A second double movement? Polanyi and shifting global opinions on neoliberalism

Abstract

Spanish

French

Karl Polanyi’s theory of the ‘double movement’ has gained great currency in recent years to explain the global growth of contemporary social movements resisting neoliberalism. However, there has been no statistical research demonstrating whether these protest movements represent a more general trend of growing discontent with ‘disembedding’ markets from public control. This article uses questions from the World Values Survey to construct an ‘embeddedness’ index measuring public opinion on the desired relationship between states and markets. Focusing on public opinion in 20 countries during the 1990s, the analysis poses three questions: First, is there evidence of increasing global support for ‘re-embedding’ markets? Second, how does such opinion vary across regions of the world? Finally, what is the class and gender composition of this latent countermovement? The results provide substantial evidence of an emerging countermovement in public opinion over the 1990s with complex class, gender, and geopolitical variation.

Keywords

Countermovements double movement embeddedness neoliberalism Polanyi

Introduction

The 1990s was by most accounts the decade of the market. The Soviet Union’s collapse opened the way for neoliberal ‘shock therapy’ across much of Eastern Europe. Many countries in the developing world began to abandon or temper statist development models in favor of market liberalization. Where liberalization had already taken hold in the 1980s, such as in many parts of Latin America and Africa, neoliberal reforms generally progressed and became further entrenched over the decade. Trade and finance were liberalized, state assets were privatized, social spending was reduced, and currencies were allowed to trade on global markets. Wherever possible, governments transferred assets and responsibilities from the public to private domain. Neoliberalism became institutionalized in the WTO, IMF, and World Bank, and protagonists of this process declared that ‘there is no alternative.’

Yet, in the late 1990s, protest movements disagreeing with this proposition began to gain attention for their increasingly coordinated resistance to neoliberal policies and institutions. The protest against the WTO in Seattle in 1999 was the most dramatic demonstration of dissent to the ‘Washington Consensus,’ and was followed by clashes outside subsequent global trade negotiations and meetings of the World Bank, IMF, and G-8. These global manifestations of dissent were underlain by a more diffuse set of local resistance movements across the globe that began to challenge the social dislocations of neoliberal policies at the local and national levels. Some of these movements forged transnational networks (Keck and Sikkink, 1998; Tarrow, 2005) and came to be seen as representing ‘counter-hegemonic globalization’ (Evans, 2000, 2005, 2008).

Most scholars theorizing these new movements (e.g. Evans, 2000; Hardt and Negri, 2000; Mertes, 2004) have agreed that they cannot be understood in classical Marxist terms: they are not primarily driven by the urban proletariat, and they have emerged in response to an extraordinary diversity of economic and social dislocations. Searching for new foundations, many have begun to resurrect Karl Polanyi’s (2001 [1944]) theory of the ‘double movement’ to understand this new wave of resistance to aggressive market expansion (Arrighi and Silver, 2003; Burawoy, 2003; Evans, 2008; Harvey, 2005; Levien, 2007; Watts, 2000). Polanyi famously argued that ‘disembedding’ markets from social controls produces tremendous social dislocations that instigate ‘countermovements’ for social protection. Polanyi’s thesis appears to be relevant once again because it helps us to identify a socially diverse constituency organized against market dislocations rather than class-based exploitation (Burawoy, 2003). Yet, while Polanyi’s theory of the ‘double movement’ provides a compelling framework for understanding this proliferation of social movements, the recent Polanyi-inspired scholarship has not yet attempted to quantify the extent and contours of discontent with neoliberal reforms. While scholars have provided compelling ethnographic accounts of politically organized movements, whether these movements express broader patterns of public opinion remains open to question.

We use data from three waves of the World Values Survey (1990–2001) to present a cross-country analysis of changing public opinion towards markets in the 1990s. For this purpose we construct a ‘pro-embeddedness’ scale, which gauges preferences for public rather than private ownership of companies, and for public rather than private responsibility for welfare. We then use this scale to both test Polanyi’s basic predictions and extend them. Two disclaimers are in order. First, while marketization entails many qualitatively complex processes, we believe that these two questions best capture the core principles of Polanyi’s concept of embeddedness. They provide valuable insight into shifting global opinion towards states and markets, which may be elaborated upon in future research. Second, we are measuring public opinion – what we call a latent countermovement – and not actually organized political action. The former is a starting point for understanding the subjective basis and future potential of the latter.

The analysis focuses on three questions. First, we test the general Polanyian hypothesis that ‘disembedding’ the market produces countermovements for ‘re-embedding’: did a latent countermovement emerge in the 1990s in response to neoliberal reforms, as measured by preferences for public ownership and state responsibility for welfare? Second, we ask a question that Polanyi never posed: to what extent does this latent countermovement take different forms in different countries and world regions? Third, we address a question that Polanyi left ambiguous: what is the social composition of the latent countermovement? In other words, what social groups are most in favor of ‘re-embedding’? Our findings reveal that a latent countermovement in global public opinion was underway in the 1990s, but was far from uniform. While ‘pro-embeddedness’ sentiment increased almost everywhere, distinctly different patterns of shifting public opinion emerged both between world regions and within countries. Before turning to these results in more detail, however, we first review the theoretical foundations for the study and develop our hypotheses.

A second double movement?

Observing the late 19th- and early 20th-century retractions from laissez-faire policies, Polanyi developed the notion of the ‘double movement’ whereby efforts to disembed markets from social relations would inevitably lead to large-scale ‘countermovements’ for social protection:

While on the one hand markets spread all over the face of the globe . . . on the other hand a network of measures and policies was integrated into powerful institutions designed to check the action of the market. . . . A deep-seated movement sprang into being to resist the pernicious effects of a market-controlled economy. (Polanyi, 2001 [1944]: 80)

With the widespread adoption of Keynesian policies and a worldwide re-extension of political constraints on the market – inaugurating what Ruggie (1982) calls the era of ‘embedded liberalism’ – Polanyi predicted in 1944 that, ‘Undoubtedly, our age will be credited with having seen the end of the self-regulating market’ (2001 [1944]: 148).

