Revisiting Organizational Legitimation Cognitive Diffusion and Sociopolitical Factors in the Evolution of Bulgarian Newspaper Enterprises, 1846-1992 – Statistical Data Included
Stanislav D. Dobrev
Abstract
Arguments about constitutive and sociopolitical legitimation of organizational forms are applied to studying the evolutionary dynamics of the Bulgarian newspaper industry as it transitioned through multiple political and institutional environments. The notion of re-legitimation is advanced in the context of comparing the cognitive diffusion of the organizational form prior to the Communist takeover in 1946, and its revival in the collective memory of the public after 1989. Variation in the rate of organizational founding is also linked empirically to the strength of political turbulence. Violent political conflict deters new foundings, while institutionally mediated political activism has a positive effect on the rate of organizational entry.
Descriptors: organizational ecology, density dependence, types of legitimation, organizational founding-rate analysis, Bulgarian newspapers
Introduction
Research on density dependent organizational evolution has emphasized the population-level mechanisms by which new forms become legitimated (Hannan and Freeman 1989; Hannan and Carroll 1992). Although researchers agree that growth in the number of organizations belonging to an emergent population drives the institutionalization of the new form, it has been largely assumed that such processes of constitutive legitimation do not differ substantially among organizational forms. At present, little attention has been given to the fact that some organizational forms reach a taken-for-granted status more easily than others. One reason for the inattention to possible variance in constitutive legitimation has to do with the fact that such variance is less likely to be observed in the evolution of organizational forms in industrialized societies that have experienced a relatively uninterrupted pattern of political and economic development. However, a conceptual tool to investigate the multifaceted nature of legitimatio n might be necessary to ensure contextual accuracy and proper definition of organizational forms in societies with tumultuous environments.
It has been widely acknowledged that the legitimation of a new form is a combined function of its persisting presence on the bureaucratic landscape which helps to disseminate the image of the form (constitutive legitimation), and the extent to which it gains acceptance by powerful social actors (sociopolitical legitimation) (Barron 1998; Carroll and Huo 1986; Aldrich and Fiol 1994). Achieving and sustaining government endorsement is obviously highly contingent on the level of environmental stability and political turmoil. Attempts at explaining the diversity of organizational forms in environments characterized by long-term political instability must necessarily incorporate an analysis of the interaction between the population and its broader environment. According to the ecological concept of legitimation, the sheer presence of an organizational form increases its public and institutional acceptance until it eventually becomes synonymous with the social action it represents. At the same time, environmental f orces have the capacity to reshape the resource niche of the population and in this way promote or impede the public acceptance of an emerging form.
In this research, I focus on examining how the intensity of political turbulence affects the rate of organizing. I rely on quantitative analyses of empirical data from the Bulgarian newspaper industry to develop and test propositions about the relationship between political activism and new organization building. I also examine the ways in which sociopolitical legitimation impacts the cognitive diffusion of an organizational form under conditions of extreme exogenous volatility. Less than 150 years after its inception, the Bulgarian newspaper industry has undergone multiple reorganizations that followed from the cataclysmic changes in its external environment. Between 1846 and 1992, Bulgaria was a stateless territory, a country split between two legally independent states, a quasi-independent Soviet satellite, part of an Empire, part of a Bloc, a kingdom, a people’s republic, and a republic. Its sociopolitical structure transitioned through: feudalism, monarchism, fascism, dictatorship of the proletariat, ‘ma ture’ socialism, and post-socialism. The peculiar evolution of the newspaper industry differs substantially from the patterns observed in most previously reported ecological studies and provides an opportunity to test the applicability of the density dependence model as well as to elaborate on its conceptual framework.
Theory and Hypotheses
Sources of Variance in Constitutive Legitimation
The ecological notion of legitimation, derived from the institutional theories (Meyer and Rowan 1977; Meyer 1983; Meyer and Scott 1983), refers to the process by which a new organizational form gains public acceptance. This process entails the formation of social and cultural rules that define the form’s identity until the form itself becomes an institutionalized blueprint for organizing and conducting social action. Unlike legalization, which has specific bureaucratic-legal implications explicitly stated in a regulatory framework, legitimation is a broader social phenomenon that takes time to develop. Its dynamics are hard to observe directly. While substantial empirical evidence exists that strongly supports the pattern implied by the ecological notion of legitimation (for a review, see Carroll and Hannan 1999), little is known about the ways in which achieving ‘taken-for-grantedness’ may vary among different forms.
At least two factors, related to differences in process and content, are worth examining further: First, horizontal integration with other already established forms can provide a major boost for organizations with emergent design and unrecognized statement of purpose by minimizing the risk associated with entering market transactions with such firms. Other social actors, collective or individual, invariably face a great deal of uncertainty when interacting with new organizations that lack prior track record. This problem is greatly alleviated, however, if the actions of those unaccustomed entities are rendered some institutional legitimacy derived from a mutualistic relationship with established organizations.
Second, in addition to the level of association between an emerging organizational form and an already established population, legitimation may connote different meanings depending on the historical circumstances upon which a new population makes its entry. A new organizational form may be one that has never existed before, but it may also be one that once existed and subsequently became extinct. Arguably, an organizational form that is resurfacing on the organizational landscape can claim a past institutional heritage by invoking landmark historic experiences. As a cognitive process, the difference between legitimation and re-legitimation may resemble the difference between learning and remembering. For obvious reasons, it is plausible to expect that re-emerging populations (that accrue taken-for-grantedness through re-legitimation) will benefit from relying on some existing pillars of institutional acceptance and will have an advantage over brand new populations.
The possible variations in type of legitimation are summarized in Table 1. Based on process differences, we draw a distinction between organizational forms affiliated with other populations (de alio, from Latin, meaning literally ‘from another’), and those that strive for institutional legitimacy on their own (de ipso, ‘from itself’). On the content dimension, de novo legitimation (‘from new’) involves the acceptance of a brand new form, and de antiquo legitimation (‘from previous’) captures the revival of a form that had once been a part of the organizational landscape. Cross referencing these two dimensions produces four types of legitimation that differ from one another in both content and process.
Undoubtedly, these four types have a highly uneven distribution of organizational legitimation. In fact, most previous research on the institutionalization of organizational forms is type A. New populations evolve by gradually asserting themselves, typically in an emergent market or in conjuncture with the formation of a new industry. Exogenous factors such as economic decline, government (de)regulation, technological advancements, social movements, and normative and aesthetic trends are powerful forces that have the capacity to alter the evolutionary course of an industry (Carroll and Hannan 1995; Dobrev 2000). However, it takes an event of historic precedence such as a political revolution or a scientific breakthrough, for an organizational form, once institutionalized, to become extinct. The resurfacing of an extinct organizational population (type B legitimation) rarely happens for reasons other than a dramatic reversal in the political discourse.
The history of the US beer brewing industry, and the period during and after Prohibition, provide an interesting case where the circumstances could be taken to resemble the operation of type B (de antiquo and de ipso) legitimation (Carroll and Wade 1991; Carroll et al. 1993; Carroll and Swaminathan 1992). Assuming the population went through a process of re-legitimation after the end of Prohibition, an argument can be made that a new population emerged de antiquo. However, the distribution of population density before and after Prohibition, which clearly displays almost an uninterrupted pattern, suggests that the organizational form never became extinct, but merely adapted to a sub-rosa operation. After the repeal of Prohibition, density rebounded to the number projected by the earlier trend in population density, almost precisely where it would have been, had Prohibition never occurred.
The characteristics of type C (de novo-de alio) legitimation suggest that it should have a higher incidence rate as it merely describes a particular industry pattern. Symbiotic relationships among populations are not surprising, especially early in an industry’s development. Indeed, such relationships have been documented and researched in the cases of the semiconductor and the life-insurance industries. Hannan and Freeman (1989) found that the growth in density of independent semiconductor firms played a role in the legitimation of subsidiary firms. Similarly, Hannan and Carroll (1992) discovered evidence showing that the legitimation of stock life-insurance companies was facilitated by the earlier existence of mutual life-insurance companies. If symbiotic relationships develop among entities with similar form and function, it is plausible that, contingent on environmental conditions, the same type of relations between separate populations can form and prosper as well.
