Election Polls: Will the Güegüense Return?
From the book by Leslie E. Anderson, Elections and Democracy in Central America,The University of Carolina Press, Chapel Hill, 1995. This chapter, "Elections and Public Opinion in the Development of Nicaraguan Democracy," was condensed and edited by envío with the author's permission.
Leslie E. Anderson
After its independence in 1821, Nicaragua experienced a long series of dictatorships as two principal parties, the Liberals and Conservatives, vied for political control.
Both were essentially clannish parties of elites, associated with wealthy families. Political control by Liberals or Conservatives was decided on the basis of military power. From time to time each party proved itself willing to call on foreign influence, particularly from the United States, in order to improve its own military strength.
As control of the state shifted from one party to the other, the majority of Nicaragua's population, most of whom were peasants, was excluded from meaningful political participation. Peasants were sometimes forced to serve in one or the other of the contending armies in a battle for control of the state. Alternatively, they were sometimes lured to "vote" in fraudulent elections with bribes of food and drink.
Military competition for state control made survival difficult among Nicaragua's peasantry. Many villages and individuals preferred to stay out of politics altogether and hopefully, thereby, stay alive. One common story recounts that Nicaraguan villages kept a portrait of each of the ruling generals of the Liberal and Conservative armies. When one army thundered through the village, peasants would hurriedly hang up the portrait of the appropriate general, only to change it when the opposing army arrived. The story illustrates a phenomenon that to some extent persists today??a chameleonlike quality of presenting the "public opinion" that the army wanted to hear. This Nicaraguan tendency to disguise public opinion around the issue of who should rule grew so entrenched that some observers argue that it became a national tradition with a name??the güegüense tradition [for the duplicitous indigenous go?between with colonial authorities made famous in the 17th?century Nicaraguan play of the same name].
In the late 1920s this pattern was broken when some of Nicaragua's peasantry expressed genuine and unsolicited political opinion by rallying behind nationalist revolutionary Augusto C. Sandino. Although he initially took up arms as part of one of the two elite?controlled armies, Sandino soon developed his own agenda, which included driving the US Marines out of Nicaragua and redistributing land to landless peasants in agrarian cooperatives in a region controlled by his "Crazy Little Army" of peasants and the poor. The strength and extent of Sandino's following were an early expression of autonomous popular political opinion and participation in a country where these were rarely expressed. The vehemence of the popular expression frightened elites, and Sandino was assassinated in 1934.
During the Somoza family rule (1936?1979), most forms of popular political participation were heavily repressed or manipulated by the regime. Despite such repression, the Nicaraguan population did express opposition both to labor conditions in the 1960s and to the regime itself through support for independent opposition parties and by joining numerous guerrilla movements, including the Sandinista National Liberation Front (FSLN).
Among Nicaraguan dictators, the Somozas were notorious for the use of fraudulent elections. Sometimes peasants were physically coerced to vote. True popular opinion of the Somoza rule, however, finally and dramatically expressed itself in the revolutionary overthrow of the dictatorship and its military defeat at the hands of an armed population. Leading the revolution, the Sandinistas took power on July 19, 1979.
Elections in the New DemocracyAlthough the 1978?79 insurrection was easily an expression of popular political opinion and an example of popular participation, it was not an election. Indeed, the Sandinistas were in no hurry to hold national elections. The Sandinista leadership maintained that the people had already expressed support for them that was much stronger than a mere vote, by militarily following them to victory in 1979. In the first postrevolutionary years there was a good deal of truth to this argument. Certainly spontaneous expressions of popular support such as demonstrations amply illustrated the popularity of the Sandinistas between 1979 and 1981.
As time passed, however, and the new government got down to the real business of transforming the economy, the extent of Sandinista popularity became less clear. Partly as a result of commitment to the population and partly as a result of US pressure, the Sandinistas decided to hold their first election in 1984. In November of that year, Nicaraguans voted in the first free election in the history of their country. Neither elites nor the population at large had any experience with an effective presidential election. Used to controlling a revolutionary regime in which the state and the FSLN party were practically inseparable, the Sandinistas had to divorce themselves and their party from the state for the purpose of the election. They had to present themselves as a party whose control of the state would benefit most Nicaraguans more than any alternative. Such self?presentation was new for the Sandinistas.
