same time period. So on March 31, 2011, European investors were

willing to pay 13.8 percent more for a share of renewable energy stock than on March 11,

2011, while U.S. investors were willing to pay only 3.8 percent more. This translates to

an increase in cumulative market capitalization of more than $3 billion for renewable

energy companies traded on European exchanges, and $1.7 billion for renewables

securities traded on U.S. markets.

The observation that investors operating on European exchanges bid up returns to

‡‡‡‡‡‡ We calculate the statistically significant increase in aggregate market capitalization of coal and renewable companies contained in our sample as follows. First, we take the aggregate event-day (day 0) market capitalization of all coal or renewable energy companies listed on a specific region. We then multiply this aggregate market capitalization by the statistically significant cumulative abnormal return observed in that region over 14 subsequent trading days.

23

solar and wind power companies more than their U.S. counterparts provides evidence that

investors active on European exchanges were likely to be more concerned about the

effects of climate change and as a consequence held more optimistic prospects for

renewable energy. This result is particularly interesting given the fact that wind and solar

power are relatively more expensive to produce (compared to coal-based electricity), less

reliable due to their non-dispatchable characteristics, and require more technological

innovation.

Refining our analysis by regional presence, we see that positive abnormalities in

returns of global renewable energy companies traded only on European exchanges

(region-specific companies) were twice as large in magnitude as those accrued by

renewables stocks traded only on U.S. markets. This difference is even more exaggerated

among firms the majority of whose operations occur outside of their listing region—

further evidence in favor of the environmental perception hypothesis.

Cross-listed renewable energy stocks demonstrated no difference in return

abnormalities, suggesting some arbitrage activity among investors in the two regions.

This result serves as evidence against the scenario of a naïve investor acting on

expectations of domestic policy changes. Naïve investors, in anticipation of a more

favorable policy response toward renewable energy in Europe following the crisis, would

have had a higher demand for all renewable energy stocks traded in Europe, even those

that may also have been traded elsewhere in the world. Since we do not observe a rise in

relative value of cross-listed renewable stocks, we are able to corroborate the idea that it

is global profit expectations stemming from climate change attitudes, and not local policy

24

expectations, that drive observed regional differences in responses to the Fukushima

crisis.

Finally, as shown in Table 4, U.S.-traded firms and mutual fund indexes with

significant exposure to nuclear activities amassed statistically significant losses over the

first two days following the Japanese earthquake, confirming investors’ pessimism

toward nuclear power as a consequence of the crisis. The magnitude of these losses was

more substantial for global nuclear indexes that experienced an initial two-day loss of

16.6 percent, followed by a 7.5 percent upward correction on days five and six, and

another loss of 4.2 percent on days 11 and 12. The majority of firms included in the first

nuclear sample (column (1) of Table 4) are electricity generators and nuclear operations

do not comprise the entirety of their business. Thus, return movements of the nuclear

ETF indexes (column (2) of Table 4) are likely to be more representative of the

Fukushima incident’s financial implications for the global nuclear industry.

5. Conclusion

In this paper we provide a novel approach to investigating the trans-Atlantic difference in

public opinion on climate change and the environment, using financial investors’ actions, rather

than self-reported opinions. We exploit the link between people’s attitudes toward climate

change and their profitability perceptions of certain energy sources and use the 2011 Fukushima

crisis as a natural experiment to evaluate responses of U.S. and European investors to a shifting

view of nuclear power. We analyze return abnormalities to stocks of energy companies that

operate in multiple global markets. We also narrow our sample to companies with major

operations outside of their listing region, whose expected profitability would not be determined

25

by a local policy response to the Fukushima crisis. Our results uncover a more optimistic

perception about coal firms among investors trading on U.S. markets, and more favorable views

of renewable energies among investors trading on European exchanges.

As concerns about nuclear power grew during the crisis, investors active on U.S.

exchanges perceived coal to be the most likely substitute for the coming dearth of nuclear

generation. This perception meant that demand for coal assets, and therefore coal stocks,

would rise and thus generate excess returns for global coal producers. Coal companies

traded on U.S. stock exchanges realized cumulative excess returns of 8.6 percent,

equivalent to a $4.2 billion increase in their combined market capitalization. Coal stocks

traded on European exchanges, by contrast, realized no excess returns. Investors trading

on European markets, on the other hand, perceived renewable energy companies as the

most viable replacement for nuclear generation. They thus bid up demand for renewable

stocks enough to generate excess returns of 13.8 percent – 10 percentage points higher

than for similar firms traded on U.S. markets. This translates to an aggregate increase in

the value of these firms of over $3 billion – nearly double the $1.7 billion increase in

aggregate value of renewable securities traded on U.S. exchanges. These results hold

even for companies with the majority of operations outside of their listing region and

suggest that investors on European exchanges at that time valued the incremental

potential of renewables to contribute to power generation in a post-Fukushima world

more than investors on U.S. exchanges did.

