Whisper sweet nothings to me Erdoğan: developmentalist propaganda, partisan emotions, and economic evaluations in Turkey

ABSTRACT Economic explanations are central to our understanding of popular support for authoritarian regimes. Yet, we do not know to what extent economic evaluations of citizens living under these regimes are amenable to regime propaganda. This article engages with this question by relying on the case of Turkey, where the authoritarian regime led by Erdoğan has been able to sustain its popular support despite years of economic decline. A national developmentalist narrative has been central to the regime’s economic propaganda in Turkey since 2011. Relying on national face-to-face survey data, I first demonstrate that economic misperceptions grounded in this narrative are widespread among supporters of the ruling coalition in Turkey. I then use an online survey experiment to show that exposure to the developmentalist narrative improves economic evaluations among ruling coalition voters. These effects are mediated through the increase in partisan emotions, and they are especially large for non-partisan voters of the ruling coalition. These results help us understand how Erdoğan’s regime could sustain its popular support despite years of economic decline. From a broader perspective, this article demonstrates that scholars of authoritarian regimes need to pay more attention to economic narratives and their affective structures.


Introduction
Authoritarian regimes have an advantage when dealing with public opinion: unparalleled control over the information space. 1 Researchers have demonstrated that enthusiasm and hope among regime supporters.As such, this article draws attention to the role of positive partisan emotions under authoritarianism, a topic that has been largely ignored in the literature. 5he rest of the article is formed of six sections.The following two sections are mainly theoretical and qualitative.First, I introduce national developmentalism as a strategic narrative and propaganda and discuss why Turkey is an appropriate case to study mass support under authoritarianism.Second, I theorize how the NDN in Turkey affects voters' economic and emotional evaluations.Two quantitative sections follow, in which I test arguments laid out in theory sections through a national sample and an online survey experiment.I conclude with the discussion and conclusion sections.

Narratives in politics and national developmentalism
From a political perspective, we can analyse narratives along three dimensions: temporal, collective, and symbolic. 6The temporal dimension is central to the definition of narratives: narratives are sequential stories, combining specific interpretations of past and current events with a vision of the future.Second, narratives include references to actors and groups.By focusing on this collective dimension, we can explain how narratives contribute to forming group identities.Finally, we should also pay attention to how narratives are inscribed and embodied in the material world through the reorganization of the environment and human practices.Capturing this symbolic dimension is essential to understand how the power of narratives is reproduced through daily practices.
Political actors can use narratives strategically as "a communicative tool to attempt to give determined meaning to past, present, and future to achieve political objectives." 7Strategic narratives are important in both democratic and authoritarian countries. 8However, it can be argued that authoritarian control over the information space and the economy creates even more room for authoritarian regimes to instrumentalise narratives for political goals.Authoritarian regimes have vast resources to propagate narratives, not only through mass media but also through the re-organization of the material space.Procedures and institutions that slow down this process, such as parliamentary and judicial oversight or independent media, do not exist under authoritarianism.Furthermore, opposition and civil society actors have a limited capacity to challenge these narratives.Hence most citizens do not have access to critical information to evaluate them.
Economic narratives play an important role in politics, mediating the relationship between objective economic conditions and political preferences. 9This article studies the developmentalist economic narratives which are especially popular in the Global South.These narratives imagine the national economy as a coherent object that competes with other national economies.According to the developmentalist perspective, fast economic development of the country and a "great leap" into the league of advanced countries is possible.As such, this should be the ultimate national goal as it can solve all country problems and bring joy and harmony to the entire nation.
Political scientists have predominantly studied developmentalism in regard to associated policy ideas and their outcomes. 10However, the developmentalist conceptualization has a utopian appeal that exists independently from specific policy ideas. 11Political leaders with various ideologies can attempt to instrumentalise this utopian core to build mass support for their political agendas.For example, it is possible to find historical examples of the use of developmentalist narratives under both communist regimes, such as the USSR under Stalin's rule 12 and China under Mao's rule 13 and democratic regimes, such as Brazil under Kubitschek's rule. 14xamples of the strategic use of developmentalist narratives by contemporary authoritarian regimes are not hard to find.Paget argues that Magafuli's regime in Tanzania used the imperative of development to justify the authoritarian rule in the country. 15According to Sinha, developmentalism is part of Modi's appeal in India, as it is grounded in his image as a "development man" and his grandiose promises, such as constructing the world's largest statue, bringing in the fastest train, making India like Singapore, and building 200 smart cities. 16 Similarly, Huda argues that contemporary authoritarian regimes use energy megaprojects to perpetuate national and transnational narratives, foster patriotism, and gain legitimacy. 17trategic developmentalist narratives share some common properties.Regarding the temporal dimension, all developmentalist narratives focus on the utopian promise of development that will take place in the future.The collective dimension of all developmentalist narratives is built on the division between the nation and advanced nations.Finally, regarding the symbolic dimension, all developmentalist narratives produce and celebrate material symbols that represent the realization of developmentalist promises.Political actors will also add new elements to their developmentalist narratives in line with their political agendas and ideologies.For example, the past can be depicted as a state of deprivation, as was the case with Stalin's communist developmentalism, 18 or a state of glory, as was the case with Magafuli's restorationist developmental nationalism. 19The competition at the international level can be depicted as benign or existential.The nation, i.e. the "us" of the narrative, can refer to the entire nation or only some parts of it.Finally, the symbolic dimension of developmentalist narratives can refer to industrialization and factories, construction and infrastructural projects, or consumer goods.