The opposition was also experimenting. Opposition in Nicaragua, of course, had very limited and unsuccessful experience with expressing itself through elections. Before 1979 Somoza had often frustrated the opposition through repression or fraud. Moreover, several of the parties that opposed the FSLN in 1984 had themselves sometimes been involved with Somoza in engineering electoral fraud during the years of the dictatorship. Thus the opposition parties in 1984 had almost no experience with bidding for popular support in a legal fashion or expressing their opposition to the incumbent through electoral channels.
This inexperience with fair elections and the Sandinistas' lingering revolutionary popularity made Nicaragua's first election in 1984 highly unusual. First, the incumbent FSLN enjoyed a major advantage in the form of revolutionary success and charisma. Second, the opposition parties seemed to spend as much energy disagreeing with one another as they did opposing the Sandinistas. As a result, the Sandinista party faced no fewer than six opposition parties, including remnants of the traditional Liberal and Conservative parties.
Despite the newness of the experiment, however, the Nicaraguan population seems to have embraced the election ??91% of the registered population voted. Given the revolutionary popularity of the FSLN, the still recent memory of FSLN courage in the revolutionary struggle and the splintered nature of the opposition, it is not surprising that the Sandinistas won the 1984 election by a landslide, taking 67% of the vote. The second strongest party, the Conservatives, took 14%, while the Liberals won 9%.
Although 1984 marked Nicaragua's first authentic election and popular support for the Sandinistas was clear, the election received very little positive attention in the United States. Coming at the height of the Reagan years, in fact only two days before his own reelection, Nicaragua's election was dismissed by both the administration and the media. Both argued that the Sandinistas had coerced the results and that the election had not been fair.
Nicaragua's 1990 ElectionNicaragua's second election, held in February 1990, was markedly different from 1984, both in the way it was handled internally and in the international attention it received. Everyone seemed to be learning about conducting elections in Nicaragua. The Sandinistas ran again as the incumbent party. The opposition, however, changed its tactics entirely. Realizing that no opposition party alone could defeat the Sandinistas, most of the opposition parties banded together in one umbrella coalition: the United Nicaraguan Opposition (UNO). The coalition was not an easy one, given the historical divisions among the parties. In an effort to appease all members and maximize its electoral prospects, UNO selected a conservative, Violeta Barrios de Chamorro, as its presidential candidate and the leader of the Independent Liberals, Virgilio Godoy, for vice president. These choices caused considerable internal tension, but UNO held together and competed successfully in the election.
The US reaction in 1990 was also different from its response in 1984. Instead of dismissing the election, the United States openly supported UNO, providing both political and financial support. Another important difference with the 1984 election lay in the amount of academic and political study the 1990 vote received. The 1984 election had been held under extensive international scrutiny on the day of the election, but in 1990 the scrutiny began earlier and was more detailed. Many public opinion polls were conducted by organizations inside and outside Nicaragua during the nine months prior to February 1990. These polls allow a depth of study of Nicaraguan public opinion that had never before been possible in the history of the country. They also reveal the atmosphere surrounding the election in a manner that was impossible in 1984 and unexpected even in 1990. Although electoral studies in advanced democracies usually give a substantial advantage to incumbents, that advantage can become a disadvantage if voters are unhappy about the national state of affairs at election time. The most important measuring stick by which voters in most societies judge incumbents is the state of the economy. As the election approached, the Nicaraguan economy, never strong after 1979, was in a shambles. Unemployment was very high and rising steadily. Between October and December 1989, inflation moved from over 50,000% to 71,572%, and real salaries had fallen precipitously in the previous few years, according to the World Bank. The Sandinista government had tried to control inflation by repeatedly devaluing the córdoba, but to no avail. Devaluation, however, made it increasingly difficult to import necessities such as gasoline, medical supplies and mechanical parts for cars and machinery. In addition, the US trade embargo against Nicaragua made many replacement parts unavailable even if the country had had the money to buy them. As a result, production fell dramatically, lowering the volume of exports and the availability of domestic items.
These various factors interacted with one another to produce a downward economic spiral such that, between 1984 and 1990, life in Nicaragua became extremely difficult. Almost everything was in scarce supply and tightly rationed. Entire families lost jobs. Workers often could not make ends meet even when they retained their salaries. Finally, pressure on Nicaragua from international financial institutions forced the government to roll back many reforms it had put in place during the first five years of the revolution. As a result, Nicaraguans lost many of the social benefits they had come to see as major gains of the revolution and on which they had come to depend as the economic situation worsened.