This study provides the first quantified financial evidence that reflects relative

climate change attitudes among investors in the two examined regions, manifested by

their observed investing actions. It is worth noting that because European investors

26

transact on U.S. exchanges and vice versa, the difference we observe in responses of

financial markets in the United States and in Europe could underestimate the difference

in views about climate change and perceptions about energy assets between the two

groups of investors. Additionally, investors may have used the Fukushima crisis to seek

out arbitrage opportunities. The arbitrage effect would attenuate our coefficient estimates

further.§§§§§§ Even with all these potential attenuations, however, our results validate and

supplement opinion polls administered to the general public. The quantification of

investors’ environmental sentiment is a helpful step in delineating regional climate

attitudes.

References

Abreu, Margarida, Victor Mendes and Joao A.C. Santos, “Home Country Bias: Does Domestic

Experience Help Investors Enter Foreign Markets?” Journal of Banking & Finance,

2011, 35, 2330-2340.

§§§§§§ Consider a hypothetical U.S. investor tracking energy stock prices in both U.S. and European markets at the same time. He would notice that coal stocks listed on European exchanges were becoming relatively cheaper than coal stocks listed on U.S. exchanges. This would make European coal assets appear to be underpriced relative to U.S. coal assets and encourage our investor to buy stock shares of European coal producers. This would raise the demand for European-listed coal stocks and increase their returns, thereby reducing the returns gap between coal stocks traded in the United States and Europe.

27

Affleck-Graves, J., Callahan, C., and Ramanan, R., “Detecting Abnormal Bid-Ask Spread: A

Comparison of Event Study Methods,” Review of Quantitative Finance and Accounting,

2004, Vol. 14(1), pp. 45-65.

Aktar, I. “A comparison of the effects of the Chernobyl and Three Mile Island nuclear accidents

on the U.S. electric utility industry,” Sosyoekonomi 2005(2): 11-24.

Binder, John J., “The Event Study Methodology Since 1969,” Review of Quantitative Finance

and Accounting, Vol. 11(1998), pp. 111-117.

Brechin, Steven R. and Bhandari, Medani, “Perceptions of Climate Change Worldwide,” WIREs

Clim Change, 2011, 2, 871-885. Doi: 10.1002/wcc.146.

Carlarne, Cinnamon, “Climate Change Policies an Ocean Apart: EU & US Climate Change

Policies Compared,” Penn State Environmental Law Review, 2006, 14(3), 435-482.

Chang, S. J. and Son-Nan Chen, “Stock-Price Adjustment to Earnings and Dividend Surprises,”

Quarterly Review of Economics and Business, Vol. 29(1), 1989, pp. 68-81.

Davis, L., “Prospects for U.S. Nuclear Power After Fukushima,” Journal of Economic

Perspectives, 26(1), pp. 49-66, 2012.

Duffy, R., “Déjà vu all over again: climate change and the prospects for a nuclear power

renaissance,” Environmental Politics, Vol. 20(5), September 2011, pp. 668-686.

Fama, Eugene F. "The Behavior of Stock Market Prices." Journal of Business, 38(1): 34–105,

January 1965.

Fama, E., Fisher, L., Jensen, M.C., Roll, R, “The adjustment of stock prices to new information,”

International Economic Review, 10(1):1 – 21, 1969.28

Ferstl, R., Uta, S., Wimmer, M., “The effect of the Japan 2011 disaster on nuclear and alternative

energy stocks worldwise: an event study,” SSRN Working Paper Series, May 2011.

Fields, M.A., Janjigian, V, “The effect of Chernobyl on electric-utility stock prices,” Journal of

Business Research, 18:81-87, 1989.

French, Kenneth and James Poterba, "Investor Diversification and International Equity Markets".

American Economic Review Vol. 81(2): 222–226, 1991.

Gilani, I.S., Shahid, R., “Japan Earthquake Jolts Global Views on Nuclear Energy,” Findings

from a Global Snap Poll in 47 countries by WIN-Gallup International, April 15, 2011.

Grinblatt, Mark and Matti Keloharju, “How Distance, Language, and Culture Influence

Stockholdings and Trades,” The Journal of Finance 56(3), pp. 1053-1073, June 2001.

Hillmer, S.C. and P.L. Yu “The Market Speed of Adjustment to New Information,” Journal of

Financial Economics, Vol. 7(4), pp. 321-345, 1979.

Joskow, P. and Parsons, J., “The Future of Nuclear Power after Fukushima,” Economics of

Energy and Environmental Policy 1(2):99-113, January 2012.

Kalra, R., Henderson, G. V. Jr., Raines, G.A., “Effects of the Chernobyl nuclear accident on

utility share prices,” Quarterly Journal of Business and Economics 32(2):52-77, 1993.