What Polanyi did not anticipate, of course, was the possibility for a neoliberal round of market disembedding. Beginning as a marginal economic doctrine, neoliberalism gained an institutional foothold with the Reagan and Thatcherite policies of the 1980s, and over the course of the 1990s became the hegemonic model for structuring the relationship between states and markets. This model principally entailed transferring domains of social life from the former to the latter: deregulation of markets; privatization of public firms, natural resources, and public utilities; the withdrawal of the state from many areas of social provision; and an increased emphasis on individual responsibility and ‘self-help.’ With the encouragement and coercion of institutions like the IMF and World Bank, neoliberal restructuring spread rapidly through the Global South and formerly Communist countries in the 1990s. This ‘rebirth of the liberal creed’ (Fourcade-Gourinchas and Babb, 2002) has led numerous scholars to compare the current era with the laissez-faire period described by Polanyi (e.g. Arrighi and Silver, 2003; Burawoy, 2003; Evans, 2008; Somers and Block, 2005; Watts, 2000).

While some scholars have focused on the ability of liberalization to deliver benefits and create ‘coalitions for reform’ (Haggard and Webb, 1994), many have emphasized the profound dislocations and countermovements that liberalization has produced (e.g. Arrighi and Silver, 2003; Burawoy, 2003; Evans, 2005; Harvey, 2005; Levien, 2007). But the aggregate picture of global public opinion towards the core principles of neoliberalism remains unexamined. We test the Polanyian hypothesis that these social movements reflect widespread discontent with neoliberal reforms. We predict that market disembedding in the 1990s (Gwartney et al., 2007)¹ produced a global shift in public opinion in favor of re-embedding the market:

Hypothesis 1: Between 1990 and 2001, ‘pro-embeddedness’ sentiment increased in all regions of the globe.

While it is important to establish whether there is a broad countermovement occurring in global public opinion, it is equally important to know who is included. Polanyi remained ambiguous and even contradictory about the composition of countermovements for social protection. We extend Polanyi’s theory by exploring two dimensions of this composition: regional specificity and sociodemographic constitution.

Polanyi was primarily concerned with the double movement as a phenomenon of the industrialized Global North. He does suggest that the lack of political autonomy of colonial societies prevented any kind of state intervention for social protection, amplifying the destructive effects of commodification (2001 [1944]: 167, 171, 192). But Polanyi did not live to see the fall of the Soviet Union, or witness the decolonization, industrialization, and eventual liberalization of the Global South. He was therefore unable to systematically compare the double movement in different regions of the globe.

While there is little to draw on in Polanyi’s account to formulate a hypothesis about the regional specificity of countermovements, we can make some broad extrapolations based on his general theory of embeddedness. For Polanyi, the double movement resembles a pendulum: progressive disembedding of the market (through the commodification of land, labor, and money) generates counter-tendencies in the opposite direction. One might reasonably extend this concept to suggest that the greater the degree of disembedding, the greater the movement for re-embedding. Polanyi is also explicit that the pace of the disembedding is crucial for the amount of dislocation it wreaks and the amount of resistance it foments (2001 [1944]: 79, 86). Polanyi’s thesis would thus predict that we should see a more dramatic countermovement where the degree and speed of disembedding has been the greatest.

While we are sensitive to complex cross-country variation in the degree, speed, and features of liberalization, we might still expect to see the largest countermovement in countries of the former Soviet bloc. These countries moved from economies that were substantially embedded within state socialism to market economies with private property in a highly compressed span of time (e.g. Burawoy, 1996; Burawoy et al., 2000; Stiglitz, 2003; Stuckler et al., 2009). In the non-Soviet countries of the industrializing world, by contrast, most countries already had market economies in the 1990s, albeit in most cases embedded within developmentalist or dirigiste states. Structural adjustment produced abrupt dislocations in many of these countries, though liberalization was still relatively less compressed and dramatic than in the post-Soviet bloc. Finally, the economies of the Global North were already relatively liberalized and experienced the most gradual marketization in the 1990s. In sum, while we expect to find latent countermovements in all countries, we expect them to be more dramatic in the former Communist countries than in the Global South, which in turn should be more pronounced than in the Global North:

Hypothesis 2: All countries showed increased ‘pro-embeddedness’ sentiment over the course of the 1990s, but the increase was greatest in the former Communist countries, followed by the Global South and then the Global North.

But we do not just want to know the global contours of contemporary countermovements. We also want to know their social composition within countries. The other major lacuna in Polanyi’s analysis regards the social origins of the countermovement. Polanyi is rather ambiguous about who will lead the charge for social protection. At various points, Polanyi identifies the working class, or the working class and the landed aristocracy, as the vanguard(s) of social protection (2001 [1944]: 139, 162). While he asserts that the trading classes ‘had no organ to sense the dangers’ of the self-regulating market, and thus could not be relied upon for the protection of society, he nonetheless believed that such protection was in their interest (2001 [1944]: 139). Further, elite-driven legislation forms a significant part of Polanyi’s description of the countermovement. The important point, however, is that for Polanyi whichever class is taking the lead is doing so not only out of narrow self-interest, but out of the functional need of society to protect itself (2001 [1944]: 139, 159, 163). Polanyi is thus ambiguous about who will lead a countermovement, leaving us to extract two potential hypotheses. In the ‘organic’ reading of Polanyi, we might expect society as a whole to react homogeneously to the disembedding of the market. In this case, we would expect to see no differentiation in the changes in desire for re-embedding between social groups, or would at least expect them to all increase. An alternative is the ‘agonistic’ Polanyian hypothesis, which is more of a reconstruction of Polanyi than a faithful reading. Here we expect the most disadvantaged social groups to be in the forefront of any countermovement for social protection (Burawoy, 2003). Based on an abundance of studies suggesting that the disproportionate brunt of neoliberal policies are borne by women (e.g. Benería, 2003; Cagatay and Erturk, 2004; Hart, 2002; Roberts, 2008; Smith, 2008) and the poor (Agenor, 2004; Dumenil and Levy, 2004; Harvey, 2005; Wade, 2004, 2005), we adopt the ‘agonistic’ Polanyian hypothesis and propose to test it against the ‘organic’ hypothesis:

Hypothesis 3: Women and persons with low income will express the greatest amounts of ‘pro-embeddedness’ sentiment. Generally, pro-embeddedness will decrease as one moves up the income ladder.