In addition to the disastrous economic situation, the war against the US?financed anti?Sandinista rebels (the contras) slowly wore out the Nicaraguan people. When the war had begun in the early 1980s, many Nicaraguans had enthusiastically supported the Sandinista effort to stop the contras, many of whom were seen by Nicaraguans as former members of Somoza's National Guard. The contras habitually attacked unarmed civilians and civil servants such as teachers, doctors, nurses and agricultural technicians as well as schools, factories and medical clinics. Over the years, however, the resolve of the Nicaraguan people understandably weakened as the war dragged on and on with more young men killed each week.
The conflict was a huge drain on a broken economy. Each soldier was a dependent rather than a productive citizen, and individual families could not afford to lose a breadwinner even temporarily to the army. Eventually the economic crisis and the growing unpopularity of government policies made some sectors increasingly receptive to the contras' position, particularly in the rural north and east and among the informal sector in Managua. Sickened by war and exhausted by the economic crisis, Nicaraguans were desperate for peace and prosperity. They wanted goods available in the stores and their loved ones home from the front. Anxious to be out from under US pressure, both economically and militarily, Nicaraguans took their exhaustion and anxiety with them to the polls in 1990. As in 1984, the population again embraced the election, this time with a voting rate of 86% of the registered population. In the election, Violeta Chamorro defeated the Sandinista candidate by 54.7% to 40.8%. After a series of negotiations of transition details and a cease?fire with the contras, Ortega quietly ceded power to the new President. Although UNO actually carried only slightly more than half of the vote, its margin of victory was a virtual landslide when compared with the Sandinista percentage. The magnitude of the Sandinista defeat is even greater when compared with the 67% it had received just six years earlier. Moreover, for the most part, both within Nicaragua and in the United States, the Sandinista defeat was a surprise and even a shock for supporters of both sides. Many observers simply had not been able to believe that the once popular revolutionary party could have been so soundly defeated at the polls.
The 1990 Election In RetrospectSince the Sandinista defeat, students of Nicaraguan politics have endeavored to understand why the FSLN lost and why its loss was so severe. In retrospect it has been so easy to find reasons for the defeat that one wonders why it came as such a surprise. Yet in the months prior to the 1990 election, few would have called the election with certainty, and many observers were sure the FSLN would win. Sandinista supporters within Nicaragua and Sandinista opponents within the Reagan administration alike expected a Sandinista victory.
Reputable public opinion polling organizations in the United States, including The Washington Post, NBC and Greenberg?Lake, had all reported survey results showing a Sandinista lead. The same conclusions emerged from repeated surveys conducted by Nicaraguan polling organizations and research institutes. A careful review of all organizations conducting polls in Nicaragua, however, reveals that several predicted a Sandinista defeat. But these predictions were not heard above the din surrounding the election either in Nicaragua or in the United States. While some of those predicting a UNO victory were directly associated with the UNO campaign and were not considered objective, others were apparently nonpartisan organizations, notably those from Venezuela and Argentina.
We will never know for certain why the majority of the voting Nicaraguan population chose Chamorro. Yet the Sandinista defeat and the failure to foresee it leave serious questions in the wake of the 1990 election about the use of public opinion polls to predict electoral outcomes in a country at war, under heavy external pressures and with longstanding traditions of repression and dissimulation; and about the possibility of democracy in a tiny and relatively powerless country in the face of powerful external opposition to the representatives of a large element of popular opinion.
The earliest surveys of Nicaraguan public opinion prior to the election were conducted in August 1989, seven months before the election. This was an unusual and interesting time to be examining preelectoral public opinion, since UNO did not yet formally exist. While it was clear that many parties opposed the Sandinistas, as they had in 1984, it was far from clear that they would be able to swallow their own differences enough to coalesce under an opposition umbrella. Thus, when these early surveys were done, the Sandinistas looked like the only cohesive and viable major party capable of entering the campaign.
In an early survey by the Nicaraguan research institute Itztani, the Sandinistas had a definitive lead among respondents who gave any partisan opinion at all. The survey showed 32% supporting the FSLN and only 19% supporting all the opposition parties combined. These general contours of public opinion continued throughout the fall of 1989 and were responsible for initiating the general impression, in both the United States and Nicaragua, that the Sandinistas were far ahead in public preference.