Kawashima, S. and Takeda, F., “The Effect of the Fukushima Nuclear Accident on Stock Prices

of Electric Power Utilities,” (January 11, 2012). Available at SSRN:

http://ssrn.com/abstract=1983138 or http://dx.doi.org/10.2139/ssrn.1983138

Kothari, S.P., and Warner, Jerold B., “Econometrics of Event Studies,” Handbook of Corporate

Finance: Empirical Corporate Finance, Volume 1, Chapter 1, pp. 3-36, B. Espen Eckbo

(ed.), Elsevier/North-Holland, Oxford, United Kingdom, 2007.29

Krivin, D., Patton, R., Rose, E., Tabak, D., “Determination of the Appropriate Event Window

Length in Individual Stock Event Studies,” November 4, 2003, NERA Economic

Consulting Working Paper Series.

Lachapelle, Erick, Christopher P. Borick and Barry Rabe, “Public Attitudes toward Climate

Science and Climate Policy in Federal Systems: Canada and the United States

Compared,” Review of Policy Research, 2012, 29 (3), 334-357.

Leiserowitz, Anthony, “Climate Change Risk Perception and Policy Preferences: the Role of

Affect, Imagery, and Values,” Climatic Change, 2006, 77, 45-72.

Lorenzoni, Irene, Nick F. Pidgeon, “Public Views on Climate Change: European and USA

Perspective,” Climatic Change, 2006, 77, 73-95.

McWilliams, Abagail and Donald Siegel, “Event Studies in Management Research: Theoretical

and Empirical Issues,” The Academy of Management Journal, Vol. 40(3), pp. 626-657,

1997.

Podsakoff, Philip M.; MacKenzie, Scott B.; Lee, Jeong-Yeon; Podsakoff, Nathan P., “Common

method biases in behavioral research: A critical review of the literature and

recommended remedies.” Journal of Applied Psychology 88(Oct 2003): 879-903.

Sandeep Nabar and K. K. Boonlert‐U‐Thai (2007) Earnings Management, Investor Protection,

and National Culture. Journal of International Accounting Research: Fall, Vol. 6, No. 2,

pp. 35-54.

Schmidt, J. and Haifly, A., “Delivering on renewable energy around the world: how do key

countries stack up?” Energy Facts, National Resources Defense Council, May 2012.

Available at http://www.nrdc.org/energy/files/delivering-renewable-energy.pdf.

30

Skjærseth, Jon B, Bang, Guri and Schreurs, Miranda A, “Explaining Growing Climate Policy

Differences Between the European Union and the United States,” Global Environmental

Politics, 2013, 13(4), 61-80.

Spence, Alexa, Wouter Poortinga, Nick Pidgeon and Irene Lorenzoni, “Public perceptions of

energy choices: The influence of beliefs about climate change and the environment”,

Energy and Environment, 2010, 21(5). pp. 385-407.

Swanson, Eric T., “Let’s Twist Again: A High-Frequency Event-Study Analysis of Operation

Twist and Its Implications for QE2,” Brookings Papers on Economic Activity, pp. 151-

188, Spring 2011.

U.S. Energy Information Administration, “U.S. Natural Gas Wellhead Price Data,” 2013,

http://www.eia.gov/dnav/ng/hist/n9190us3m.htm.

Wittneben, B., “The impact of the Fukushima nuclear accident on European energy policy,”

Environmental Science and Policy 15(2012), pp. 1-3.

31

Figure 1: Daily abnormal turnover of U.S.-listed securities, by portfolio

-10

0

10

20

30

40

50Coal

Trading day

Dai

ly a

bnor

mal

turn

over

, %

-10-505

10152025

Renewables

Trading day

Dai

ly a

bnor

mal

turn

over

, %

-4-202468

10Nuclear

Trading day

Dai

ly a

bnor

mal

turn

over

, %

Figure 2: Cumulative abnormal returns of coal firms

-0.125

-0.075

-0.025

0.025

0.075

0.125

a) Region-specific coal firms

EU CoalUS CoalTrading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.125

-0.075

-0.025

0.025

0.075

0.125

b) Region-specific coal firms with major operations outside listing region

EU CoalUS CoalTrading day

Cum

ulati

ve a

bnor

mal

retu

rn

Figure 3: Cumulative abnormal returns of renewable firms

-5 -3 -1 1 3 5 7 9 11 13

-0.1

-0.05

-2.77555756156289E-17

0.05

0.1

0.15

0.2a) All renewable firms

EU RenewablesUS Renewables

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.1

0

0.1

0.2b) Cross-listed renewable firms

EU RenewablesUS Renewables

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.1

0

0.1

0.2c) Region-specific renewable firms

EU RenewablesUS Renewables

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.1

-0.05

0

0.05

0.1

0.15

0.2

d) Region-specific renewable firms with main operatins outside listing

region

EU RenewablesUS Renewables

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

32

Table 1: Daily abnormal turnover of U.S.-listed securities, by portfolio

(1) (2) (3)Trading day Coal Renewables Nuclear firms

-5 11.602 1.232 -0.667(1.36) (0.218) -(0.372)

-4 15.002 * 6.919 0.419(1.758) (1.224) (0.234)

-3 1.501 4.353 -0.563(0.176) (0.77) -(0.314)

