State-society relations and the “re-industrialization” of the developed world: Power relations, coordination and growth paths

1. Introduction

The Covid-19 crisis heated the discussion that had been running for more than a decade in the developed world. The growing interest in the developed world on industrial development, or “re-industrialization”, went beyond academic debate and knew its way to government policies and strategies. The Fourth Industrial Revolution (Industry 4.0)- with its focus on digital production technologies, management processes based on information technology (IT), and new materials-promises more flexibility, efficiency, and limited labor-cost component. Such benefits encourage the reshoring of manufacturing activities to the developed world (Kinkel, 2020; Kinkel, Pegoraro, & Coates, 2020; Pegoraro, Propris, & Chidlow, 2020). Moreover, the de-industrialization in manufacturing increased the developed world vulnerability to external shocks and harmed their economic diversity (Pegoraro et al., 2020, p. 155). Investing in manufacturing activities should address such shortcomings. The reshoring process is now a reality, especially for US and to a lesser degree EU firms. The USA is increasingly encouraging such a development to create more jobs (Kinkel et al., 2020). The re-industrialization of the advanced economies seems to reverse the decades-long de-industrialization process that marginalized the sector to the benefit of the presently-dominant service sector.

Reindustrialization calls on the new-old debates on how to boost industrial innovation and productivity, as well as how automation and the new technologies could lead to loss of jobs and harm equality of distribution to the favor of capital owners. This brings industrial policy back to the forefront and with it the usual actors that mostly shape and benefit or lose from it, businesspeople and workers. Investigating state-society relations, and particularly state-business-labor relations (SBLR), should lay at the center of any trial to understand and predict the growth path of industrialization, in the past and present alike.

Many bodies of literature would be helpful in this regard. The first is the literature on varieties of capitalism (VoC). It differentiates between countries according to the way various actors coordinate their actions, categorizing developed countries into coordinated market economies (CME) and liberal market economies (LME) (see Hall & Soskice, 2001; Hall & Thelen, 2008; Martin, 2005). Another important body of literature is that on state-business relations (SBR). It differentiates between formal (e.g.: public-private dialogues-PPD) and informal (e.g.: family or political connections) relations connecting state officials and businesspeople and how these forms affect economic outcomes (see Hausmann, Rodrik, & Sabel, 2008; Hausmann & Rodrik, 2003; Enderwick, 2005, and Haggard, Maxfield, & Schneider, 1997) see as well Sabry (2017, 2018). A third relevant body of literature is the one on industrial relations (e.g. Glassner & Keune, 2012; Hayter, 2018; Streeck, 2008), and especially the pluralist perspective that accepts the presence of conflict between businesspeople and labor but believes in the possibility of containing it, for instance, by collective bargaining (Hayter, 2018, p. 3–4) [1]. SBLR seek to integrate these bodies of literature in order to build a more comprehensive understanding the dynamics of state-society relations and their economic consequences.

This paper is using the SBLR theoretical framework as introduced in Sabry (2022, 2023) for studying the effect of state-society relations on the industrial growth paths of developed countries. The novelty of this work is mainly the use of this framework in understanding why SBLR lead to different policy choices and empirically investigating the outcome of these choices on the growth path of developed countries. Four main actors are identified: the state, big businesspersons or tycoons, owners and managers of small and medium enterprises (SMEs) or Entrepreneurs, and labor. Different SBLR categories or modes are introduced, depending on levels of coordination and power relations between the actors. Then, the paper investigates how these SBLR modes shape industrial policy and the economic outcomes from different policies, leading to different growth paths. A growth path here is perceived as a matrix of (industrial-related) levels of long-run growth and equality of distribution. Given that the level of coordination is an important dimension in drawing the lines between different modes, this paper focuses on developed countries, which are at the center of the literature on VoC. Although these countries are predominantly from Europe and European settled former colonies in North America and Oceania, two East Asian economies are also included, Japan and South Korea.

