REPORT Global Satisfaction with Democracy · Global Satisfaction with Democracy Report 2020 4. The Dataset Figure 1: Countries for which data on satisfaction with democracy exists

REPORT

Global Satisfaction

with Democracy

2020

This report was prepared at the Bennett Institute for Public Policy at the

University of Cambridge and forms part of the work of the new Centre for

the Future of Democracy.

Suggested citation:

Foa, R.S., Klassen, A., Slade, M., Rand, A. and R. Williams. 2020. “The Global Satisfaction with

Democracy Report 2020.” Cambridge, United Kingdom: Centre for the Future of Democracy.

Report published January 2020

The Bennett Institute for Public Policy at the University of Cambridge aims to become a

world leader in achieving successful and sustainable solutions to some of the most pressing

problems of our time.

Our goal is to rethink public policy in an era of turbulence and growing inequality. Our research

connects the world-leading work in technology and science at Cambridge with the economic and

political dimensions of policy making. We are committed to outstanding teaching, policy

engagement, and to devising sustainable and long lasting solutions.

www.bennettinstitute.cam.ac.uk

The Centre for the Future of Democracy was launched in January 2020 to explore the

challenges and opportunities faced by democratic politics over the coming century.

Based at the Bennett Institute for Public Policy, the goal of the Centre is to understand the

prospects for democracy in broad historical and international perspective, getting beyond the

immediate crisis to identify different possible trajectories for democracy around the world. This

means distinguishing what is essential to democracy, what is contingent and what can be

changed. That requires taking the long view, drawing on the big picture and expanding our

imaginative horizons. This is what the Centre hopes to achieve, and in doing so it will connect

with work being done across Cambridge in a wide variety of fields, from computer science and

environmental science to history and philosophy.

The Centre’s aim is to move away from a fixation on the here and now, and beyond the who and

what of democratic politics – who is going to get elected, what are they going to do? – to look at

the how. How do democratic decisions get made and how can they be made differently? How can

the consent of losers and outsiders be achieved? How can new social divisions be bridged? How

can the use of technology be brought under democratic control? And if we can’t do these things, how will democracy not merely survive but flourish in the future?

Contents

1. Executive Summary 1

2. Key Findings 2

3. Introduction – The Democratic Malaise 3

4. The Dataset 4

5. The Global Picture 9

6. Regional Trends and Comparisons 14

The Anglo-Saxon Democracies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Latin America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Sub-Saharan Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Northeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37The Middle East . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

7. Conclusion: Why Are Citizens Unhappy with Democracy? 42

xx.Methodology I: Data Sources 44

xx.Methodology II: Aggregation Methodology 46

xx.Methodology III: Testing Semantic Equivalence 47

xx.Methodology IV: Sensitivity Analysis 54

xx.Methodology V: Detrending the Economic Cycle 56

Global Satisfaction with Democracy Report 2020

1. Executive Summary

• We use a new dataset combining more than 25 data sources, 3,500 country surveys, and4 million respondents between 1973 and 2020 asking citizens whether they are satisfiedor dissatisfied with democracy in their countries.

• Using this combined, pooled dataset, we are able to present a time-series for almost 50years in Western Europe, and 25 years for the rest of the world.

• We find that dissatisfaction with democracy has risen over time, and is reaching anall-time global high, in particular in developed democracies.

Page 1


2. Key Findings

Across the globe, democracy is in a state of malaise. In the mid-1990s, a majority ofcitizens in countries for which we have time-series data – in North America, Latin America,Europe, Africa, theMiddle East, Asia and Australasia – were satisfied with the performance oftheir democracies. Since then, the share of individuals who are “dissatisfied” with democracyhas risen by around +10% points, from 47.9 to 57.5%.

This is the highest level of global dissatisfaction since the start of the series in 1995.After a large increase in civic dissatisfaction in the prior decade, 2019 represents the highestlevel of democratic discontent on record.

The rise in democratic dissatisfaction has been especially sharp since 2005. The yearthat marks the beginning of the so-called “global democratic recession” is also the highpoint for global satisfaction with democracy, with just 38.7% of citizens dissatisfied in thatyear. Since then, the proportion of “dissatisfied” citizens has risen by almost one-fifth of thepopulation (+18.8%).

Many of the world’s most populous democracies – including the United States, Brazil,Nigeria, and Mexico – have led the downward trend. In the United States, levels of dis-satisfaction with democracy have risen by over a third of the population in one generation.

As a result, many large democracies are at their highest-ever recorded level for demo-cratic dissatisfaction. These include the United States, Brazil, Mexico, the United Kingdom,South Africa, Colombia, and Australia. Other countries that remain close to their all-timehighs include Japan, Spain, and Greece.

Citizens of developed democracies have also experienced a large increase in demo-cratic dissatisfaction. While in the 1990s, around two-thirds of the citizens of Europe,North America, Northeast Asia and Australasia felt satisfied with democracy in their coun-tries, today a majority feel dissatisfied.

While it goes beyond the scope of this report to explain the cause of this shift, we observethat citizens’ levels of dissatisfaction with democracy are largely responsive to objec-tive circumstances and events – economic shocks, corruption scandals, and policy crises.These have an immediately observable effect upon average levels of civic dissatisfaction.

However, the picture is not entirely negative. Many small, high-income democra-cies have moved in the direction of greater civic confidence in their institutions. InSwitzerland, Denmark, Norway, the Netherlands and Luxembourg, for example, democraticsatisfaction is reaching all-time highs. These countries form part of the “island of content-ment” – a select group of nations, containing just 2% of the world’s democratic citizenry, inwhich less than a quarter of the public express discontent with their political system.

Comparison by region shows a number of other bright spots, above all in Asia. Indemocracies in South Asia, Northeast Asia, and above all in Southeast Asia, levels of civiccontentment are significantly higher than in other regions. For now, much of Asia hasavoided the crisis of democratic faith affecting other parts of the world.

Page 2


3. Introduction – The Democratic Malaise

Across the globe, democracy is in a state ofdeep malaise. In the West, growing politi-cal polarisation, economic frustration, andthe rise of populist parties, have eroded thepromise of democratic institutions to offergovernance that is not only popularly sup-ported, but also stable and effective. Mean-while, in developing democracies the eupho-ria of the transition years has faded, leavingendemic challenges of corruption, intergroupconflict, and urban violence that underminedemocracy’s appeal.Yet how does our current predicament

compare with earlier periods of democraticdissatisfaction – such as the “governabil-ity crises” of 1970s’ Western Europe, or theemerging market financial crises of the late1990s? Does the current anxiety represent apunctuated equilibrium – or is it part of a se-ries of cyclical troughs, from which eventualrecovery is likely?This report sheds empirical light on these

questions. It does so using a new dataset thatcomprises data from over four million surveyrespondents collected during half a centuryof social science research. For the first time,we are able to provide a truly global answerto the question of democracy’s “performance”legitimacy – using data from democracies inall regions of the world.

The Research Background

Much of the existing academic research upontrends in satisfaction with democracy hasderived from two sources. The first are sur-veys from the United States, such as Gallupor the National Election Study, and showAmerican satisfaction and trust in govern-ment in long-term decline.1 The second aresurveys from Western Europe, notably theEurobarometer, which began from a signif-icantly lower level but show only unclearfluctuation since.2 These two data sources

have, in turn, supported two correspondingliteratures, the former centred on America’s“crisis of trust”, and the latter on Europe’schronic – though, not necessarily worsening– democratic deficit.3

While a number of recent studies have be-gun to take advantage of new data from LatinAmerica, Sub-Saharan Africa, and Asia,4 un-til now, a genuinely global perspective onlong-term changes in citizen evaluation ofdemocratic performance has been lacking.The fragmentation of relevant data acrossdisparate surveys and sources has made it dif-ficult to draw generalisable inferences fromany one single dataset.

Our Approach

This report is built upon a simple method-ological premise: to combine questions onsatisfaction with democracy from the widestpossible range of available sources, in orderto generate a global “mega-dataset” – con-sisting of more than 3,500 unique countrysurveys – fromwhich to analyse global trendsover time.The results suggest cause for deep con-

cern. Since the mid-1990s, the proportion ofcitizens who are “dissatisfied” with the per-formance of democracy in their countries hasrisen by almost 10 percentage points globally.The deterioration has been especially deepin high-income, “consolidated” democracies,where the proportion has risen from a thirdto half of all citizens. Yet also among manyemerging democracies – in Latin America,Sub-Saharan Africa, and the Middle East –confidence in the capacity of democratic in-stitutions has eroded. Many countries inthese regions are at or near an all-time low,including systemically important democra-cies such as Brazil, Nigeria, or Mexico.

1 Marc Hetherington (2005)Why Trust Matters: Declining Political Trust and the Demise of American Liberalism.2 Carolien van Ham, Jacques Thomassen, Kees Aarts and Rudy Andeweg (Eds) (2017)Myth and Reality of the LegitimacyCrisis: Explaining Trends and Cross-National Differences in Established Democracies.

3 For a good overview, see TomW.G. van der Meer (2017) “Political Trust and the ‘Crisis of Democracy”’, in the OxfordResearch Encyclopedia of Politics.

4 E.g. Marc F. Plattner and Larry Diamond (Eds.) (2008) How People View Democracy.

Page 3


4. The Dataset

Figure 1: Countries for which data on satisfaction with democracy exists in the HUMAN Surveys dataset, by numberof years of available data. Both democracies and non-democracies are shown in the above graphic; forthis report, only data from electoral democracies are included in the analysis.

In recent years, discussions of public faith indemocracy frequently have generated moreheat than light. In part, this is due to dif-ferences between scholars in their choice ofsurvey sources, country coverage, and pe-riods of observation. The wide variety ofsurvey sources now available for compara-tive analysis – with over 25 different sourcesidentified by this study alone – can gener-ate widely varying impressions of individualcountry and regional trends.This report aims to provide a comprehen-

sive answer to questions regarding one mea-sure of democratic legitimacy – satisfactionwith democracy – by combining data fromalmost all available survey sources, usingconsistent, constant country coverage that isregionally and globally representative, fromthe earliest possible period to surveys thatwere fielded in recent months.The data used in this report represents

the views of almost all individuals livingin a system of electoral democracy. Thesurveys have been gathered and standard-ised by the Human Understanding MeasuredAcross National (HUMAN) Surveys project(www.humansurveys.org), with additionaldata for 2017–2019 added from supplemen-

tary survey sources, including prerelease datafrom the seventh round of the World ValuesSurvey, and individually commissioned sur-veys for October to December 2019 providedby the YouGov-Cambridge Centre.

Satisfaction with Democracy

This report examines one indicator of demo-cratic legitimacy – satisfaction with democ-racy – across the vast majority of publicdatasets in which such questions have beenasked.It is important to acknowledge upfront

what such questions do, and do not, tell usabout civic attitudes to democracy. The an-swers to such questions primarily tell us howwell citizens perceive their political systemto be performing. They offer a weaker basisfor inferring support for liberal or democraticvalues: individuals may be strong believersin liberal democracy and yet dissatisfied withthe performance of such institutions in prac-tice – or on the flipside, be satisfied with theinstitutions under which they are governed,even though such institutions fall well shortof accepted democratic standards.5

That said, there is value in knowing how,5 See Jonas Linde and Joakim Ekman (2003) “Satisfaction with Democracy: A Note on a Frequently Used Indicator inComparative Politics”. European Journal of Political Research, 42: 391–408; and Pippa Norris (2011) “Does DemocraticSatisfaction Reflect Regime Performance?” in How Democracy Works: Political Representation and Policy Congruence inModern Societies. Ed. Martin Rosema, Bas Denters, and Kees Arts.

Page 4


why and when citizens are losing faith in theability of democracy to deliver. While sub-jective feelings of satisfaction with democ-racy may in part reflect higher or lower civicstandards, the cross-country evidence makesclear that in countries where political insti-tutions are transparent, responsive, and freeof corruption, civic satisfaction with democ-racy is overwhelmingly high. In Denmark,Switzerland, and Luxembourg, political in-stitutions are held to high standards – andmeet those standards.6 There is no evidencethat rising expectations have led to a dete-rioration in democratic legitimacy in suchsocieties. On the other hand, societies wheresatisfaction with democracy is at its lowestare uniformly characterised by political in-stability, corruption scandals, and ingrainedsocietal conflict.Second, even if democratic satisfaction

and support for democratic values are con-ceptually separate, they are empirically re-lated. Studies show that individuals whoare dissatisfied with democracy are morelikely to support populist political partiesthat eschew liberal democratic norms.7 Atthe cross-country level, there is a strong as-sociation between democracies in which thepublic is dissatisfied, and those in which thepublic express lukewarm support for demo-cratic principles. And as we shall see, manyof the countries in the 1990s with the lowestlevels of democratic faith – such as Russia,Venezuela, and Belarus – are exactly thosewhich experienced democratic erosion overthe following decade, often due to electedstrongmen who in office proceeded to under-mine civil rights and liberties.8

Figure 2: Cumulative number of surveys gathered in to the dataset, 1973–2020.

6 Mónica Ferrín, (2016) “An Empirical Assessment of Satisfaction with Democracy” in Mónica Ferrín and Hanspeter Kriesi(eds.), How Europeans View and Evaluate Democracy.

7 Pew Research Center (2018) “Many Across the Globe Are Dissatisfied With How Democracy Is Working”.8 Roberto Stefan Foa and Yascha Mounk (2019) “Democratic Deconsolidation in Developed Democracies, 1995-2018”,Harvard Centre for European Studies Open Forum Series.