Data and methods

We use the World Values Survey (WVS), which is based on face-to-face interviews and includes public opinion and attitude data for individuals throughout the globe. We focus on the first three waves of the survey, administered in 1990 and 1991 (wave 1), 1995–1998 (wave 2), and 1999–2001 (wave 3). Because our intent is to examine the extent to which ‘pro-embeddedness’ sentiment increased with the onset of neoliberal reforms, the analysis focuses on how such attitudes change over time. We therefore limit our sample to those countries where data are available on the relevant questions (see below) for all three waves. This includes 20 countries, which we organize into three groups: (1) Global North, including Finland, Germany, Japan, Spain, and the United States; (2) former Communist countries of Eastern Europe, including Belarus, Czech Republic,² Estonia, Lithuania, Poland, Romania, and Russia; and (3) Global South, including Argentina, Chile, China, India, South Korea, Mexico, South Africa, and Turkey. We recognize that these larger groupings obscure a great deal of diversity among countries within each group, and for this reason we present the results for both the larger groups and the individual countries.

Variables

Our concern in this article is whether or not there is evidence of a ‘latent’ Polanyian countermovement, expressed in the form of growing public discontent with neoliberal reforms. We are not measuring the extent to which this discontent is actually organized into political opposition, but are measuring public opinion that may or may not be activated in the present or future. The dependent variable in our analysis is an index of ‘pro-embeddedness’ sentiment. This index taps into two oppositions that are essential to neoliberal restructuring, and capture important elements of Polanyi’s notion of ‘embeddedness’: private versus public ownership of business, and individual versus state responsibility for welfare. We use two questions that asked respondents to rank their views on each issue on a scale from 1 to 10:

How would you place your views on this scale: Private ownership of business should be increased (1) vs. Government ownership of business should be increased (10)?

How would you place your views on this scale: People should take more responsibility to provide for themselves (1) vs. Government should take more responsibility to ensure that everyone is provided for (10)?

To construct a ‘pro-embeddedness’ score we simply add the two rankings, creating a scale that ranges from 2 (strongly against embedding the market) to 20 (strongly in favor of embedding the market).

It is important to note that we are using ‘embeddedness’ in a very specific way here. As Block (2003) has demonstrated, Polanyi hints at the idea of an ‘always embedded market.’ For Polanyi the self-regulating market is a utopian possibility because market societies are always embedded within institutional structures, state policies, and social relations. In this sense even highly liberalized markets are dependent on the coordinating role of firms, states, and global governance institutions (Ó Riain, 2000). In this study, however, we operationalize embeddedness in a more limited sense as the degree of state control over the economy and state responsibility for public welfare. While this obscures qualitative shifts in the character of market embeddedness, it captures the overall direction of neoliberal policies towards minimizing state ownership and social protection.

In considering variation in pro-embeddedness sentiment between groups within countries, we focus on gender and relative income level. Gender is a simple male/female binary and relative income level includes three groups – low, middle, high – indicating the positioning of respondents’ self-reported income relative to the top, middle, and bottom thirds of the national distribution. Self-reported income level is admittedly not the ideal measure of what we are interested in, which is socioeconomic class. However, it is not possible to generate consistent class categories from the WVS data. We thus settle on income level as a rough measurement of how opinions on market embeddedness vary across economic strata. Table 1 provides descriptive statistics for the entire sample.

Table 1.

Descriptive statistics.

Country	N				Relative income
	Wave 1	Wave 2	Wave 3	Female	Low	Mid	High	Missing
Total	31,071	27,837	28,741	49.9	31.3	34.2	26.4	8.2
Global North	9328	6074	5978	49.9	27.8	32.2	26.7	13.2
Finland	530	940	933	49.8	30.2	31.2	32.0	6.7
Germany	3107	1913	1783	53.0	29.1	29.4	26.7	14.9
Japan	663	727	1031	46.0	26.7	30.8	32.6	9.9
Spain	3320	1028	1038	48.2	22.8	35.1	21.7	20.4
United States	1708	1466	1193	49.6	31.3	34.5	26.8	7.4
Former Communist	10,159	9118	8610	52.8	28.9	38.2	27.4	5.5
Belarus	960	1961	854	55.1	32.1	40.3	24.6	3.1
Czech Republic	2929	1083	1831	51.8	25.2	39.5	28.0	7.3
Estonia	892	987	883	54.5	21.2	39.4	34.9	4.6
Lithuania	857	942	862	49.9	25.3	36.8	28.9	9.0
Poland	1744	1062	997	51.7	33.7	45.8	18.0	2.5
Romania	1054	1137	1014	49.2	32.2	35.8	29.6	2.4
Russia	1723	1946	2169	55.6	30.8	31.9	29.4	7.9
Global South	11,584	12,645	14,153	47.8	35.0	32.3	25.5	7.2
Argentina	873	997	1110	50.5	24.1	34.1	29.2	12.5
Chile	1390	955	1148	51.7	35.0	33.9	27.3	3.8
China	968	1241	838	43.5	31.2	37.6	29.6	1.6
India	2340	1631	1377	41.5	25.8	32.9	33.6	7.8
Korea	1221	1230	1198	51.1	35.5	30.9	31.7	1.8
Mexico	1350	2143	1270	47.9	32.7	30.1	26.2	10.9
South Africa	2462	2622	2742	48.6	43.6	26.6	17.1	12.8
Turkey	980	1826	4470	48.7	40.0	36.3	20.6	3.1