Finally, type D legitimation refers to a process whereby an extinct organizational form is revived through a symbiotic relationship with another population. Inasmuch as de antiquo legitimation would probably occur as a consequence of reversal in the political discourse (namely, when a regulatory ban on a specific economic transaction related to the core function of an organizational form is lifted), a resurgent population might benefit from its connectedness to another population and quickly rebound to its previous level of public acceptance.
In some research cases where tumultuous environments have interfered with the natural organizational processes (as in the case of the Bulgarian newspaper industry), the legitimation taxonomy is a valuable conceptual tool that enhances the substantive validity of research on organizational evolution by eliminating problems arising from ahistoricism and contextual imprecision.
Density-dependent Legitimation and Competition
The standard density dependence model treats the rate of founding in the population as directly proportional to constitutive legitimation, and inversely proportional to competition:
[lambda](t)[varies][L.sub.t]/[C.sub.t]
Ecologists have suggested that at low counts, population density drives legitimation and thus can serve as an observable variable to measure legitimation. The addition of every new organization to the population contributes to spreading its stated goals and purpose, but the contribution of each new founding varies across density. At very low density, the effect is very strong, and it decreases with increasing density and eventually wears off as the form becomes institutionalized. In other words, population density increases legitimation at a decreasing rate.
The ecological conception of competition differs from its mainstream social science counterpart in that it does not treat tangible social relations as a prerequisite to the build-up of competitive forces (Carroll and Hannan 1999; Dobrev and Carroll 1999; Hannan and Carroll 1992). Instead, diffuse competition takes effect at the point at which the number of existing organizations exceeds the carrying capacity of the population. Thus, two organizations are considered potential competitors as long as they depend on resources from the same niche, even though they may not even be aware of each other’s existence. When the number of existing organizations thriving on resources from the same niche is small, and the boundaries of their population unsettled, competition is of little importance. But as population density grows, competition for resources intensifies with an increasing propensity. At high levels, population density escalates competition at an increasing rate.
Expressing legitimation and competition as particular parametric representations in terms of density yields the generalized Yule model:
[lambda](t) = [k.sub.[lambda]] (t)[[N.sup.[alpha]].sub.t] exp([beta][[N.sup.2].sub.t]), 0 [less than] [alpha] [less than] 1 and [beta] [less than] 0,
where the first-order density effect captures the positive effect (0 [less than] [alpha] [less than] 1) of legitimation on the rate of founding, and the second-order density effect represents the negative effect of competition ([beta] [less than] 0) on the rate. Overall, the relationship between density and the population’s rate of founding has an inverted U-shape with legitimation dominating lower density and competition expressed as a function of high density. Following the generalized Yule model, the hypotheses are straight forward:
H1a: A rise in density at low density levels will increase the rate of founding at a decreasing rate.
H1b: A rise in density at high density levels will decrease the rate of founding at an increasing rate.
Previous research on populations of newspaper organizations has provided strong support for the theory of density dependence (Carroll and Hannan 1989a 1989b; Olzak and West 1991). Note that H1a will hold true regardless of whether legitimation is of type A, B, C, or D, because the only possible outcome of the process by which organizational forms become embedded in the relevant social and market structures is the eventual accrual of taken-for-grantedness. Once attained, legitimacy becomes an essential and indivisible part of the form’s identity. Introducing a typology of constitutive legitimation aims to improve the contextually accurate modelling of density dependence and does not modify the theory in any way.
Organization Building and the Political Environment
According to institutionalists and ecologists alike, sociopolitical legitimation operates independently of constitutive legitimation. Regardless of the type of legitimation involved in the diffusion of an organizational form, its evolution is affected by large-scale institutional and political developments. Although, in general, scholars agree that vital rates of organizational populations are considerably influenced by such external developments, arguments exist regarding the extent to which strength, variability, and direction of effect change.
An interpretation of the relationship between the rate of organizational foundings and periods of political upheavals (Aldrich 1979: 69) suggests that ‘externally induced changes in the nature of environmental selection criteria’ confuse potential venture founders by impairing their ability to foresee future market developments. Political instability is likely to lead to economic recession, which in turn diminishes the willingness of entrepreneurs to commit resources and invest capital when the promise of future returns is weak. In ecological terms, unstable environments lower the population entry rate.
This argument is countered by a different stream of thought which suggests that environmental transformations inevitably lead to the addition of new organizations (Stinchcombe 1965). In his research on several newspaper populations, Carroll (1987) found that political turbulence significantly increased the rate of founding. His interpretation, based on previous research by Tilly (1978), implies that newly founded newspapers appear to fulfill the needs and tastes of newly formed social groups emerging in response to environmental restructuring.
The common thread between these two arguments lies in their general treatment of type and degree of political turbulence. Variation in the intensity of environmental restructuring will probably have a non-linear relationship with the rate of founding. A specific case of this argument was developed and tested by Olzak and West in their research on the effect of racial conflict on the vital rates of ethnic newspapers (1991). They found an inverted U-shape relationship between the intensity of such conflict and the founding rate of ethnic newspaper organizations. If these findings were generalized, they would suggest that institutional politics (low-scale political turbulence) will increase the rate of entry to the population, whereas political turmoil (high-scale political turbulence) will decrease it.
Previous research suggests that both types of political turbulence would result in an increased number of foundings. For example, Hannan and Freeman conclude that ‘the rate of organization building will rise during periods of political crisis and transformation’ (1989: 127). According to this reasoning, the proposition holds true both in the case of social revolutions and when response to political and social crises is mediated through the institutions. Other ecological research has acknowledged the different effects of violent and institutional political turbulence on the vital rates of organizational populations. In his study of newspaper enterprises in Ireland, Argentina, and San Francisco, Carroll (1987) found support for the argument that political turbulence systematically increases the founding rate, whereas regime changes do not significantly affect it. Similar results showing that, unlike institutional politics, political turmoil does affect vital rates were reported by Carroll and Huo (1986), Delacr oxic and Carroll (1983), Carroll and Delacroix (1982), and Carroll and Hannan (1989b). The exact opposite argument based on the model developed by Olzak and West is tested by the following hypotheses:
H2a: institutional politics will increase the rate of founding.
H2b: Political turmoil will decrease the rate of founding.
Thus formulated, these hypotheses reconcile Stinchcombe’s idea that environmental reshuffling stimulates the entry of new organizations (H2a) with Aldrich’s notion that serious environmental shake-ups deter entrepreneurial activity (H2b).
Overview of the Bulgarian Newspaper Industry
Around the mid-19th century when the first newspaper appeared, Bulgaria was a territory on the map of the Ottoman Balkans. The lack of essential political freedoms, such as freedom of speech, and minimal resources at that time had profound consequences on the early years of the industry. Scarce availability of readers (the majority of Bulgarians at that time were uneducated), the large gap between literary Bulgarian language and the plethora of locally spoken dialects, the unavailability of print shops capable of printing the Slavic alphabet, and the lack of demand for up-to-date news coverage due to the nature of economic relations in the country were all major impediments in the early development of the industry. Feudal relations were deeply entrenched in 19th century Bulgaria, and the practical need for information on nearby trade activity — a need contributing to the emergence of the newspaper in Western Europe — was an insignificant factor. The emergence of Bulgarian newspapers was less a response to a n existing economic opportunity than a deliberate attempt to awaken the collective consciousness of the nation.
Most of the early attempts to establish a newspaper enterprise, including the first newspaper and the first daily, were undertaken abroad. The difficulties faced by pioneering entrepreneurs were far greater than the usual start-up costs. A heavily underdeveloped infrastructure made the building of a sales distribution network unrealistic. In addition, since the industry was not formally legalized or regulated (the first Press Law was not passed until 1881), it was not possible to sanction delinquent subscribers (Andreev 1946). Indeed, at this early stage of the industry, newspapers were not conceived as economic profit-making enterprises, but as propellers of the national self-consciousness.