Page 5


Data Selection and Validity

In order to ensure that our data representsa valid measurement of satisfaction withdemocracy, the survey questions aggregatedin our dataset are subject to strict standardsof semantic equivalence. First, they mustask citizens about their degree of satisfactionwith democracy in their country; items usingsimilar, yet related terminology are excluded.So too are items asking people their views re-garding democracy in general. Second, theymust be coded on a response scale that allowsfor verifiable equivalence with other surveyresponse scales (see Appendix Section III, on

Testing for Semantic Equivalence, and Sec-tion IV on Sensitivity Analysis).Having recoded responses into satisfied

and dissatisfied and ascertained the percent-age for each category, rolling averages aregenerated by country, while regional aver-ages are generated by merging country sur-veys to a quarterly or annual data series, andtaking the population-weighted average ofthe most recent observation for all countriesin that region over time (see Appendix Sec-tion II, on Aggregation Methodology).

Survey Data Points Countries Years

Afrobarometer 135 36 1999–2015American National Election Studies 7 1 1996–2016AmericasBarometer 169 33 2004–2016Arab Transformations Project 6 6 2013–2014Asia Barometer 41 25 2003–2007Asian Barometer Survey 51 20 2001–2016Australian Election Study 9 1 1996–2016Australian National Political Attitudes Surveys 2 1 1969–1979Australian Survey of Social Attitudes 3 1 2005–2015Comparative Study of Electoral Systems 133 54 1996–2018Consolidation of Democracy 25 14 1990–2001... in Central and Eastern EuropeEU Neighbourhood Barometer 94 16 2012–2014Eurobarometer: Applicant and Candidate Countries 52 13 2002-2004Eurobarometer: Central and Eastern 10 10 1997Eurobarometer: Standard and Special 1471 34 1973–2019European Social Survey 194 35 2002–2017European Values Study 77 45 1999–2018Pew Global Attitudes and Trends 75 50 2007–2018International Social Survey Programme 189 48 1995–2016Latinobarómetro 355 19 1995–2017New Europe Barometer 30 13 1992–2005New Zealand Election Study 6 1 1996–2017South African Social Attitudes Survey 11 1 2003–2013Voice of the People Series 121 69 2005–2007World Values Survey 84 66 1996–2019YouGov Surveys 12 10 2019

Table 1: Data sources used in this study, showing number of survey observations, countries covered, and years ofavailable data.

Page 6


The HUMAN Surveys Project

www.humansurveys.org

The Human Understanding Measured Across National (HUMAN) Surveys project com-bines multiple sources of freely available and nationally representative public opinionsurveys. Available surveys have traditionally been difficult and time-consuming tocompare because the same variables have different names and values across differentsources. HUMAN Surveys uses scripts to format selected variables, merge datasets,and harmonise target variables – including satisfaction with democracy, one of themost frequently-included items in comparative social research.

The scripts work by creating “data warehouses” for micro-level individual responses,as well as macro-level aggregated country-survey and country-year scores. The respon-dent dataset currently includes ten million individuals from 169 countries, combiningdata from over thirty sources, and spanning a period from 1948 to the present. Thoughthis report only includes items on satisfaction with democracy, current target variablesalso include social trust, attitudes towards democracy and elections, and confidencein political institutions.

There are many benefits to using merged multi-survey public opinion datasets. Asthis report illustrates, merged data enables unprecedented geographical and tem-poral coverage, allowing for a better understanding of trends across regions of theworld. HUMAN Surveys saves time in managing large amounts of public opinion data,allowing scholars to focus more attention on key research questions. The scriptingframework is designed to facilitate additional data to eventually include all variablesacross all publicly-accessible surveys.

Page 7


The Presentation of Data in this Report

The purpose of this report is primarily de-scriptive – to lay out in as clear, uncontro-versial, and systematic means as possible –fundamental trends in global public opinion.For that reason, certain standards in visualcommunication that are adhered to in thepresentation of data in this report.

1. First, all data for regional or global av-erages are averaged based on population-weighting. This ensures that figures reflectan estimated average for the pool of all in-dividuals in the region under consideration,and are not disproportionately influenced bytrends in small- or micro-states (see insetbox, “The Importance of Population Weight-ing”).

2. Second, we always use a constant-country sample when presenting aggre-gated data.9 This is to ensure that changeson charts are not due to countries droppingin and out of the dataset, but are only due tochanges in actual collected data. We ensurethis by only including country cases whichare covered by survey data at the start of theobservation period and the end, and “rollingover” survey results in periods in which nonew survey data was collected – in effect,using the “most recent” survey observationfor each country. Fortunately, because thedataset includes such a vast number of ob-servations, for many regions – East and WestEurope, Latin America, North America, andNortheast Asia – there are few countries thatlack consistent data, and many countrieswith multiple observations per year.10

3. Third, when presenting regional aver-ages at the start of each section, rollingaverages are used in order to smooth be-tween years. This is done for the regional

averages to smooth over cases where a “roguepoll” in a large country can cause a suddenyet temporary shift in the weighted mean. Itis more important for regions where surveydata may be collected on a less than annualbasis (e.g. Southeast Asia or the Middle East).On the other hand, in cases where the dataare of sufficiently high quality and frequencywe also present the raw (non-smoothed) se-ries on a quarterly basis – e.g. for the globalseries in Figure 3 and for Western Europe inFigure 17.

4. Fourth, when presenting regional av-erages, we show the full possible range ofthe data on the y-axis (from 0 to 100% ofcitizens who are estimated to be dissatis-fied with democracy), but then highlightthe “relevant range” of the data withinwhich most variation across the worldcan be found (between 25% and 75%).While it is possible to exaggerate changeby narrowing scales, it is also possible tounderstate change by widening scales be-yond a substantively meaningful degree ofvariation. We therefore highlight the areathat corresponds to variation in political out-comes in the real world. In the 25% to 75%range, four-fifths of countries can be found:this is the range that separates Sweden andNew Zealand, at one end, and Venezuela andGreece, at the other.

5. Fifth, when presenting country aver-ages over time, we show all of the in-dividual polling results for that country,together with a rolling average line be-tween them. Where possible we display“raw” individual polling in countries, to allowthe reader to infer the reliability of rollingaverages and means.

9 The one exception to this principle is Figure 16, showing average levels of dissatisfaction in the European Union. This issimply because the country membership of the European Union itself changes over time: an accurate representation ofpublic opinion within the European Union requires country representation to alter in line with the bloc’s membership.However, changes in country membership are clearly indicated underneath this chart, and Figure 17 presents a secondfigure for Europe based solely on the Western European data for countries surveyed since the 1980s, and for whichcountry sample is constant following the entry of Spain and Portugal in 1985 (plus the incorporation of Eastern Germanyin to the Federal Republic of Germany in 1991).

10 In the United Kingdom, for example, we now have an observation rate averaging four surveys a year for the last decade,and a similar level among other major European democracies. But also in many developing democracies, we now havemultiple annual observations over the same period – two surveys per year over the last decade in Brazil, Mexico andSouth Africa, for example, and at least once per year in Nigeria.

Page 8


5. The Global Picture

Figure 3: Rising dissatisfaction with democracy across the world, in democracies representing 2.43 billion individualsacross Latin America, Africa, the Middle East, Europe, North America, East Asia, and Australasia. Basedon 3,218 separate individual country surveys, classified by date of fieldwork and grouped on a quarterlyannual basis, with the total averaged on a population-weighted basis. A constant country sample ismaintained in all periods by rolling forward observations in country quarters lacking new data, to preventsample bias from affecting the changes.

Across the world, satisfaction with democ-racy has fallen, and dissatisfaction risen, overthe past quarter-century. In the mid-1990s,citizens in a majority of countries for whichwe have data felt satisfied with the perfor-mance of democracy in their countries. Bar-ring a brief dip following the Asian and LatinAmerican financial crises of the late 1990s,this remained broadly the case until 2015,when amajority of citizens turned negative intheir evaluation of democratic performance.

Since then, dissatisfaction has continued togrow.Overall, we estimate that the number of

individuals who are “dissatisfied” with thecondition of democracy in their countrieshas risen by 9.7 percentage points, from 47.9to 57.5%. This observation is based on aconstant-country, population-weighted sam-ple of 77 democracies for which data existsfrom themid-1990s to today. This represents2.43bn individuals across the span of Europe,

Page 9


Latin America, Sub-Saharan Africa, the Mid-dle East, North America, East Asia, and Aus-tralasia.

Figure 4:Weighting of regions in the world aggregatestarting in 1995, based on countries availablefor the period under observation. Note thatIndia, for which our survey data begins onlyin 2002, is not included in this sample.

Naturally, there are large differences be-tween regions. In some parts of the world,in particular in North America, Southern Eu-rope, Latin America, and Sub-Saharan Africa,the fall has been acute. In other regions, suchas Northeast Asia, their are no clear positiveor negative trends over the study period. Andin other parts of the world – notably South-east Asia, Scandinavia, and Eastern Europe– there is even a silver lining: satisfactionwith democracy has been rising, rather thanfalling, in recent years.

Explaining Change

What, then, can explain the shifts in publicopinion that we observe in Figure 3? Whyhave people, in general, become more dis-satisfied with the democracies in which theylive? These questions are already the subjectof a vast research literature; yet the annota-tion of the quarter by quarter shifts give ussome initial basis for inference.Once survey data on satisfaction with

democracy are aggregated to a quarterlyannual series, it is clear that specific eco-nomic and political events – the financial

crisis of 2008, the eurozone crisis begin-ning in 2009, the European refugee crisisof 2015 – have had a profound and imme-diate effect on public opinion. Followingthe onset of the global financial crisis andthe collapse of Lehman Brothers in Octo-ber 2008, for example, global dissatisfactionwith the functioning of democracy jumped byaround 6.5 percentage points. Much of thisincrease, moreover, appears to have beendurable. Conversely, signs of democratic gov-ernments working together to resolve policycrises appear to have a positive effect. In thewake of the European Council’s agreement toform a European Stability Mechanism, andthe resultant waning of the sovereign debtcrisis, dissatisfaction with democracy fell by10 percentage points in Western Europe.Further trends from individual countries –

shown in the “country in focus” sections ofthis report – provide further country-specificcases. In the United Kingdom, for example,dissatisfaction with democracy has soaredin the period of the “Brexit crisis”, approxi-mately from the aftermath of the 2017 Gen-eral Election until the most recent surveysin November of last year. And in Brazil, theseries of scandals exposed by the “Lava Jato”corruption probe has seen public dissatisfac-tion reach record highs.This implies that citizens are ultimately

rational in their assessment of democraticgovernance, updating their views in responseto the flow of information. A constant seriesof negative events will push baseline eval-uations of democratic performance lower,while a stream of positive events will pullthat evaluation back up. If citizen viewsof democracy have deteriorated in recentdecades, then there is an least one simple ex-planation: democratic governments simplyhave not been seen to provide effective policysolutions to pressing societal problems. Themore visibly democratic governments appearto be failing to address problems of publicaccountability, economic governance, andtransnational dilemmas such as migrationor climate change, the greater the degreeto which citizens perceive – with some jus-tification – that their institutions are notdelivering results.

Page 10


How do these differences correspond with po-litical changes in theworld over time? Follow-ing the rapid advance of democracy aroundthe world in the 1970s, 1980s, and 1990s, theperiod since 2005 has been characterised as a“democratic recession” in which democraticinstitutions around the world have faced set-backs ranging from military coups, to do-mestic crises, to the election of populist orauthoritarian leaders willing to use their of-fice to erode the independence of parliament,courts and civil society.11 And we find thatmany of those countries which have experi-

enced democratic backsliding since the 1990s– such as Russia, Venezuela, and Belarus –were those most susceptible to doing do (Fig-ure 5).Meanwhile, in many countries that have

retained free and fair elections, there is asense, in thewords of Francis Fukuyama, that“the performance of democracies around theworld has been deficient in recent years”,12 asense reflected by the growing despondencyvisible across struggling democracies in theAmericas, Africa, and Southern Europe.

Satisfaction with Democracy in the 1990s.

Satisfaction with Democracy Today.

Figure 5: Satisfaction with democracy across the world in the mid- to late-1990s, and today (average of most recentthree surveys).

11 Larry Diamond (2015) “Facing Up to the Democratic Recession”. Journal of Democracy.12 Francis Fukuyama (2015) “Why Is Democracy Performing So Poorly?”, Journal of Democracy.

Page 11


We can also examine these differences overtime by country. Figure 6 shows the changeby country in satisfaction with democracyfrom the mid-1990s to the most recent setof observations, and allows us to distinguishfour broad groups.The first zone is that of democratic con-

tentment: where more than three quartersof citizens are satisfied with their democ-racy. These cases, which include Switzerland,the Netherlands, Denmark and Luxembourg,might be considered the world’s “island” ofdemocratic stability – some 2% of the world’sdemocratic citizenry, or 46 million people,who have thus far avoided the malaise else-where.The second group can be labelled as “cases

of concern”, where more than a quarter, yetless than half, of the population are dissatis-fied with the country’s institutions. This isa much larger subset of countries, contain-ing 393 million people, and includes many ofthe new democracies of central Europe, alongwith Germany, Canada and Australia. How-ever many of these societies, in particular ineastern Europe, have also seen democraticsatisfaction rise in recent decades.Next, by far the largest group by popula-

tion – accounting for 1.09 billion people –are the democracies in “malaise”: countrieswhere a majority (though less than three-quarters) feel dissatisfied with democraticperformance. It is here that we find manyof the large democracies of the world – theUnited States, Japan, the United Kingdom,Spain and France. With the exception ofFrance, these are all new entrants to this cat-egory, having formerly possessed majoritiesthat were satisfied with their democratic in-stitutions.Finallly, there are 377 million people liv-

ing in democracies that face a genuine legiti-macy “crisis”, insofar as for every citizen whois satisfied with how the democratic system isworking, there are three who have lost faithin the process. These countries include Mex-ico, Brazil, and Ukraine. This is the subset ofcountries for which the term “crisis of democ-racy” may be considered neutral and descrip-tive.