Nonrespondents Important In NicaraguaOmens of trouble for the FSLN appeared in these early surveys, but were generally ignored. Many, and sometimes even most, respondents did not give a partisan opinion at all. They refused to answer questions about candidate preference or avoided such questions by evasive tactics. The Nicaraguan güegüense tradition seemed to be reappearing, unnoticed by electoral observers.
The general approach to evaluating public opinion polls is to treat nonrespondents (persons who refuse to answer) as "undecided" and omit them from the analysis. Analysts consider that nonrespondents/undecideds will eventually decide whom they prefer and divide their vote largely along the same lines as those who gave a partisan response. Since partisan respondents are considered to "represent" nonrespondents, the latter may safely be ignored. This general approach led many pollsters to predict the election based only on respondents who had given an opinion and thus to call a Sandinista lead.
In retrospect, however, this practice, while possibly defensible in established democracies, was highly questionable in Nicaragua. First, the number of nonrespondents on the expected vote questions was nigh. Nonresponse rates of 10?15% are considered normal in advanced democracies, yet these early surveys drew nonresponse rates of 30?50%! Why did so many decline to name their presidential choice? In the politicized atmosphere of revolutionary Nicaragua, were that many people really undecided? Even more important, in the first Itztani survey in September 1989, when UNO did not yet exist, why did only 32% of respondents say they would support the Sandinistas? Pollsters, however, were not asking these questions. They were ignoring the 30?50% of nonrespondents, assuming that respondents "represented" them. Early opinion polls are normally considered only preliminary indicators of public opinion. However, in view of the Sandinista loss, early polls such as Itztani's may have been more valuable than pollsters realized precisely because they did not show a strong Sandinista lead among all respondents. In these surveys, large percentages of respondents declined to answer party?preference or candidate?choice questions. Viewed in this way, these early nonrespondents were an indicator that Sandinista support was not strong among the general population of survey participants. As the election drew near and more surveys were conducted, the nonresponse rate declined and the opinion indicator that nonresponse constituted disappeared.
In the polls conducted by Nicaraguan and US organizations, the Sandinista lead grew steadily as the nonresponse rate dropped, continuing to lead analysts to believe that the Sandinistas held an electoral lead. These later results, however, may have been less accurate than earlier polls showing substantial nonresponse rates that might have been read as lack of definitive support for the Sandinistas. Lacking such high nonresponse rates, the later ones incorrectly lured analysts into believing the partisan divisions that showed in the survey results.
My analyses of these varied opinion polls have led me to believe that the ancient Nicaraguan tradition of concealing one's partisan opinion may have misled pollsters into incorrectly forecasting the election. Indeed, it may have been responsible for giving the impression that the Sandinistas were in the lead when in fact they never were.
The DOXA polling organization, located in Caracas, Venezuela, conducted five public opinion surveys in the three months prior to the election (early November 1989 to early February 1990) which received relatively little attention in either Nicaragua or the United States. Like many other polling organizations, DOXA received a high nonresponse rate on the expected vote question and, also like others, did not consider the nonrespondents when predicting electoral results. Unlike Nicaraguan and US?based organizations, however, DOXA correctly forecast the UNO victory.
In response to a question on who respondents thought would win the election, the nonresponse rates declined substantially between the first survey and the last. In addition, both FSLN and UNO support rates rose, with UNO support surpassing FSLN support by the third poll in early January 1990 and retaining its lead. DOXA based its forecast on these results.
However, DOXA's data contained even more reliable information than DOXA analysts realized. At least two other precise predictions could have been made by including nonrespondents. In the first, more conservative one, a prediction of electoral outcome can be made based on average slope figures for the FSLN, UNO and nonrespondents over the three?month period. The slope for the FSLN showed an average increase of 0.46 and for the UNO an average increase of 1.64, while the nonrespondent slope decreased an average 2.12. Simply extrapolating these figures yields a prediction that the Sandinista vote would have been approximately 41% and the UNO vote would have been 43%. This prediction assumes that the actual election would have had a voter abstention rate as high as the nonresponse rate in the DOXA survey (over 19%), while we know that only 14% of the eligible adult population did not vote. Thus, while correctly forecasting the election, this prediction would have been off in both its estimate of voter participation rate and its prediction of UNO support.