Models

In order to test our three hypotheses we use a series of simple ordinary least squares regression models. The analysis uses three independent variables: wave (wave 1 = 1990–1991; wave 2 = 1995–1998; wave 3 = 1999–2000); place, represented either by geopolitical region (Global North, Former Communist, Global South) or country (e.g. Spain, Belarus, Chile); and gender-income group (low-income female, mid-income female, high-income female, low-income male, mid-income male, high-income male). The dependent variable is respondent’s score on the pro-embeddedness scale.

Model 1 is designed to test the first two hypotheses, and posits pro-embeddedness as a function of time and place:

[1] PE = α + β_{1} [wave] + β_{2} [place] + β_{3} [wave * place] + ε

where PE is the pro-embeddedness scale, α is the constant, wave includes dummy variables for the second and third waves, place is a series of dummy variables for either the major geopolitical groups or specific countries, wave*place is the full array of interactions between the wave dummies and the place dummies, ϵ is an error term, and the β’s are the coefficients representing the statistical effects. We estimate two equations: one using two dummy variables for the Former Communist group and the Global South group, respectively, with the Global North as the reference group; and one using a series of dummy variables for 19 of the 20 countries, with the United States as the reference group. The key coefficient is β₃, which indicates the extent to which pro-embeddedness sentiment changed between waves for each place.

Models 2 and 3 test our third hypothesis regarding the gender and income composition of the countermovement. Model 2 examines pro-embeddedness sentiment at a single point in time, and has the following form:

[2] PE = α + β_{1} [gender / income] + ε

where gender/income is a series of five dummy variables for each of the gender-income groups, with high-income males as the reference group. We estimate separate equations for each wave and place. This includes 24 equations (1 for the entire sample, 3 for each of the geopolitical groups, and 20 for each of the countries) for each of the three waves, for a total of 72 equations. Model 3 examines changes in pro-embeddedness sentiment between waves, and has the following form:

[3] PE = α + β_{1} [wave] + β_{2} [gender / income] + β_{3} [wave * gender / income] + ε

where wave*gender/income is the full array of interactions between the wave dummy variables and the gender/income dummy variables.

Findings

The main results for Model 1 are presented in Table 2. The top panel presents results for the larger geopolitical groups, and the bottom panel presents results for the more detailed country classification. The coefficients and significance tests are presented on the left side of the table. The first column for the ‘dummy’ variables indicates the absolute level of pro-embeddedness for each geopolitical group or country at wave 1. The right side of the table summarizes the changes in pro-embeddedness sentiment between waves, and indicates whether or not these changes are statistically different from zero. This table is relevant for our first two hypotheses.

Table 2.

OLS regression of pro-embeddedness scale on wave and group/country.

OLS regression I: Geopolitical groups (N = 87,649)				Change between waves
				1 to 3	1 to 2	2 to 3
Constant	8.96*	(Reference group: Global North)		0.49*	0.68*	−0.20*
Wave 2	0.68*
Wave 3	0.49*
	Dummy	Dummy	Dummy
		*Wave 2	*Wave 3
Former Communist	1.08*	1.84*	0.19	0.68*	2.53*	−1.85*
Global South	1.31*	0.57*	1.13*	1.61*	1.25*	0.36*
OLS regression II: Countries (N = 87,649)				Change between waves
				1 to 3	1 to 2	2 to 3
Constant	6.58*	(Reference group: United States)		1.22*	0.35*	0.87*
Wave 2	0.35*
Wave 3	1.22*
Wave 3	Dummy	Dummy	Dummy
		*Wave 2	*Wave 3
Global North
Finland	1.23*	1.34*	−0.26	0.96*	1.69*	−0.73*
Germany	1.09*	1.93*	−0.67*	0.55*	2.28*	−1.73*
Japan	5.09*	−0.53*	−1.89*	−0.67*	−0.18	−0.49*
Spain	4.44*	0.31	0.30	1.52*	0.67*	0.85*
Former Communist
Belarus	5.33*	1.04*	−2.43*	−1.21*	1.39*	−2.60*
Czech Republic	1.71*	3.27*	0.10	1.32*	3.62*	−2.31*
Estonia	3.06*	2.87*	0.88*	2.10*	3.23*	−1.13*
Lithuania	4.00*	0.52*	−1.59*	−0.37	0.88*	−1.24*
Poland	4.70*	0.14	−0.96*	0.26	0.50*	−0.24
Romania	3.26*	0.34	−1.89*	−0.67*	0.69*	−1.36*
Russia	4.20*	3.05*	−0.23	0.99*	3.41*	−2.42*
Global South
Argentina	2.33*	1.58*	2.06*	3.28*	1.94*	1.35*
Chile	5.88*	−0.58*	−0.64*	0.58*	−0.22	0.81*
China	5.44*	−0.09	−1.31*	−0.09	0.26	−0.35
India	3.09*	1.43*	1.45*	2.67*	1.79*	0.88*
Korea	2.42*	3.22*	2.48*	3.70*	3.57*	0.13
Mexico	3.18*	0.73*	0.46	1.68*	1.08*	0.60*
South Africa	3.01*	1.52*	0.91*	2.13*	1.87*	0.26
Turkey	5.46*	−1.13*	−1.90*	−0.68*	−0.78*	0.10

p < .05.