Figure 6: Change in percentage who are dissatisfiedwith democracy, from mid-1990s (average ofsurveys) to latest observation (average of thethree most recent surveys). Points proportion-ate to country population. While most largedemocracies have seen rising rate of discon-tentment, a number of small democracies haveimproved, notably in Scandinavia and post-communist Europe.

Page 12


Methodological Note

The Importance of Population-Weighting

All of the regional averages shown in thisreport are population-weighted averagesof the countries they contain. This ensuresthat the reported levels of democratic satis-faction or dissatisfaction are representativeof the total population within that region.Otherwise, using an equal-weighted aver-age of the countries in a region, we arriveat extreme distortions. Each citizen ofLuxembourg, for example, with a popula-tion of 0.62m, would count for 134 citizensof Germany, with its population of 83m.

By equal-weighting countries, we under-weight people.Failure to weight by population can there-fore create a highly misleading impressionof regional and global trends. As the figurebelow illustrates, the trend among smallcountries is opposite to that found in theworld as a whole: yet it reflects the politicalreality of an extremely small minority ofindividuals, predominantly concentratedin small, wealthy, high-income countries.

Figure 7: Changes in levels of democratic dissatisfaction within large democracies (population above10m) and small democracies (population below 10m). The thickness of the lines is relative tothe total population of each group. While each grouping contains a similar number of countries– 34 vs. 41 – large democracies here contain 93% of the total population of the two groups.

Page 13


6. Regional Trends and Comparisons

Figure 8: Change over time in democratic dissatisfaction by region from mid-1990s to latest set of observations. ForSoutheast Asia, changes are only shown since 2001; for South Asia, since 2002. Regional points relativeto size of region’s population. Dissatisfaction has risen in every region except in South Asia, where it hasremained stable, and Southeast Asia, where it has fallen.

Until recent decades, the vast majority ofresearch on political attitudes was based ondata from countries that are, as the acronymgoes, “weird” – that is to say, western, ed-ucated, industrialised, rich and democratic.Yet the phenomenal advance in collabora-tive, global survey projects – starting withthe World Values Survey and the Interna-tional Social Survey Programme in the 1980s,moving to the Global Barometer projects inthe 1990s, and finally more recent initiativessuch as the Gallup World Poll, Pew GlobalAttitudes Survey, and the YouGov-CambridgeGlobalism Project – makes it possible to con-duct truly global comparative research.One of the key findings of such research,

not surprisingly, is that regions differ. Obser-vations of trends occurring in one region, arenot always representative of the global pic-

ture. Moreover, rarely is data simply presentormissing at random: often it is the countrieswith the best institutions (including socialscience research institutions) that have pro-vided the most data.13 In the rest of thisreport, therefore, we move beyond the globalaggregates in order to examine each regionon an individual basis. What is the patternin each major cluster of world democracies?And what specific factors could, in principle,help to understand the trends that we ob-serve?

The Overall Picture

Figure 8 shows change over the past gener-ation in seven major regions: Europe, LatinAmerica, Southeast Asia, South Asia, Sub-Saharan Africa, the “Anglo-Saxon” democra-

13 Joseph Henrich, Steven J. Heine and Ara Norenzayan (2010) “The Weirdest People in the World?”, Behavioral and BrainSciences.

14 Two additional regions, the Caribbean and the Middle East, are excluded due to low population – in the former case dueto low population in absolute terms, and in the latter case as the population of the democratic Middle East is so low.

Page 14


Figure 9: Distribution of democratic dissatisfaction by region (latest data), showing how dissatisfaction variesamong countries within each geographic area. Country points relative to size of country population. Insome regions, levels of dissatisfaction are similar among countries: for example in Northeast Asia, around50%, or in Southeast Asia, at just 25%. Yet other regions contain high variation – in particular Europe,which contains some of the highest and some of the lowest levels of dissatisfaction in the world.

cies (North America, Australasia and theUnited Kingdom), and Northeast Asia.14

Most regions are moving toward greater dis-satisfaction with democracy, though, not all.In South Asia, the data appear flat since thefirst observations in our dataset (2002), whilein Southeast Asia, public satisfaction withdemocracy has improved a great deal since2001, the year of our first survey observationfor Indonesia. The overall ranking of regionshas also changed markedly over time: in the1990s, the Anglo-Saxon countries were themost satisfied with democracy and SoutheastAsia the least so, whereas today, their posi-tions are almost reversed. Finally, whereasLatin America was almost alone in the 1990sfor having a majority of its citizens dissatis-fied with the democratic process, now otherregions are catching up.

Variation Within Regions

Figure 9 also shows the variation within eachregion by country. Broadly speaking, coun-tries within each region tend to cluster to-gether. One significant exception, however,is Europe – which contains both the world’smost satisfied, but also many of its least sat-isfied democracies. This is due to a growingdivide between northern and southern Eu-rope, which we detail and explore furtherin the Europe section of this report. Else-where in the world, there are clearer regionalpatterns: Latin America overwhelmingly con-sists of democracies where citizens are dis-satisfied with their institutions, while Southand Southeast Asia contain democracies inwhich civic satisfaction is robust.

Page 15


Regional Convergence

Developed and Developing Democracies

While dissatisfaction with the performanceof democracy exists across both developedand developing democracies, there are im-portant differences between the two. Indeveloping democracies in Latin America,Africa, and Eastern Europe, levels of discon-tent are high – with more than half of re-spondents typically dissatisfied – yet withonly a slight average increase in the lastquarter century.In developed democracies, by contrast,public satisfaction has eroded since the1990s, with levels of discontent rising froma third to half of all respondents. To

the extent that the performance legiti-macy of democratic institutions is a featureof democratic consolidation, this impliesthat some countries that were previouslythought to be consolidated, such as Greece,Chile, or even the United States, may havepartly “deconsolidated” in recent years.As expected during the early post-ColdWar years, in this respect developed anddeveloping democracies have converged.However, they have done so more in thedirection of the latter than in the directionof the former.

Figure 10: Changes in levels of democratic dissatisfaction among developed and developing democracies,since the mid-1990s. Consistent country sample.

Page 16


Regional Divergence

The Asian Exception

While satisfaction with democratic perfor-mance has declined across the world overthe past generation, one region standsas a notable exception: Asia. Whetheramong the emerging democracies of South-east Asia, the developed democracies ofnortheast Asia, or in the world’s longest-established major developing democracy –India – satisfaction with democratic gov-ernance in the East seems higher than inother parts of the world.On average almost two-thirds of Asians

are satisfied with the performance ofdemocracy in their countries, comparedto just two-fifths of individuals elsewhere.Not only has this gap existed since thestart of widespread surveys across Southand Southeast Asia in the early 2000s, butit has widened since. The observation ofdemocratic contentment in Asia, togetherwith the rise of Asia in general, offer theprospect of a radically different interpre-tation of democratic prospects and legiti-macy in the twenty-first century.

Figure 11: Changes in levels of democratic dissatisfaction among democracies in Asia, and democraciesoutside of Asia.

Page 17


The Anglo-Saxon Countries – the End of Exceptionalism?

Figure 12: Satisfaction with Democracy in the “Anglo-Saxon” Countries: the United States, Australia, Canada, NewZealand, and the United Kingdom.

In recent years, there has been an especiallyacute crisis of democratic faith in the “Anglo-Saxon” democracies – the United States, Aus-tralia, Canada, New Zealand, and the UnitedKingdom. Overall, the proportion of citizenswho are “dissatisfied” with the performanceof democracy in these countries has doubledsince the 1990s, from a quarter, to half of allindividuals.Though much of this increase is accounted

for by the United States, public levels ofconfidence have also slipped in the UnitedKingdom, Australia, and Canada. While theproportion of Americans who are dissatis-fied with democracy has increased by overone-third of the population (+34 percentagepoints) since the mid-1990s, this amount hasalso risen by one-fifth of the population inAustralia (+19 percentage points) and Britain(+18 percentage points), and by almost a

tenth of Canadians (Figure 13).What can explain this synchronised down-

turn in public sentiment across high-income,English-speaking democracies? First, giventhe concurrence of the shift with the tim-ing of the global financial crisis, economicfactors may play an important role. Yet thisexplanation, while a part of the story, wouldstruggle to explain why Australia, whichlargely avoided an economic downturn af-ter 2008, appears as negatively affected asBritain and the United States. An alternativethough related view is that the financialisa-tion of the U.S., British, Canadian, and Aus-tralian economies has led to this outcomeby exacerbating spatial inequality betweena handful of successful, globally-integratedcosmopolitan cities – New York, London,Toronto, or Sydney – and the rest of theirsocieties. Evidence suggests that rising in-

15 Sung Min Han and Eric C.C. Chang, (2016) “Economic Inequality, Winner-Loser gap, and Satisfaction with Democracy”,Electoral Studies, 44: 85–97.

Page 18


Figure 13: Change in satisfaction with democracy since 1995 in the Anglo-Saxon democracies, from baseline level.

come inequality also decreases satisfactionwith democracy,15 and the effect may beespecially strong where entire regions of acountry feel left behind – and whose needshave been ignored by political parties dueto the prevalence of either gerrymanderedor “safe” seats.16 This sense of exclusion andfrustration with political elites is only madestronger when the other effect of incomeinequality is to skew influence over the po-litical system, providing increased resourcesfor lobbyists and rendering politicians moredependent upon securing donor campaigncontributions.A second literature that is pertinent to ex-

plaining the trajectory of the Anglo-Saxondemocracies suggests that satisfaction withdemocracy is lower in majoritarian “winner-takes-all” systems than in consensus-based,proportionally representative democracies,17

and this could explain why New Zealand –the lone member of this group with electionsby proportional representation – appears tohave avoided a trajectory of soaring public

discontent (Figure 13).This still leaves the question, of course, as

to why discontent in majoritarian democ-racies has been on the rise, rather thansimply higher on average. One intriguinghypothesis is that there may be an interac-tion between the confrontational, two-partymodel of Anglo-Saxon politics, and the ef-fect of social media in siloing society intoopposing “tribes”.18 Rising political polarisa-tion between Democrats and Republicans inthe United States is one example, though aclearer demonstration is themanner inwhichBritain’s referendum on whether to remainin or leave the European Union subsequentlysplit society into the hitherto unknown cat-egories of “Leavers” vs. “Remainers”. Com-bined with social media, the winner-takes-allnature of political competition in Anglo-Saxon democracies contributes to polarisa-tion, which in turnmakes citizens less willingto compromise or accept the legitimacy of arival’s electoral mandate.

16 Lawrence Ezrow and Georgios Xezonakis (2011) “Citizen Satisfaction with Democracy and Parties’ Policy Offerings: ACross-National Analysis of Twelve European Party Systems, 1976–2003”, Comparative Political Studies.

17 Kees Aarts and Jacques Thomassen, (2008) “Satisfaction with Democracy: Do Institutions Matter?”, Electoral Studies, 27,5–18.

18 Andrea Ceron and Vincenzo Memoli (2016) “Flames and Debates: Do Social Media Affect Satisfaction with Democracy?”Social Indicators Research.

Page 19


Country in Focus

The United States

In few countries has the decline in satisfac-tion with democratic performance been asdramatic or as unexpected as in the UnitedStates. For much of its modern history,America has viewed itself as a “shiningcity on a hill” – a model democracy, andone that can serve as an example unto oth-ers that wish to emulate its success. Sur-vey data shows that the American publiclargely cohered with such a viewpoint, withmore than three-quarters of U.S. respon-dents regularly expressing their “satisfac-tion” with the state of democracy in Amer-ica.Following the 2008 financial crisis, how-ever, that has begun to change, with Ameri-cans’ evaluation of the functioning of theirpolitical system continuing to deteriorateyear on year. Rising political polarisation,government shutdowns, the widespreaduse of public office for private gain, a costlywar in Iraq, and growing spatial and inter-generational inequality have all weighed

against Americans’ view of the ability oftheir democracy to deliver. Now, for thefirst time on record, polls show a majorityof Americans dissatisfied with their systemof government – a system of which theywere once famously proud.Such levels of democratic dissatisfactionwould not be unusual elsewhere. But fortheUnited States, itmarks an “end of excep-tionalism” – a profound shift in America’sview of itself, and therefore, of its place inthe world. It is a reflection of just how re-markable this shift in sentiment has beenthat a presidential candidate – Donald J.Trump – could arrive at theWhite House af-ter a presidential campaign that denouncedAmerican political institutions as corrupt,and promised to step back from promot-ing democracy abroad in favour of putting“America First”, treating all countries trans-actionally based on a spirit of realism, re-gardless of their adherence to or deviationfrom democratic norms.

Figure 14: Rising dissatisfaction with democracy in the United States from the mid-1990s to the presentday. Shown here are individual survey observations, plus their rolling average.