The Turkish case: Erdoğan's national developmentalism
Under the leadership of Recep Tayyip Erdoğan, the Justice and Development Party (AKP) came to power in Turkey in 2002 through a free and fair election.Since then, the country has gone through a period of autocratization and personalization, following a trajectory similar to countries such as Russia and Venezuela. 20Today's Turkey is a typical case of electoral authoritarian regimes, which combine some competitive electoral properties with the de facto authoritarian control of the government over the political and social space. 21Given the strength of popular support of Erdoğan's regime in the face of a deep economic crisis, as discussed in the Introduction, Turkey is a likely case to document the relationship between authoritarian propaganda and performance evaluations. 22he power of Erdoğanist propaganda is usually explained with reference to populism and nationalism. 23While these are important to understand the regime's policies and popular appeal, they should not overshadow the significance of performance evaluations for the survival of the authoritarian regime in Turkey.Academic researchers repeatedly find that economic evaluations are significant predictors of support for the AKP, even when accounting for partisanship, socio-demographic factors, and ideological self-positioning. 24Importantly, however, the gap between economic perceptions of government voters and opposition voters is increasingly growing, 25 and media preferences are predictive of economic assessments. 26These findings suggest that, as Turkey's economic crisis deepens, Erdoğan benefits from the authoritarian media space to preserve his image of competency.The AKP has become increasingly more dominant in the media since it came to power, and the media coverage of the party has become increasingly more positive. 27Currently, none of the popular TV channels in Turkey can openly criticize the regime in its news programmes; the only exception is Fox TV, owned by Fox Networks Group.However, media domination is not enough to explain the success of authoritarian propaganda.It is also necessary to pay attention to the content of the message and how its audience receives it.
Erdoğan's economic propaganda is built on a national developmentalist narrative, which gradually emerged throughout the first decade of the AKP's rule in Turkey.To begin with, the promise of historical economic development is at the core of the NDN.The utopian articulation of "development" in the AKP discourse has crystallised with the 2011 election campaign. 28"Turkey is ready, Target: 2023" was the official slogan of this campaign and making Turkey one of the top 10 economies by 2023 was the most significant promise of the campaign period. 292023 is a carefully selected date with high symbolic and aesthetic significance; it is the centenary of the foundation of the Turkish Republic.The goal itself, on the other hand, is a grandiose one.According to World Bank data, Turkey was the eighteenth biggest economy in the world by 2011, and India, the world's tenth biggest economy by that time, had an economy that was more than twice the size of Turkey's economy.While the goal itself was unrealistic, the appeal was grounded in grandiosity.The promise of historic economic growth, along with references to 2023, has become a central part of the AKP's propaganda since then.Most tellingly, all campaigns for general elections after 2011 used similar slogans, such as "National Will, National Power, Target 2023," "New Turkey, New Power, Target 2023," and "It is Turkey's Time." 30 The NDN includes references to political actors and groups, which turn it into a collective partisan narrative.According to the NDN, the sole political will behind the 2023 vision is Erdoğan; he is the one who first dreamt of it, and he is the one who has pushed for it since then.Thus, the NDN is important in building Erdoğan's image as a competent and visionary leader.On the other hand, references to political actors that work against Turkey's development have also gradually become a central part of the narrative.These references were added when the regime started to be faced with economic and political setbacks.In those moments, the regime's propaganda machine reversed the developmentalist frame to explain why such a successful government would face problems and discontent.For example, "foreign conspirers that are unhappy about Turkey's economic development" was presented as the reason behind the Gezi protests, both by Erdoğan and the other regime actors. 31ublic surveys conducted during that period show that this narrative was successful; around 80% of AKP supporters believed that foreign conspirers planned Gezi protests. 32At other times, foreign states were ridiculed.For example, to exaggerate the significance of the new airport, which was another symbol of the NDN, the regime media frequently claimed that Germany was jealous and dismayed about this development. 33inally, the symbolic dimension of the NDN relies on a set of grandiose construction projects, such as two suspension bridges in Istanbul, a new airport in Istanbul, an artificial water leeway between the Marmara Sea and the Black Sea, as well as two new "cities" in Istanbul.These projects were first introduced as election promises in 2011, being the backbone of the AKP's newly announced 2023 vision.They helped voters imagine what a top-ten Turkey would look like, hence turning the abstract goal of "Target: 2023" into a concrete representation.Their function as symbols of the developmentalist promise still continues.Several of these projects have since been completed and were opened to service.The AKP's propaganda machine vividly celebrated these openings, arguing that they heralded that the 2023 vision was becoming a reality. 34o be clear, it is impossible to deny the economic significance of these megaprojects, especially in terms of the political economy of the authoritarian regime in Turkey. 35or example, the new Istanbul airport, which opened in 2019, has already become the busiest airport in entire Europe in terms of total passenger traffic.However, it is also crucial to note how the authoritarian regime in Turkey has always presented these megaprojects to the Turkish public within a narrative framework.
For example, one of the symbols of Erdoğan's national developmentalist narrative (NDN) is the Osmangazi Bridgea suspension bridge built over the Gulf of Izmit.The bridge was one of the AKP's campaign promises in the 2011 election period, and it was later named after Osman Gazi, the founder of the Ottoman Empire.The bridge's construction started in 2013, and it was ceremoniously opened to service on 30 June 2016.Pro-regime TV channels made live broadcasts from the bridge throughout the week, dubbing 30 June "a historic day" and "the day of pride."The speech that Erdoğan gave during the opening ceremony includes references to all dimensions of the NDN: We will continue to be the biggest friend of those who sees us as a friend.Similarly, we won't hesitate to take any measure against those who nurture enmity toward us.No one can stand in our way in the new era.No one will be able to stop us from achieving our goals for 2023. 36