After the country regained its independence in 1878, economic growth in Bulgaria led to a rapid expansion of the industrial infrastructure; in the first half of the 20th century, the development of the Bulgarian press was an integral part of the growing national economy. The adoption and implementation of current technological developments contributed significantly to the quick evolution of the newspaper organization. By the time of the Balkan wars in 1912, the newspapers had turned into profit-making capitalist enterprises. In 1914, the for-profit enterprise ‘Courier’ was founded to handle the sales and distribution of all periodicals published in the capital city, Sofia. Meanwhile, the journalists there had already established the Organization of the_Journalists from the Capital, followed, in 1925, by the Union of the Provincial Journalists in_Bulgaria (Nikolov 1932). The first news telegraph agency stationed in Sofia was founded in 1911 under the name The Balkan Agency and helped to further consolidate the industry. The number of Bulgarian newspapers grew from 33 in 1890 to 117 in 1900, and 868 in 1933 (Figure 1).
Bulgaria’s inability to chart its own political and economic fate during and immediately after World War II led to its annexation to the Soviet Bloc. The communist regime did not merely institute strict censorship on the newspapers operating at the time of the regime takeover. It shut down all newspapers except for the few that sympathized with the socialist party, and even they were completely re-organized. After the forceful nationalization completed by the regime in 1948, the overlap between party and state virtually denied everyone but the communists the legal authority to publish newspapers. The communist party became the monopolist of all material resources: commercial paper, publishing and other physical equipment, sales and distribution network, etc. New editions were founded on the basis of administrative regulations rather than readers’ interests and preferences. Although the 1960s and 70s were marked by relative political liberalization, the 1980s saw escalating political confrontation and the suff ocating state-planned economy which led to tighter control over the press and created a decline. In 1989, the number of newspapers dropped to 267 — almost half of the 456 operating a decade earlier (Figure 1).
The weakening of the communist regime did not become apparent until the fall of 1989, when the smeared-letter print-outs posted hastily on the building walls in downtown Sofia turned out to be the early harbinger of the independent press in post-socialist Bulgaria. The formal legalization of the anti-communist opposition in the December of 1989 paved the way for publishing non-communist editions. With the state’s continued monopoly over the rest of the mass media, the press became the most exhaustively and sometimes excessively used public forum for political debate. It is not surprising that, under such conditions, political newspapers comprised most of the new foundings. On the one hand, political newspapers were the means through which new political figures and agendas were introduced and legitimated to the constituents; on the other hand, there was a high public demand for them (Naidenov 1995a). More than a year after the founding of the first independent newspaper Svoboden Narod (Free People) in February 1990, political newspapers continued to be the dominant form in the industry. In fact, most newspapers during that year were founded as an organizational extension of yet another newly founded political party or coalition.
The tight coupling between political parties and their official newspapers had profound consequences on the further development of the industry. During the second half of 1991, the degree of social politicization eventually decreased, and public exaltation with the political discourse gave way to growing concerns for economic and social welfare. Most political newspapers found it hard to make a transition from serving party propaganda to serving the needs of their readers and functioning as profit-making enterprises. The strong dependence between politics and journalism, which prevented the newspapers from developing as fully independent entities, produced internal conflicts that broke many of the early enterprises. In one year, Svoboden Narod, one of the most prominent dailies of 1990, was reduced to an unimpressive weekly edition with less than 10 percent of the circulation of the year before.
Attempts to make the transition to a market economy by applying shock therapy, without having passed the appropriate legislation in advance, only worsened the deep economic crisis. The 2000-percent increase in the price of commercial paper in January 1991 (Velinova 1992) raised an almost insurmountable barrier to entry for many potential entrepreneurs. In the absence of an appropriate credit system to provide entrepreneurs with fair access to resources, the formation of a private sector did not entail market competition. Moreover, while the state continued to have an almost exclusive ownership of the means of newspaper production, taxes applied to the private sector of the Bulgarian press were higher than in any other European country (Naidenov 1995b). The exorbitant interest rates imposed by the central bank virtually denied small- and medium-sized private owners the opportunity to borrow money to cover start-up costs (Alfandari 1995). The magnitude of political, economic, and cultural changes in the environ ment of the Bulgarian newspapers has eroded the population’s natural course of development and has produced dramatic disruptions that make it unrealistic to speak of one population of newspaper organizations. As suggested by Polos et al. (1998), forms are cultural rules based on social identities. Undoubtedly, the cultural rules enforcing the model for organizing a newspaper enterprise differed substantially in the period before and after the communist takeover. Further, as Hannan and Freeman acknowledge, ‘periods of political crises and social revolution seem to be peak times for building new forms of organizations’ (1989: 126). This observation clearly applies in the case of Bulgarian newspapers and suggests the definition of three organizational populations: pre-socialist (1846-1948), socialist (1949-1989), and post-socialist (1990 and after). Thus, understanding the historical landmarks that guided the evolution of the industry not only enhances contextual interpretation of events, but also guards against the inadequate specification of organizational forms (Baum and Powell 1995).
Research Design
Data
I used a great variety of sources to collect information on the foundings and density of Bulgarian newspaper organizations dating back to the start of the industry in 1846. Because a law requiring the cataloging of newspapers was not passed until 1891, I have relied on several historical studies of the early years of the industry (Ivanov 1893; Bobchev 1894; Nikolov 1932; Andreev 1946; Topencharov 1963; Borshukov 1957). Since information about the number of newspapers in this early period varies significantly among different reports, I counted all newspapers that were mentioned at least once in any one of these publications. The data that cover the period from 1892 onwards was collected by counting newspapers listed in the annual directories Bibliographicheski Buletin (Bibliographical Bulletin) and Bulgarski Periodichen Pechat (Bulgarian Periodicals), both of which contain information on the founding and failure of each newspaper. I was able to recover all missing data by using the national annual statistics v olume and the directory Knigoizdavane I Pechat (Book-Publishing and Press) published by the Central Department of Statistics. The large variability of sources that provided information on founding events and the scarcity of information for specific years necessitated that the time of founding be made specific only to the year.
To estimate the founding rates of the three populations (pre-socialist, socialist, and post-socialist) separately, I split the data accordingly. Estimating separate models for the three populations also resolves the problem arising from the possibility that the effects of density are not symmetric over time. Fortunately, since more precise data were available for the period after 1989, it became possible to determine the timing of founding events specific to the month. The dataset containing the entries for new foundings of Bulgarian newspapers covers the period from January 1990 to September 1992. The number of records in this datafile corresponds to the 33 months that comprise the observation period. The main source for the data collection was the book 1000 Vestnika (1000 Newspapers), which is a complete directory of all newspapers that at least made an attempt to establish a newspaper organization, even though such an attempt proved to be unsuccessful. That is, I include newspapers that published at least a pilot issue — evidence that they intended to continue production. However, I exclude all newspapers that appeared only once or twice with the sole purpose of commemorating a specific event or that were designed as election campaign bulletins. Having cleaned and reorganized the data, I counted a total of 734 new foundings for the period of interest. 1000 Vestnika contains information about the time of founding and on whether the newspaper had ceased to exist (including the time of disbanding) or whether it was still in production at the end of the observation period.
I also used the 1990, 1991, and 1992 volumes of Bulgarski Periodichen Pechat and Knigoizdavane I Pechat to gather information about the number of state newspapers (i.e., those founded prior to 1990) that continued to exist until the end of 1992. These additional data were necessary in order to calculate total population density. The data for the density of political parties and organizations were drawn from two directories (Danevski 1990; Todorova 1991) and a detailed chronology of the events that occurred on the political arena between 1988 and 1991 (Jackowicz 1992).
Methods
Following Hannan and Carroll (1992), I conceive of organizational founding as an arrival process. In this formulation, each instance of founding is equivalent to the arrival of a given unit to some system of components similar to the arriving unit. Thus, the instantaneous rate of entry ([[lambda].sub.t]) is expressed as:
[[lambda].sub.y](t) = [lim.sub.[delta]t[right arrow]0] Pr{Y(t + [delta]t) – Y(t) = 1Y(t) = y}/[delta]t,
where Y(t) is a random variable denoting the cumulative number of foundings by time t, [[lambda].sub.y](t) is the rate of arriving at state y + 1 at (or just after) time t, and Pr{*} is the probability that an instance of organizational founding will occur during the period [delta]t.