Page 20


Country in Focus

The United Kingdom

The United Kingdom has one of the richestdatasets on satisfaction with democracy,starting with the first surveys completedin the early 1970s by Eurobarometer, andsupplemented by more recent surveys com-pleted by survey organisations includingGallup, Pew, and YouGov.For thirty years from the 1970s to the earlytwenty-first century, average levels of satis-faction with democracy in the U.K. were,in fact, rising. The 1970s marked a mo-ment of deep crisis for Britain, with generalstrikes, power cuts, periods ofminority gov-ernment, an embarrassing IMF bailout, andthe start of “the troubles” in Northern Ire-land. By the end of the 1990s, when TonyBlair’s government had brokered the GoodFriday agreement in Northern Ireland, de-volved power to Scotland and Wales, andrebranded the country as a “cool Britannia”that could reconcile Margaret Thatcher’s

market reforms with increased social in-vestment in health and education, it hadseemed that “things” – to borrow NewLabour’s ownpoliticalmantra – “could onlyget better”. And indeed, had done so.In retrospect, this period representeda high point for satisfaction in Britain’sdemocratic institutions which would neverbe reattained. Following brief dips after theIraq War and the parliamentary expensesscandal, satisfaction with democracy hasplunged during the “Brexit” stalemate of2016-19. The sudden onset of Britain’s re-cent legitimacy crisis holds out the hopethat a recovery may follow as swiftly as thebreakdown. Yet this will depend upon howthe United Kingdom finally negotiates itsway out of the Brexit impasse, and movesbeyond the divisions sowed by the 2016 ref-erendum.

Figure 15: Dissatisfaction with democracy in the United Kingdom from the mid-1970s to the present day.Shown here are individual survey observations, plus their rolling average. For much of thisperiod, a majority of respondents have expressed contentment with the outcome of the Britishpolitical process. However, in 2019, for the first time since the mid-1970s, more than half ofBritish respondents were dissatisfied with democracy in the UK.

Page 21


Europe – Complacency and Internal Divergence

Figure 16: Satisfaction with democracy in European Union member states, 1975-2020. Addition of countries to theEEC/EC/EU shown by country flags.

Overall, satisfaction with democracy in Eu-rope has fluctuated in cycles over the pasthalf century. We are now at a third peak inrelative dissatisfaction, with the first dur-ing the “governability crises” of the 1970s,the second following the 1990s recessionand the final phase beginning in 2009 withthe eurozone crisis. Interspersed with theseepisodes, there have been periods of relativecivic contentment – for a few years followingthe collapse of the Berlin Wall, and for thefirst decade following the launch of the euro.Previous episodes were experienced as

genuine crises at the time, capable of disrupt-ing democratic stability within individualEuropean countries. The Trilateral Commis-sion report of 1975 warned of a threat todemocratic institutions in Italy, the UnitedKingdom, West Germany, and France, asstrikes, riots, and terrorism became common-

place, and support for the communist Leftgrew.19 Meanwhile, though democratic ide-als emerged unassailable from the end of theCold War, democratic institutions seemedless robust: in 1992 the Italian party systemimploded following corruption scandals, andthe collapse of the communist parties wasfollowed by the growth of the extreme rightin countries such as France, Austria, and theNetherlands.Sowhy then, does Europe’s currentmalaise

“feel” worse than past episodes? The first dif-ference, perhaps, is one of duration: The pe-riod from 1992 to 1994 may have containedthe ERM crisis, the collapse of the Italianparty system, and the 1990s recession, yetwas a blip between two periods of relativeoptimism and euphoria (Figure 17).Secondly, there is the fact of divergence:

Europe’s average level of satisfaction masks

19 Michel Crozier, Samuel Huntington and Joji Watanuki (1975) “The Crisis of Democracy: Report on the Governability ofDemocracies to the Trilateral Commission”.

Page 22


a large and growing divide within the con-tinent, between a “zone of despair” acrossFrance and southern Europe, and a “zoneof complacency” across western Germany,Scandinavia, and the Netherlands. Becausethe former group began from a lower base-line, the result is a wide and growing “sat-isfaction gap” – much in evidence from the2019 elections to the European Parliament,where anti-system parties swept seats acrossFrance, Italy, Spain, and Greece, yet were lesssuccessful in northern Europe.The length of the current malaise also ex-

plains why this time it has led to a wave ofpopulism, a wave that began some five yearsafter the onset of the eurozone sovereign debtcrisis. At first, European publics were pre-pared to give established parties a chance toaddress the continent’s mounting economicand migration challenges. For a numberof years in Greece, successive governmentsfollowed the strictures of the IMF bailoutpackage, while in Italy, Silvio Berlusconi re-signed tomake way for a technocratic govern-

ment. In the United Kingdom, deep austeritypolicies were initially supported by the elec-torate, while in France, François Hollandecampaigned on a platform of increased taxesand reduced public spending. By the end ofthe decade, however, electorates had lost pa-tience. A left-populist government in Greecewon (then ignored) a referendum rejectingthe terms of the country’s bailout agreement,the Cameron government collapsed after los-ing a referendum to remain in the EuropeanUnion, and Italy’s technocrats were sweptfrom office by a populist coalition of leftand right. Whether and when Europe canescape its longest period of institutional dis-satisfaction on record, will depend in largepart upon the capacity of governments toescape its underlying pressures – economicstagnation,20 regional inequality within andbetween countries,21 demographic anxieties,and imbalances of power between nationstates in the post-Lisbon Treaty EuropeanUnion.22

Figure 17: Dissatisfaction with democracy in Western Europe, divided into three phases: 1974–1992; 1992–2008;2008–today. Population-weighted, constant country sample barring additions of: (i) Greece (data seriesbegins in 1980), (ii) Spain and Portugal (data series begin in 1985), (iii) Eastern Germany (from 1990,included in the Federal Republic of Germany). Europe has experienced legitimacy crises in the past; yetin episodes of 4-6 years, rather than a decade-long malaise.

20 Klaus Armingeon and Kai Guthmann (2013) “Democracy in crisis? The declining support for national democracy inEuropean countries, 2007–2011”, European Journal of Political Research.

21 Lewis Dijkstra, Hugo Poelman and Andrés Rodríguez-Pose (2019), “The geography of EU discontent”, Regional Studies;Roberto Stefan Foa and Jonathan Wilmot (2019) “The West Has a Resentment Epidemic”, Foreign Policy.

22 Sonia Alonso and Rubén Ruiz-Rufino, (2018) “The costs of responsibility for the political establishment of the Eurozone(1999–2015)”, Party Politics.

Page 23


Divergence within the Eurozone

Is democratic dissatisfaction in Europesimply a cyclical economic effect, or doesit reflect a deeper sense of discontent withhow institutions are performing?In order to shed light on this ques-

tion, we produced a second version of ourdataset that was “detrended” to removethe cyclical effect of economic growthupon democratic satisfaction. This is de-tailed in section V of the Appendix, butin short, we estimated the portion of thechange in satisfaction with democracythat covaries with recent economic per-formance, and adjusted the survey resultsto reflect what might have happened “asif” the level of GDP growth had remainedconstant.After detrending the growth rate effect

from European data, divergence withinthe eurozone since the launch of the sin-gle currency reduces – but only moder-ately. In many countries, though notablyin France and Spain, popular discontenthas continued growing throughout the pe-riod of economic recovery. To the extentthat economic factors matter, they arelikely to be mediated by factors that dis-play less of a clear cyclical pattern, suchas wages or regional inequality.Meanwhile, how does the evolution of

dissatisfaction appear over time? Dur-ing the first decade following the euro’slaunch, growth-detrended satisfactionwith democracy in southern Europe wasmarginally lower than in northern Europe(Figure 19). However, this gap widenssignificantly in the decade following thesovereign debt crisis of 2009-11. It islikely that, beyond personal feelings ofeconomic dissatisfaction, the crisis hasbrought forward a broader sense of polit-ical discontent that is tied to economicsovereignty, national pride, and angerover the use of public resources.

Figure 18: Change in satisfaction with democracysince launch of the euro, across mem-ber countries: raw (faded) and growth-adjusted (blue) figures. Points are pro-portional in area to country population.Sluggish economic growth has depressedsatisfaction with democracy in most eu-rozone countries, but cannot explain di-vergence between regions.

Figure 19: Change in satisfaction with democracysince launch of the euro, 5-year rollingaverages after detrending for short-termeconomic effects.

Page 24


Optimism in the East, Gloom in the West

If Western Europe is disillusioned, the for-mer communist countries of the Eastern Bloc,many of which joined the European Unionin 2004 and 2007, are experiencing a steadyconsolidation of faith in their new politicalinstitutions.23 Admittedly, these countriesbegan from a very low base, with only a smallminority of respondents – between a fifth anda third – expressing “satisfaction” during theeconomic transition years of the 1990s.While the combination of economic recov-

ery and European Union accession might ap-pear the most plausible explanations for ris-ing democratic contentment, in fact, theseshow very little relation to trends in the data.By 2010, only Poland had seen a substantialrecovery in political contentment, with othercountries barely up from their 1990s levels(Figure 20). Only since the onset of the globalfinancial crisis has satisfaction with democ-

racy in postcommunist Europe improved in auniform fashion – a period that is well afterthe completion of the accession process, andduring which economic growth has slowedacross the region, in line with the rest of thecontinent.It is also a period that in most countries

displayed in the figure below – notably Hun-gary, Poland, Slovakia and the Czech Repub-lic – the first generation of liberal postcom-munist elites has been swept aside by theelection to high office of populist politiciansand parties, often on a platform of nation-alism, social welfare, and anti-immigration.The concurrence of populism and democraticsatisfaction reminds us, perhaps, that satis-faction with democracy is not the same as abelief in liberal principles or values – but isas much due to congruence between popu-lar sentiment and the attitudes expressed bythe political class, whatever those sentimentsmay be.

Figure 20: Evidence of democratic consolidation in post-communist Europe? With the notable exception of Romania,the major new democracies of central and eastern Europe have seen a gradual strengthening of civicconfidence in their political institutions since the “dual transition” to democracy and the market economyin the 1990s. Notably, central Europe is one of the few regions to have witnessed an increase in satisfactionwith democracy since the global financial crisis. Rolling averages are displayed.

23 Christian Haerpfer and Kseniya Kizilova (2015) “Support for Democracy in Post-communist Europe and Post-Soviet Eura-sia” in The Civic Culture Transformed; Roberto Stefan Foa and Grzegorz Ekiert (2017) “The Weakness of PostcommunistCivil Society Reassessed,” European Journal of Political Research, Vol. 56(2).

Page 25


Figure 21: Declining dissatisfaction with democracy in the Former Soviet Union countries for which time-seriesdata are available. The Former Soviet Union average includes Russia, Ukraine, Belarus, Kyrgyzstan, andMoldova; but excludes the EU accession countries (the Baltic states).

Further East: The Former Soviet Union

Beyond the EuropeanUnion, time-series dataon satisfaction with democracy is availablefor a number of former Soviet republics, in-cluding the Russian Federation, Ukraine, Be-larus, Armenia, Georgia, Kyrgyzstan, andMoldova.The story of democratic transition in the

post-Soviet space is more chequered thanamong the EU accession states. While mostformer Soviet states attempted some form ofdemocratisation in the 1990s, in a numberof cases they relapsed toward authoritarianrule – relatively rapidly in Kazakhstan andBelarus, somewhat more slowly in Russia –while those countries which maintained gen-uinely competitive elections have struggledwith economic stagnation, corruption, andcivil strife.Public opinion across the post-Soviet space

shows democracy to be struggling, and au-thoritarianism resurgent. Among the re-gion’s remaining democracies – in Moldova,Georgia, Armenia and Ukraine – satisfactionwith democracy in the last two decades has

stagnated or declined from an extremely lowlevel, with 3 in every 4 citizens expressing dis-content with the democratic system as theycurrently experience it.Meanwhile in Russia and Belarus, regu-

lar data on civic satisfaction with “democ-racy” continues to be collected – as it hasbeen since the mid-1990s – in spite of theincreasingly autocratic governance of thesecountries. In general, such surveys show acontinuing recovery of confidence in the po-litical system. Figure 21 shows this trendin Russia, for example, compared to thepopulation-weighted average among non-EU, post-Soviet nations. The contrast be-tween democratic and authoritarian politiesin the post-Soviet space indicates an impor-tant point: citizens evaluate the performanceof their polity not only by its adherence toliberal-democratic norms, but also for its abil-ity to offer valued outputs such as politicaland social stability, economic growth, and asense of collective purpose and pride. To theextent that emerging democracies fail to dothis, the legitimacy of the democratic systemitself may be eroded.

Page 26


Country in Focus

Greece

Greece joined the European Union on the1st of January, 1981, and began fielding Eu-robarometer surveys the previous year. Asa result there is consistent data on demo-cratic attitudes for 40 years – from the ini-tial years following the country’s return todemocracy in 1975 to the period of the eu-rozone crisis that began in 2009.By the early years of the twenty-first cen-tury, with two-third of citizens satisfiedwith the political system, it appeared asif Greece had finally achieved democratic“consolidation” – a condition that politi-

cal scientists define as one in which demo-cratic legitimacy is assured, and alterna-tives to democratic governance have be-come unthinkable.However, the sovereign debt crisis hasshown the fragility of this condition. Sev-eral years in to the eurozone debt crisis,Greece saw the revival of an openly “fascist”political party, in the form of Golden Dawn,and the electoral victory of a left-populistgovernment prepared to challenge the in-dependence of the media and the courts.

Figure 22: Dissatisfaction with democracy in Greece, from 1980 to the present day. Public opinion inGreece has fluctuated in cycles, with only brief periods – under the Andreas Papandreougovernment of 1981–1989, and then again in the 2000s – when a majority of the publicapproved of how democracy was performing. Since the onset of the sovereign debt crisis,discontent has reached a record high, with 4 out of 5 Greeks expressing dissatisfaction. Inrecent years, however, this level is starting to fall back once more.

Page 27


Latin America – the Cycle of Instability

Figure 23: Satisfaction with Democracy in Latin America. Population-weighted average of Brazil, Argentina,Colombia, Chile, Mexico, Uruguay, Venezuela, Nicaragua, Bolivia, Paraguay, Ecuador, El Salvador, Peru,Panama, Costa Rica, Guatemala, and Honduras.