How do developmentalist narratives affect regime support: the theory
How does the NDN affect voters' economic evaluations in Turkey?This section aims to develop a theoretical explanation for this question, drawing on the literature in public opinion, political psychology, and cultural studies.

National developmentalist narrative and emotions
Emotions are central to various forms of political behaviour, such as the formation of collective identities, political mobilization, and motivated reasoning. 37There is a very close relationship between narratives and emotions.By simplifying and dramatizing the complex reality surrounding us, narratives can play an important role in generating emotional reactions.Emotions, in return, can explain why some narratives are more powerful and attractive than others to the audience. 38 argue that the NDN cultivates affective ties between the ruling coalition voters and Erdoğan thanks to the powerful vision of the future that it provides.According to the appraisal theory of emotions, goal-orientation underlies all emotional reactions; positive emotions are derived from the cognitive appraisal of approaching a desirable goal, whereas negative emotions are associated with the appraisal of obstacles. 39In parallel with this insight, studies of charismatic leadership list "offering a powerful future vision" as one of the key elements of charismatic bonds between the leader and the follower, as well as being central to the elicitation of positive emotions from the followers. 40Similarly, McLaughlin et al. demonstrate that political ads that have a narrative format evoke emotional reactions from voters by stimulating them to imagine future scenarios. 41The NDN offers an appealing future vision for government voters, promising welfare, prosperity, and status.The goal of being one of the leader countries in the world may be grandiose, but it is the grandiosity of promises that arouses the audience, so long as these promises are made by "credible" leaders. 42he NDN also draws on existing identities to strengthen its affective appeal.For voters with strong national identification, the NDN offers a vision in which Turkey has a higher status in the international arena.International status hierarchies are closely associated with emotional reactions, and they are especially central to Turkish collective identity. 43For voters of the AKP, the NDN offers the pride of being a part of the political party that has achieved improving Turkey's international status.

