Books in Review

The Churchill Hypothesis

Krzysztof Jasiewicz

Everyone knows that Winston Churchill said something to the effect that democracy is the worst form of government except for all the others, but few people check the exact citation. It goes as follows: "Many forms of Government have been tried and will be tried in this world of sin and woe. No one pretends that democracy is perfect or all wise. Indeed, it has been said that democracy is the worst form of Government except all those other forms that have been tried from time to time." Richard Rose, William Mishler, and Christian Haerpfer did check. Moreover, they saw this statement as not just a clever bon mot, but also an empirically testable hypothesis. If people who have personally experienced the shortcomings of both democratic and undemocratic regimes opt in favor of the former, then the "Churchill hypothesis" stands. If, however, those people prefer the undemocratic alternatives to the imperfections of democracy, then the Churchill hypothesis is false: Democracy falls short of even the "lesser evil" threshold.

Recent history has provided the authors with an excellent testing ground for their hypothesis. The communist regimes in Central and Eastern Europe fell, one after another, within a span of two years, and new democracies began to emerge, albeit not without difficulties. Rose, Mishler, and Haerpfer recognize these difficulties. They recognize also that communism was once perceived by many as a viable (if not outright superior) alternative to democracy. It failed because it did not fulfill its promise of an efficient, prosperous economy and a more just society. This supply-and-demand analytical framework can also be used, the authors claim, to assess democratic regimes. "[J]ust as macro-economic theories have no relevance to everyday life if they cannot be related to micro-economic activities of individuals, so constitutional forms are lifeless or irrelevant if they do not have the support of the people. That is why even though the elites propose, the masses dispose" (p. 8). Policies that alienate significant segments of the population may not only bring down a particular government, they may even undermine popular support for democracy itself. Earlier this century, democracies across the region gave way to authoritarian governments precisely because they failed to meet popular expectations. Will history repeat itself?

To answer this question, the authors performed elaborate analyses of public-opinion data for nine postcommunist states in East-Central Europe: [End Page 169] Belarus, Bulgaria, the Czech Republic, Hungary, Poland, Romania, Slo-vakia, Slovenia, and Ukraine. The data have been gathered as part of the New Democracies Barometer (NDB) project, initiated in 1991 by the Paul Lazarsfeld Society of Vienna, in parallel with the well-established Euro-barometer. Since its inception, the NDB has conducted public-opinion surveys in individual Central and Eastern European states (similar projects also cover Russia and the Baltic states). For their investigation, the authors chose the survey conducted in the fall of 1993 and winter of 1994.

The authors selected two core dependent variables, assessment of the current regime (adding, when appropriate, evaluation of the communist past and hopes for the future) and opinions on the alternatives to democ-racy (a composite of five variables), which they examined in relation to a host of independent variables. This procedure allowed the authors to measure the impact that factors such as social structure, the political lega-cies of the past, the performance of the new regimes, the microeconomic and macroeconomic dimensions of economic transformation, and finally, historical context and country-specific idiosyncrasies have had on opin-ions about democracy and its alternatives.

In their presentation of cross-national survey data, the authors go far beyond the usual comparisons of apples to oranges. All too often, researchers limit their discussion of data collected in comparative studies of postcommunist countries to the cross-national comparison of frequen-cies and means, without any regard to historical, cultural, social, economic, or political context. By using multiple regression analysis (a relatively simple statistical procedure), the authors were able to demonstrate not only the intensity of particular opinions, but also (on a regionwide level) the complex web of factors shaping and influencing these opinions. Their conclusions are moderately optimistic. Democracy seems to have taken root in East-Central Europe, and to have proven more attractive than its alternatives. The Churchill hypothesis has been conditionally confirmed. I say "conditionally" because of the variance among the cases: Some countries in the NDB sample have already become consolidated democ-racies, while others have much further to go, and remain merely the "lesser evil" in comparison with nondemocratic alternatives. One country, Bela-rus, has slipped back into outright dictatorship since the study was completed.

According to the conventional wisdom ("It's the economy, stupid!"), popular support for the new regimes in East-Central Europe depends on their success in building a market economy. This relationship is usually expressed in a negative form: poor economic performance (or even just the short-term hardships of transition) undermines confidence in new political regimes. Yet Rose, Mishler, and Haerpfer show that, although economic considerations do count in the popular assessment of the new order, "politics matter more" (p. 176)--more, that is, than opinions on the economic situation (both of the household and of the national economy) or [End Page 170] "hard" economic indicators like income or deprivation. The authors also found that structural variables such as education, urbanization, gender, and age--the backbone of sociological theories of political change--are poor predictors of support for the new regime. Only education is relatively strongly--and positively--associated with rejection of nondemocratic alternatives.

A closer look, however, at the table (p. 193) in which these conclusions are summarized raises some interesting questions. First, the two dependent variables, support for current regime and rejection of nondemocratic alternatives, are explained by different sets of independent variables. Although both dependent varia-bles are influenced by political opinions, rejection of nondemocratic alternatives is also influenced by social and economic structural factors like education, urbanization, deprivation, and income, while assessment of the economy has no significant impact. By contrast, in the case of support for current regime, structural factors don't count, while economic opinions matter even more than political opinions. Attempting to explain that disparity would take us beyond the intended scope of this book (and, needless to say, this review).