In no other region of the world has theprospect of democratic consolidation ap-peared as elusive as it has in Latin Amer-ica. In the nineteenth century, countries inthe region were among the pioneers of rep-resentative government, with competitiveelections occurring in Chile, Colombia, andPeru, among other cases. Yet for more than acentury since, democratic institutions havebeen wracked by cyclical crises, culminatingin military coups, civil wars, revolutionarypopulism, and single-party dictatorships. Inevery period when democratic institutionsappeared to be taking root once more – inthe 1920s, 1950s, and 1980s – another periodhas followed in which their legitimacy haseroded and the cycle has repeated anew.We are currently in the midst of precisely

such a “reverse wave.” More than 3 in ev-ery 4 citizens across the region register dis-content with the performance of democracy– the highest level since the start of publicopinion surveys. Concurrent with this legiti-

macy crisis, we have also seen the steady ero-sion of democratic institutions in Venezuela,Ecuador, and Bolivia, the recent electionof populist leaders in Mexico and Brazil,and widespread violent protests and riots inColombia and Chile.This has proven a sharp reversal following

the period of relative optimism that prevailedin the first decade of the twenty-first cen-tury. At that time, a series of leaders – fromPresident Lula in Brazil, to Michele Bacheletin Chile, and Cristina Fernandez de Kirch-ner in Argentina – were elected to office aspart of a “pink wave” of social-democraticpoliticians, promising to use the proceeds ofa commodities boom to invest in overcom-ing entrenched inequalities. Coming just twodecades after Latin America’s rapid transi-tion to liberal democracy – as recently as thelate 1970s, 17 out of 20 Latin American na-tions had been ruled by dictators – it seemedas if Latin America was following the socialdemocratic path of Spain, Italy, and Portugal.

Page 28


Figure 24: From recovery to reversal – the cycle of instability. Major democracies in Latin America experiencedeuphoric recovery from the financial crisis of the late 1990s, with democratic satisfaction rising by up to30 percentage points. Yet in the last decade, the region has plunged into an even worse legitimacy crisis.

What has since gone wrong in Latin Amer-ica, and why has the region failed – yet again– to achieve democratic stability? In many re-spects, the promise of the pink wave provedunsustainable because even its more moder-ate members – for example, president Lulain Brazil or Cristina Fernandez de Kirchnerin Argentina – remained trapped in the pop-ulist tradition of clientelism and procycli-cal spending, leaving countries with limitedspace for borrowing once the commodity cy-cle turned. Meanwhile, whereas Europeansocial democracies combine universal publicservices with institutions and policies thatsupport private sector growth and invest-ment, in Latin America a deteriorating busi-ness environment has hobbled the meansof rekindling economic growth. Nearly ev-ery major economy fell in its ranking on theWorld Economic Forum Global Competitive-ness Index in the decade to 2018.Yet the failures of Latin American democ-

racies are not those of one decade or one setof leaders but rather, issues that are chronic

and structural. Corruption is endemic, andhas deteriorated in recent decades accordingto Transparency International’s CorruptionPerceptions Index. Violent crime is out ofcontrol, as countries from Mexico to Brazilreport homicide rates that are double or triplethe level from two decades prior. And socialinjustices persist due to stubbornly high in-equality, regressive tax rates, and informallabour markets that exclude the poor frombasic legal protections and services.All of this has contributed to Latin Amer-

ica’s current “year of rage”, with violentprotests from Buenos Aires to Bogotá, andSantiago to San Juan, Puerto Rico. While it ispossible that record high levels of dissatisfac-tion and resultant civic protest could lead todemocratic renewal, other survey measurespaint a less optimistic interpretation: sup-port for democracy over other forms of gov-ernment fell belowhalf of survey respondentsin 2018, while the proportion who believethat high levels of crime and corruption jus-tify a military coup reached 37%.24

24 See Latinobarometro (2018) Informe Latinobarometro 2018, http://www.latinobarometro.org/lat.jsp., and Mollie J. Cohen,Noam Lupu, and Elizabeth J. Zechmeister (2017) “The Political Culture of Democracy in the Americas, 2016/17”.

Page 29


Country in Focus

Brazil

In 1985, Brazil returned to democratic rulefollowing two decades of military dictator-ship. However, formuch of the period since,the country’s institutions have struggled togain credibility, with persistent corruption,urban poverty, and rising criminal violencemarring the impression that democracy candeliver.A brief exception to this malaise occurredduring the first decade of the twenty-firstcentury, under the government of Luiz In-ácio Lula da Silva from 2003 to 2010. En-joying the fruits of a global commoditiesboom, the Lula administration invested inprogrammes to relieve widespread povertyand reduce inequality, while maintaininga commitment to reduce inflation, holddown public debt, and attract foreign di-rect investment. By the time that Brazilwas awarded the right in 2007 to hold FIFAWorld Cup tournament, it appeared that deGaulle’s eternal “country of the future” hadfinally arrived – if not exactly in the future,

then at least in the democratic present day.For the first time on record, on average amajority of Brazilians expressed satisfac-tion with their political system – a feelingof contentment that was to last until thetournament itself was hosted in 2014.In retrospect, however, this was only ahiatus between two periods of instabil-ity – one marked by the aftermath of theemerging markets financial crisis of thelate 1990s, and another that began withthe Lava Jato (“car wash”) investigationswhich exposed widespread clientelism andnepotism graft on the underbelly of theBrazilian political system. With the formerBrazilian President Lula da Silva in jail, inlate 2018 rightwing populist candidate JairBolsonaro won the presidency on a plat-form that included support for vigilantismagainst petty and organised crime, and nos-talgia for the country’s former military dic-tatorship. For Brazil, it seems, the futurehas been deferred once more.

Figure 25: Dissatisfaction with democracy in Brazil from the mid-1990s to the present day.

Page 30


Latin America: The Pink Tide Goes Out

Howmuch of Latin America’s swelling po-litical discontent is tied to cyclical eco-nomic reversal, and how much to deeperinstitutional factors? We have exam-ined this using a growth-detrended ver-sion of the satisfaction with democracydata, which strips out the effect of re-cent economic growth (see Appendix sec-

tion V). During the “pink tide”, dissatis-faction fell substantially in country aftercountry. Very little of this drop was ex-plained by economic growth per se, exceptin Venezuela where “growth-detrended”dissatisfaction with democracy remainedaround 50% even at the height of theChávez years.

Brazil Venezuela

Argentina Ecuador

Figure 26: Raw and detrended data series in Brazil, Venzuela, Argentina, and Ecuador. Detrended 5-yearrolling averages in black lines; raw series in dotted lines.

Page 31


Country in Focus

Mexico

After Brazil, Mexico is Latin America’ssecond-largest democracy, as well as thesixth-largest democracy in the world. Yetunlike most Latin American nations, whichtransitioned to democracy very rapidly inthe 1980s following the collapse of mil-itary juntas, Mexico experienced a moregradual democratisation. The dominantPartido Revolucionario Institucional (PRI)governed the country uninterrupted andlargely unchallenged for sixty years from1929, though began to face oppositionfrom splinter parties in the 1980s. ThePRI responded to initial challenges withwidespread electoral fraud, yet in 1994, al-lowed international observers to monitorthe election for the first time. In 2000, thefirst peaceful handover of power occurredas the PRI was defeated by opposition can-didate Vicente Fox.At no point during this transition, how-ever, has a consistent majority of the Mex-ican public expressed confidence in the

democratic process. In large part, this isdue to endemic problems of drug violence,organised crime, political corruption, andthe frequent assassination of journalistsand human rights activists. Added to thesefactors, sluggish economic growth sincethe 1980s has failed to lift a sufficient num-ber of individuals from poverty, and in-equality, both within cities and betweenthe capital Mexico City and the poorer re-gions of the south, remains structurally em-bedded.In 2018, amidst widespread political angerand discontent, Andrés Manuel LópezObrador was elected President on a left-populist platform. For now, it remains tobe seen whether he will end up more of aMexican Lula da Silva – capable of marry-ing social justice and economic stability -or something closer to Venezuela’s HugoChávez, whose decade of mismanagementhas plunged the country in to its deepestever economic, political, and social crisis.

Figure 27: Dissatisfaction with democracy in Mexico from the mid-1990s to the present day. At nopoint has a clear majority of the Mexican public felt satisfied with the country’s democraticinstitutions, though in recent years discontent has reached a record high.

Page 32


Region in Focus

The Caribbean Region – Lost at Sea?

Figure 28: Satisfaction with Democracy in the Caribbean. Key countries by population include Jamaica,Trinidad and Tobago, Haiti, and the Dominican Republic.

The Caribbean region has a strong tradi-tion of liberal democratic politics: sinceindependence in the 1960s, ten of itstwelve Anglophone countries have heldconsistently free and fair elections, as hasSpanish-speaking Puerto Rico. AnotherHispanophone country (the Dominican Re-public) returned to democracy in the 1970s,years before comparable states in LatinAmerica.However, rising levels of narcotics traffick-

ing, crime, and corruption have threatenedthese outcomes. Civic confidence in domes-tic institutions is eroded, at its lowest levelin two decades.As figure 29 below shows, this leaves widevariation within the region. In Haiti, theDominican Republic and Jamaica, only aquarter to one-third of respondents ex-press contentmentwith their systemof gov-ernment – while in Barbados, Antigua andthe Bahamas, a clear majority do so.

Figure 29: Levels of dissatisfaction with democracy in Caribbean countries – latest available survey.

Page 33


Sub-Saharan Africa – Fading Euphoria, Growing Challenges

Figure 30: Satisfaction with Democracy in Sub-Saharan Africa. Population-weighted rolling average for Nigeria,South Africa, Uganda, Tanzania, Senegal, Namibia, Malawi, Mozambique, Mali, Lesotho, Kenya, Ghana,Botswana, and Zambia.

Sub-Saharan Africa is among the last ofthe major world regions to experiencewidespread democratisation. Many countriesonly underwent free and fair elections in the1990s, following a post-independence tra-jectory which saw so-called “big men” dom-inate politics in single-party, clientelisticregimes.25

Africa’s various transitions to democracyhave been accompanied by a sense of pub-lic enthusiasm that is evident from the firstround of surveys conducted by Afrobarome-ter in the late 1990s. In every country firstsurveyed at that time – South Africa, Nigeria,Kenya, Namibia, Botswana, Zambia, Mozam-bique, Tanzania, Mali, and Ghana – a major-ity of the public expressed a positive view ofthe performance of democracy in their coun-try. Such enthusiasm is unsurprising: giventhe deeply corrupt, personalistic, and brutalsystems of rule which prevailed across Sub-Saharan Africa prior to democratisation.However, as the dream of democracy has

become a reality, it has been supersededby the need to address pressing social chal-lenges. Crime, urban poverty, and the persis-tence of corruption have eroded democraticlegitimacy, and now half of individuals in theregion express dissatisfaction with the stateof democracy in their countries.For the moment, however, democratic in-

stitutions across the region remain robust.Only in Mali has there been a major rever-sal, following a military coup in March 2012,with more limited signs of erosion in coun-tries such as Tanzania and Senegal, whereruling administrations have intervened tolimit political opposition. Meanwhile, thepublic’s sense of discontent appears to havestabilised around the global average, with nofurther growth in dissatisfaction over the lastdecade. With new governments promisingreform in South Africa and Nigeria, and ten-tative signs of opening in Angola, Ethiopia,and Zimbabwe, Sub-Saharan African democ-racies may yet rise to the challenge.

25 Michael Bratton & Nicolas van de Walle (2011) “Neopatrimonial Regimes and Political Transitions in Africa”,WorldPolitics.

Page 34


Country in Focus

Nigeria

Accounting for a fifth of the population ofSub-Saharan Africa, Nigeria is a bellwetherfor the region’s economic, political, and so-cial future. During the 1960s, as Africa fellundermilitary dictatorships, so didNigeria;as Africa experienced brutal civil conflict,the country did likewise. Split equally be-tween Christian and Muslims, Nigeria em-bodies Africa’s sectarian faultline. Sincethe 1990s, the country has also exemplifiedthe challenging and incomplete nature ofAfrican democratisation.Nigeria’s most recent pathway to democ-racy began in 1998, as the military dictator-ship of Sani Abacha gave way to free andfair elections under the country’s fourthrepublic. In 1999, free elections were heldfor the national assembly and president, re-sulting in the victory of general OlusegunObasanjo. This also marked a high pointfor civic optimism – the one and only timewhen an unambiguous majority of Nigeri-ans felt positively regarding the state ofdemocracy in their country.

Since then, however, Nigeria has had tostruggle with the realities of governing alarge, poor-albeit-resource-rich, and di-vided country. Following the transitionfrommilitary rule, freedomof the press andhuman rights havemarkedly improved, anddecentralisation has led to greater powerfor the country’s state governors. Yet com-munal violence has erupted recurrently,corruption has remained persistent, andurban poverty remains widespread, as thecountry struggles to provide services to itsgrowing city slums. Elections have alsobeen marred by fraud, and since the late2000s, a clear majority in Nigeria have ex-pressed their discontent with the function-ing in democracy in their country.With a rapidly growing population of over200 million, Nigeria is not only the world’sfifth-largest democracy, butwill potentiallybe the second -or third- largest within thenext generation. Its future represents notonly the future of Africa – but that of theworld as a whole.

Figure 31: Dissatisfaction with democracy in Nigeria from the late 1990s to the present day.