National developmentalist narrative and economic evaluations
I argue that the NDN improves economic evaluations both by offering cognitive shortcuts to evaluate the performance of the economy and through its emotional effects.
To begin with, the NDN provides a framework to evaluate the economic performance of the regime, and this framework heavily favours the regime.A large amount of literature in political science suggests that people use national economic indicators more than personal economic indicators to judge the incumbent government's economic performance. 44We know less about how people decide on something as complex as the "national economic situation." 45Recent research shows that international comparisons, such as cross-national performance comparisons and references to foreign socioeconomic conditions, can influence how people judge domestic performance. 46he NDN, by definition, draws attention to international comparisons of the economy.Crucially, the narrative also provides "benchmarks" that can be used to judge Turkey's comparative economic performance.These are the symbols of the NDN: bridges, airports, fast train lines etc.These projects are frequently described in comparison to their counterparts in the Western world.For example, when Istanbul airport was completed, the regime media compared this with Frankfurt Airport, one of the largest airports in the world.The comparison underlay the claim that Germans were envious of Turks and the new Istanbul airport. 47Given that building these infrastructural projects is easier than creating millions of new jobs, the NDN provided a cheaper way to construct the image of competency in the eyes of Turkish people who regarded the regime media and Erdoğan as credible and who engaged with the narrative.
Second, it is also important to establish the link between the narrative structure of the NDN, which generates emotional reactions, and the NDN's impact on economic evaluations.According to the affective intelligence theory, individuals who are angry or enthusiastic are more likely to rely on their existing dispositions during the reasoning process. 48In politics, this means that these individuals will rely more on the source of the message rather than its content when they process a message, and thus, they will stick more closely to their partisan and ideological commitments. 49If the NDN is evoking partisan emotions among certain individuals, as argued above, we would expect these people to bring their economic evaluations in line with their partisan preferences and adopt a motivated form of reasoning.

Study 1: national survey data
I have so far argued that the national developmentalist narrative (NDN) has been the backbone of the Erdoğan regime's economic propaganda and has successfully affected Turkish voters' economic and affective evaluations.This section provides an initial test of these arguments with a national and face-to-face survey.

Introducing the data
The data used in this section comes from a survey conducted by Konda Survey Company, one of Turkey's most reputable survey companies, in October 2018.The survey was conducted face-to-face with 2676 respondents from 148 districts, randomly chosen from a total of 49000 districts through a multi-stage, stratified cluster sampling design. 50Age and sex quotas were implemented at the district level.The survey took place four months after the 2018 general election, in which Erdoğan was re-elected as president in the first round with a vote share of 53%.That election year was a challenging year for the Turkish economy.The Turkish lira depreciated 72% against the dollar between January and October 2018, and most of the decline happened after the election.Steinberg shows that the currency crisis reduced support in the government. 51he Konda survey was conducted immediately after this sudden depreciation.It, therefore, allows us to study public opinion in the face of an economic crisis.
In addition to a wide range of questions, survey participants were also asked to express the extent to which they agreed with two statements that were directly linked to the NDN: "Turkey will be one of the top ten economies in the world by 2023" and "Germany is jealous of our third airport."As detailed in the theory section, these two questions capture core dimensions of the NDN: the promise of historical economic development by 2023, Istanbul airport as a symbol of the NDN, and Western states' uneasiness about and hostility towards Turkey's economic development.The correlation between these two statements is 0.65.

How widespread is the belief in the NDN?
I start my analysis by exploring the proportion of the Turkish electorate who believe in the NDN.Table 1 lists the proportions for three groups: respondents who voted for the ruling coalition (AKP and MHP) in the 2018 general election, 52 respondents who did not vote for these parties in the same election, and the entire sample.I will call the first group of voters "the ruling coalition voters" in the rest of this article.
Table 1 reveals that nearly half of the ruling coalition voters believe in both statements. 53The results are striking, given that these statements reflect severe misperceptions about the economy in Turkey.Results also reveal that Turkish voters are polarized over the NDN, as two equally sized groups of strong believers and strong deniers of the NDN formed along partisan lines.48% of respondents who voted for the AKP or MHP in the last election agreed with both statements, while only 11% of other voters agreed with them both.Symmetrically, only 9% of the ruling coalition voters express disagreement with both statements, while half of the remaining respondents reject both statements.Given that the statements asked in the survey did not include any partisan cues, the partisan divide in responses verifies the partisan nature of the NDN.