Would it have been possible to have come closer to the actual outcome based on the DOXA data? By making only one far?fetched assumption??that all of DOXA's respondents voted in the election itself??we can calculate how much of the total downward slope among the nonrespondents would go to UNO and how much to the FSLN. That is done by dividing each of their average slopes (.46% for the FSLN and 1.64 for UNO) by the average slope of the nonrespondents (?2.12). By doing this, we see that 77.36% of nonrespondents may have supported UNO and only 22.64% the FSLN. If we assign these percentages based on the 19.3% who did not respond in the final DOXA survey, we come surprisingly close to the actual electoral results of 54.7% to 40.8%, since these additions predict a 55.73% UNO vote and a 44.17% FSLN vote. Of course, in actually forecasting an election one cannot make as many assumptions as we have made here, for example that the only shifts were from a nonrespondent to a partisan supporter and that all who made an initial partisan choice stuck with it. However, the surprising closeness we achieve to actual electoral results when we merely speculate about partisanship among nonrespondents is remarkable. It shows that there probably was a strong relationship between nonresponse rates and UNO support. While that relationship was probably not as direct and pure as we have assumed here, it is still likely that at least some substantial proportion of UNO support showed itself as nonresponse rather than as outright partisan support for UNO.
If this relationship is so clear in the surveys of a Venezuelan organization not connected to either party or to the United States, how much stronger must that relationship have been in surveys in which respondents perceived interviewer bias? We will never know the answer with certainty, but we can be reasonably sure that the surprise UNO victory came directly from the part of the electorate that refused to answer the intended vote question.
Herein lies the key to the forecasting debacle in 1990. Nonrespondents were ignored in predictions. Their partisan preferences were not even considered. Indeed, the fact that they might even have partisan preferences was never mentioned. While polling organizations would have been foolish to make definitive predictions based on nonrespondent partisanship, they might at least have speculated about the partisanship of this large and silent group, toying with reasonable assumptions and making simple extrapolations as we have done here. Had pollsters speculated in such an informed fashion, their preelectoral estimates would have been closer than they were.
In retrospect, then, the 1990 election results might not have been such a shock if survey data had been analyzed in cognizance of the Nicaraguan güegüense tradition and the long?standing public tendency to camouflage political opinions or appear to "change" them depending on who is listening.
Polls as a Gauge of Nicaraguan DemocracyWhat can we learn from the 1990 elections? What can we say about democracy in Nicaragua based on the polling debacle and the importance of nonrespondents in the election?
Although Nicaragua has now witnessed two authentic elections and currently lives with a democratically elected civilian regime, the depth, breadth, health and stability of Nicaraguan democracy are still at question. I believe that the viability and success of survey research in Nicaragua are to some degree a measure of the country's democracy and of the extent to which democratic rules have become a part of political reality for the average person. Indeed, the success or failure of survey research may be a more accurate reflection of the health of Nicaraguan democracy than is the formal institution of national elections. Nicaraguans may hold and vote in national elections. They may place civilian regimes in power. If this practice continues over a long period of time, Nicaragua may be said to enjoy a formal, constitutional democracy. If, however, members of the public sampled at random by survey researchers do not feel comfortable expressing their opinions and do not trust that their opinions will remain confidential, Nicaraguan democracy is more shallow than these institutional aspects imply. If such secretiveness and camouflage of opinions remain common in election after election, the health and progress of democracy in Nicaragua will have been seriously constrained. It will suggest that elections are not being conducted in a free and open atmosphere. In the long run, such conditions will not be conducive to the stabilization of formal democracy in Nicaragua because regimes may be unstable, unpopular, and elected only as the result of pressure. In such circumstances elections will be only formal and superficial and will not have contributed to a popular culture of openness, dialogue and participation under democratic rules.