We begin with the simple Polanyian prediction about the double movement: given the global spread of neoliberal reforms throughout the 1990s, is there evidence of a corresponding global countermovement? The results suggest that a latent countermovement did emerge in the 1990s. At the aggregate level, pro-embeddedness sentiment increased significantly between waves 1 and 3 for all three geopolitical groups: by 0.49 for the Global North, by 0.68 in the Former Communist countries, and by 1.61 in the Global South. At the more detailed country level, pro-embeddedness sentiment was significantly greater in wave 3 than in wave 1 for 13 of the 20 countries (see Figure 1). This means that desire for public ownership and social protection was increasing precisely as governments pursued policies moving in the opposite direction. With the exception of Chile and Germany, each of these 13 countries had increases of at least 0.2 standard deviations.

Figure 1.

Change in pro-embeddedness between wave 1 and wave 3, by country.

Chile is an anomaly that helps to prove the rule, as it undertook an exceptionally early and intensive liberalization program in the 1970s and 1980s. Thus, its 1990s reforms were not as dramatic as those which were occurring in Mexico and Argentina during that decade (Mesa-Lago, 1997). This might explain the smaller increase in public demand for re-embedding in Chile. Further, Chile’s absolute level of support for re-embedding is the highest in the sample, both at the beginning and end of the decade, suggesting a more consistent opposition to neoliberalism. Germany is one of the two countries in the sample that arguably undertook very little liberalization in the 1990s (Gwartney et al., 2007). But its pro-embeddedness score does not stay flat. While Germany showed only a small net increase in pro-embeddedness, this is because of a sharp drop in the late 1990s after a very large spike in the early 1990s. We postulate that this pattern, which we observe universally in post-Soviet countries and discuss below, is due to the inclusion of East Germany in the sample.

Among those countries that did not show a significant increase, Japan, Belarus, Romania, and Turkey all had significant decreases between waves 1 and 3, while Lithuania, Poland, and China did not have a statistically significant change. Four of these cases are in the Former Communist group (Belarus, Lithuania, Poland, Romania), and all four did have significant increases in the first half of the 1990s. This increase was smallest in Poland, which may be due to the more gradual and smaller scale of its privatization reforms in the 1990s (EBRD, 2007; Stuckler et al., 2009).

This means that Japan, China, and Turkey (discussed in more detail below) are the only three countries that did not show signs of a latent countermovement emerging in the 1990s. While the evidence is not complete and universally overwhelming, it does strongly suggest that the Polanyian prediction holds true for the 1990s. This is especially true for the first half of the decade, when 16 of the 20 countries and all three of the geopolitical groups had significant increases in pro-embeddedness sentiment. As we show in the next section, the difference between the first and second halves of the decade highlights the varying trajectories taken by latent countermovements in different regions of the globe.

Countermovement in three worlds

The above analysis confirms the Polanyian prediction that global movement towards a self-regulating market has produced a global countermovement in public opinion for state regulation and social protection. But to what extent does the countermovement take different forms in different parts of the world?

The results show stark differences between countries, and noteworthy patterns associated with the broader geopolitical groupings. There is evidence of a latent countermovement in the Global North, with increasing pro-embeddedness sentiment over the course of the 1990s. However, the evidence also suggests that challenges to neoliberal reform are perhaps least likely to originate in this part of the world. As Figure 2 reveals, initial levels of pro-embeddedness sentiment in the Global North at the beginning of the 1990s are by far the lowest in the world, and increases over the course of the 1990s are still smaller than those in the Global South and former Communist countries. This suggests that countermovements for social protection are more likely to emerge outside of the Global North.

Figure 2.

Pro-embeddedness score by geopolitical group and wave.

Nevertheless, there is important variation within the Global North. The overall pattern is largely driven by the low initial levels of pro-embeddedness sentiment in the United States, Finland, and Germany, and the small increases and decreases shown by Germany and Japan. Japan’s decrease is explicable by the fact that it is one of the only countries in the sample that undertook no significant liberalization in the 1990s (Gwartney et al., 2007). Germany’s small increase is the result of a spike pattern – a significant increase in the first half of the 1990s followed by a significant decrease in the second half of the 1990s – that looks quite similar to the pattern found in the former Communist countries. As mentioned above, this makes sense given the incorporation of East Germany after 1989. Spain stands out as a Polanyian prototype with increases across the entire 1990s, which is all the more remarkable given its initially high level of pro-embeddedness at the beginning of the decade.

Although there is also internal variation within the other geopolitical groupings, perhaps the most intriguing finding is the stark contrast between the former Communist countries and the Global South. Here it is useful to analyze the first and second halves of the 1990s separately. With only three exceptions (China, Chile, Turkey), all countries outside of the Global North had increasing pro-embeddedness views in the first half of the 1990s, though the gains were somewhat greater in the former Communist countries than in the Global South. This is especially true of the Czech Republic, Estonia, and Russia, where pro-embeddedness increases between waves 1 and 2 were in the neighborhood of 3.0 points, or 0.7 standard deviations. It is also consistent with our second hypothesis, which predicts that the more abrupt market transitions in the post-Soviet countries produced a stronger latent countermovement in these places. The only country in the Global South to match these gains was Korea, though Argentina, India, and South Africa all had substantial gains that were larger than those in Belarus, Lithuania, Poland, and Romania.

The contrast in the second half of the 1990s is both more dramatic and more robust. With the exception of Poland, where pro-embeddedness sentiment remained stable, all of the former Communist countries had declining pro-embeddedness views in the second half of the 1990s. The declines were especially great in Belarus, Czech Republic, and Russia. These second half declines were not large enough to ‘undo’ the gains made by the Czech Republic, Estonia, and Russia during the first half of the 1990s, but they did outweigh the smaller first half increases in Belarus, Lithuania, and Romania. The situation was quite different in the Global South, where all eight countries had increasing or stable pro-embeddedness views in the second half of the 1990s. These increases were for the most part smaller than those shown in the first half of the 1990s, with small but statistically significant increases in Argentina, Chile, India, and Mexico, and no statistically significant difference between waves 2 and 3 in China, Korea, South Africa, and Turkey (see Figure 3).