The SBLR framework depends on the above-highlighted bodies of literature. It differentiates between tycoons and the rest of businesspeople, an important distinction that often goes unnoticed. It is proposed in this research that the effect on growth paths goes beyond the simple dichotomy between CMEs and LMEs usually present in the literature of VoC and that power relations provide an essential complementary dimension in explaining this causality. Arguably, the economic outcomes of higher coordination when the different actors have a balanced power relation would not be the same as when one actor, such as tycoons, is more dominant than other actors. Such power relations would affect government strategies, policies, and regulations which will ultimately affect various economic outcomes. In other words, levels of coordination together with inter-actors power relations strongly influence the growth path of different countries.

The paper reaches empirical findings that suggest that the presented SBLR modes lead to the formulation of different policies and to varying outcomes from the implementation of these policies. The Co-Balanced mode-having higher levels of coordination, more balanced state-society relations, and lower favoritism-leads to the best growth path among the four introduced modes. To the contrary, the Lib-Capture mode-characterized by lower coordination and higher favoritism-seems to have the worst growth path.

The rest of the paper is structured as follows. Section 2 introduces the theoretical perspective of this paper, introducing different SBLR modes, how they lead to the selection of different policies, and how the interaction of SBLR modes and policies lead to different growth paths. This is followed by the methodology section, where statistical and regression analysis is conducted to empirically investigate the suggestions of the previous section. The next section has a discussion on the results of the empirical study. The paper ends with a conclusion that summarizes the empirical findings, brings them together with theory, and then provides policy and research insights.

2. Theoretical perspective

• a- The SBLR Framework

State-society relations take various forms with different characteristics. Those forms depend on the dynamics governing the interaction between the main actors in these relations. In industrialized societies, the main actors in state-society relations are the state, businesspeople, and labor. Accordingly, state-business-labor relations (SBLR) represent the part of state-society relations relevant to the industrial sector. Big business managers and owners, or business tycoons, could be introduced as a distinctive actor, where they could be identified by the market power that they have in their respective sectors. Entrepreneurs, on the other hand, will refer to managers and owners of small and middle enterprises (SMEs). Thus, it is assumed here that there are four main players: the government, tycoons, entrepreneurs, and labor.

The literature on the varieties of capitalism (VoC) suggests a dimension of state-society relations based on which countries could be categorized. This dimension is the level of coordination between different actors in SBLR. Hall and Soskice (2001, pp. 6–7) (the foundation stone of the VoC literature) studied how different state-society actors rationally pursue their interests “in strategic interaction with others”, with a special focus on firms as central players in a capitalist economy. Studying the level of coordination among actors is at the heart of this approach. In this regard, VoC investigates how firms manage their relations internally (with their workers) and externally (with labor unions, the government, business associations…etc.). Firms’ coordination mechanisms differ in five major relation spheres: industrial relations, vocational training and education, corporate governance, inter-firm relations, and management-employees relations. Industrial relations, for instance, studies how entrepreneurs (and tycoons) coordinate with their labor force the process of wage bargaining and working conditions, while inter-firm relations investigate, among others, inter-firm collaboration in technology transfer and research and development (R&D).

According to the coordination mechanism used by firms, countries are categorized into CMEs (e.g.: Scandinavian countries, Germany, and Japan) and LMEs (e.g.: USA, UK, Australia) [2]. Coordination in LMEs is achieved by market mechanisms: “arms-length market exchange”, formal contracting, and price signals (Hall & Soskice, 2001; Ahlborn, Ahrens, & Schweickert, 2014). In CMEs, to the contrary, coordination is more intensively realized by non-market relations, or “institutional coordination among societal stakeholders” (Herrigel & Zeitlin, 2010, pp. 668–670). This includes widespread “relational or incomplete contracting”, intra-network exchange of private information, and a higher tendency to build firms’ competencies through collaborative relations (Hall & Soskice, 2001, p. 8). Coordination-oriented policies are more feasible in CMEs whose political system permits businesspeople to have “substantial structural influence” (Hall & Soskice, 2001, p. 48). Institutional complementarities exist, where using a coordination mechanism in one sphere would urge the development of the same mechanism in others (Hall & Soskice, 2001, p. 18 and Hall & Gingerich, 2009).

