8
The investigation
The health of marine ecosystems is determined by various factors in a complex and interlinked
system (UN-DESA 2008). More specifically, in order to operationalize this concept, we use a well-
established indicator, the Marine Trophic Index (MTI). This measurement captures to what extent
countries “fish down the food chain” within their exclusive economic zones. Pressure on fisheries
from harvesting tends to affect fish at the top of the food chain as humans often target larger pred-
atory fishes (Pauly 2005; Pauly and Watson, 2005; Pauly and Palomares, 2005). The MTI is calculat-
ed by assigning a number to each species according to its location in the food chain, where carni-
vores receive higher and herbivores receive lower numbers. The measure averages the trophic levels
from the overall catch, based on a dataset of commercial fish landing compiled by the Food and
Agricultural Organization of the United Nations (FAO). Lower values of the index mean that
catches consist of smaller fish. A negative trend in this measurement is thus a proxy measure for
overfishing and that “fisheries are not being sustainably managed” (Sea Around Us, 2011). Over-
fishing affects the marine ecosystem health as overexploited fish stocks lead to the loss of biodiver-
sity and ecosystems stability. The index has been criticized for not adequately reflecting the true
situation in marine ecosystems as it is built on the catches of commercial species, excluding the
impact of unregistered fishing (Branch et al., 2010; Caddy et al., 1998). However, there exist few
alternative measures of overfishing. The MTI is widely used by researchers and remains the most
well-established measure for marine trophic stability across countries and time (Clausen and York,
2008a; Emerson et al., 2010; Pauly and Watson; 2005). The MTI is also considered to be “a meas-
ure for overall ecosystem health and stability” and was included as such in the 2010 Environmental
Performance Index (Emerson et al., 2010).
In order to measure the main independent variable of the study, i.e., the degree of democ-
racy in a country at a given point in time, we use one of the most established regime type indicators
– the Freedom House/Polity index. This index reflects two important composites of regimes –
political rights and civil liberties (Freedom House, 2010). Political rights measure whether elections
in the country are free and fair, whether political rights are equal to all members of the society and
the competitiveness of political participation. The civil liberties value includes an assessment of
freedom of the press, of academic freedom, of freedom of public and private discussions, of free-
dom for NGOs’ operations, of rule of law, of an independent judiciary and other relevant aspects
(Lonardo, 2011). The average value of political rights and civil liberties in turn serves as an approx-
9
imation of the level of democracy in a country. In the present study we will use an imputed version
of this index, designed especially for time-series analysis, covering a broader sample using imputed
values for the cases where data was initially missing. The imputed version of the index is available
for the period 1972-2009 and varies from 0 to 10, where 10 corresponds to the most democratic
regimes (Teorell et al., 2011).
Following the reasoning of, for example, Li and Reuveny (2006), we include a measure of a
country’s openness to and engagement in world trade as a control variable. A country’s openness to
world trade is held to relate to environmental outcomes in several ways. For example, it has been
argued that trade and globalization encourages establishment of higher environmental standards
according to the demands from markets and also promotes technologies and innovations of a high-
er standard (Esty and Gentry, 1997; Vogel, 1995; Porter and Linde, 1995; Braithwaite and Drahos,
2000). However, others have argued in line with the hypotheses of the “race to the bottom,” hold-
ing that countries fearing to lose competitiveness will dismantle environmental standards (Sheldon,
2006). In addition, Daly (1993) and Meadows et al. (1972) conclude that trade has negative effects
on the environment, since it raises production levels and GDP, which in turn negatively affects the
environment. Indeed, empirical investigations show both positive (Frankel and Rose, 2005; Antwei-
ler et al., 2001) and negative (e.g., Managi, 2004) correlations between openness to trade and envi-
ronmental quality and they also find different effects of trade openness on different pollutants be-
tween country groups (Managi et al., 2008). The indicator of openness to trade is taken from Penn
World Trade (Heston, Summers and Aten, 2009), and measures total trade as a percentage of GDP
in constant 2005 prices. The data covers the years 1950-2007. The variable required log-
transformation to correct for its skewed distribution.
In addition, following Delgado et al. (2003), who discuss the impact from growing human
populations on the pressure put on fisheries, we include a control variable for the size of a coun-
try’s population. The data on population is taken from the World Bank database for the years 1971-
2010, and is measured in numbers of inhabitants. The variable is logarithmically transformed due to
its skewed distribution.
Of all the gears used in harvesting marine fish resources, bottom trawls and dredges are
recognized as considered to be the most destructive ones (Watson et al., 2004, 2006). They cause
chronic disturbances in coastal waters and lead to changes in trophic structures (Jennings et al.,
2001). We therefore include a control for trawling intensity in our analysis. We use the Coastal Shelf
Fishing Pressure Index, developed by the Environment Performance Index (2012). The index