Three different methods are typically used to estimate event count models. The most common approach is to use Quasi Likelihood (QL) estimation when modelling the founding rate of organizational populations (Hannan et al. 1995; Torres 1995; Han 1998). Unlike Poisson regression, QL estimation allows the addition of a stochastic error term required to control for overdispersion; unlike negative binomial regression, the QL method does not require an assumption about a specific form of the distribution. In fact, the only requirement in QL estimation is that the relationship between the mean and the variance of the entry process be specified. I assume that this relationship is linear:
Var([Y.sub.t]) = (1 + [delta]) E ([Y.sub.t]).
I also assume that the mean of the process is equivalent to the entry rate:
E ([Y.sub.t]) = [[lambda].sub.t].
Provided the function of the mean has been specified correctly, the QL estimators are consistent, asymptotically Gaussian, and robust (McCullagh 1983). I use the statistical module for QL estimation in Gauss written by David Barron.
Following prior research on organizational founding that relies on QL estimation, I use Haberman’s chi-square statistic for model comparison (Haberman 1977). This statistic is analogous to the likelihood ratio test statistic and has been shown to have an asymptotic chi-square distribution, even when expected counts are small (Agresti 1990).
I estimate founding models with the following general form:
[[lambda].sub.t] = [theta]([N.sub.t], [n.sub.t]) exp([pi]'[x.sub.t] + [varphi.sub.p])[[epsilon].sub.t],
where [theta]([N.sub.t], [n.sub.t]) is one of two possible specifications for the density dependence model of organizational evolution, [x.sub.t] is the vector of covariates and dummy variables, [pi]’ indexes all relevant period effects, and [[epsilon].sub.t] is the stochastic error term.
Variables
The dependent variable, the ‘rate of entry’, is defined as a function of the sum of all organizational foundings in the newspaper industry at any point during the year (or month, in the case of post-socialist foundings). Population ‘density’ is measured as the total count of newspapers that existed during any given year, whereas in the foundings analysis of postsocialist newspapers, the same measure is taken on a monthly basis. All product-level information is aggregated to the level of the organization. Density measures are lagged by one observation period. The same rules are used to calculate the density of political parties and organizations in 1990 and 1991, a variable used to model interpopulation density dependence.
Two dichotomous variables capture major events in the political discourse: ‘political turmoil’ has the value of one in years when various political crises have spurred militant conflicts such as large-scale uprisings and wars, whereas ‘institutional politics’ distinguishes years when national elections resulted in government changes, as documented in Bulgarian History (Bozhilov et al. 1993) and Bulgaria 1879-1946: The Challenge of Choice (Kostadinova 1995). Adding these two variables produces a dummy variable measuring ‘political turbulence’. This definition of political turbulence includes election periods and creates a broad variable that refers to ‘a state of tension and action affecting the distribution of power in society’ (Tilly and Rule 1965: 4). Despite Tilly and Rule’s suggestion that such a definition is too general, it is appropriate for the purpose of the analysis presented here, especially considering the scant availability of historical data. In the founding analysis of post-socialist newspapers , the variable ‘election’ refers to the three months immediately preceding national parliamentary or presidential elections. Effect coding (a modelling technique based on overlapping dummy variables) is used to compare eleven period effects. The period between 1846 and 1878, during which Bulgaria was still part of the Ottoman Empire, is omitted. The interpretation of the milestone events in the history of Bulgaria of the past two centuries is based on an in-depth reading of a comprehensive volume entitled Istorija na Bulgaria (Bulgarian History). The volume was published in 1993, four years after the system of party censorship was dismantled, and thus represents the first truly academic effort to examine the sociopolitical developments in the country from the second half of the 20th century. In the founding analysis of state-socialist newspapers, the variable ‘number of TV licences’ refers to the number of households owning a TV set (PRB: Annual Statistical Directory) and is included as a control variable capturing the emerging resource competition between the TV and newspaper industries.
Results
I begin the empirical analysis with a test of the effect of political environment on the rate of organization building by modelling the founding rate of newspaper organizations over the whole range of the industry. Eleven period effects in the baseline model (Table 2, Model 1) capture the major milestone events in the history of the country that paralleled the evolution of the newspaper industry. Consistent with H2a, the coefficients show that periods of widespread non-violent social movements, such as the one for unified statehood (after the disintegration of the Ottoman Empire in 1878, the former Bulgarian province was split into two autonomous kingdoms, Bulgaria and East Romelia, which were united in 1885) and for democracy and liberal reforms after the collapse of the communist regime, were by far the most influential in increasing the founding rate. Periods of economic growth and political stability in the early 20th century and prior to World War II also increase the rate. Not surprisingly, periods of a uthoritarian and despotic governing significantly decrease the number of entries to the population.
Model 2 includes a covariate (which improves model fit significantly: Haberman’s [X.sup.2] = 92.22; [delta]d.f. = 1; p[less than].01) measuring the effect of political turbulence indiscriminate of whether it is institutional or turmoil. In accordance with earlier findings (Hannan and Freeman 1989), its coefficient is positive and highly significant. Models 3 and 4 aim to distinguish between political turmoil and institutional politics. The results offer strong support for H2a and H2b: the coefficients are in the predicted direction, and the effect of institutional politics is highly significant. Directly compared to Model 1, Model 4 provides an improvement in the fit of the data relative to Models 2 and 3 (Haberman’s [X.sup.2] = 184.17; [delta]d.f. = 1; p[less than].01).
Model 5 presents results from the generalized Yule specification of the density dependence model. Note that, as in any prior research on density dependence, I assume that type A legitimation was the driving force behind the process that led to the cognitive diffusion of the organizational form. This model includes the first- and second-order effects of the lagged number of foundings along with the two density terms. The results support the argument of rate-dependence (Delacroix and Carroll 1983), and both density terms are highly significant and in the predicted direction, suggesting that, controlling for the effect of sociopolitical factors, low counts of density triggered legitimation and high count density drove competition. Including the prior foundings and density terms significantly improves the fit of the data (Haberman’s [X.sup.2] = 1084.58; [delta]d.f. = 4; p[less than].01).
As suggested in the previous section, estimating a single effect of density over the whole history of the industry is implausible because it violates the scope conditions of the density dependence model (the theory applies to industries where a pattern of relative growth in density is observed). As is evident from Figure 1, density declined sharply in the 1940s, experienced a peak in the 1960s, and again declined in the 1980s, only to resurge again in the early 1990s. Such a complicated pattern clearly points to a process other than the early evolution of a single organizational form. The density dependence theory also only makes sense for organizations operating in market environments, and thus should not be used to explain the dynamics of state-socialist newspaper enterprises. Consequently, the results from Model 5 in support of the density dependence model are either substantively implausible or, more likely, agree with the theoretical predictions, because they actually apply to the population of pre-socia list newspaper organizations but spill off into the next two, much more short-lived, populations. This possibility is tested next.
Results from modelling the rate of pre-socialist Bulgarian newspaper organizations (1846-1948) are presented in Table 3. Again, the best fitting baseline model is the one including institutional politics rather than political turmoil, or the more general variable reflecting political turbulence. In accordance with H2a, the effect is positive and highly significant. The period effects also agree with estimates from the model based on the complete dataset. The addition of the covariates for the plain and squared terms density in Model 2 offers strong support for H1a and H1b — low counts of density increase the founding rate and high counts decrease it. Further, including the density terms significantly improves model fit (Haberman’s [X.sup.2] = 122.28; [delta]d.f. = 2; p[less than].01). Support for H1a confirms the operation of type A (de novo–de ipso) legitimation.
Next, I test the effects of covariates on the entry rate of state-socialist newspaper enterprises (1949-1989). The baseline model (Table 4, Model 1) includes two period effects, lagged foundings, and the resource competition variable measuring the number of TV licences in the country. Only the latter variable has a significant (negative) effect on the rate, suggesting the existence of diffuse competition for resources between the TV and the newspaper industries. The two terms for population density are included in Model 2, and not surprisingly, their effects are insignificant. During this period, centrally drafted and enforced government regulations met little resistance in overriding any natural process of cognitive legitimation or market resource competition. According to the density dependence model, density controls legitimation and is not merely an indicator of it (Hannan and Carroll 1992: 69), but in the case of the population of state-socialist newspapers, the opposite is true. The regulatory manner of imposing an organizational blueprint functions as a cause rather than a consequence of the number of existing organizations.