Figure 3.

Change in pro-embeddedness, by country and wave.

These patterns go beyond those predicted by our hypotheses. Though we expected the large initial increases in the former Communist countries, due to the dramatic social dislocations and economic involution accompanying market reforms in the early 1990s (Burawoy et al., 2000; Stuckler et al., 2009; UNDP, 1999), we did not anticipate the subsequent decline. We suspect that the decline is related to a settling after the initial shock of the reforms, and the partial economic recovery experienced in most post-Soviet countries in the latter part of the 1990s. By 2000 there was evidently less concern about social protection than in the middle of the decade. It is important to note, however, that in three of the seven former Communist countries pro-embeddedness sentiment was still greater at the end of the 1990s than it was at the beginning, suggesting that there was a latent countermovement underway even if it waned somewhat towards the end of the decade. We also did not predict the deceleration of the countermovement in the Global South, but our general hypothesis nonetheless holds true, as pro-embeddedness sentiment remained stable or increased in most cases.

The real puzzles to be explained are China and Turkey. Unlike the other Global South countries (other than Turkey), each of which had increasing pro-embeddedness views during at least one half of the 1990s, China did not have any statistically significant increases at all. Why is there no evidence of a latent countermovement in China, where market reforms continued throughout the 1990s? We offer three potential explanations that we leave to China scholars to adjudicate. First, it is important to note that China begins the 1990s with already high levels of pro-embeddedness views. Market reforms really took hold in the 1980s with the onset of liberalization policies, so it may be that these initially high levels reflect an already developed countermovement in public opinion. Second, China is a special case due to its status as the new ‘workshop of the world.’ A potential countermovement may have been offset or slowed down as some people came to reap the benefits associated with the country’s high rate of economic growth. Third, and probably most important, China is the only socialist state in our sample. Although the Chinese state has certainly implemented neoliberal reforms, the continued strength of the Communist state apparatus means that the state has remained more clearly involved in regulating businesses and promoting social welfare than what occurred, for example, in post-reform Russia (Burawoy, 1996, 2001; Harvey, 2005). Likewise, a Communist ideology whereby the state is expected to undertake some measure of social protection remains strong, even in the context of strong support for market reforms (Blecher, 2002; Lee, 2002).

Turkey is more puzzling. It is the only country in the Global South with a significant decrease in pro-embeddedness sentiment between waves 1 and 3. While this decrease is small in absolute terms and concentrated in the early 1990s, and Turkey’s level of pro-embeddedness in wave 3 (11.4) remained comparable to Global South countries such as South Africa (11.7) and Mexico (11.4), the decline is nonetheless significant. Moreover, the absence of an increase in pro-embeddedness requires explanation given the near unanimous presence of latent countermovements in the Global South during the 1990s. A possible explanation is that the main thrust of Turkey’s liberalization occurred in the 1980s and then again after 2000, with relatively ineffective reforms in the 1990s (Tugal, 2009). Moreover, though there was significant financial liberalization in 1989, the early years of the 1990s were characterized by relatively high growth rates and the continuance of populist economic policies (Onis and Aysan, 2000). This may in part explain the decrease in pro-embeddedness sentiment in the early 1990s. Yet financial crisis, structural adjustment policies, employment decline, and land commodification all occurred (Bugra and Keyder, 2006; Onis and Aysan, 2000). Why this did not generate a significant countermovement in public opinion is a question that summons further investigation.

The social composition of contemporary countermovements

Table 3 summarizes the results from the gender and income analysis, which has two parts – one cross-sectional (Model 2) and the other longitudinal (Model 3). The cross-sectional piece examines the extent to which there are significant gender and income differences within waves, whereas the longitudinal analysis examines the extent to which change between waves varies by gender and income.

Table 3.

Summary of gender/income comparisons.

	Wave 1				Wave 2				Wave 3				All waves
Within-wave comparisons	+	−	0	+ as %	+	−	0	+ as %	+	−	0	+ as %	+	−	0	+ as %
Female (vs. male)
Low-income	9	1	14	37.5	8	0	16	33.3	9	0	15	37.5	26	1	45	36.1
Mid-income	12	1	11	50.0	9	0	15	37.5	11	0	13	45.8	32	1	39	44.4
High-income	12	1	11	50.0	6	0	18	25.0	11	0	13	45.8	29	1	42	40.3
All	33	3	36	45.8	23	0	49	31.9	31	0	41	43.1	87	3	126	40.3
Low-income (vs. high-income)
Female	17	0	7	70.8	19	0	5	79.2	16	0	8	66.7	52	0	20	72.2
Male	17	0	7	70.8	18	0	6	75.0	20	0	4	83.3	55	0	17	76.4
All	34	0	14	70.8	37	0	11	77.1	36	0	12	75.0	107	0	37	74.3
	Wave 1 to Wave 2				Wave 2 to Wave 3								All changes
Between-wave comparisons	+	−	0	0 as %	+	−	0	0 as %					+	−	0	0 as %
Female (vs. male)
Low-income	2	0	22	91.7	2	1	21	87.5					4	1	43	89.6
Mid-income	1	2	21	87.5	2	0	22	91.7					3	2	43	89.6
High-income	2	4	18	75.0	2	0	22	91.7					4	4	40	83.3
All	5	6	61	84.7	6	1	65	90.3					11	7	126	87.5
Low-income (vs. high-income)
Female	8	2	14	58.3	1	0	23	95.8					9	2	37	77.1
Male	4	2	18	75.0	1	0	23	95.8					5	2	41	85.4
All	12	4	32	66.7	2	0	46	95.8					14	4	78	81.3

Key:

+ = number of comparisons that are statistically significant at p < .05 and positive.

