Temperatures, food riots, and adaptation: A long-term historical analysis of England
| Published date | 01 March 2020 |
| Author | Tim Wegenast,Alexander De Juan |
| DOI | 10.1177/0022343319863474 |
| Date | 01 March 2020 |
| Subject Matter | Research Articles |
Temperatures, food riots, and adaptation:
A long-term historical analysis of England
Alexander De Juan
School of Cultural Studies and Social Sciences, University of Osnabru
¨ck
Tim Wegenast
Department of Politics and Public Administration, University of Konstanz
Abstract
A large body of research indicates that environmental conditions can influence the risk of social unrest. However, we
know little about how these effects may change in the long run. Are they likely to remain constant or do they change
over time – for example as a consequence of human adaptation? To investigate this question, we rely on a
disaggregated analysis of England over a period of more than 300 years. Combining data on geo-referenced food
riots with reconstructed climate data, we first assess the impact of annual temperatures on social unrest over the
period 1500–1817. We then use our long-term time-series dataset to assess the temporal heterogeneity of year-to-
year associations between temperatures and social conflict. Our models show a substantive negative correlation
between temperatures and food riots in the aggregate. This association, however, seems to be highly inconsistent over
time and largely confined to the 18th century. In addition, we find evidence of decadal processes of adaptation: past
exposure to adverse weather conditions dampens the effect of current exposure. Taken together, these findings
underline the importance of considering temporal heterogeneities when assessing the climate–conflict nexus and
caution against any simple extrapolations of observable present-day effects of environmental conditions into the
future.
Keywords
adaptation, climate change, England, food riots, wheat prices
Introduction
Adverse environmental conditions can negatively impact
agricultural production, increase food prices, threaten
people’s livelihoods, and trigger social unrest. In 2016,
for example, riots erupted in the Indian state of Karna-
taka. The region was suffering a severe drought that had
reduced drinking water supplies and ignited disputes
over the distribution of water resources (Lodaya &
Mukherjee, 2016). In 2017, northeastern Nigeria saw
deadly clashes between farmers and herders as a conse-
quence of a drought that hit Lake Chad basin and esca-
lated distributional conflicts over fertile land (Akinwotu,
2017). In 2016, Bolivia experienced the worst drought
in 25 years, which triggered violent protests in several
urban centers (Reuters, 2016). These and other examples
not only indicate that environmental conditions can in
fact be conducive to social unrest but also raise the
important question of whether we can expect to see an
increase in similar instances of violence as the frequency
and/or intensity of adverse weather events increases in
the long run.
There are two possible responses to this question,
which differ in terms of how they expect the marginal
effects of weather conditions to change over time: the
first, rather grim answer is based on the assumption that
these effects are likely to remain constant or increase.
Consequently, we could simply interpolate the results
Corresponding author:
alexander.dejuan@uni-osnabrueck.de
Journal of Peace Research
2020, Vol. 57(2) 265–280
ªThe Author(s) 2019
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/0022343319863474
journals.sagepub.com/home/jpr
of previous studies stressing positive associations between
weather conditions and violent unrest. Given that the
frequency and/or intensity of extreme weather events is
likely to accelerate in the future, we would expect an
increase in environmentally induced violence.
But what if the effect of each individual extreme
weather period on social unrest changes over time? For
example, effects may decrease due to adaptation pro-
cesses. As a reaction to past exposure to harsh weather
conditions, people may adapt their behavior in order to
reduce their vulnerability to similar conditions in the
future. Consequently, even if we were to see more
extreme weather events as a result of global warming,
we would not necessarily observe any substantial increase
in environmentally induced violence.
This article aims to assess the plausibility of these two
competing scenarios by his torically investigating their
basic underlying assumptions: have associations between
environmentalconditions and social unrestchanged in the
long run? Do past weather conditions influence the sub-
sequent effects of weather conditions on social unrest?
Numerous studies have analyzed the effects of year-to-
year weather variation on large-scale violence using rela-
tively short time periods spanning ‘only’ a few decades.
While this research has contributed greatly to our under-
standing of the potential associations between weather
conditions and violence in the short run, the underlying
approach is limited in terms of contributing to the main
research interest of this article. In particular, the time
frames investigated are simply too short to effectiv ely
assess whether weather effects remain constant or change
over longer time periods.
Another research strand has focused on analyzing
whether large-scale instances of violence such as inter-
state or civil war have historically been more frequent in
particularly cold or warm decades or centuries. This
research strand has added important insights on truly
long-term associations between environmental condi-
tions and violent conflict. However, it has relied on units
of analysis that are temporally and geographically highly
aggregated, thereby hampering the identification of het-
erogeneities on smaller scales.
We combine elements of both research strands to assess
the extent to which associations between environmental
conditions and social unrest may change over time. We
focus on yearly weather conditions at a relatively high
spatial resolution (*50 x 50 km grid cells) and for an
extended periodof more than 300 years. Englandserves as
our empirical case. We combine gridded information on
yearly temperatures with geo-referenced historical food
riots between 1500 and 1817. In line with previous
research, we assume that (H1) low yearly temperatures
increase the risk of food riot occurrence. Most impor-
tantly, (H2) we expect substantial variation of effects over
time. Specifically, (H3) weargue that previous exposureto
particularly cold years dampens the effects of subsequent
low temperatures on the risk of social unrest.
Our analysis proceeds through three main steps. We
start by estimating the effects of temperatures on the
probability of food riot occurrence using grid-years as
our unit of analysis. Our models show that there was
more unrest in colder years than in years with more
favorable weather conditions. Next, we assess heteroge-
neities of effects over time. We find associations between
weather conditions and unrest to be highly inconsistent
over time and driven primarily by individual decades of
the 18th century. Third, we use our long-term time-
series dataset to assess the mediating effects of past
weather conditions. We find that the number of very
cold years in the past 30 to 40 years substantially dam-
pens the effects of current exposure to low temperatures.
These results contribute to the literature in two main
ways. First, they lend support to previous research in that
they show that associations between weather conditions
and unrest can also be observed in other historical time
periods (i.e. the 18th century) and are not confined to
recent and rather short time periods that most research
has focused on (i.e. the 1990s and early 2000s). How-
ever, we also find substantial variation in the strength
and direction of these correlations – whether we find a
short-term association between weather conditions and
unrest seems to be highly dependent on the specific time
periods we investigate. Second, we provide tentative evi-
dence of adaptation processes over the long term. As
people are exposed to repeated instances of adverse
weather, they appear to adapt in a way that reduces their
vulnerability to any subsequent unfavorable climatic
conditions. Taken together, both findings underscore
the temporal heterogeneity of effects and caution against
any simple extrapolation of current effects of environ-
mental conditions into the future.
Associations between weather conditions and
conflict in the long run
Vulnerability to weather conditions is highly context-
dependent. It is characterized by a complex interplay
between particular climatic factors, crop types, socio-
economic capabilities, farm-based innovation processes,
and institutional policies. Most research on the climate–
conflict nexus acknowledges that this heterogeneity of
weather effects prevents any direct extrapolation of
266 journal of PEACE RESEARCH 57(2)
To continue reading
Request your trial