-2 9.656 7.174 0.444(1.132) (1.269) (0.247)

-1 19.737 ** 12.552 ** 1.673(2.313) (2.221) (0.932)

0 10.650 9.954 * 0.970(1.248) (1.761) (0.541)

1 17.699 ** 15.910 *** 8.080 ***(2.075) (2.815) (4.503)

2 14.404 * 21.980 *** 6.529 ***(1.688) (3.889) (3.638)

3 40.249 *** 14.623 *** 6.303 ***(4.718) (2.587) (3.513)

4 17.134 ** 5.763 3.699 **(2.008) (1.02) (2.062)

5 30.950 *** 5.883 2.149(3.628) (1.041) (1.197)

6 7.700 -2.749 -0.776(0.902) -(0.486) -(0.433)

7 18.294 ** -4.016 -0.162(2.144) -(0.71) -(0.09)

8 11.236 -2.398 -2.225(1.317) -(0.424) -(1.24)

9 15.397 * -0.030 -2.151(1.805) -(0.005) -(1.199)

10 2.555 -5.428 -2.359(0.3) -(0.96) -(1.315)

11 -3.989 1.891 -2.235-(0.468) (0.335) -(1.245)

12 -2.398 -0.489 -1.477-(0.281) -(0.087) -(0.823)

13 0.508 -1.918 -1.454(0.06) -(0.339) -(0.81)

14 0.056 -2.759 0.332(0.007) -(0.488) (0.185)

Note: z-statistics in parentheses. H0: the Fukushima crisis had no effect on daily stock turnover of publically traded energy companies. Significance levels: * 10%, ** 5%, *** 1%.

33

Table 2: Daily abnormal returns, COAL FIRMS

All firms, all of which are also region-specific firms

Region-specific firms with major operations outside of

the listing region(1) (2) (3) (4)

Trading day U.S. Europe U.S. Europe-5 0.015 0.019 0.010 0.006

(1.06) (0.985) (0.661) (0.188)-4 0.001 -0.006 0.001 -0.009

(0.057) (-0.307) (0.066) -(0.280)-3 -0.019 0.002 -0.022 -0.009

(-1.287) (0.111) -(1.380) -(0.282)-2 -0.030** -0.006 -0.031* -0.007

(-2.088) (-0.295) -(1.942) -(0.206)-1 -0.019 -0.010 -0.024 -0.012

(-1.326) (-0.530) -(1.512) -(0.370)0 0.006 -0.013 0.012 -0.008

(0.388) (-0.675) (0.788) -(0.240)1 0.038 *** -0.011 0.043 *** -0.025

(2.644) (-0.601) (2.749) -(0.775)2 0.018 -0.023 0.016 -0.027

(1.232) (-1.198) (1.019) -(0.859)3 0.048 *** 0.0001 0.066 *** 0.003

(3.293) (0.004) (4.187) (0.10)4 0.014 -0.007 0.014 -0.012

(0.964) (-0.380) (0.921) -(0.394)5 -0.010 0.015 -0.008 0.009

(-0.705) (0.781) -(0.485) (0.290)6 -0.020 -0.022 -0.018 -0.004

(-1.386) (-1.174) -(1.158) -(0.122)7 0.013 0.018 0.001 -0.001

(0.869) (0.958) (0.050) -(0.026)8 0.018 0.014 0.022 0.000

(1.254) (0.751) (1.370) (0.008)9 -0.012 -0.025 -0.017 -0.013

(-0.832) (-1.320) -(1.053) -(0.417)10 -0.004 -0.001 -0.004 -0.005

(-0.282) (-0.062) -(0.258) -(0.170)11 -0.011 -0.005 -0.017 -0.004

(-0.765) (-0.274) -(1.106) -(0.122)12 -0.009 -0.006 0.000 -0.007

(-0.621) (-0.311) (0.016) -(0.222)13 -0.008 -0.006 -0.006 0.006

(-0.521) (-0.321) -(0.392) (0.202)14 0.002 0.017 -0.003 0.033

(0.106) (0.899) -(0.216) (1.054)

34

Note: z-statistics in parentheses. H0: the Fukushima crisis had no effect on daily returns of publically traded energy companies. Significance levels: * 10%, ** 5%, *** 1%.

35

Table 3: Daily abnormal returns, RENEWABLE FIRMS

All Firms Cross-Listed firms Region-specific firms Region-specific firms with major operations outside of

the listing region(1) (2) (3) (4) (5) (6) (7) (8)

Trading day U.S. Europe U.S. Europe U.S. Europe U.S. Europe-5 0.002 0.010 0.006 0.004 0.001 0.016 0.001 0.013

(0.267) (0.819) (0.414) (0.222) (0.072) (1.003) (0.094) (0.533)-4 -0.016 -0.013 -0.012 -0.019 -0.017** -0.007 -0.024 *** 0.000

(-1.736) (-1.044) (-0.790) (-1.185) (-2.111) (-0.453) -(2.724) -(0.017)-3 -0.012 -0.002 -0.005 -0.0003 -0.015* -0.005 -0.018 ** 0.000