Is the belief in the NDN correlated with regime support?
The Konda survey also allows us to measure the strength of the relationship between the NDN and regime support while controlling for other demographic and political variables.
I use six variables created from six different questions as outcome variables.First, the survey includes two questions to measure future economic expectations: "the expectation of personal economic hardship in the coming months" and "the expectation of an economic crisis in the coming months."Furthermore, one question asks respondents about the level of satisfaction with democracy in the country; one question asks respondents to evaluate Erdoğan's performance; and a series of questions ask respondents how likely they are to vote for each of the political parties in Turkey, including the AKP and the main opposition party CHP, in an immediate election.These are the six questions that are used to measure regime support.In order to simplify the analysis, I combine two NDN items into a single variable taking their row average.This variable, called "belief in the NDN," is used as the independent variable in this analysis.
Two groups of control variables are added to the models.The first group is formed only of demographic variables: age, level of education, household income, and gender.The second group includes a set of political and attitudinal variables: religiosity, populism scale, conspiratorial thinking scale, ideology, political knowledge, experiencing personal economic hardship in the last year, voting for regime parties in the last election, and voting for opposition parties in the last election.A complete list of survey questions used to create these variables and descriptive statistics summarizing the properties of these variables can be found in the Online Appendix Section 3. Results are presented in Figure 1.
Figure 1 shows that the belief in the NDN is strongly associated with all indicators of regime support, except the likelihood of voting for the opposition party.These relationships are substantially significant as well.For example, the belief in developmentalism increases Erdoğan's performance evaluations by 0.3 standard deviations, even when we control for a wide range of demographic and political variables.
To summarize, believing that Turkey is experiencing historic economic growth and the developed world is envious of this growth is a vital component of regime support in Both statements 48% 11% 28% 9% 50% 31% Note: "Strongly agree" and "agree" categories and "strongly disagree" and "disagree" categories are combined.
"Neither agree nor disagree" category is skipped.Survey weights provided by the company are applied.
Turkey.However, the observational nature of the data limits the reliability of causal claims we can make over these relationships.As explained above, observational data also does not allow us to directly study emotions' mediating role.I tackle these issues in the next section with the help of an experimental design.

Study 2: online survey experiment
I conducted an online survey experiment in Turkey during the Fall of 2021 to explore the causal effects of the national developmentalist narrative (NDN). 54The goal of the experimental design was to make people randomly assigned to the treatment group think about the NDN and then measure what kind of attitudinal and emotional changes would occur among those people.If the belief in the NDN were simply a result of partisan cue-taking from the AKP elites and Erdoğan, without any further political implications, we would not see any spillover effects of thinking about the NDN on questions that were not directly about the content of the NDN, such as political emotions about political actors.However, as explained below, I have found significant attitudinal and emotional changes in respondents assigned to the treatment condition.

Introducing the data
Participants in the experiment were recruited through Facebook paid advertisements, which then directed them to a survey page hosted on Qualtrics. 55The survey was registered on 21 September. 56The data collection lasted from 22 September to 1 October.In total, 1543 people participated in the survey, 773 of which had voted for the AKP or MHP in the last general election in 2018.The advantage of recruiting participants online and letting them complete the survey through their own devices is increasing the treatment's power.This was especially important as I intended to evoke emotions through a set of questions and images.The challenge, on the other hand, is to reach a representative sample.The demographic distribution of respondents in this study is available in the Online Appendix Section 4.3, along with the comparison to the probability-based Comparative Study of Electoral Systems (CSES) 2018 Turkey sample.I used targeting tools offered by Facebook and incentive-based advertisements to increase the diversity of participants in the sample. 57My online sample diverges from the national population, measured through the CSES sample, primarily in education levels.Compared to the national population, my sample under-represents people with less than a secondary school degree (57% to 23%) and over-represents people with a university degree (36% to 18%).Overall, I find the differences between my sample and the national population acceptable, as my causal inferences in this section rely on random assignment into treatment groups.
The survey lasted less than ten minutes, and a complete list of questions asked in the survey is provided in Online Appendix Section 4.4.Outcome variables will be introduced in the relevant parts of this section; a complete list of all questions and outcome variables used in the study, along with descriptive statistics, is available in the Online Appendix Section 4.6.