In the following section, I present the results from analyzing the effect of covariates on the founding rate of post-socialist newspaper organizations (January 1990-September 1992). Treating the post-socialist newspaper enterprise as representative of a distinct organizational form puts to use the concept of type B (de antiquo-de ipso) legitimation. Modelling density dependence in the industry after January 1990 is built on the assumption that a new organizational form, that of the independent commercial newspaper enterprise, was re-emerging on the market after having been forced out almost half a century ago. Results (Table 5, Model 2) provide strong support for H1a and H1b and for the density dependence model: both coefficients are significant and in the predicted direction. Model fit improves significantly (Haberman’s [X.sup.2] = 14.36; [delta]d.f. = 2; p[less than].01). The covariate capturing election-campaign periods also has a strong and significant positive effect which speaks for the close relations hip between the mass media and the political process. It also provides additional support for the argument that institutional politics escalates the rate of organizational foundings.
To deal with the fluctuation in population density, I test an extended version of the density dependence model (Carroll and Hannan 1999; Dobrev et al. 1999; Hannan 1997; Hannan and Carroll 1995) where important industry milestones, such as the development of industry structure and the gradual institutional embeddedness of organizational dynamics, are captured by the interaction terms between density and industry age. Similarly, processes of resource-partitioning and the emergence of peripheral market niches (Dobrev et al. 1999) are modelled by estimating the effect of interaction terms between density and the squared term of population age. The reformulated model of density dependence can be expressed in terms of the effects of legitimation and competition on the rate of founding by interacting the first- and second-order density terms with the first- and second-order population age terms:
[lambda](t) = [k.sub.[lambda]](t)[[N.sup.[[alpha].sub.0]+[[alpha].sub.1]t+[[alpha]. sub.2][t.sup.2]].sub.t] exp [([[beta].sub.0] + [[beta].sub.1]t + [[beta].sub.2][t.sup.2])[[N.sup.2].sub.t]],
0 [less than] [[alpha].sub.0] [less than] 1, [[alpha].sub.1] [less than] 0, [[alpha].sub.2] [greater than] 0,
[[beta].sub.0] [less than] 0, [[beta].sub.1] [greater than] 0, [[beta].sub.2] [less than] 0,
Thus specified, the model predicts that, initially, low density triggers legitimation and increases the rate (0 [less than] [[alpha].sub.0] [less than] 1), then becomes irrelevant ([[alpha].sub.1] [less than] 0), and resumes its original effect later in the population’s history ([[alpha].sub.2] [greater than] 0). Similarly, high density offsets competition when the population is young ([[beta].sub.0] [less than] 0). Then the effect wears off ([[beta].sub.1] [greater than] 0), and sets in again at a later period, following a resurgence in density ([[beta].sub.2] [less than] 0). (For empirical tests of this model, see Hannan 1997; Torres 1995; Hannan et al. 1998; Han 1998.) In this analysis, the non-proportional-effects specification of the density dependence model is tested in Model 3.
The estimated coefficients (Table 5, Model 3) do not offer support for the full model, but nevertheless point to an interesting observation: the interactions between the first- and second- order density terms and the squared term of population age improve model fit significantly (Haberman’s [X.sup.2] = 39.96; [delta]d.f. = 6; p[less than].01). Both coefficients are highly significant and in the predicted directions. This finding suggests that population density became relevant to the founding rate only in the latter part of the observation period. Other factors most probably shaped the demographics of the newspaper industry during 1990 and for the most part of 1991.
The most strongly pronounced environmental characteristic of this period was the extreme politicization of Bulgarian society. In fact, most newspaper organizations founded during this period were merely outgrowths of recently established political parties and organizations. The dynamics of the early post-socialist newspaper industry were intimately intertwined with the processes occurring in the population of post-socialist political parties and organizations. I hypothesize that the re-legitimation of newspaper organizations was to a large degree driven by the revival of the system of political pluralism. Perhaps the most appropriate way to model the institutionalization of post-socialist newspapers is to consider it as type D (de alio-de antiquo) legitimation where a once existing organizational form is re-establishing itself on the organizational landscape, largely through its association with another population. Furthermore, because of the close interdependence between the two populations, it is likely tha t increasing competition among political parties will deter foundings of newspaper organizations.
Modelling this relationship implies that low counts of density of political parties and organizations will increase the founding rate of newspaper organizations, whereas high counts of density will decrease it. I estimate the following model, using the log-quadratic specification of density dependence:
[lambda](t) = [k.sub.[lambda]] (t) exp([[theta].sub.1][N.sub.ti] + [[theta].sub.2][N.sub.tj] + [[theta].sub.3][[N.sup.2].sub.ti] + [[theta].sub.4][[N.sup.2].sub.tj]), [[theta].sub.1], [[theta].sub.2] [greater than] 0 and [[theta].sub.3], [[theta].sub.4] [less than] 0,
where [N.sub.ti] refers to the density of population i at time t, and [N.sub.tj] refers to the density of population j at time t.
Results from modelling interpopulation dynamics of newspapers and political parties (Table 5, Model 4) confirm the prediction of the model. Both density terms capturing the impact of political entities are significant and in the expected direction, but the effects of newspaper density are insignificant. Compared to a baseline model without any density covariates, the fit of the data improves significantly (Haberman’s [X.sup.2] = 10.06; [delta]d.f. = 4; p[less than].05). The number of political parties not only contributed to the legitimation of post-socialist newspaper enterprises, but by shaping the newspapers’ resource niche in this early period also acted as a catalyst to the rate of entry to the newspaper population by triggering competition on the political landscape.
Discussion
Several important implications stem from the fact that the models tested here were applied to the evolutionary dynamics of an industry whose history was far more turbulent and unpredictable than any other considered in previous tests of the theory of density-dependent legitimation and competition.
The history of the newspaper industry in Bulgaria paralleled the emergence of three separate populations: pre-socialist, state-socialist, and post-socialist. As argued earlier, the mechanical application of the density-dependence model combined with lack of substantive knowledge would have produced quantitatively sound results, yet would have significantly overlooked the intricate dynamics of organizational evolution. Using a taxonomy of legitimation based on differences in content and process, a comparison was drawn between the legitimation of the pre-socialist and the post-socialist population.
In 1990, the demise of state socialism revealed the remnants of a social and cultural phenomenon that was once forcefully suppressed. Post-socialist newspapers did not start from scratch, but built on the traditions and the heritage of their precursors. The organizational processes accompanying the early evolution of the post-socialist form both did and did not occur for the first time. Rather, they repeated themselves a century and a half later under completely different conditions. As a result, the processes were compressed in a much shorter time-frame than the one typical of a brand new population. While, in the second half of the 19th century, Bulgarians took several decades to embrace the idea of the newspaper and to accept its unquestionable place in society, at the beginning of 1990, it was only a few months before readers realized that the free press was making a successful comeback and that no political persecution would follow if they took advantage of it.
In a similar fashion, the exploding number of newspapers that appeared only four months after independent newspapers were allowed to exist led to a sharp increase in density and expedited the onset of competition. Such differences in the time-frame of legitimation for pre-socialist and post-socialist newspapers are understood in the context of the difference between type A and type D legitimation. The difference is twofold: in terms of process, post-socialist newspapers emerged closely related to political parties, which helped validate their claims of legitimacy. In terms of content, whereas pre-socialist newspaper enterprises struggled to define and assert a social image, post-socialist newspapers benefited from merely evoking that dormant image. Re-legitimation is comparable to a cognitive process through which memory is invoked and revitalized.
Figure 2 plots the estimated effects of density on the founding rates of the pre-socialist and post-socialist populations. Both curves follow the predictions of the density-dependence model and exhibit an inverted U-shape relationship between density and entry rate. The plot also indicates that the maximum increase in the rate of post-socialist newspapers occurs after a density increase of only 129, whereas it takes a differential of 352 for the same point to be reached by pre-socialist newspapers.