Second, explaining the dependent variables, which themselves are merely political opinions, in terms of other political opinions (evaluation of communist regime, more perceived freedom now, patience with new regime, new regime perceived fairer, etc.) dangerously borders on tautology, and begs a more interesting question. If the structural and contextual factors do not matter much, how can we explain the very real differences in political opinions and attitudes documented in this book and elsewhere? Why is it that some well-educated individuals were members of the communist party and today support--or even lead--its postcom-munist successor organizations, while others were, and remain, staunch anticommunists? Why do some workers vote for socialists or social democrats, while others vote for Christian democrats or nationalists? Why do some pensioners long for the social safety net of the communist past, while others are happy to accept austerity and suffering as the price for being a part of the new order? This is an issue that certainly deserves further inquiry, but extends beyond the limited scope of public-opinion surveys.

This book is more than just another comparative public-opinion project, or another exercise in comparative transitology. Perhaps more than any other academic discipline, political science faces a challenge stemming from the need to incorporate data on and from East-Central Europe into the body of scientific evidence available to scholars attempting comparative [End Page 171] cross-national analyses. Such analyses, so popular and important in mainstream political science, rarely include evidence coming from Central Europe. For example, the immense literature on elections or on institutional design of democratic regimes is based on data coming almost exclusively from Western Europe and the Americas, sometimes with the addition of Southeast Asia and Oceania. A decade after the establishment of democratic regimes in Central and Eastern Europe, evidence from this region is still usually ignored, and the rare references that appear in the literature are often incorrect. Democracy and Its Alternatives is a serious attempt to fill this gap. The project is well designed, the data are analyzed in a creative and disciplined manner within the context of the broader literature on democratic theory, and the conclusions are convincing. This is, arguably, the most solid comparative empirical study of postcommunist transitions to date.

Like democracy itself, however, this book is not perfect. There are some flaws that could have been easily avoided with better research and editing. In providing general statistics on the region, the authors seem to rely too much on general handbooks (Eurostat, Political Handbook of the World, 1995-1996, etc.), when more detailed and reliable information is available in specialized publications (like those issued annually by OMRI or the East-West Institute). Consulting the latter would have helped to avoid lapses like those found in the data on ethnic minorities in Figure 6.2: For Poland, only Germans and Jews are listed while the much more numerous Byelorussian and Ukrainian minorities are not; and for the Czech Republic, the minorities listed include Moravians, who differ from Czechs no more than, say, Bavarians do from Germans.

Second-hand citations are another flaw spoiling the book's elegance. While I admit that Ken Jowitt's analyses are not only penetrating, but also a good read, I see no reason to cite Karl Marx from Jowitt's work (p. 144). I find it hard to believe that no copy of The Eighteenth Brumaire could be found in the libraries and bookstores of Glasgow, Vienna, Tucson, or Baltimore.

My more serious reservations, however, concern methodological issues. As I have already indicated, the authors' methodology is generally sound, and they should be praised for their conscious efforts to avoid sweeping generalizations. Their analysis uses the "tenth country" model, based on the mean values calculated for respondents from all nine countries, but they do not ignore national peculiarities. They not only report the frequencies and means for particular countries, but also introduce country-specific and contextual variables to their explanatory models. The problem is that they offer only regionwide models to explain the origins of opinions about democracy. The variables used in a model can vary from case to case not only in their intensity (which is obvious--otherwise they would be constants), but also in their mutual relationships. For instance, religiosity may be related to support of a nondemocratic [End Page 172] alternative in a positive way in country A and in a negative way in country B, while it may be irrelevant in country C. These differences from country to country can be obscured, however, if the analysis is performed only on aggregate regional numbers. Such analysis is likely to produce "mean"--and therefore weak--relationships among variables. I suspect that is why many of the reported relationships in this book barely surpass the accepted levels of statistical significance. I would be most interested in seeing the optimal explanatory models for each country separately.

To some extent, this criticism is offset by the report that regression analyses were also performed using countries as dummy variables (p. 191), and that the general model was tested independently for each country (p. 194), without any striking results. Not surprisingly, the two most deviant cases were Belarus and Ukraine. This observation prompts another question: How would the explanatory model have changed if these two former Soviet republics were dropped from the sample? I suspect that the Churchill hypothesis would have found a more forceful confirmation. Yet would it still stand if the sample consisted only of Belarus and Ukraine? If the confirmation of a hypothesis depends on the composition of the aggregated sample, why not look at each case separately? The conclusion that support for democracy depends positively on religiosity in country A, but negatively in country B, is more meaningful than the conclusion that there is no relationship between these two variables in countries A and B taken together.

I am also a bit baffled by the authors' choice of contextual variables. One of them, "democratic traditions," lists the Czech Republic and Slovakia as the only positive cases. One might question whether the Czechoslovak experience of parliamentary democracy between the wars is more relevant for current attitudes than, say, traditions of democracy among the nobility in Poland, but my main point is a different one. Why include this variable and exclude another so often mentioned in historical-comparative analysis: The great divide between Western and Eastern Christianity? Samuel Huntington and others make a great deal of this variable, so why not test it?

Fortunately, opportunities to do so are not precluded. The NDB data are available in the public domain. The excellent website (http://rs2.tarki.hu:90/ndb-html) provides easy access to frequencies and simple cross-tabulations for all variables and all countries covered in the annual surveys from 1991 to 1995. It is my understanding that interested researchers may also obtain the raw data files from the Paul Lazarsfeld Society. If some readers are not satisfied with the job Rose, Mishler, and Haerpfer did, let them try to do better.

Krzysztof Jasiewicz is professor of sociology at Washington and Lee University and research fellow at the Institute of Political Studies of the Polish Academy of Sciences.

[an error occurred while processing this directive]