− = number of comparisons that are statistically significant at p < .05 and negative.

0 = number of comparisons that are not statistically significant.

The top half of Table 3 summarizes the results from the cross-sectional analysis. For each wave we ran 24 sets of comparisons: one for the entire sample, three for each of the geopolitical groups, and 20 for each of the countries. Each set of comparisons tests for statistically significant gender differences as well as income differences. The former identifies the effect of being female for low-income, middle-income, and high-income respondents; the latter identifies the effect of being low-income (compared to high-income) for females and males. In other words, for each place there are five comparisons: three gender comparisons and two income comparisons.

The results from the cross-sectional analysis show overwhelming support for our third hypothesis that the latent countermovement will be strongest among women and the poor. With only three exceptions (all three pertaining to Korean male advantages in wave 1) out of 360 comparisons – that is, across all three geopolitical groupings, all 20 countries, and all three waves – all of the statistically significant differences run in the same direction: pro-embeddedness views are greater for women than men, and greater for low-income respondents compared to high-income respondents. Of the 360 cross-sectional comparisons, more than half (54%) show a statistically significant difference in the expected direction. There is, however, more evidence of an income effect than there is of a gender effect. Only 40% of the gender comparisons show greater pro-embeddedness among women, whereas almost three-quarters (74%) of the income comparisons show greater pro-embeddedness among low-income respondents. This is consistent with the ‘agonistic’ Polanyi hypothesis that the least privileged groups will lead the countermovement.

The longitudinal analysis paints a different picture. The bottom half of Table 3 replicates the format of the cross-sectional analysis, but examines changes in pro-embeddedness sentiment between waves instead of average levels during a given wave. There are two sets of comparisons for each of the transitions: the change between waves 1 and 2, and the change between waves 2 and 3. This analysis reveals the extent to which female and low-income respondents had greater increases in pro-embeddedness over time. In contrast to the cross-sectional analysis, in this analysis only 15% of the contrasts show a statistically significant difference. Most of the statistically significant differences (roughly two-thirds) are in the hypothesized direction, with women and low-income respondents showing the greatest increases. The most prevalent effect is associated with greater increases among low-income women as compared to high-income women between waves 1 and 2, and is most prominent in the former Communist countries.

Table 4 summarizes the changes in embeddedness views for each country-gender-income group. Low-income and middle-income women, as well as low-income and high-income men, all had significant increases between waves 1 and 3 in more than half of the 20 countries. Likewise, at least four of the six gender-income groups had significant increases between waves 1 and 3 in 10 of the 20 countries. These results provide partial support for the ‘organic’ Polanyi hypothesis in which the countermovement is the response of society as a whole.

Table 4.

Summary of changes between waves 1 and 3, by gender-income group and country.

Country	Female			Male
	Low-income	Middle- income	High-income	Low-income	Middle- income	High-income
Finland		+		+		+
Germany
Japan		+		−
Spain	+	+	+	+	+	+
United States		+	+	+		+
Belarus	−	−	−			−
Czech Republic	+	+		+	+	+
Estonia	+	+		+	+	+
Lithuania
Poland	+
Romania			−		−
Russia	+	+		+	+	+
Argentina	+	+	+	+	+	+
Chile						+
China			−
India	+	+	+	+	+	+
Korea	+	+	+	+	+	+
Mexico	+	+	+	+	+	+
South Africa	+	+	+	+	+	+
Turkey	−				−
Countries with + (N)	10	12	7	11	9	12
Countries with + (%)	50	60	35	55	45	60

Key:

+ = significant increase in pro-embeddedness between wave 1 and wave 3.

− = significant decrease in pro-embeddedness between wave 1 and wave 3.

Conclusion

The above analysis presents a broad overview of the global countermovement in public opinion towards states and markets during the 1990s. To our knowledge it is the only systematic quantitative exploration to date that has been done through a Polanyian lens. Our analysis provides compelling evidence for the existence of a latent global countermovement, in the form of a widespread increase in desire for re-embedding the market at a time of global disembedding. This finding provides systematic evidence for the existence of widespread and generally increasing opposition to neoliberalism across all regions of the globe during the 1990s. Our study, of course, does not itself show the existence of politically organized countermovements. But it suggests that there was a significant increase in latent opposition to a disembedded market during the 1990s that was, and may still remain, available for organization by political actors.

We also reconstruct and extend Polanyi by suggesting where and among whom these countermovements might emerge in practice. The latent potential for organized movements seems to lie most significantly with the women and low-income people of the Global South. Low-income women of the Global South appear to be the strongest advocates of re-embedding and thus might be expected, to fill the ranks of contemporary ‘countermovements.’ But our data also suggest that more dominated groups might increasingly find allies in other classes.

The general conclusion, then, is that neoliberalism appears to be producing its own potential re-embedders. Whether this latent potential becomes manifest through organized political opposition remains an open question. What seems clear is that Polanyi’s ‘double movement’ has some objective basis in global public opinion.

Footnotes

Acknowledgements

We would like to thank Michael Burawoy, Peter Evans, participants at the 2008 annual meetings of the Society for the Advancement of Socio-Economics in San Jose, Costa Rica, the International Sociology editors and two anonymous reviewers for helpful comments on earlier versions of the article.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

Author biographies

Michael Levien is a PhD candidate in the Sociology Department at the University of California-Berkeley. His dissertation, based largely on ethnographic research, examines the political economy of land dispossession for Special Economic Zones in India.

Marcel Paret is a PhD candidate in the Sociology Department at the University of California-Berkeley. His dissertation is a comparative historical study of citizenship and class formation in the United States and South Africa between 1950 and 2010.