The density of the pre-socialist population took 76 years to increase by 352, while the post-socialist population took only 20 months to increase by 129. Though the early histories of these two populations bear many similarities (for example, the early prevalence of the political newspaper), they differ substantially in terms of the sources of legitimation on which each population relied. Organizations from the second half of the 19th century were built on the claim of providing leadership for the modernization of the nation. Organizational founders in 1990 proudly announced the rehabilitation of newspapers that had been shut down by force in the late 1940s. In fact, all of the once outlawed newspapers that reappeared in 1990 announced their founding year to be the time of the founding of their predecessors. This claim of continuity was displayed proudly on top of their front pages. Such distant references to previously existing enterprises obviously did not in any way imply regeneration of existing organizat ional structures. In 1990, the emerging newspaper enterprises were exposed to liability of newness for all the reasons discussed by organizational scholars in so much detail (Hannan et al. 1998a). However, while these early post-socialist organizations stumbled upon difficulties associated with establishing simple working routines and practices and mobilizing resources, their claim on the legacies of pre-existing newspapers was a very important, albeit purely symbolic, attempt to infer legitimacy from an organizational form that had left its lasting mark on the collective memory.
Figure 2 also indicates that, for the pre-socialist population, the impact of density (of newspaper organizations) on the rate was much stronger than for the post-socialist population. At its peak, legitimation increased the founding rate of pre-socialist organizations 7.35 times, whereas re-legitimation increased the entry rate of post-socialist newspapers 2.31 times. One reason for this disparity might be the observed close relationship between newspapers and political parties after the collapse of state socialism. The findings point to a strong mutualistic relationship between the two populations. Political parties and organizations helped to legitimate the rebirth of the organizational form of the private newspaper enterprise. This close relationship, however, eventually proved detrimental for the press, because political parties began to evolve into coalitions and to compete for constituents. The status of purely ideological newspapers further deteriorated when the politicization of society began to wear off. At that point, many political newspapers found it extremely hard to achieve relative independence from their protagonists. The environmental shift from political idealism to pragmatic materialism evident as early as the end of 1991 found many politically affiliated newspaper organizations unprepared and unable to reorganize their structures in order to match exogenous transformations. The political dependency of such newspaper enterprises, strongly imprinted in their organizational cores, suddenly turned from a vital resource providing competitive advantage to a source of liability to environmental selection.
Another important implication of the results presented in this research relates to the curvilinear relationship between intensity of political conflict and the rate of founding. Political violence and large-scale militant conflicts reduce the entry rate, because environmental instability yields uncertainty which, in turn, scares potential investors away from entrepreneurial activity as expectations are unclear and there is a fear of diminished returns. Furthermore, under conditions of political crisis, the general state of the economy deteriorates (e.g., workers are at war, import/export trade is impeded, etc.) and depletion of resources hinders additional foundings. A third reason that substantiates this finding and one that is especially relevant for newspaper organizations, is related to the weakening of democratic institutions and the impeachment on freedom of speech in periods of militant activity.
The results presented here do not offer a test among these alternative explanations. Further research is necessary to address the issue in more detail. The results presented in this research cannot easily be generalized, because newspaper enterprises are clearly a distinct type of organization. More importantly, the findings from modelling the dynamics of Bulgarian newspapers over the whole course of the industry reinforce the argument about the complementary (rather than opposing) nature of ecological and institutional theories. The analysis clearly revealed that to take full advantage of the explanatory power of the density-dependence model of organizational evolution, it is imperative to investigate both constitutive and sociopolitical legitimation with a great deal of historical precision and contextual accuracy.
Inversely, low-scale institutionally mediated political upheaval increases the rate of organizational founding. Political activity is an environmental resource that expands the carrying capacity of the population by simply providing newspaper enterprises with more material to report. In addition, in periods of intensified political activity, newspapers emerge as vehicals of propaganda. The majority of foundings are of explicitly political newspapers that serve organizations contending for power. From a different perspective, in periods when the established configuration of power is reshuffled, the founding of generalist newspapers mirrors the goal of emerging social groups to become a permanent part of the organizational landscape (Carroll 1987).
Stanislav D. Dobrev
Graduate School of Business, University of Chicago, USA
Stanislav D. Dobrev
Stanislav D. Dobrev holds a Ph.D. in organizational sociology from Stanford University and is currently an assistant professor of strategy in the Organizations and Markets Group at the Graduate School of Business of the University of Chicago. Dobrev’s primary research interests include the dynamics of organizational populations, industrial change and evolution, corporate demography, and transition to entrepreneurship. His research has been published in the American Journal of Sociology, the European Sociological Review, Organization Studies, and Industrial and Corporate Change. His current projects include modelling the effects of scale and scope in organizational evolution (with Glenn Carroll and Tai-Young Kim), developing a collective action approach to the study of business strategy (with Glenn Carroll and Tai-Young Kim), and investigating the career dynamics of professional managers (with Bill Barnett).
Mailing Address: Graduate School of Business, University of Chicago, 1101 East 58th Street, Chicago, IL 60637, USA.
E-mail: SDOBREV@GSB.UCHICAGO.EDU
Note (*.) This paper is based on my dissertation research, and as such has benefited greatly from the ideas and guidance of my advisors Michael Hannan, Glenn Carroll, Susan Olzak, and John Meyer. I thank Bill Barnett, Joon Han, Tai-Young Kim, the participants in the Organizational Ecology workshop at Stanford, the audiences at the Organizational Behavior seminar at the Stanford GSB and at the EGOS 14th Colloquium, and the Organization Studies editor and reviewers for comments on earlier drafts of the paper. I am also grateful to Dimitar Dobrev and Velichka Dobreva for having devotedly provided excellent, albeit uncompensated, research assistance. As always, the flaws are all mine.
References
Agresti, Alan
1990 Categorical data analysis. New York: Wiley.
Aldrich, Howard E.
1979 Organizations and environments. Englewood Cliffs, NJ: Prentice Hall.
Aldrich, Howard E., and C. Marlene Fiol
1994 ‘Fools rush in? The institutional context of industry creation’. Academy of Management Review 19/4: 645-670.
Alfandari, Eliezer
1995 ‘Pazarnoto i Politicheskoto v Novata ni Zhurnalistika’ in Zhurnalistikata v Totalitarnoto i Posttotalitarnoto Obshtestvo. Penka Karaivanova et al. (eds.), 113-122. Sofia: Poligraficheski Kombinat.
Andreev, Boris M.
1948 Nachalo, Razvoi i Vazhod na Bulgarskia Pechat, Vol. 2. Sofia: Globus.
Andreev Boris M.
1946 Nachalo, Razvoi i Vazhod na Bulgarskia Pechat, Vol. 1. Sofia: Globus.
Barron, David N.
1998 ‘Pathways to legitimacy among consumer loan providers in New York City, 1914-1934’. Organization Studies 19/2: 207-233.
Baum, Joel A.C., and Walter W. Powell
1995 ‘Cultivating an institutional ecology of organizations: Comment on Hannan, Carroll, Dundon, and Torres’. American Sociological Review 60: 529-538.
Bobchev, S.S., editor
1894 Iubileen Sbornik pa Sluchai Petdeset-Godishninata na Bulgarskata Zhurnalistika, Sofia: Pechatnitza Valkov.
Borshukov, Georgi
1957 Istoria no Bulgarskata Zhurnalistika ot Zarazhdaneto na Bulgarskiia Pechat do Osvobozhdenieto prez 1878. Sofia: Nauka i Izkustvo.
Bozhilov, Ivan, Vera Mutafchieva, Konstantin Kosev, Andrei Pantev, and Stoicho Grancharov
1993 Istorija no Bulgaria. Sofia: Hristo Botev.
Carroll, Glenn R.
1987 Publish and perish: The organizational ecology of newspaper industries. Greenwich, CT: JAI Press.
Carroll, Glenn R., and Jacques Delacroix
1982 ‘Organizational mortality in the newspaper industries of Argentina and Ireland: An ecological approach’. Administrative Science Quarterly 27: 169-198.
Carroll, Glenn R., and Michael T. Hannan
1999 The demography of corporations and industries. princeton, NJ: Princeton University Press.
Carroll, Glenn R., and Michael T. Hannan, editors
1995 Organizations in industry. Strategy, structure, and selection. New York: Oxford University Press.
Carroll, Glenn R., and Michael T. Hannan
1989a ‘Density delay in the evolution of organizational populations: A model and five empirical tests’. Administrative Science Quarterly 34:411-430.