Page 35


Country in Focus

South Africa

Of all the democratic transitions that oc-curred from the 1970s to 1990s as part ofthe “Third Wave”, few raised as much hopeor excitement as the fall of apartheid inSouth Africa and its succession by a mul-tiracial, pluralist democracy.Barring the period following Nelson Man-dela’s election in 1994, for which no datais available, optimism regarding the healthof democracy in South Africa subsequentlypeaked in the mid-2000s, towards the endof Thabo Mbeki’s presidency. At that timethe South African economy was growingat an average rate of 4.5% per year, whilequota policies helped to form a black mid-dle class. Similar to the situation in Brazilunder President Lula da Silva, there wasa sense that South Africa had managed tosquare the circle – combining a freemarket,globally-integrated economy with socialwelfare and redistribution of income. Nordid the analogy end there, as in 2010 – fouryears prior to Brazil – South Africa hostedthe FIFA World Cup, showcasing its eco-nomic growth and social transformation tothe world at large.

In the last decade, however, such opti-mism has faded as South Africa has hadto reckon with problems of economic mis-management, crime, and clientenlism. Al-ready under Mbeki, electricity rationingand widespread blackouts in 2008 revealeda deeper legacy of underinvestment and en-ergy overallocation for political ends. Thenext year Jacob Zuma, who had previouslybeen charged for corruption, won the presi-dency as candidate for the African NationalCongress. Under his administration from2009 to 2018 a majority of South Africansturned negative in their evaluation of theirdemocracy for the first time.The Zuma years have been characterisedby economic stagnation, with per capitaincome declining for six successive yearsand unemployment levels rising. Yet be-hind this lies a more fundamental politi-cal decay – misuse of resources, indebted-ness in state-owned entities, rising pub-lic debt, and outflows of foreign capital –which Zuma’s successor, Cyril Ramaphosa,has vowed to reverse.

Figure 32: Dissatisfaction with democracy in South Africa from the late 1990s to the present day.

Page 36


Northeast Asia – Stability Amidst Change

Figure 33: Satisfaction with Democracy in Northeast Asia: average for Japan, South Korea, and the Republic ofChina (Taiwan).

The Northeast Asia region compromisesthose countries forming the boundaries ofthe East China Sea – Korea, Japan, and China,including the Republic of China (Taiwan).Though Japan has practised free and fair elec-tions since 1946, South Korea’s democratictrajectory was interrupted by a period of mili-tary rule from 1971 to 1987, while Taiwanwasunder single-party rule by the Kuomintanguntil the 1980s.Throughout most of the last twenty-five

years, a majority of citizens in NortheastAsia have expressed satisfaction with the waydemocracy works in their country, yet therehave been fluctuations over time. Follow-ing the East Asian financial crisis of the late1990s, dissatisfaction with democracy spiked,concurrent with a period of political turmoil,with the dominant Liberal Democratic Party(LDP) in Japan forming a series of weak andunstable coalitions, and Taiwan electing itsfirst non-Kuomintang government. Togetherwith political upset came both polarisation

and political scandal, including corruptionallegations against the new presidents ofTaiwan and South Korea – both of whomwere eventually prosecuted (though the lat-ter committed suicide before charges couldbe brought).However dissatisfaction has begun to rise

again in the past decade, principally in Japan,which following a period of relative content-ment under the administration of JunichiroKoizumi is now at an all-time high, with 55%of the public dissatisfied with democracy’sperformance. This has come in the wake ofcorruption scandals, dissatisfaction over thegovernment’s handling of nuclear and flood-ing disasters, and polarisation over constitu-tional reform. It has also come amidst a back-drop of steady economic and demographicdecline – a factor which augurs ill for neigh-bouring democracies in South Korea and Tai-wan, which in the coming decades must in-creasingly grapple with the same dilemmas.

Page 37


The Broader Middle East – Hope and Disappointment

Figure 34: Satisfaction with democracy among countries surveyed in the broader Middle East region. IncludesLebanon, Turkey, Israel, and Tunisia.

The Middle East has been one of the regionsmost resistant to democratisation, and gen-uinely competitive multiparty elections oc-cur in only a handful of countries. Two of thecountries with the longest tradition of polit-ical pluralism are both non-Arab societies –Turkey and Israel – with a historical footholdin Europe, while in Lebanon, the one countryin the Arab world with a history of politicalpluralism, elections since independence havebeen marred by recurrent violence and civilstrife.Though there is little data from the non-

democratic Arab Middle East prior to 2010,the data from the region that is available –shown in Figure 34 above – suggests risingdiscontent in the five years building up to

the Arab Spring, during which regimes fromTunisia and Libya to Egypt and Syria werechallenged by mass protest movements.The years following the Arab Spring were

concurrent with a wave of enthusiasm acrossthe region that is also evident in comparativesurvey data. Yet this brief window of opti-mism has now passed, and now exhaustionand cynicism are prevalent. When prompted,half of all Arabs agreewith the sentiment thatpeople in their country are “not yet ready fordemocracy”, with clear majorities endorsingthis view in Tunisia and Palestine: the twopredominantly Arab Muslim societies withthe most experience of the democratic pro-cess.26

26 Natalya Rahman (2018), “Democracy in the Middle East and North Africa: Five Years after the Arab Uprisings”, ArabBarometer, Wave IV Topic Report.

Page 38


Country in Focus

Tunisia

Of all the countries that made up the “ArabSpring” of 2010, only Tunisia experienceda transition to free and fair elections. InEgypt, elections were held but succeededby a military coup. In Libya, Syria, andYemen, peaceful protest was followed bydescent into brutal civil war. In Jordan, Mo-rocco, and Saudi Arabia, entrenchedmonar-chies were able to co-opt and defuse dis-sent. This makes the fate of democracy inTunisia symbolic for the region, if not for allstates undergoing democratisation in thesecond decade of the twenty-first century.Almost a decade after Tunisian fruit-sellerMohamed Bouazizi set himself on fire andwith him the entire Arab Middle East,Tunisia alone remains a multiparty democ-racy, with its most recent elections inSeptember 2019 seeing Kais Saied win thepresidency on a platformof anti-corruptionand political reform. Yet Tunisian democ-

racy is by no means consolidated. Corrup-tion is widespread, economic growth hasbeen stagnant, and a state of emergencyhas been in place since terror attacks in2015. In an Arab Barometer survey from2016, over half of Tunisians responded thatdemocracy is “ineffective in maintainingorder”. Three-quarters of the public alsostated that people in their country were“not yet ready for democracy,” and a thirdof the population now consider emigrating,including a majority of youth.Trends in satisfaction with democracy areconsistent with this sentiment: across 14surveys conducted in the past seven years,a clear majority of the public has expresseddissatisfaction with the way democracy isdeveloping in Tunisia on all but two ocass-sions, with that figure reaching 72% in themost recent 2018 survey conducted by thePew Research Center.

Figure 35: Dissatisfaction with democracy in Tunisia from 2013 to the present day. Following the transi-tion to democracy, discontent in Tunisia began to moderate. In one poll in 2015, a majorityreported satisfaction with the condition of democracy in the country.

Page 39


Southeast Asia – For Today, A Continued Faith in Tomorrow

Figure 36: Satisfaction with Democracy in Southeast Asia: average for Indonesia, the Philippines, Cambodia, andMalaysia.

The Southeast Asia region is home to 650million people, making it comparable in pop-ulation to Europe, Latin America, or Sub-Saharan Africa. It also contains some of themost extreme diversity in political institu-tions to be found anywhere on earth: rangingfrom newly emerging democracies in Indone-sia and the Philippines, episodic military rulein Thailand, a communist regime in Vietnam,and Singapore’s one-party-dominant tech-nocracy.If we consider only the emerging democra-

cies from this cluster – Indonesia, the Philip-pines, Malaysia and, until recently, Cambo-dia – we find a surprising degree of con-tentment and optimism with the conditionand performance of the region’s democraticinstitutions. Surprising, perhaps, becauseof the manifest challenges: In Indonesia,which transitioned to democracy followingSuharto’s resignation in 1998, political liber-alisation has been jolted by ethnic and sectar-

ian conflict, rampant low-level clientelism,andmore recent concerns over freedom of ex-pression, yet the broader social background isone in which the country has recovered firmlyfrom the East Asian financial crisis and ex-perienced two strong decades of economicgrowth – raising living standards for rich andpoor alike. A similar picture prevails in thePhilippines, where in spite of persistent cor-ruption and a ruthless drug war led by Pres-ident Rodrigo Duterte claiming thousandsof innocent lives, steady economic growth istransforming the lives of the next generation.As a result, there is optimism in Southeast

Asia regarding the performance of electedgovernments. In no other region of the worldis contentment with the functioning of thepolitical system presently as high, with morethan two-thirds of survey respondents in In-donesia, the Philippines, and Malaysia ex-pressing their satisfaction with the currentstate of democracy in their countries.27

27 See See Yun-han Chu, Michael Bratton, Mark Tessler, Marta Lagos and Sandeep Shastri (2008) “Public Opinion andDemocratic Legitimacy”, Journal of Democracy; and Richard Nadeau, Vincent Arel-Bundock and Jean-François Daoust(2019) “Satisfaction with Democracy and the American Dream”, The Journal of Politics, 81(3).

Page 40


South Asia – Democracy Continuing to Survive Against the Odds

Figure 37: Satisfaction with Democracy in South Asia: average for India, Pakistan, Bangladesh, and Sri Lanka.

South Asia’s experience of democratisationdiffers from other former colonies in Asia,Africa and the Middle East, as under the pres-sure of the Indian National Congress, elec-tions were organised during the final decadesof British colonial rule. Pakistan, Bangladesh,and India therefore each have a long experi-ence with electoral institutions, which in thelatter case has been unbroken barring twoyears of emergency rule in the 1970s.Satisfaction with democracy is also com-

paratively high across the region today, notonly in India but also in neighbouring coun-tries, where consistent majorities have ex-pressed satisfaction in all but two surveys(Bangladesh in 2013, amidst widespread un-rest, and Pakistan in 2001, preceding theMusharraf coup). The one exception to thispicture of contentment is Nepal, where ineach of three surveys since 2005, betweenhalf and three-quarters of respondents ex-press dissatisfaction with how their democ-racy is performing.

As for India, scholars have long asked whythe country’s experience of democracy hasbeen so stable relative to that of other for-mer colonies. These explanations range froma deferential political culture rooted in thecaste system,28 colonial-era socialisation ofdemocratic norms,29 and the dominance ofthe Congress Party – a dominance, however,that has recently come to an end.30

However, though majorities in India haveconsistently expressed satisfaction with thecountry’s institutions, the data is sparse andthere have been fluctuations over time. Ontwo ocassions – in 2013, a year before thelandslide election of the BJP under Naren-dra Modi, and several years later, followinga botched currency reform – dissatisfactionspiked at close to half of all respondents. Nordo we yet have data on public opinion duringthe current period of protests and instability,which may prove to be a fresh dip of confi-dence in India’s democratic process.

28 Barrington Moore (1966) The Social Origins of Dictatorship and Democracy.29 Steven Wilkinson & Massimiliano Onorato (2010) “Colonial Democratic Legacies”, APSA Annual Conference.30 Ashutosh Varshney (1998), “Why Democracy Survives”, Journal of Democracy.

Page 41


7. Conclusion: Why Are Citizens Unhappy with Democracy?

The results of our analysis suggest that dis-satisfaction with democracy among the cit-izens of developed countries has increasedfrom a third to half of all individuals, while intransitional democracies in Africa and LatinAmerica, a growing majority is also discon-tent with their system of governance. Whatcan explain this trend?One factor, undoubtedly, is the point from

which we have fallen. During the 1990s,western liberal democracy had emerged fromthe Cold War with its legitimacy renewed,while across Latin America, Sub-SaharanAfrica, and East Asia, it was believed that theadoption of multiparty competitive electionswould not only bring out greater civic partic-ipation and influence, but also help solve di-verse problems of corruption, state fragility,and the absence of infrastructure or welfareprovision.Today, a quarter of a century later, both

western legitimacy and the merit of democ-racy as a development panacea have comeunder strain. In the established democra-cies of the western world, the financial andeurozone crises, foreign policy failures, andthe rise of populism have eroded the per-ception that democratic institutions producegovernance that is balanced, far-sighted andeffective. Meanwhile, in new democraciesin Latin America, Africa and Asia, electedgovernments have struggled to overcome en-demic problems of corruption, criminality,and state fragility, disappointing the earlyhopes of the transition years. If anything, therise of populism may be less a cause, than asymptom of this weakening legitimacy: with-out it, it would have been unthinkable fora U.S. presidential candidate to denounceAmerican democracy as rigged and corrupt,or for the winning presidential candidate inLatin America’s largest democracy to openlyentertain nostalgia for military rule.At the same time, there are positive sto-

ries amongst our findings which must notbe lost. Countries such as Switzerland, Den-mark, Norway and Luxembourg are at all-time highs for contentment with their democ-racies, and may have lessons to offer regard-ing the role of electoral systems in enhanc-

ing democratic responsiveness and represen-tativeness. In addition the global scope ofour analysis has revealed significant regionaldifferences, with civic contentment muchhigher in Asia than in the established democ-racies of the West. Part of the democraticmalaise, perhaps, may simply be a westernmalaise.Finally, might it be that rising levels of dis-

satisfaction reflect higher civic standards, asbetter educated and informed citizens “raisethe bar” on their expectations regarding theprobity and integrity of public officials? It ispossible, though also, unlikely: were it true,we would expect high quality of governmentand critical citzenship to go hand in hand,when in reality the most politically satisfiedsocieties are those in which institutions aretransparent, responsive, and free of corrup-tion, while societies where satisfaction withdemocracy is at its lowest are uniformly char-acterised by political instability and conflict.Thatmeans that if satisfactionwith democ-

racy is now falling across many of the world’slargest mature and emerging democracies– including the United States, Brazil, theUnited Kingdom, and South Africa – it isnot because citizens’ expectations are exces-sive or unrealistic, but because democraticinstitutions are falling short of the outcomesthat matter most for their legitimacy, includ-ing probity in office, upholding the rule oflaw, responsiveness to public concerns, en-suring economic and financial security, andraising living standards for the larger ma-jority of society. Our analysis suggests thatcitizens are rational in their view of politi-cal institutions, updating their assessmentin response to what they observe. If confi-dence in democracy has been slipping, thenthemost likely explanation is that democrati-cally elected governments have not been seento succeed in addressing some of the majorchallenges of our era, including economic co-ordination in the eurozone, the managementof refugee flows, and providing a credible re-sponse to the threat of global climate change.The best means of restoring democratic legit-imacy would be for this to change.

