in the spread of democracy ( 31 , 32 ) and have
provided strong evidence that trends in demo-
cratic and autocratic transitions follow major
changes in the international environment
( 2 , 20 , 33 ). Gleditsch and Ward, for example,
have shown that the probability that an autoc-
racy in a predominantly autocratic region
will transition to democracy in a given year is
just 0.015, but when the proportion of demo-
cratic neighbors exceeds two-thirds and when
at least one neighbor transitioned recently to
democracy, the probability that a given autoc-
racy will become democratic increases by a
factor of 10 ( 31 ). Research on electoral revolu-
tions and prodemocracy protest movements
suggests that democracy activists learn trans-
nationally and adapt strategies from each other,
sometimes through internationally supported
horizontal networks of activists ( 28 , 34 ), and that
mass movements aimed at bringing down dicta-
tors are often regionally contagious ( 28 , 34 – 36 ).
Figure 2 illustrates regional clustering in the dif-
fusion of formal democratic institutions ( 37 , 38 , 39 ).
Research on the international diffusion of
democracy has highlighted coercion, compe-
tition, emulation, and learning as the central
mechanisms by which democracy spreads ( 32 ).
Recent studies have linked such processes to
great power transitions at the system level,
arguing that whether the most powerful country
in the world is democratic or authoritarian
causes global waves of regime transitions (al-
though regimes that result from these transitions
are not necessarily durable) ( 2 , 40 ). Additionally,
other studies have shown that countries were more
likely to democratize when they had more eco-
nomic and political linkages with Western coun-
tries ( 41 ), were members of intergovernmental
organizations with more democratic members
( 1 , 7 ), or were geographically proximate to other
democracies ( 31 ). Relatedly, democracy promo-
tionandforeignaidconditionedondemocratic
reforms have been shown to have meaningful,
positive effects on democratization ( 2 , 6 , 42 ).
The prodemocracy international environment
supported the global diffusion of democratic
institutions and ideas, even if it did not always
bring about genuine, durable democracy. Even
partial political liberalization gave citizens more
exposure to democratic practices, freedom of
expression, individual rights, and experience
with less overt repression than would have been
thecaseabsentaprodemocracyinternational
environment—a counterfactualperspective
that is in contrast to much of this literature’s
characterization of hybrid regimes as failed
democratic transitions [e.g., ( 41 )]. Democracy
promotion undoubtedly contributed to many
cases of successful democratic transitions, but
it also likely led to the diffusion of hybrid regime
types, which combine aspects of democracy
and authoritarianism. Many of these hybrid
regimes could be characterized as pseudo-
democracies, in which leaders were willing to
adopt democratic institutions—like multiparty
elections, a pluralistic media environment, and
universal suffrage—but had no intention of
givinguppoweriftheyweretoloseanelection
( 3 , 43 ), a form of exposure that pseudodemo-
cratic leaders worked to avoid. The fact that
some leaders have displayed more skill than
others at maintaining plausible deniability that
they were moving toward democracy has stood
as a fundamental roadblock for researchers in
measuring the extent of pseudodemocracy and
its consequences ( 12 ).Emergent threats to democracy from above
and below
Current threats to democracy around the
world come from both international and do-
mestic sources, and even internal threats to
democracy, like antidemocratic political parties,
often have global ties. Although the Trump
administration has accelerated the decline of
U.S. support for democracy abroad, including
its“admiration”of dictators and“disdain for
traditional democratic allies”( 15 ), the most
recent trends undermining democracy promotion
have been ongoing since 2001, amid subsequent
U.S.-led anti-terrorism foreign policy efforts. This
includes declining enthusiasm for democracy
promotion by a variety of Western actors ( 15 ),
such as the European Union ( 44 ). These changes
“have turned a world that was once relatively
favorable to the spread of democratic norms into
one where authoritarians can push back—and
havelearnedtodosoininnovativeways”( 18 ).
Recent research documents at least three ways
in which the international environment has
shifted away from democracy that have implica-
tions for understanding democratic backsliding.
First, although many forms of democracy
assistance continue, the decline in more overt
forms of democracy promotion and the rhetor-ical embrace of authoritarian leaders are likely
to ease pressure on elites to appear plausibly
democratic and simultaneously decrease sup-
port for mass movements that favor democracy
( 16 , 18 ). Prodemocracy citizen movements are
nowmorelikelytofaceleaderswhoarewilling
to engage in blatant targeting of political op-
position and willing to commit violence against
citizen demonstrations, and these movements
are receiving less support from foreign allies. In
Cambodia for example, members of the opposi-
tion Cambodian National Rescue Party (CNRP)
were arrested on false charges, harassed, and,
in some cases, forced into exile ( 45 ). By 2017,
Hun Sen’s government abandoned any pre-
tense of democracy and banned the CNRP. As
one prominent human rights commentator
stated,“Prime Minister Hun Sen’srepression
of the opposition, media, and activist groups
has effectively turned the country’sdemocracy
into a one-party state”( 45 ). Up until at least
2013, the Cambodian government was some-
what responsive to Western pressure, allowing
opposition parties to organize and permitting
foreign democracy promotion programs to oper-
ate. However, the increasing ability of Hun Sen’s
government to rely on Chinese support instead of
Western aid, combined with more-tepid Western
democracy promotion efforts, have substantially
diminished the Cambodian government’s will-
ingness to allow any political opposition ( 45 ).
Second, mirroring patterns in the diffusion of
democracy, authoritarian practices and strat-
egies to undermine democratic institutions are
more freely diffusing across borders by means
of emulation and learning ( 46 , 47 ). There is not
yet much evidence that backsliding has been
caused by overt coercive autocracy promotion
by either China or Russia or by explicit rewards
for autocratic partners ( 48 ). However, emulation
of strategies of repression and the surveillanceHyde,Science 369 , 1192–1196 (2020) 4 September 2020 2of5
0102030405060708090100Rate of all sovereign states1950 1960 1970 1980 1990 2000 2010 2020Recent national elections (5-year moving window)
Recent potentially competitive elections (5-year moving window)Fig. 1. The global diffusion of national elections and formal political competition, 1950 to 2018.As democracy
grew in global prominence over time, nearly all sovereign states in the world introduced direct national elections and
allowed multiple political parties to compete. At the peak of this trend, all but seven countries held national elections,
95% of which allowed for the possibility of multiparty competition. Data are from the National Elections Across
Democracy and Autocracy (NELDA) project ( 9 ). Recent national elections are considered to be those from the current
year or any of the 4 prior years. The potential for formal political competition is defined as in ( 9 ) and includes elections
in which electoral opposition was allowed [the variable referred to as NELDA3 in the NELDA dataset ( 9 )], a choice
of candidates appeared on the ballot (NELDA4), and at least one opposition party was legal (NELDA5). The full data and
codebook are available at https://nelda.co/. Replication data and code are available from Harvard Dataverse ( 62 ).