Introducing the treatment
Around half of the survey participants were randomly assigned to a treatment condition after answering pre-treatment questions.Randomization was conducted through Qualtrics.Balance tests, presented in Online Appendix Section 4.2, demonstrate that randomization worked as planned with only minor differences between treatment and control groups. 58he treatment aimed to make respondents think about the NDN by showing propaganda images and asking questions to increase engagement.Respondents in the treatment condition were asked five questions about the NDN before they were directed to post-treatment questions.Images of Osmangazi Bridge and Istanbul Airport accompanied the questions in the treatment group.These images were taken from pro-regime websites.Two of them also included logos of pro-regime media outlets and propaganda phrases pasted on the images, such as "Germans are in dismay."On the other hand, survey questions were objectively phrased, e.g."Some media outlets claim that Germans are envious of the third airport.Do you agree with this claim?"Online Appendix Section 4.5 includes the complete treatment, including images and questions.Respondents in the control group did not see any questions or images.
The ruling coalition and other voters were separated at the data analysis stage based on the registration plan.The ruling coalition voters are respondents who said that they voted for the AKP or MHP in the last election, i.e. the 2018 legislative election, during the pre-treatment part of the survey.Other voters are all respondents who either voted for other parties or did not vote.Pre-registered hypotheses referred to ruling coalition voters, but I have also tested the influence of the NDN on other voters.
Finally, I conducted moderation and mediation analyses to interpret the causal effects of the NDN better.Moderation analyses explored whether the effects of the NDN depended on partisanship strength.Mediation analysis tested whether emotions mediated the relationship between the treatment and the change in economic and political preferences.

5.
3.1.Does exposure to the NDN affect partisan emotions?One of the main arguments of this article is that the NDN evokes affective reactions from the ruling coalition voters.Respondents' affective reactions were measured through twelve self-reported questions.I asked respondents to what extent Erdoğan and Kilicdaroglu (the leader of the main opposition party, CHP) evoked emotions of enthusiasm, pride, hope, fear, anger, and hatred from them.As pre-registered, I only analysed the emotions of enthusiasm, pride, and hope for Erdoğan and the emotions of fear, anger, and hatred for Kilicdaroglu.The emotions scale ranged from 0 to 10.I analysed the ruling coalition and opposition voters separately following the registration plan.
Results, shown in Figure 2, show that the ruling coalition voters who were exposed to the NDN were more likely to associate enthusiasm and hope with Erdoğan.More specifically, the ruling coalition voters' association of Erdoğan with hope and enthusiasm increased by around 0.1 standard deviations when asked whether they agreed with the premises of the NDN.These results verify that the narrative structure of the propaganda is effective at building affective and charismatic ties between Erdoğan and his supporters.
Unlike hope and enthusiasm, we do not find any relationship between the treatment and pride.There are two alternative explanations here, which are probably both true.Firstly, among the ruling coalition voters in the control group, the level of pride was higher than levels of enthusiasm and hope, as can be seen in Online Appendix Section 4.6.Thus, there was less room to manipulate pride.In addition to this, it can be argued that the origin of pride is different from the origin of enthusiasm and hope among the ruling coalition voters.Pride might be associated with partisan group identities and feelings of superiority vis-a-vis other partisan groups in Turkey.In contrast, hope and enthusiasm are associated with the vision of the future.
We also do not find any relationship between the NDN and negative partisan emotions.This finding is in line with the finding from the national study, which showed no relationship between belief in the NDN and the likelihood of voting for the opposition party.Finally, we do not find any effect of the treatment on voters who did not vote for regime parties in the last election.This is an important finding which shows the limitations of narrative-based economic propaganda.

Does exposure to the NDN affect economic evaluations?
I use four variables to measure economic evaluations.The first two variables are future economic expectations on an individual and national level.Each averages responses to one positively worded and one negatively-worded and reverse-coded question on a scale of 1 to 5. The third and fourth variables measure to what extent respondents believe that Erdoğan or ministers of the economy are to blame for the country's economic troubles.These variables are produced from two separate questions, asking respondents to express blame assignment on a scale of 1 to 5.
The first two columns of Figure 3 present the results, which consistently demonstrate that exposure to the NDN improves economic evaluations among the ruling coalition voters.The ruling coalition voters not only express positive economic expectations for the future direction of the economy and their personal economic situation, but they also put less blame on Erdoğan and his ministers for the current economic troubles in the country.

Does exposure to the NDN affect political preferences?
Finally, I use two variables to measure political preferences.The first variable measures how likely respondents were to vote for the AKP if there was an immediate election.The scale ranged from 1 to 5. The second variable was built on the partisanship item used in the CSES.I asked respondents whether they felt close to any political party and how close they felt (if they felt at all), from 1 to 3.This set of questions was asked both before and after the treatment.As indicated in the registration plan, I created a variable ranging from -3 to 3, measuring the individual-level change in partisanship due to the treatment.
Results are presented in the third column of Figure 3.An unexpected result is that we do not find any relationships between the treatment and political preferences.As I have already established that the NDN improves both economic and affective evaluations, my only explanation is that the room for experimental manipulation was more limited in these questions than in other questions.The analysis of heterogeneous treatment effects in the following subsection allows us to explore if this was the case.