Page 42


Methodological Annex

Page 43


Methodology I: Data Sources

Survey Sources

Public opinion on satisfaction with democracyis drawn from twenty-five different nationallyrepresentative survey sources, most of which wasformatted and merged by the HUMAN Surveysproject. We supplemented this with aggregatedmeasures of satisfaction with democracy fromthe most recent years of Pew Research’s GlobalAttitudes and Trends series, since the respondentsurvey data was not available to the public at thetime of publication. We additionally includedsurvey data from YouGov for recent years to getto most up-to-date picture of global satisfactionwith democracy.

The report draws upon the aggregated responsesof almost four million survey respondents fromover 150 countries between 1969 and 2019. Themerged data represents over 3500 country-surveyobservations. Many countries were surveyed mul-tiple times a year by different survey sources,providing greater reliability from repeated mea-surements. Data comes from the rounds, waves,years, or modules containing selected satisfac-tion with democracy variables that were listed inTable 1 at the start of this report.

Survey Items

The following questions and answers are usedon different surveys to measure satisfaction withdemocracy. There may be minor differences fromthe version asked on surveys, such as the layoutof the questions and direction or order of theanswers. HUMAN Surveys reordered answer val-ues for consistency to facilitate harmonizationof target variables. The formatted versions aredisplayed here, but all original responses weremaintained. All non-valid and unusable answerswere recoded into four standard missing valueswherever possible (do not know, refused, not ap-plicable, and missing), but these were all treatedas missing data when aggregating national scoresfor analysis.

AmericasBarometer, IntUne - Integrated andUnited:

“In general, would you say you are very satisfied, sat-isfied, unsatisfied or very unsatisfied with the way

democracy works in [Country]? 0 = very dissatisfied,1 = dissatisfied, 2 = satisfied, 3 = very satisfied”

Asian Barometer Survey, European ValuesStudy, Latinobarómetro, World Values Sur-vey, Comparative Study of Electoral Systems,Australian Election Study, Eurobarometer:Central and Eastern, Eurobarometer: Appli-cant and Candidate Countries, Eurobarome-ter: Standard and Special, Global Attitudesand Trends, American National ElectionStudies, EU Neighbourhood Barometer:

“On the whole, how satisfied or dissatisfied are youwith the way democracy works in [Country]. Are you...? 0 = not at all satisfied, 1 = not very satisfied, 2 =fairly satisfied, 3 = very satisfied”

Voice of the People Series, New EuropeBarometer:

“Please tellmewhether you agree or disagreewith thefollowing statement. Is that strongly or slightly? Ingeneral, I am satisfied with democracy. 0 = disagreestrongly, 1 = disagree slightly, 2 = agree slightly, 3 =agree strongly”

American National Election Studies:

“On the whole, are you satisfied, fairly satisfied, notvery satisfied, or not at all satisfied with the waydemocracy works in [Country]? 0 = not at all satis-fied, 1 = not very satisfied, 2 = fairly satisfied, 3 =satisfied”

Arab Transformations Project:

“How satisfied are you with the following: The waydemocracy is developing in our country? 0 = defi-nitely dissatisfied, 1 = rather dissatisfied, 2 = quitesatisfied, 3 = definitely satisfied”

Afrobarometer:

“Overall, how satisfied are you with the way democ-racy works in [Country]? 0 = [Country] is not ademocracy, 1 = not at all satisfied, 2 = not verysatisfied, 3 = fairly satisfied, 4 = very satisfied”

Asia Barometer:

Page 44


“Please tell me how satisfied or dissatisfied you arewith the following aspects of your life. The demo-cratic system. 0 = very dissatisfied, 1 = somewhatdissatisfied, 2 = neither satisfied nor dissatisfied, 3= somewhat satisfied, 4 = very satisfied”

Afrobarometer:

“Overall, how satisfied are you with the way democ-racy works in [Country]? 0 = [Country] is not ademocracy, 1 = very dissatisfied, 2 = somewhat dis-satisfied, 3 = neutral, 4 = somewhat satisfied, 5 =very satisfied”

Consolidation of Democracy in Central andEastern Europe:

“Are you completely satisfied or completely dissat-isfied with the way in which democracy is workingin [Country] today? 0 = completely dissatisfied, 9 =completely satisfied”

Eurobarometer: Standard and Special:

“On the whole, to what extent would you say you aresatisfied with the way democracy works in [Coun-try]? 0 = completely dissatisfied, 9 = completelysatisfied”

European Social Survey:

“And on the whole, how satisfied are you with theway democracy works in [Country]? 0 = extremelydissatisfied, 10 = extremely satisfied”

International Social Survey Programme, Aus-tralian Survey of Social Attitudes:

“On the whole, on a scale of 0 to 10 where 0 is verypoorly and 10 is very well: How well does democracywork in [Country] today? 0 = very poorly, 10 = verywell”

Eurobarometer: Standard and Special:

“Now I would like you to indicate on this scale towhat extent you are satisfied with your present situ-ation in the following respects: The way democracyis functioning in [Country]? 0 = completely dissatis-fied, 10 = completely satisfied”

World Values Survey Wave 7, European Val-ues Study:

“On a scale from 1 to 10 where “1” is “not satisfied atall” and “10” is “completely satisfied”, how satisfiedare you with how the political system is functioningin your country these days? 1 = not at all satisfied,10 = completely satisfied” (note: question followstwo previous items on democracy: the importance ofliving in a democracy, and how democratically therespondent feels the country is being governed)

Page 45


Methodology II: Aggregation Methodology

Before aggregating data from individual surveysources, we first recode responses to satisfactionwith democracy questions in to a binary classifi-cation: as either “satisfied” or “dissatisfied.” Thisallows us to state the percentage of respondentsin a given country in a given month who are sat-isfied with the performance of democracy in theircountry.The overwhelming majority of observations in

our dataset derive from survey indicators basedupon a symmetrical 4-point scale, asking respon-dents directly about their degree of satisfactionwith democracy in their country. As such, thesepose few dilemmas regarding possible differencesin meaning (absence of semantic equivalence).For other indicators we had to first test for se-mantic equivalence, which is dealt with in thenext section.After reducing indicators to a binary (satis-

fied/dissatisfied) classification, we then took themean average for each polling observation: thepercentage who are satisfied (or dissatisfied) withthe condition of democracy in their country, atthat exact point in time.The data presented in the individual country

charts in this report shows the outcomes for eachpoll within that country, at the point (month andyear) in which the survey was conducted.For regional charts, population-weighting by

country was used to generate a weighted average“as if” we had conducted a stratified random sur-vey sample in that region, sampling based on thepopulation of each constituent country unit.In addition, for regional charts we also ensured

a constant country sample in each year (or quar-ter, for quarterly annual charts) by “rolling for-ward” country observations in periods in whichthere was no new survey, thus using the “mostrecent” information for each country before ag-gregation. For the 1995 series used in Figures 3and 5, some countries entered the time seriesshortly after 1995 but not in 1995 itself: thesecases data were “rolled back” to 1995 to ensureconstant country representation, in this case withthe most recent observation being in the futurerather than in the past.

Lorem ipsum dolor sit amet, consectetuer adip-iscing elit. Ut purus elit, vestibulum ut, plac-erat ac, adipiscing vitae, felis. Curabitur dictumgravida mauris. Nam arcu libero, nonummy eget,consectetuer id, vulputate a, magna. Donec vehic-ula augue eu neque. Pellentesque habitant morbitristique senectus et netus et malesuada fames acturpis egestas. Mauris ut leo. Cras viverra metusrhoncus sem. Nulla et lectus vestibulum urnafringilla ultrices. Phasellus eu tellus sit amettortor gravida placerat. Integer sapien est, iac-ulis in, pretium quis, viverra ac, nunc. Praesenteget sem vel leo ultrices bibendum. Aenean fau-cibus. Morbi dolor nulla, malesuada eu, pulvinarat, mollis ac, nulla. Curabitur auctor semper nulla.Donec varius orci eget risus. Duis nibhmi, congueeu, accumsan eleifend, sagittis quis, diam. Duiseget orci sit amet orci dignissim rutrum.Nam dui ligula, fringilla a, euismod sodales,

sollicitudin vel, wisi. Morbi auctor lorem nonjusto. Nam lacus libero, pretium at, lobortis vi-tae, ultricies et, tellus. Donec aliquet, tortor sedaccumsan bibendum, erat ligula aliquet magna,vitae ornare odio metus a mi. Morbi ac orci et nislhendrerit mollis. Suspendisse ut massa. Cras necante. Pellentesque a nulla. Cum sociis natoquepenatibus etmagnis dis parturientmontes, nasce-tur ridiculus mus. Aliquam tincidunt urna. Nullaullamcorper vestibulum turpis. Pellentesque cur-sus luctus mauris.Nulla malesuada porttitor diam. Donec felis

erat, congue non, volutpat at, tincidunt tristique,libero. Vivamus viverra fermentum felis. Donecnonummy pellentesque ante. Phasellus adipisc-ing semper elit. Proin fermentummassa ac quam.Sed diam turpis, molestie vitae, placerat a, mo-lestie nec, leo. Maecenas lacinia. Nam ipsumligula, eleifend at, accumsan nec, suscipit a, ip-sum. Morbi blandit ligula feugiat magna. Nunceleifend consequat lorem. Sed lacinia nulla vitaeenim. Pellentesque tincidunt purus vel magna.Integer non enim. Praesent euismod nunc eu pu-rus. Donec bibendum quam in tellus. Nullam cur-sus pulvinar lectus. Donec et mi. Nam vulputatemetus eu enim. Vestibulum pellentesque felis eumassa.

Page 46


Methodology III: Testing Semantic Equivalence for Non-Standard Sat-isfaction Items

The vast majority of polling observations in ourdataset derive from 4-point scale questions re-garding satisfaction with democracy, wherebytwo points indicate some degree of satisfaction(e.g. “fairly satisfied” and “very satisfied”) andtwo points indicate some degree of dissatisfac-tion (e.g. “not very satisfied” and “not satisfiedat all”). Because they are ordered scales withfour items around a midpoint – using the samesubstantive question keywords (democracy andsatisfaction) – few concerns arise regarding thesemantic equivalence of the resultant averages.

However, the same cannot be said for a rangeof additional satisfaction with democracy surveyitems that depart from a 4-step scale, for exampleusing a 3-step scale with a single middle category,or asking respondents to rate their degree of satis-faction on a 10-point scale. Which of these can berecoded to a “semantically equivalent” satisfiedvs. dissatisfied dichotomy – and if so, how shouldthe items be recoded such that they appropriatelymatch the result that a 4-step scale would haveattained?

Fortunately, because so many surveys are con-ducted at the same time as other surveys withineach country, we have a simple means of check-ing for semantic equivalence: to examine thecommon sample of country-year observations foreach indicator, and see which survey recodings,if any, correlate sufficiently with our baseline4-item response scale results.

The Results

We present in this section the results of the se-mantic equivalence tests for the items that wereultimately included in our final dataset. They in-clude scatter plots of country-year observationsusing only the 4-point scales against country-year results of a range of recoding possibilitiesfor our non-standard items, together with a 45-degree line – representing what we should expectto see if there is perfect equivalence. A weightedregression line of fit through the actual sharedobservations is also shown for comparison.

While we were able to find semantic equivalentsfor the large majority of survey items, severalmeasures failed our tests and were eventually ex-cluded from the dataset.

1. All 3-item survey questions were ex-cluded from the dataset. In these cases,no possible recoding produced an unbiasedrange of values equivalent to those of the4-item scale. This is most likely becausethe language of the middle value was notneutral, but leaned towards or away fromsatisfaction (e.g. “somewhat satisfied”).

2. Several items with insufficiently equiv-alent question wording were excluded.For example, items on satisfaction with gov-ernment, which did not specifically men-tion democracy, failed tests of semanticequivalence, as did a survey question on“pride” in one’s democracy.

Page 47


The 11-Point Satisfaction With Democracy Scale (A)

“And on the whole, how satisfied are you with the way democracy works in [Country]? 0 = extremelydissatisfied, 10 = extremely satisfied”

This is the 11-point satisfaction with democ-racy scale is used in the European Social Survey(ESS) dataset. Instead of respondents being askedwhether they are satisfied or dissatisfied with thefunctioning of democracy in their country, theywere asked to rate their satisfaction on an 11-point scale from 0 to 10 inclusive.To find a semantically equivalent recoding, we

test five alternative recodings of the satisfactionscale, starting with the intuitive split of 5 and be-low as dissatisfied, and 6 and above as satisfied.This is found to be negatively biased; yet recod-ing 5 as a midpoint (N/A) value produces a closeapproximation.

(1) Semantic equivalence test for 11-point satisfactionwith democracy scale and other items, with 5 as N/Amidpoint (0-4 dissatisfied, 6-10 satisfied).

(2) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 3-5as N/A midpoints (0-2 dissatisfied, 6-10 satisfied).

(3) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with nomidpoint: i.e. 0-5 as dissatisfied, 6-10, satisfied.

(4) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 4 asmidpoint (0-3 as dissatisfied, 5-10 satisfied).

(5) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 4-5as midpoints (0-3 as dissatisfied, 6-10, satisfied).