Carroll, Glenn R., and Michael T. Hannan
1989b ‘Density dependence in the evolutions of newspaper populations’. American Sociological Review 54:524-541.
Carroll. Glenn R., and Yangchung Paul Huo
1986 ‘Organizational task and institutional environments in ecological perspective: Findings from the local newspaper industry’. American Journal of Sociology 91: 838-873.
Carroll, Glenn R., Peter Preisendorfer, Anand Swaminathan, and Gabrielle Wiedenmayer
1993 ‘Brewery and brauerei: The organizational ecology of brewing’. Organization Studies 14/2: 155-188.
Carroll, Glenn R., and Anand Swaminathan
1992 ‘The organizational ecology of strategic groups in the American brewing industry from 1975 to 1990’. Industrial and Corporate Change 1: 65-97.
Carroll, Glenn R., and James B. Wade
1991 ‘Density dependence in the organizational evolution of the American brewing industry across different levels of analysis’. Social Science Research 20: 217-302.
Danevski, Valentin
1990 Politicheski Partii Dvizhenija I Organizatzii v Balgaria I Tehnite Lideri. Sofia: Sofia Press.
Delacroix, Jacques, and Glenn R. Carroll
1983 ‘Organizational foundings: An ecological study of the newspaper Industries of Argentina and Ireland’. Administrative Science Quarterly 28: 274-291.
Dobrev, Stanislav D.
2000 ‘Decreasing concentration and reversibility of the resource partitioning model: Supply shortages and de-regulation in the Bulgarian newspaper industry, 1987-1992’. Organization Studies 21/2: 383-404.
Dobrev, Stanislav D.
1999 ‘The dynamics of the Bulgarian newspaper industry in a period of transition: Organizational adaptation, structural inertia, and political change’. Industrial and Corporate Change 8/3: 573-605.
Dobrev, Stanislav D., and Glenn R. Carroll
1999 ‘Size (and competition) among organizations: Modelling scale-based selection among automobile producers in four major countries, 1885-1981’. Working paper, Freeman School of Business, Tulane University.
Dobrev, Stanislav D., Tai-Young Kim, and Michael T. Hannan
1999 ‘Dynamics of niche width and resource partitioning’. Working paper, Freeman School of Business, Tulane University.
Haberman, Shelby J.
1977 ‘Log-linear models and frequency tables with small expected cell counts’. Annals of Statistics 5: 1148-1165.
Han, Joon
1998 ‘The evolution of the Japanese banking industry: An ecological analysis, 1873-1945’. Ph.D. Dissertation, Department of Sociology, Stanford University, CA.
Hannan, Michael T.
1997 ‘Inertia, density, and the structure of organizational populations’. Organization Studies 18/2: 193: 228.
Hannan, Michael T., and Glenn R. Carroll
1995 ‘Theory building and cheap talk about legitimation: Reply to Baum and Powell’. American Sociological Review 60: 539-544.
Hannan, Michael T., and Glenn R. Carroll
1992 Dynamics of organizational populations: Density, legitimation and competition. New York and Oxford: Oxford University Press.
Hannan, Michael T., Glenn R. Carroll, Stanislav D. Dobrev, and Joon Han
1998a ‘Organizational mortality in European and American automobile industries. Part I: Revisiting the effects of age and size’. European Journal of Sociology 14: 279-302.
Hannan, Michael T., Glenn R. Carroll, John Torres, Stanislav D. Dobrev, and Joon Han
1998b ‘Organizational mortality in European and American automobile industries. Part II: Coupled clocks’. European Journal of Sociology 14:303-3 13.
Hannan, Michael T., Glenn R. Carroll, Elizabeth A. Dundon, and John C. Torres
1995 ‘Organizational evolution in a multinational context: Entries of automobile manufacturers in Belgium, Britain, France, Germany, and Italy’. American Sociological Review 60: 509-528.
Hannan, Michael T., and John Freeman
1989 Organizational ecology. Cambridge, MA: Harvard University Press.
Ivanov, Iurdan
1893 Bulgarskii Periodicheski Pechat, Vol. 1. Sofia: Darzhavna Pechatnitza.
Jackowicz, Jerzy
1992 Bulgaria od Rzadow Komunistycznych do Demokracji Parlament-arnej 1988-1991. Warszawa: Typografika.
Kostadinova, Tatiana
1995 Bulgaria 1879-1946. The challenge of choice. New York: Columbia University Press.
Knigoizdavane i Pechat
1965 Statisticheski Sbornik Sofia: Tzentralno Statistichesko Upravlenie.
McCullagh, P.
1983 ‘Quasi-likelihood functions’. Annals of Statistics 11: 59-67.
Meyer, John W.
1983 ‘Institutionalization and the rationality of formal organizational structure’ in Organizational environments: Ritual and rationality. John W. Meyer and W. Richard Scott (eds), 26 1-282. Beverly Hills, CA: Sage.
Meyer, John W., and Brian Rowan
1977 ‘Institutionalized organizations: Formal structure as myth and ceremony’. American Journal of Sociology 83: 340-363.
Meyer, John W., and W. Richard Scott
1983 Organizational environments: Ritual and rationality. Beverly Hills, CA: Sage.
Naidenov Dimitar
1995a ‘Prehodat kum Pazarno Regulirane na Presata’ in Zhurnalistikata v Totalitarnoto i Posttotalitarnoto Obshtestvo. Penka Karaivanova et al. (eds.), 186-212. Sofia: Poligraficheski Kombinat.
Naidenov Dimitar
1995b Preszhurnalistika. Sofia: Poligraphicheski Kombinat.
Narodna Biblioteka v Sofia
1903 Opis na Balgarskite Periodichni Izdania 1844-1900. Sofia: DP.
Narodna Biblioteka v Sofia
1901 – Bibliograficheski Biuletin, Vols. 1901 1944 to 1944. Sofia: DP.
Narodna Biblioteka ‘Kiril I Metodii’
1975 Bulgarski Periodichen Pechat 1944-1969, Vol. III. Sofia: NB.
Narodna Biblioteka ‘Kiril I Metodii’
1975 Bulgarski Periodichen Pechat, Vol. 1972 to 1992. Sofia: NB.
Nikolov, Georgi
1932 Istoriia na Balgarskia Vsekidneven Pechat, 1877-1932, Vol I and II. Sofia: Pechatnitza Rodopi.
Olzak, Susan, and Elizabeth West
1991 ‘Ethnic conflicts and the rise and fall of ethnic newspapers. American Sociological Review 56: 458-474.
People’s Republic of Bulgaria
Annual Statistical Directory, Volumes 1947-1989. Sofia: DP.
Polos, Laszlo, Michael T. Hannan, and Glenn R. Carroll
1998 ‘Forms and identities: On the structure of organizational forms’. Paper presented at the 14th Colloquium of the European Group on Organization Studies.
Stinchcombe. Arthur L.
1965 ‘Organizations and social structure’ in Handbook of Organizations. James G. March (ed.), 153-193. Chicago: Rand McNally.
Tilly, Charles
1978 From mobilization to revolution. Reading. MA: Addison-Wesley.
Tilly, Charles, and James Rule
1965. Measuring political upheaval. Princeton University: Research Monograph No. 19.
Todorova, Sofia
1991 Koi Kakuv E ? Who’s who in elections? Sofia: Informatzionno Obsluzhvane.
Topencharov, Vladimir
1981 Bulgarskata Zhurnalistika, 1903-1917. Sofia: Nauka I Izkustvo.
Topencharov, Vladimir
1963 Bulgarskata Zhurnalistika, 1885-1903. Sofia: Nauka I Izkustvo.
Torres, John C.
1995 ‘The dynamics of the U.K. motor industry: An ecological analysis, 1885-1981’. Ph.D. Dissertation, Department of Sociology, Stanford University, CA.
Velinova, Miglena
1992 1000 Vestnika: Spravochnik na Bulgarskata Presa sled 10.XI.1989. Sofia: Faculty of Journalism and Mass Communications.
Table 1
Type of Constitutive Legitimation Based on Process and Content
Differences
Content Differences Based
on Origin
de novo de antiquo
Process Differences de ipso Type A Type B
Based on External
Affiliation de alio Type C Type D
Table 2.