5.3.4.
Exploring causal effects: the role of non-partisans and emotions I ran moderation and mediation analyses over the ruling coalition voters' responses to better understand causal effects.
Firstly, I explored how partisanship and its strength moderated the effect of the treatment on emotional, economic and political responses.Partisanship is a significant factor in Turkish politics, and it is essential to see whether there are any distinctions between strong partisans and non-partisans concerning the effect of the NDN.I created aggregate measures to simplify the analysis by taking row averages of relevant variables.All three variables of positive partisan emotion (hope, enthusiasm, and pride) were combined to form an aggregate positive emotions variable; personal and national economic evaluations were combined to form an aggregate economic evaluations measure; blame assigned to Erdoğan and ministers were combined to form an aggregate blame assignment variable, and vote and partisanship were combined to create an aggregate political preferences measure.
The results, presented in Figure 4, consistently demonstrate that treatment effects were strongest among non-partisan ruling coalition voters.This means that the NDN helps the regime to sustain its support among one of the most crucial voter groups in terms of the regime's survival, that is, voters who voted for regime parties in the last election but do not feel close to those parties currently.Given that these people are probably more likely to abstain or defect in the next election than strong partisans, these results reveal that the NDN serves an important political function.
Interestingly, causal effects were nearly absent among strong partisan ruling coalition voters, which formed around 30% of ruling coalition voters in the sample.We can interpret these results as an indication of ceiling effects.For example, while the average level of positive partisan emotions among strong regime partisans was nine over 10, the average level of positive partisan emotions among the rest of the ruling coalition voters was 6.2.Similarly, these results suggest that our inability to find the effects of the NDN on political preferences was because of the lack of room for experimental manipulation among strong partisans.Exposure to the NDN does move political preferences among non-partisans.
Second, I ran three different mediation analyses, exploring how the change in aggregate positive emotions variable mediates the effect of the treatment on three different variables: Aggregate economic evaluations measure, aggregate blame assignment variable, and aggregate political preferences measure.Results are presented in Table 2.
These results demonstrate that for each of the three outcome variables, there is a statistically significant effect between the treatment and the outcome variable that is mediated through partisan emotions.The size of this effect is around 0.1 standard deviation.This corresponds to more than 40% of the effect on economic evaluations and 30% of the effect on blame assignment.It is important to note that positive emotions fully mediate the significant and positive relationship between the treatment and pro-regime political preferences.Thus, we can conclude that evoking positive emotions is crucial to economic propaganda in Turkey.

Discussion: the political function of developmentalist propaganda
Empirical findings in this article raise two questions crucial to understanding the political function of developmentalist propaganda in Turkey.First, if the national developmentalist narrative (NDN) promises economic development for the entire nation, why is it only effective among voters of the ruling coalition?Second, if the NDN's effects are limited to people who voted for ruling parties in the previous election, how does the NDN promote the regime's political agenda?
To begin with the first question, it is not surprising that exposure to the NDN does not affect opposition voters in Turkey.Regime critics living in authoritarian countries develop distrust towards the regime's propaganda channels and any news spread by them. 59As a result, it is harder to influence the political opinions of these voters through propaganda.Given that the NDN promotes serious economic misperceptions, such as the claim that Turkey will be one of the top ten economies in the world by 2023, it is probable that opposition voters are more cautious when engaging with these claims.Furthermore, the NDN includes partisan elements, such as "the heroic role of Erdoğan" and "obstacles laid out by opposition parties."These partisan elements might discourage opposition voters from engaging with the narrative.As the concept of "narrative proximity" describes, a certain level of resonance between the narrative and the audience is necessary for the narrative to be influential. 60On the other hand, partisan elements of the NDN can make it easier to evoke pro-AKP emotions among voters close to the regime for sociological, ideological, or economic reasons.
This brings us to our second question: how does the NDN help the regime if its effects are concentrated among people who already voted for ruling parties in the last election?To answer this question, it may be helpful to underline the difference between partisan cue-taking and the emotional mobilization of partisans.It is a well-established finding in political science literature that partisans in democracies follow cues from in-party elites to form their opinions. 61This article contributes to the literature by demonstrating that authoritarian propaganda, built on developmentalist narratives, also can influence how partisans evaluate the economic situation.
Beyond this, however, results also reveal the emotional effects of propaganda.Partisans exposed to developmentalist propaganda become more likely to associate Erdoğan with hope and enthusiasm.These emotional effects talk to the broader literature on emotions and partisanship. 62According to this literature, emotions influence voters' reliance on their partisan identities while forming political attitudes and evaluations.People who feel enthusiastic or angry are likelier to develop a political opinion aligning with their partisan identity.Second, emotional appeals play an essential role in mobilizing the partisan base.For example, partisans who feel positive or negative emotions are likelier to share partisan information or vote than partisans who do not feel the same emotions. 63Emotional appeals are essential for mobilizing weak partisans. 64he findings in this article point in the same direction.First, the mediation analysis suggests that the emotional nature of developmentalist propaganda plays a vital role in convincing voters of the ruling coalition to update their economic expectations in a positive direction.Furthermore, developmentalist propaganda sustains and strengthens partisan ties among voters of the ruling coalition by evoking partisan emotions.These findings should be considered in the context of the ongoing economic decline in Turkey.As shown above, the effects of developmentalist propaganda are most substantial among voters who voted for regime parties in the last election but do not define themselves as partisans.These non-partisan voters would normally be the voters most likely to defect in an upcoming election because of the economic crisis.Thus, developmentalist propaganda helps the regime to preserve its support base under difficult economic and political conditions.