Page 48


The 11-point Satisfaction with Democracy Scale (B)

“On the whole, on a scale of 0 to 10 where 0 is very poorly and 10 is very well: How well does democracywork in [this country] today? 0 = very poorly, 10 = very well”

The 11-point scale used by the International So-cial Survey Programme (ISSP) asks respondentsabout how well democracy is performing. Wefind that the shift in question formulation – fromone’s own personal feeling of satisfaction to anobjective assessment of democratic performance– makes respondents more positive, on average:recoding the mid-point (5) to N/A in this instancebiases results upwards. The equivalent recodingfor this item is to exclude the midpoint, and re-code 0-5 as dissatisfied, and 6-10 as satisfied.

(1) Semantic equivalence test for 11-point satisfactionwith democracy scale and other items, with no midpoint:i.e. 0-5 as dissatisfied, 6-10, satisfied.

(2) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 3-5as N/A midpoints (0-2 dissatisfied, 6-10 satisfied).

(3) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 5 asN/A midpoint (0-4 dissatisfied, 6-10 satisfied).


(5) Semantic equivalence test for 11-point satisfac-tion with democracy scale and other items, with 4-5as midpoints (0-3 as dissatisfied, 6-10, satisfied).

Page 49


The 11-point Satisfaction with Democracy Scale (C)

“Now I would like you to indicate on this scale to what extent you are satisfied with your present situationin the following respects: The way democracy is functioning in [Country]? 0 = completely dissatisfied, 10 =completely satisfied”

This 11-point satisfaction with democracy scalewas featured in the early Eurobarometer surveysof the 1970s, for a limited number of countriesin Western Europe. In common with the later 11-point scale used by the European Social Survey(ESS) – with which it shares a common phraseol-ogy – we find that a recoding of the middle value(5) to N/A is the most equivalent in its results tothe 4-point scale used in later surveys. This im-plies, again, that respondents use the 5 value as aneutral or non-response, and express satisfactionor dissatisfaction with democracy at values aboveor below this point, respectively.

(1) Semantic equivalence test for 11-point (0-10) satis-faction with democracy scale and other items, with 0-4as dissatisfied and 6-10 satisfied.

(2) Semantic equivalence test for 11-point (0-10) sat-isfaction with democracy scale and other items, with0-3 as dissatisfied and 5-10 satisfied.


Page 50


5-Point and 6-Point Satisfaction With Democracy Scales

5-point scale“Please tell me how satisfied or dissatisfied you are with [...] The democratic system. 0 = very dissatisfied, 1= somewhat dissatisfied, 2 = neither satisfied nor dissatisfied, 3 = somewhat satisfied, 4 = very satisfied”

6-point scale“Overall, how satisfied are you with the way democracy works in [this country]? 0 = [Country] is not ademocracy, 1 = very dissatisfied, 2 = somewhat dissatisfied, 3 = neutral, 4 = somewhat satisfied, 5 = verysatisfied”

Whereas the conventional 4-point satisfactionwith democracy scale (used since the 1970s by Eu-robarometer) offers a clean recoding into “satis-fied” and “dissatisfied” respondents, the 5-pointsatisfaction with democracy scale, used only byAsia Barometer from 2003-7, leaves a dilemma ofhow to code the middle value. A neutral responseto a satisfaction prompt could be coded as notbeing satisfied, or could be coded as equivalentto N/A (undecided).To find a semantically equivalent recoding, we

test both of these. Perhaps unsurprisingly, therecoding of the middle value to not satisfied in-troduces a negative bias. However, recoding themidpoint to N/A produces a close approximation

to a satisfied/dissatisfied recoding of the 4-itemscale.A further variant is a 6-point satisfaction with

democracy scale that also offers an additional“negative” response: to say that the country is“not a democracy”. Such an item was fielded onlyby Afrobarometer during their first wave. Thisintroduces a semantic dilemma, as it could eithertreated as the most negative possible responseon an ordinal scale, or as equivalent to a non-response – a refusal to answer the question.In practice, however, only a very small number

(< 2%) of interviewees offer this response whenasked, such that it is best coded as N/A.


(2) Semantic equivalence test for 5-point (0-4) satis-faction with democracy scale and other items, with0-2 as dissatisfied and 3-4 satisfied.

Page 51


The 10-point Satisfaction with Democracy Scale (A)

“Are you completely satisfied or completely dissatisfied with the way in which democracy is working in [thiscountry] today? 0 = completely dissatisfied, 9 = completely satisfied”

“On the whole, to what extent would you say you are satisfied with the way democracy works in [Country]?0 = completely dissatisfied, 9 = completely satisfied”

The 10-point satisfaction with democracy scaleasks respondents to rate their satisfaction on anumerical scale from 0 to 9 inclusive, and appears(in slightly different forms) in both the Consoli-dation of Democracy in Central and Eastern Eu-rope (CDCEE) surveys, fielded in 1990-2 and 1998-2001, and in a single round of the Eurobarometersurveys in 1988.While an intuitive approach might be to split

the first 5 and last 5 points – that is, recoding to0-4 dissatisfied, and 5-9 satisfied – we found thatthis resulted in estimates that were heavily biasedto lower reported satisfaction. We suspect thismay be due to a tendency for respondents to haveused “4” as a neutral category; after consideringa range of recodings, we found the most equiv-alent results when either omitting the responsecategory of 4, or recoding it as positive.

(1) Semantic equivalence test for 10-point (0-9) satisfac-tion with democracy scale and other items, with 0-3 asdissatisfied and 5-9 satisfied.



Page 52


The 10-point Satisfaction with Democracy Scale (B)

“On a scale from 1 to 10 where “1” is “not satisfied at all” and “10” is “completely satisfied”, how satisfiedare you with how the political system is functioning in your country these days? 1 = not at all satisfied, 10= completely satisfied” (this question follows two previous items on democracy: the importance ofliving in a democracy, and how democratically the respondent feels the country is being governed)

This 10-point satisfaction scale is used in thecurrent round of the European Values Study andthe World Values Survey, fielded from 2017 to2019. Due to the large number of non-democraticregimes in the World Values Survey sample, theitem refers to “the political system” rather thanto the “democratic system”, as in earlier surveys.However it follows two preceding questions thatask directly about the condition of democracy inone’s country, setting a contextual frame for anevaluation of democratic performance.In order to check whether the question formu-

lation has affected its interpretation in a way

that deviates significantly from other satisfac-tion with democracy items, we check the satisfa-tion/dissatisfaction coding of this item againstthe results of standard satisfaction with democ-racy items in the same country-years. In this case,a simple recoding “down the middle” of 1-5 (dis-satisfied) and 6-10 (satisfied) provides the highestcorrespondence with standard satisfaction withdemocracy items. We suspect that using a 10-point scale starting from 1 (rather than from 0)proved less confusing for respondents, who couldmore easily interpret themidpoint range betweensatisfaction and dissatisfaction.

(1) Semantic equivalence test for 10-point (1-10) sat-isfaction with democracy scale and other items, on asimple split coding (1-5 dissatisfied, 6-10 satisfied).

(2) Semantic equivalence test for 10-point (1-10) sat-isfaction with democracy scale and other items, with5 recoded to NA (1-4 dissatisfied, 6-10 satisfied).

Page 53


Methodology IV: Sensitivity Analysis

To test that our results are robust to alternativedecisions regarding inclusion or exclusion of dif-ferent satisfaction with democracy question for-mulations, we conducted a series of alternativeaggregations: excluding eachmeasure one by onefrom the dataset and checking to see how thischanges the resultant values, and checking theplot of all country-year observations with andwithout the source question. If a measure wasfound to have a disproportionate or biasing effectupon index scores, it was excluded.31

As indicator selection for inclusion within thefinal dataset had to first pass a semantic equiv-alence test, the sensitivity analysis reveals thatthe inclusion or exclusion of individual surveysources has only a marginal effect on the resul-tant satisfaction with democracy averages.

Sensitivity to exclusion of 11-point satisfaction withdemocracy scale. “Now Iwould like you to indicate on thisscale to what extent you are satisfied with your presentsituation in the following respects: The way democracy isfunctioning in [this country]? 0 = completely dissatisfied,10 = completely satisfied.” 5 treated as N/A midpoint.

Sensitivity to exclusion of 11-point satisfaction withdemocracy scale. “On the whole, how satisfied areyou with the way democracy works in [Country]? 0 =extremely dissatisfied, 10 = extremely satisfied.” Overa large number of observations, there are no outliersand a high correlation.

Sensitivity to exclusion of 5-item satisfaction withdemocracy scale, with mid-point values recoded toN/A. “Please tell me how satisfied or dissatisfied youare with [the democratic system]. 0 = very dissatis-fied, 1 = somewhat dissatisfied, 2 = neither satisfiednor dissatisfied, 3 = somewhat satisfied, 4 = very sat-isfied.” No outliers.

31 Michaela Saisana, Andrea Saltelli and Stefano Tarantola, (2005) “Uncertainty and Sensitivity Analysis Techniques asTools for the Quality Assessment of Composite Indicators”, Journal of the Royal Statistical Society, 168, 307–323.

Page 54


Sensitivity Analysis Plots (Continued)

Sensitivity to exclusion of 6-point satisfaction ques-tion: “Overall, how satisfied are you with the waydemocracy works in [Country]? 0 = [Country] is nota democracy, 1 = very dissatisfied, 2 = somewhat dis-satisfied, 3 = neutral, 4 = somewhat satisfied, 5 = verysatisfied.” Midpoint and “not a democracy” recodedN/A. Few cases but no outliers.

Sensitivity to exclusion of 10-point (0-9) satisfactionwith democracy scale. “Are you completely satis-fied or completely dissatisfied with the way in whichdemocracy is working in [Country] today? 0 = com-pletely dissatisfied, 9 = completely satisfied.” With“4” recoded as N/A midpoint. Minor discrepanciesexist but no substantial outliers.

Sensitivity to exclusion of 10-point satisfaction withdemocracy scale. “On the whole, to what extentwould you say you are satisfied with the way democ-racy works in [Country]? 0 = completely dissatisfied,9 = completely satisfied.” 4/5 treated as midpointN/A values. No outliers.

Sensitivity to exclusion of 11-point satisfaction withdemocracy scale. “On the whole, on a scale of 0 to10 where 0 is very poorly and 10 is very well: Howwell does democracy work in [Country] today? 0 =very poorly, 10 = very well.” (1-5, dissatisfied, 6-10, satisfied). Many observations, no substantialoutliers.

Page 55


Methodology V: Detrending Satisfaction with Democracy from the Eco-nomic Cycle

In two of the insets in this report – on divergencewithin the eurozone, and the end of the “pinktide” in Latin America – we present cylically de-trended estimates of democratic satisfaction, inan attempt to remove the influence of short-termeconomic trends upon trends in satisfaction withdemocracy. This is clearly indicated within eachinset, and these are the only instances in the re-port where detrended (rather than raw) data arepresented.Why generate a cyclically-detrended satisfac-

tion with democracy series? The underlying as-sumption is that in times of economic prosper-ity, citizens’ satisfaction with the political systemwill be higher, whereas in times of recession itdeclines. This has important consequences forour understanding of democratic dissatisfactionfollowing financial crises, and in particular theGreat Recession. Insofar as we are interested inlong-term or structural change, we may wish tostrip out that portion of variation that is due totemporary economic factors, in order to focus onthe underlying shift, if any, that remains.The detrended satisfaction series produce a

number of important results. First, there is lit-tle evidence that the decline in satisfaction withdemocracy in the West, and in particular theUnited States, is due to cyclical economic factors.There is at best weak correlation between thedecline in satisfaction and periods of economicgrowth or recession. Other economic factors mayplay a role – but not in a way that is clearly linkedto the economic cycle.Second, there is however some evidence of a

cyclical pattern in emerging democracies, in par-ticular in Latin America. After detrending foreconomic growth effects, we have seen that lev-els of dissatisfaction in Brazil, Chile, and evenVenezuela are at a similar level today to their pre-vious lows of the 1990s. This also has the implica-tion that the currently high levels of democraticsatisfaction in Southeast Asia may be dependentupon continued economic growth in this region.

Detrending the Data

We detrend the data by estimating a simpleregression that controls for each country’saverage satisfaction (the country “fixed” ef-fect), and then the effect on satisfaction ofthe current rate of economic growth, plus theper capita economic growth rate in each ofthe six years preceding the survey.

Table 2: Estimated coefficients for association be-tween satisfaction with democracy and re-cent economic growth.

Estimated Effect of Per CapitaGDP Growth (on % Satisfied) in:

Current Year +0.5***Previous Year +0.4***2 Years Prior +0.4***3 Years Prior +0.2*4 Years Prior +0.3**5 Years Prior +0.026 Years Prior +0.02

*** significant at p < 0.001;** significant at p < 0.01;

* significant at p < 0.1.

What we find is rather as one would expect:the strongest effect is growth in the year ofthe survey, with a steadily weaker effect foreach of the preceding years. Thus if a countryhas an economic shock of -6%, this reducessatisfaction by -3% pts: but if a country ex-periences five years of -6% growth the cumu-lative effect is around -11% pts. To producea detrended series around the country mean,we first adjust the satisfaction with democ-racy data to remove the growth effect, be-fore “adding back” effects using the averagegrowth rate of that country during the periodof observation.

Page 56

www.bennettinstitute.cam.ac.uk

Department of Politics and International Studies

Alison Richard Building

7 West Road, Cambridge, CB3 9DT

[email protected]

@BennettInst

Documents

REPORT Global Satisfaction with Democracy · Global Satisfaction with Democracy Report 2020 4. The Dataset Figure 1: Countries for which data on satisfaction with democracy exists