(-1.339) (-0.186) (-0.313) (-0.019) (-1.888) (-0.293) -(2.045) -(0.007)-2 -0.011 0.005 -0.001 -0.003 -0.016* 0.012 -0.017 * 0.003

(-1.260) (0.404) (-0.048) (-0.218) (-1.993) (0.8) -(1.948) (0.109)-1 0.001 -0.008 0.00001 -0.007 0.001 -0.009 0.000 -0.016

(0.091) (-0.640) (0.001) (-0.453) (0.145) (-0.576) -(0.053) -(0.694)0 -0.008 -0.004 -0.011 -0.0003 -0.007 -0.007 -0.003 -0.008

(-0.893) (-0.305) (-0.703) (-0.021) (-0.831) (-0.479) -(0.372) -(0.360)1 0.026 *** 0.050 *** 0.037 ** 0.032 ** 0.020 ** 0.067 *** 0.015 * 0.070 ***

(2.860) (4.147) (2.459) (2.019) (2.462) (4.311) (1.676) (2.974)2 0.034 *** 0.068 *** 0.067 *** 0.070 *** 0.018 ** 0.066 *** 0.019 ** 0.112 ***

(3.850) (5.621) (4.498) (4.379) (2.233) (4.220) (2.146) (4.750)3 -0.002 0.008 0.002 0.007 -0.004 0.008 -0.006 -0.001

(-0.198) (0.649) (0.155) (0.434) (-0.477) (0.509) -(0.678) -(0.024)4 -0.005 -0.006 -0.017 -0.011 0.0002 -0.001 0.003 0.001

(-0.614) (-0.493) (-1.106) (-0.706) (0.027) (-0.049) (0.374) (0.030)5 -0.011 -0.014 -0.021 -0.038** -0.006 0.009 -0.010 -0.002

(-1.280) (-1.139) (-1.436) (-2.376) (-0.778) (0.559) -(1.169) -(0.086)6 -0.003 -0.021 * -0.010 -0.005 0.001 -0.035** 0.003 -0.034

(-0.318) (-1.718) (-0.645) (-0.307) (0.066) (-2.263) (0.347) -(1.426)7 -0.003 -0.005 -0.002 -0.007 -0.004 -0.004 -0.004 -0.023

(-0.368) (-0.444) (-0.102) (-0.446) (-0.514) (-0.272) -(0.421) -(0.989)8 0.005 0.009 0.005 0.010 0.005 0.007 0.006 0.000

(0.541) (0.708) (0.302) (0.643) (0.606) (0.424) (0.642) -(0.006)9 0.003 -0.003 -0.003 -0.006 0.006 -0.001 0.006 -0.007

(0.350) (-0.272) (-0.207) (-0.383) (0.757) (-0.050) (0.680) -(0.305)10 -0.004 0.002 -0.003 -0.000003 -0.004 0.003 -0.004 -0.007

(-0.426) (0.145) (-0.199) (-0.0002) (-0.520) (0.189) -(0.441) -(0.293)11 0.013 0.031 ** 0.016 0.021 0.012 0.040 *** 0.014 0.038

(1.475) (2.572) (1.036) (1.307) (1.468) (2.588) (1.633) (1.613)12 0.003 -0.004 0.004 0.006 0.003 -0.014 0.005 -0.011

(0.367) (-0.347) (0.269) (0.392) (0.353) (-0.898) (0.576) -(0.447)13 -0.006 -0.004 -0.010 -0.015 -0.004 0.006 -0.003 0.003

(-0.685) (-0.357) (-0.685) (-0.963) (-0.503) (0.364) -(0.352) (0.143)14 0.005 0.014 0.018 0.015 -0.002 0.013 -0.001 0.003

(0.548) (1.170) (1.225) (0.918) (-0.215) (0.862) -(0.104) (0.124)

36

Note: z-statistics in parentheses. H0: the Fukushima crisis had no effect on daily returns of publically traded energy companies. Significance levels: * 10%, ** 5%, *** 1%

37

Table 4: Daily abnormal returns, NUCLEAR PORFTOLIOS

(1) (2)Nuclear FIRMS Nuclear INDICES

Trading day U.S. U.S.-5 0.0003 0.001

(0.051) (0.101)-4 0.011 ** -0.009

(1.966) (-1.210)-3 0.004 -0.009

(0.775) (-1.214)-2 0.014 ** -0.002

(2.430) (-0.203)-1 0.004 -0.011

(0.788) (-1.488)0 -0.001 -0.007

(-0.232) (-0.917)1 -0.010 * -0.126 ***

(-1.854) (-16.476)2 -0.014 ** -0.040 ***

(-2.478) (-5.265)3 -0.009 -0.008

(-1.623) (-1.009)4 -0.008 0.014 *

(-1.505) (1.855)5 -0.001 0.043 ***

(-0.126) (5.651)6 0.001 0.032 ***

(0.253) (4.176)7 0.007 -0.002

(1.195) (-0.202)8 -0.002 -0.007

(-0.377) (-0.896)9 -0.002 -0.008

(-0.287) (-1.048)10 -0.003 -0.006

(-0.584) (-0.849)11 -0.001 -0.025 ***

(-0.249) (-3.325)12 0.006 -0.017 **

(1.048) (-2.264)13 0.006 0.010

(1.153) (1.259)14 -0.004 -0.007

(-0.678) (-0.963)Note: z-statistics in parentheses. H0: the Fukushima crisis had no effect on daily returns of publically traded nuclear companies. Significance levels: * 10%, ** 5%, *** 1%.