QL Estimates of the Effects of Sociopolitical Factors and Density on
the Entry Rate of Bulgarian Newspaper Organisation, 1846-1992
(1) (2)
Constant .81 (6.13) .77 (5.71)
Unification movement (1879-1885) 1.89 (8.91) 1.75 (7.76)
Transition to capitalism (1886-1912) .66 (3.60) .72 (3.77)
Balkan wars and WW I (1913-1919) -.03 (-.16) -.13 (-.72)
Developing capitalism (1920-1940) 1.94 (11.27) 2.09 (11.01)
WW II (1941-1944) -.74 (-3.68) -.87 (-4.04)
Communist cleansing (1945-1948) -.72 (-3.68) -.67 (-2.39)
Stalinist regime (1949-1956) .01 (.05) .17 (.67)
Liberal reforms (1957-1978) -.04 (-.25) -.04 (-.24)
Cold war re-freeze (1979-1989) -.94 (-6.10) -.94 (-5.89)
Post-socialism (1990-1992) 2.81 (11.46) 2.61 (9.68)
Political turbulence: .21(2.50)
turmoil
institutional
Foundings
[Foundings.sup.2] x [10.sup.-3]Log (density)
[Density.sup.2] x [10.sup.-3][sigma] .127 .139[X.sup.2] 364.14 337.51Habermans’s [X.sup.2] 92.22 (vs. 1)
N 147 147
(3) (4)
Constant .82 (6.16) .81 (5.99)
Unification movement (1879-1885) 1.89 (8.91) 1.65 (6.98)
Transition to capitalism (1886-1912) .66 (3.58) .74 (3.73)
Balkan wars and WW I (1913-1919) -.01 (-.07) -.13 (-.72)
Developing capitalism (1920-1940) 1.91 (9.71) 2.09 (11.11)
WW II (1941-1944) -.67 (-2.76) -.63 (-2.82)
Communist cleansing (1945-1948) -.80 (-2.62) -.99 (-3.32)
Stalinist regime (1949-1956) .01 (.05) .28 (1.06)
Liberal reforms (1957-1978) -.04 (-.25) -.04 (.51)
Cold war re-freeze (1979-1989) -.94 (-6.09) -.94 (-5.67)
Post-socialism (1990-1992) 2.81 (11.44) 2.46 (8.52)
Political turbulence:
turmoil -.07(-.50)
institutional .35 (3.29)
Foundings
[Foundings.sup.2] x [10.sup.-3]Log (density)
[Density.sup.2] x [10.sup.-3][sigma] .127 .154[X.sup.2] 362.40 309.71Habermans’s [X.sup.2] 1.85 (vs. 1) 184.17 (vs. 1)
N 147 147
(5)
Constant .21 (1.08)
Unification movement (1879-1885) .97 (3.74)
Transition to capitalism (1886-1912) .11 (.50)
Balkan wars and WW I (1913-1919) -.05 (-.27)
Developing capitalism (1920-1940) .71 (2.80)
WW II (1941-1944) -.01 (-.06)
Communist cleansing (1945-1948) -1.22 (-4.14)
Stalinist regime (1949-1956) .76 (3.00)
Liberal reforms (1957-1978) .07 (.32)
Cold war re-freeze (1979-1989) -.70 (-4.21)
Post-socialism (1990-1992) 1.56 (5.09)
Political turbulence:
turmoil
institutional .42 (4.12)
Foundings .01 (4.73)
[Foundings.sup.2] x [10.sup.-3] -.01 (-1.62)Log (density) .38 (4.41)
[Density.sup.2] x [10.sup.-3] -.003 (-4.52)[sigma] .130[X.sup.2] 265.33Habermans’s [X.sup.2] 1084.58(vs. 4)
N 147
Figures in parentheses are t-statistics.
Omitted period is lack of political independence (1846-1878).
Table 3
QL Estimates of the Effects of Density on the Entry Rate of
Pre-Socialist Bulgarian Newspaper Organizations, 1846-1948
(1) (2)
Constant .78 (5.99) .12 (.61)
Unification movement (1879-1885) 1.56 (7.29) .92 (3.74)
Transition to capitalism (1886-1912) .53 (2.95) .09 (.48)
Balkan wars and WW I (1913-1919) .03 (.18) -.11 (-.65)
Developing capitalism (1920-1940) .79 (2.72) .79 (2.71)
WW II (1941-1944) .01 (.06) -.02 (-.70)
Communist cleansing (1945-1948) -1.05 (-4.12) -1.04 (-3.77)
Institutional politics .33 (3.65) .41 (4.41)
Foundings .01 (4.13) .01 (1.89)
[Foundings.sup.2] x [10.sup.-3] -.02 (-2.88) -.004 (-.50)Log (density) .45 (5.00)
[Density.sup.2] x [10.sup.-3] -.002 (-2.46)[sigma] .102 .103[X.sup.2] 222.63 195.39Habermans’s [X.sup.2] 122.28 (vs. 1)
N 103 103
Figures in parentheses are t-statistics.
Omitted period is lack of political independence (1846-1878).
Table 4
QL Estimates of the Effects of Density on the Entry Rate of State-
Socialist Bulgarian Newspaper Organizations, 1949-1989
(1) (2)
Constant 3.24 (11.78) 5.22 (2.04)
Number of TV licences -.82 (-4.14) -.67 (-3.33)
Liberal reforms (1957-1978) .36 (1.58) .77 (2.35)
Cold war re-freeze (1979-1989) .27 (1.12) .11 (.48)
Foundings .01 (1.18) .02 (1.08)
[Foundingss.sup.2] x [10.sup.-3] -.03 (-.54) -.06 (-1.10)Log (density) -.38 (-.72)
[Density.sup.2] x [10.sup.-3] -.001 (-.32)[sigma] .175 .138[X.sup.2] 53.07 60.24Habermans’s [X.sup.2] 22.33 (vs. 1)
N 41 41
Figures in parentheses are t-statistics.
Omitted period is Stalininst regime (1949-1956).
Table 5
QL Estimates of the Estimates of Density on the Entry Rate of Post–
Socialist Bulgarian Newspaper Organizations
(1) (2)
Constant 2.38 (9.13) -4.98 (-1.89)
Election campaign .28 (2.19) .33 (2.63)
Foundings .03 (1.36) -.02 (-.63)
[Foundings.sup.2] x [10.sup.-3] -.12 (-.27) .64 (1.33)Density .04 (2.82)
[Density.sup.2] x [10.sup.-3] -.05 (-2.80)Density x t
[Density.sup.2] x t x [10.sup.-3]Density x [t.sup.2][Density.sup.2] x [t.sup.2] x
[10.sup.-3]Population age (t)
Population [age.sup.2] ([t.sup.2])
Density PPO
Density [PPO.sup.2] x [10.sup.-3][sigma] .058 .043
[X.sup.2] 35.68 34.08Habermans’s [X.sup.2] 14.36 (vs.1)
N 33 33
(3) (4)
Constant 27.49 (1.43) -.13 (-.02)
Election campaign .62 (3.90) -.007 (-.10)
Foundings -.02 (-.85) -.007 (-.10)
[Foundings.sup.2] x [10.sup.-3] .40 (.91) -.02 (-.09)Density -.18 (-1.48) -.01 (-.52)
[Density.sup.2] x [10.sup.-3] .31 (1.66) .003 (.69)Density x t -.01 (-.48)
[Density.sup.2] x t x [10.sup.-3] -.002 (-.12)Density x [t.sup.2] .001 (3.34)
[Density.sup.2] x [t.sup.2] x -.001 (-2.19)[10.sup.-3]Population age (t) 1.99 (1.21)
Population [age.sup.2] ([t.sup.2]) -.15 (-3.99)
Density PPO .13 (1.95)
Density [PPO.sup.2] x [10.sup.-3] -.08 (-2.03)
[sigma] .000 .002[X.sup.2] 27.10 23.4Habermans’s [X.sup.2] 39.96 (vs.2)
N 33 21
Figures in parentheses are t-statistics.
[Graph omitted]
[Graph omitted]
COPYRIGHT 2001 Walter de Gruyter und Co.
COPYRIGHT 2001 Gale Group