References

Agenor

P-R

(2004) Macroeconomic adjustment and the poor: Analytical issues and cross-country evidence. Journal of Economic Surveys 18: 351–408.

Arrighi

Silver

(2003) Polanyi’s ‘double movement’: The belle epoques of British and US hegemony compared. Politics and Society 31: 325–355.

Benería

(2003) Gender, Development, and Globalization: Economics as if All People Mattered. New York: Routledge.

Blecher

(2002) Hegemony and workers’ politics in China. The China Quarterly 170: 283–303.

Block

(2003) Karl Polanyi and the writing of The Great Transformation. Theory and Society 32: 275–306.

Bugra

Keyder

(2006) The Turkish welfare regime in transformation. Journal of European Social Policy 16: 124–134.

Burawoy

(1996) The state and economic involution: Russia through a China lens. World Development 24: 1105–1117.

Burawoy

(2001) Transition without transformation: Russia’s involutionary road to capitalism. East European Politics and Societies 15: 269–290.

Burawoy

(2003) For a sociological Marxism: The complementary convergence of Antonio Gramsci and Karl Polanyi. Politics and Society 31: 193–261.

10.

Burawoy

Krotov

Lytkina

(2000) Involution and destitution in capitalist Russia. Ethnography 1: 43–65.

11.

Cagatay

Erturk

(2004) Gender and Globalization: A Macroeconomic Perspective. Geneva: International Labor Organization.

12.

Dumenil

Dominique

(2004) Neoliberal income trends. New Left Review 30: 105–133.

13.

EBRD (European Bank for Reconstruction and Development) (2007) Transition Report 2007: People in Transition. London: EBRD.

14.

Evans

(2000) Fighting marginalization with transnational networks: Counter-hegemonic globalization. Contemporary Sociology 29: 230–241.

15.

Evans

(2005) Counter-hegemonic globalization: Transnational social movements in the contemporary global political economy. In: Janoski

(ed.) The Handbook of Political Sociology: States, Civil Societies, and Globalization. Cambridge: Cambridge University Press, pp. 655–671.

16.

Evans

(2008) Is an alternative globalization possible? Politics and Society 36: 271–305.

17.

Fourcade-Gourinchas

Babb

(2002) The rebirth of the liberal creed: Paths to neoliberalism in four countries. American Journal of Sociology 108: 533–579.

18.

Gwartney

Lawson

Sobel

Leeson

(2007) Economic Freedom of the World: 2007 Annual Report. Vancouver: The Fraser Institute.

19.

Haggard

Webb

(1994) Voting for Reform: Democracy, Political Liberalization, and Economic Adjustment. Washington, DC: The World Bank.

20.

Hardt

Negri

(2000) Empire. Cambridge, MA: Harvard University Press.

21.

Hart

(2002) Disabling Globalization: Places of Power in Post-Apartheid South Africa. Berkeley, CA: University of California Press.

22.

Harvey

(2005) A Brief History of Neoliberalism. Oxford: Oxford University Press.

23.

Keck

Sikkink

(1998) Activists Beyond Borders: Advocacy Networks in International Politics. Ithaca, NY: Cornell University Press.

24.

Lee

(2002). From the specter of Mao to the spirit of the law: Labor insurgency in China. Theory and Society 31: 189–228.

25.

Levien

(2007) India’s double movement: Polany and the National Alliance of People’s Movements. Berkeley Journal of Sociology 51: 119–149.

26.

Mertes

(2004) A Movement of Movements: Is Another World Really Possible? London: Verso.

27.

Mesa-Lago

(1997) Social welfare reform in the context of economic-political liberalization: Latin American cases. World Development 25: 497–517.

28.

Onis

Aysan

(2000) Neoliberal globalisation, the nation-state and financial crises in the semi-periphery: A comparative analysis. Third World Quarterly 21: 119–139.

29.

Ó Riain

(2000) States and markets in an era of globalization. Annual Review of Sociology 26: 187–213.

30.

Polanyi

(2001 [1944]) The Great Transformation: The Political and Economic Origins of Our Time. Boston, MA: Beacon Press.

31.

Roberts

(2008) Privatizing social reproduction: The primitive accumulation of water in an era of neoliberalism. Antipode 40: 535–560.

32.

Rowbotham

Linkogle

(2001) Women Resist Globalization: Mobilizing for Livelihood and Rights. London: Zed Books.

33.

Ruggie

(1982) International regimes, transactions and change: Embedded liberalism in the postwar economic order. International Organization 36: 195–231.

34.

Smith

(2008) Neoliberalism, welfare policy, and feminist theories of social justice. Feminist Theory 9: 131–144.

35.

Somers

Block

(2005) From poverty to perversity: Ideas, markets, and institutions over 200 years of welfare debate. American Sociological Review 70: 260–287.

36.

Stiglitz

(2003) Globalization and its Discontents. New York: WW Norton.

37.

Stuckler

King

McKee

(2009) Mass privatization and the post-communist mortality crisis: A cross-national analysis. TheLancet 373: 399–407.

38.

Tarrow

(2005) The New Transnational Activism. Cambridge: Cambridge University Press.

39.

Tugal

(2009) Passive Revolution: Absorbing the Islamic Challenge to Capitalism. Stanford, CA: Stanford University Press.

40.

UNDP (United Nations Development Program) (1999) Transition 1999: Human Development Report for Eastern Europe and the CIS. New York: UNDP REBEC.

41.

Wade

(2004) Is globalization reducing poverty and inequality? World Development 32: 567–589.

42.

Wade

(2005) The causes of increasing world poverty and inequality: Or, why the Matthew effect prevails. International Journal of Health Services 35: 631–653.

43.

Watts

(2000) Poverty and the politics of alternatives at the end of the millennium. In: Pieterse

(ed.) Global Futures: Shaping Globalization. London: Zed Books, pp. 133–147.