38

Appendix A: List of firms included in the study

Table A1: Nuclear energy firms included in empirical analysis

U.S. – PORTFOLIO 1: NUCLEAR FIRMS U.S. – PORTFOLIO 2: NUCLEAR INDICESCONSTELLATION ENERGY GROUP INC GLOBAL X FUNDS

DOMINION RESOURCES INC ISHARES S&P GLOBAL NUCLEAR ENERGY INDEX

DUKE ENERGY CORP MARKET VECTORS URANIUIM+NUCLEAR ENERGY ETF

ENERGY SOLUTIONS INC POWERSHARES GLOBAL NUCLEAR ENERGY ETF

ENTERGY CORP

EXELON CORP

NEXTERA ENERGY INC

PROGRESS ENERGY INC

SOUTHERN CO

Table A2: Coal firms included in empirical analysis

U.S. EUROPEALLIANCE HOLDINGS ^ ATLANTIC COAL PLC

ALLIANCE RESOURCE PARTNERS ^ BISICHI MINING PLC

ALPHA NATURAL RESOURCES INC ^ GCM RESOURCES PLC

ARCH COAL INC ^ NCONDEZI COAL CO

CLOUD PEAK ENERGY INC NEW WORLD RESOURCES PLC

INTERNATIONAL COAL GROUP INC ^ STRATEGIC NATURAL RESOURCES

JAMES RIVER COAL CO UK COAL PLC

MASSEY ENERGY CO

OXFORD RESOURCE PARTNERS

PATRIOT COAL CORP

^ PEABODY ENERGY CORP

PENN VIRGINIA RESOURCE PARTNERS

RHINO RESOURCE PARTNERS

WESTMORELAND COAL CO

^ YANZHOU COAL MINING CO LTD

^ Indicates firms with major operations outside the listing region.