What were Nicaraguans afraid of? Partisans of each color blame the other side. UNO supporters say that the Sandinista regime was repressive and always pressured the public. Sandinistas, in contrast, argue that US pressure influenced the election in far more fundamental but indirect ways. Sandinista supporters say Nicaraguans were afraid to vote for the FSLN because they dreaded a continuation of the economic blockade, other economic pressure and the contra war. It is likely that both sides are correct, that Nicaraguans felt unjustly pressured by both sides and from many directions. It is also possible that these fears were exaggerated or even unjustified. Perhaps many Nicaraguans felt pressured and fearful because they were used to feeling repressed and fearful??a sort of political?cultural vicious circle. Yet regardless of the source of the reluctance to answer or of whether and how far feelings of fear were justified, clearly the perception of pressure or risk was there for some of the electorate. All of this would seem to indicate that the 1990 Nicaraguan election was a good deal less democratic than its formal appearance would lead us to believe. Whether or not the public actually was pressured is secondary to the fact that it felt pressured. True or not, justified or not, that perception reflects something of the nature of democracy in Nicaragua.
Would a less pressurized atmosphere have resulted in different percentages and different results? We will never know, but viewed from this perspective, the electoral surveys uncovered something far more important than precise electoral results. They revealed an uncomfortable and potentially ominous underpinning to the election. In this sense, then, the surveys and polls revealed much more than they concealed. They uncovered fear, distrust, caution and discomfort and a population that continued to act as if it were living under repressive circumstances even as it participated in a technically clean and open election. The polls thus uncovered what the actual election itself concealed: the perception of a less than fully democratic atmosphere surrounding a democratic election.
Nicaraguan Democracy In PerspectiveIn Nicaragua, elections and democracy constitute new experiences, forms of politics still being explored and understood. Formal electoral democracy in Nicaragua is only beginning and, at present, has only two authentic elections to its credit. Participatory democracy, a democratic atmosphere, and a general feeling of openness and freedom "in the air" are also new and may be even less developed than is formal electoral democracy. The survey results indicate this, particularly surrounding the 1990 election. In this sense, I have argued that the surveys have provided an indirect measure of democracy that pollsters had not anticipated.
Political scientists sometimes equate elections with formal democracy and with at least a beginning of democratic popular participation. At one level the Nicaraguan elections, particularly the 1990 one, accomplished this purpose. Citizens voted. Participation was broad. Turnout rates were high. Elections have also carried on the revolutionary pattern of calling for popular demonstrations, thus continuing a relatively wide range of political participation. And, of course, they have allowed citizens to democratically choose leaders and enjoy electoral regimes for the better part of a decade now. In these ways elections have been conducive to democracy in Nicaragua.
In other ways, however, elections have inadvertently provided opportunities for the continuation of long?standing non democratic traditions in Nicaragua. These conditions and traditions have not advanced democratization in Nicaragua. When fear, distrust and nonresponse are prevalent surrounding an election, the depth and quality of political participation are limited among those who feel fearful and do not respond. When elections occur in a pressurized atmosphere, including heavy?handed US involvement, they lack an environment conducive to the exercise of full, participatory rights by everyone. If a president has been elected in an atmosphere in which citizens felt pressured to vote in one way or another, the election may not have helped consolidate a stable regime at all. If elections have created an excuse for domestic or foreign actors to pressure or intimidate the population, they have not contributed to a participatory political culture. And finally, if the United States or other foreign power has had a negative, dominant role in domestic economics and politics for years before the election, Nicaraguan freedom to exercise democratic rights has been seriously compromised. The survey results surrounding the 1990 elections indicate that all of the above nondemocratic influences were or may have been present in 1990. Elections in Nicaragua, particularly in 1990, therefore offer a mixed indicator about the development of democracy in Nicaragua. The elections themselves are a step in a democratic direction, but the atmosphere and pressure (real and perceived) surrounding the 1990 election make that step a good deal smaller than it otherwise might have been. This mixed assessment of democracy in Nicaragua in the past leaves us with uncertainty about democracy in Nicaragua in the future.
Nicaragua has continued under Chamorro's civilian regime for five years now. Although there have been threats to Chamorro's presidency, she has fended them off successfully so far. The future health and stability of democracy in Nicaragua will improve only if minimal conditions are met: Nicaragua will need repeated elections at regular intervals. The next should be held on schedule in 1996. The national and international context will need to allow the FSLN to compete and win, as well as to take and hold office. If it regains office, the FSLN will need to pay more careful attention to democratic rules than it did in the past. Democracy in Nicaragua has taken a small step in the right direction against overwhelming odds. Whether or not that progress continues is uncertain and bears careful watching.