Conclusion
Arguably, the AKP in Turkey has been more successful at managing the perceptions of the economy than managing the economy itself.Despite the economic decline that was ongoing for nearly a decade, the AKP and its leader Erdoğan secured another electoral victory in 2023.The national developmentalist propaganda preserved its central place in the AKP discourse during the election period, as the AKP and Erdoğan replaced temporal references to "2023" with an ostentatious slogan serving the same purpose: "The Century of Turkey." This article demonstrates that the national developmentalist propaganda in Turkey has built affective ties between Erdoğan and his voters, improved economic evaluations, and increased political support for the regime.Observational and experimental data support these arguments.Observational survey data confirms that a significant portion of ruling coalition voters in Turkey hold economic misperceptions originating from the national developmentalist narrative.It could be argued that this relationship was a result of the partisan cue-taking, with no further political consequences.However, the survey experiment overcomes this argument by showing that exposure to the developmentalist narrative results in emotional and attitudinal changes that can be observed through questions that are not directly related to the developmentalist narrative.
This article improves our understanding of support for authoritarianism in Turkey and other countries.Whether authoritarian regimes could improve their performance evaluations through propaganda has been an open question.This study shows that the narrative structure, which offers a powerful and appealing vision for the future, can help authoritarian regimes escape the responsibility for current economic troubles and strengthen affective bonds with their voters.As such, this article offers a new perspective to understand the development of partisan and affective attachments towards authoritarian parties.
Going beyond the case of Turkey and authoritarianism, this article also contributes to studying ideological formations in the Global South.Historians and anthropologists have long pointed to the role of utopian developmentalist ideologies in developing countries. 65However, political scientists' engagement with this discourse has been limited to studying developmentalist policy ideas and their results. 66Similarly, political scientists have overwhelmingly focused on negative emotions and ignored the role that can be played by positive emotions, such as hope and enthusiasm. 67This article introduces developmentalism as a popular narrative, and it reveals the mediating role that positive emotions play in the formation of economic evaluations.
This study also opens new avenues for future research.One such area is the variation in the success of certain authoritarian narratives.Many leaders in the Global South attempt to use promises of fast economic development to build public support and legitimacy.Under what conditions do these promises evoke enthusiasm and hope among the population and build popular support for the regime?The previous credibility of the regime and the temporal order of economic developments can play important roles here.When Erdoğan made the developmentalist promise the centre of his electoral campaign in 2011, he could boast about years of fast economic growth under his leadership.When the economic decline started in the mid-2010s, his supporters had already developed affective investments in the NDN.While this temporal explanation seems plausible to me, we need more systematic comparisons to understand political and economic conditions facilitating the success of authoritarian narratives.

Figure 1 .
Figure 1.OLS results measuring the relationship between the belief in the NDN and political and economic outcomes.

Figure 2 .
Figure 2. The effect of the NDN treatment on emotions, based on voting behavior in 2018.

Figure 3 .
Figure 3.The effect of the NDN treatment on economic and political evaluations, based on voting behavior in 2018.

Figure 4 .
Figure 4.The moderating effect of partisanship on the relationship between developmentalism treatment and aggregate measures.

Table 1 .
Proportion of the Turkish electorate agreeing and disagreeing with the NDN, with samples divided based on past vote choice.

Table 2 .
Mediation analysis: Treatment, Emotions, Economic and Political Preferences.All control variables are added to all models.Analysis conduct in Stata with mediation package.