39

Table A3: Renewable energy firms included in empirical analysis

U.S. EUROPE^ A POWER ENERGY GENERATION SYS LT ^ HANWHA SOLARONE CO LTD # AEROVIRONMENT INC # MICREL INC

^ ADVANCED ENERGY INDUSTRIES INC ^ HEXCEL CORP ALEO SOLAR AG # POWER ONE INC

# AEROVIRONMENT INC HOKU CORP # AMTECH SYSTEMS INC POWERBAGS AG

# AMTECH SYSTEMS INC # IDACORP INC ^ ARISE TECHNOLOGIES CORP ^ POWERFILM INC

^ APPLIED MATERIALS INC ^ ITRON INC # BTU INTERNATIONAL INC PV CRYSTALOX SOLAR

# BTU INTERNATIONAL INC ^ J A SOLAR HOLDINGS CO LTD # CANADIAN SOLAR INC ^ RENEWABLE ENERGY CORP

BEACON POWER CORP # JINPAN INTERNATIONAL LTD # CAPSTONE TURBINE CORP ROTH & RAU AG

CVD EQUIPMENT CORP # LDK SOLAR CO LTD # CLEANTECH SOLUTIONS INTL ^ SOLAR ENERGY INITIATIVE

# CANADIAN SOLAR INC MEMC ELECTRONIC MATERIALS # CYPRESS SEMICONDUCTOR CORP ^ SOLAR ENERTECH CORP

# CAPSTONE TURBINE CORP # MAGNETEK INC DAY4 ENERGY INC ^ SOLAR POWER INC

^ CHINA MING YANG WIND POWER GRP # MAXWELL TECHNOLOGIES INC # DAYSTAR TECHNOLOGIES INC ^ SOLAR THIN FILMS INC

^ CHINA SUNERGY CO LTD # MICREL INC ^ ENTECH SOLAR INC SOLAR2 AG

^ CHINA WIND SYSTEMS INC ^ NATIONAL SEMICONDUCTOR CORP # EVERGREEN SOLAR INC # SPIRE CORP

# CLEANTECH SOLUTIONS INTL # POWER ONE INC # FIRST SOLAR INC # SPX CORP

# CYPRESS SEMICONDUCTOR CORP ^ RENESOLA LTD ^ GT ADVANCED TECHNOLOGIES INC # SUNPOWER CORP

# DAYSTAR TECHNOLOGIES INC # SPIRE CORP ^ HANSEN TRANSMISSIONS INT ^ TIMMINCO LTD

ENERGY CONVERSION DEVICES INC # SPX CORP # IDACORP INC ULTIMA NETWORKS PLC

^ ENTEGRIS INC # SUNPOWER CORP # JINPAN INTERNATIONAL LTD VESTAS WIND SYSTEMS

# EVERGREEN SOLAR INC ^ SUNTECH POWER HOLDINGS CO # LDK SOLAR CO LTD ^ XINJIANG GOLDWIND SCIENCE & TECHNOLOGY

# FIRST SOLAR INC ^ TRINA SOLAR LIMITED # MAGNETEK INC

GT SOLAR INTERNATIONAL ^ YINGLI GREEN ENERGY HLDG CO # MAXWELL TECHNOLOGIES INC

# Indicates cross-listed firms (i.e. those listed on both U.S. and European exchanges).^ Indicates firms with major operations outside the listing region.

40

Appendix B: Analysis of Natural Gas Firms

It is difficult to identify publically listed and globally operating firms that focus solely on

natural gas production due to the co-production relationship of petroleum and natural gas, which

leads to a difficulty in decomposing these companies’ stock outcomes into petroleum- and

natural gas-driven components. Moreover, the supply chain of natural gas is complex and there

are many other types of midstream companies present in gas markets that specialize in field

services and construction and operation of pipeline infrastructure. Here, we use Yahoo Finance

and Google Finance to supplement CRSP and Bloomberg data with daily information on prices

and volumes of 20 natural gas companies traded in the United States (five natural gas producers

and 15 midstream firms) and 37 natural gas companies traded on European exchanges (13

producers and 24 midstream companies). Additionally, we collect quarterly data on the number

of shares outstanding for all natural gas firms in the sample from YCharts’ historical database.

We exclude oil and gas majors from our analysis because of the above confounding effect.

Natural gas firms contained in our sample are listed in Table B2.

We replicate the analysis described in Section 2 of this manuscript for natural gas

portfolios and present results in Figure B1 and Table B1 below. From Figure B1, it appears that

U.S.-listed natural gas firms fared better than European-listed natural gas companies. However,

if we add up the daily return abnormalities that are statistically significant at the 5 percent level

or better during the 14 trading days following the earthquake, the combined magnitude of all

statistically significant abnormalities in U.S. and European markets is 1.6 and 0.1 percent,

respectively, which is rather negligible. In the United States, an abundance of economically

recoverable shale gas reserves has driven prices to record low levels, and it is surprising that we

do not observe U.S. investors rallying behind natural gas as the favored future source of energy.

One possible explanation is that U.S. natural gas prices began to decline in late 2008 (U.S. EIA),

and so U.S. financial markets likely absorbed the appeal of low-priced natural gas for electricity

generation well before the 2011 Fukushima incident. European markets, however, have long

remained dependent on pipeline supplies from Russia and the Middle East, as well as on more

expensive LNG deliveries from the Middle East and South East Asia. These sources are costly

in both political and economic terms, so investors in Europe might in general hold a less

optimistic view about natural gas-based electricity.

41

Figure B1: Cumulative abnormal returns of natural gas firms

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02a) All Natural Gas Firms

EU Natural GasUS Natural Gas

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02b) Cross-Listed Natural Gas Firms

EU Natural GasUS Natural Gas

Trading dayCu

mul

ative

abn

orm

al re

turn

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02c) Region-Specific Natural Gas Firms

EU Natural Gas

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

d) Region-specific natural gas firms with major operations outside listing region

EU Natural Gas

Trading day

Cum

ulati

ve a

bnor

mal

retu

rn

42

Table B1: Daily abnormal returns, NATURAL GAS FIRMS

All Firms Cross-Listed firms Region-specific firms Region-specific firms with major operations outside of the listing

region(1) (2) (3) (4) (5) (6) (7) (8)

Trading day U.S. Europe U.S. Europe U.S. Europe U.S. Europe-5 0.005 0.003 0.005 -0.005 - 0.003 - -0.015

(0.768) (0.264) (0.768) (-0.403) (0.294) -(0.934)-4 0.002 -0.002 0.002 -0.006 - -0.002 - -0.012

(0.405) (-0.245) (0.406) (-0.526) (-0.216) -(0.777)-3 -0.009 0.003 -0.009 -0.001 - 0.003 - -0.001

(-1.573) (0.286) (-1.572) (-0.119) (0.300) -(0.089)-2 -0.008 0.002 -0.008 -0.001 - 0.002 - 0.003

(-1.351) (0.198) (-1.350) (-0.118) (0.210) (0.164)-1 0.002 0.005 0.002 -0.005 - 0.006 - 0.006

(0.311) (0.509) (0.313) (-0.392) (0.540) (0.412)0 -0.001 -0.014 -0.001 -0.015 - -0.014 - -0.015

(-0.216) (-1.427) (-0.216) (-1.325) (-1.374) -(0.963)1 0.001 -0.013 0.001 -0.005 - -0.014 - 0.023

(0.128) (-1.329) (0.128) (-0.406) (-1.326) (1.438)2 -0.003 -0.0001 -0.003 -0.007 - 0.0002 - -0.022

(-0.506) (-0.013) (-0.505) (-0.580) (0.018) -(1.402)3 0.016 *** -0.004 0.016 *** -0.007 - -0.004 - -0.007

(2.647) (-0.421) (2.646) (-0.558) (-0.397) -(0.449)4 0.007 -0.009 0.007 -0.001 - -0.009 - 0.002

(1.215) (-0.874) (1.214) (-0.104) (-0.876) (0.102)5 -0.002 0.023 ** -0.002 0.001 - 0.024 ** - 0.004

(-0.281) (2.250) (-0.281) (0.059) (2.292) (0.249)6 0.003 -0.011 0.003 -0.002 - -0.011 - 0.005

(0.472) (-1.041) (0.470) (-0.167) (-1.043) (0.317)7 0.001 -0.004 0.001 0.010 - -0.004 - -0.002

(0.186) (-0.349) (0.186) (0.897) (-0.403) -(0.113)8 -0.006 -0.022** -0.006 0.0002 - -0.023 ** - -0.014

(-0.955) (-2.199) (-0.954) (0.021) (-2.235) -(0.883)9 -0.002 0.007 -0.002 -0.0002 - 0.007 - -0.005

(-0.260) (0.664) (-0.261) (-0.015) (0.683) -(0.293)10 0.003 -0.002 0.003 0.003 - -0.002 - 0.004

(0.571) (-0.214) (0.571) (0.234) (-0.230) (0.231)11 -0.006 0.003 -0.006 0.015 - 0.003 - -0.041 ***

(-1.012) (0.326) (-1.011) (1.283) (0.259) -(2.628)12 -0.009 -0.002 -0.009 -0.016 - -0.001 - 0.013

(-1.522) (-0.204) (-1.521) (-1.388) (-0.124) (0.841)13 0.003 0.017 * 0.003 0.002 - 0.018 * - 0.015

(0.581) (1.648) (0.580) (0.141) (1.674) (0.984)14 0.003 -0.015 0.003 0.001 - -0.015 - 0.005

(0.569) (-1.449) (0.568) (0.101) (-1.477) (0.343)43

Note: z-statistics in parentheses. H0: the Fukushima crisis had no effect on daily returns of publically traded energy companies. Significance levels: * 10%, ** 5%, *** 1%.

Table B2: Natural gas firms included in empirical analysis

U.S. EUROPE# ATLAS PIPELINE PARTNERS L P ^* AJ LUCAS GROUP LTD

# BOARDWALK PIPELINE PARTNERS LP # ATLAS PIPIELINE PARTNERS LP

# CENTERPOINT ENERGY INC # BOARDWALK PIPELINE PARTNERS LP

# CHENIERE ENERGY PARTNERS L P ^ CANADINA UTILITIES LTD

# CHESAPEAKE MIDSTREAM PARTNERS LP # CENTERNPOINT EENERGY INC

# COPANO ENERGY L L C * CENTRICA PLC

# CRESTWOOD MIDSTREAM PARTNERS L P # CHENIERE ENERGY PARTNERS LP

# DCP MIDSTREAM PARTNERS L P # CHESAPEAKE MIDSTREAM PARTNERS LP

#* EL PASO CORP # COPANO ENERGY LLC

# ENERGY TRANSFER PARTNERS L P # CRESTWOOD MIDSTREAM PARTNERS LP

#* ONEOK INC # DCP MIDSTREAM PARTNERS LP

# ONEOK PARTNERS L P #* EL PASO CORP

#* QUESTAR CORP ENAGAS SA

# REGENCY ENERGY PARTNERS LP ^ ENERGY TRANSFER EQUITY LP

# SOUTHERN UNION CO # ENERGY TRANSFER PARTNERS LP

# SPECTRA ENERGY CORP FLUXYS

# TARGA RESOURCES PARTNERS LP ^ GAIL INDIA LTD

#* TRANSPORTADORA DE GAS DEL SUR SA GAS NATURAL SDG

# WESTERN GAS PARTNERS LP * GAS PLUS

#* WILLIAMS COS * GASOL PLC

# WILLIAMS PARTNERS L P ^ GREKA DRILLING LTD

LATVIJAS GAZE

* LIQUEFIED NATURAL GAS

#* ONEOK INC

# ONEOK PARTNERS LP

#* QUESTAR CORP

^# REGENCY ENERGY PARTNERS LP

# SOUTHERN UNION CO

# SPECTRA ENERGY CORP

# TARGA RESOURCES PARTNERS LP

#* TRANSPORTADORA DE GAS DEL SUR SA

^# WESTERN GAS PARTNERS LP

#* WILLIAMS COMPANIES

# WILLIAMS PARTNERS LP

* Indicates natural gas producers; non-starred entries denote midstream service firms in natural gas sector.# Indicates cross-listed firms (i.e. those listed on both U.S. and European exchanges).^ Indicates firms with major operations outside the listing region.

44