A good proxy for measuring the level of coordination is the indicator introduced by Witt and Jackson (2016). To construct this indicator, the authors combined data on the above-mentioned five relation spheres presented by Hall and Soskice (2001). The average of the last three years for which data on this indicator is available (2001–2003) is used here as the coordination indicator. Given data limitation for this indicator, this average will be assumed to be fixed and applicable for the empirical analysis done later in this paper which covers the period between 2007 and 2018.

Little is said in the VoC literature, however, on power relations between the four actors. Power relations should point to who benefits from more coordination? Are tycoons and entrepreneurs benefiting equally from government involvement and coordination? Does lack of coordination always leave labor at a disadvantage? These and other similar questions point to the need of integrating an additional SBLR dimension. This is suggested here to be a power dimension, resting on the relative power of state-society actors vis a vis each other. Such dimension is based on the literature on state-business relations (SBR) and industrial relations. This is suggested here to be favoritism, which usually is to the advantage of tycoons and at the expense of other social actors such as entrepreneurs and labor and at times even at the expense of state autonomy. Such relations are usually referred to in the literature as state-capture (see Adly, 2010; Enderwick, 2005). In developed democratic countries (as is the case in the studied 22 countries according to the World Bank’s voice and accountability indicator) [3], favoritism is more likely translated into state capture, rather than crony capitalism that is more common in developing and authoritarian states (Sabry, 2023). A more open political system with an “incomplete process of political liberalization” encourages state capture through various means, including the financing of parliamentary election campaigns (see Hellman, Jones, & Kaufmann, 2003; Innes, 2013). Although the open political system permits labor and entrepreneurs to organize and aggregate their power, they are unable to balance the political and economic power that tycoons have. Entrepreneurs face considerable challenges in their trial to forge collective action even in a democratic system. This includes their isolation that hinders monitoring participation and preventing free-riding (unlike factory workers) and their inability to provide large electoral votes (like workers) or significant financial support to political candidates (like tycoons) (Shadlen, 2002). Moreover, tycoons might dominate business associations or entrepreneurs might be so reliant on tycoons, as the discussion below would reveal. On the other hand, in the post-Fordist order-dominant since the 1980s-developed countries have witnessed diminishing union membership, decentralization of the collective bargaining mechanism, and a fall in bargaining coverage (see Eurofound, 2016; Glassner & Keune, 2012, p. 368). The literature on labor market segmentation attributes diminishing unionized labor power to workers’ segmentation and fragmentation into different groups based on skills, industrial sector, and other criteria (see Streeck, 2009; Reich, Gordon, & Edwards, 1973). Acemoglu, Aghion, and Violante (2001) argued that de-unionization took place because of “skill-biased technical change”. This bias widened productivity differentials between workers and provided incentives for skilled labor to seek the “competitive market return” rather than joining forces with unskilled labor and taking part in collective bargaining through labor unions.

To the contrary, when the organizational power of the different actors brings a more balanced form of power relations among them and limits chances for favoritism, one can refer to this as “balanced-power” relations. In balanced-power relations, entrepreneurs and labor can organize and their organizations are less fragmented or filled with rivalry. For labor, this could be realized by having labor unions that represent workers on the national, sectoral, and firm levels, as is the case in Germany (Traub-Merz, 2010, p. 7). These conditions should limit the power of tycoons and favoritism.

A good indicator that helps in differentiating between the highlighted-above two forms of power relations is the favoritism indicator of the Global Competitiveness Indicators’ (GCI) dataset. This measures “to what extent do government officials show favoritism to well-connected firms and individuals when deciding upon policies and contracts?” (Schwab & Sala-i-Martín, 2017, p. 342). Data for the indicator is available for the whole period studied later in the empirical section of this paper (2007–2018).

Differentiating between various SBLR modes depending on levels of coordination and power relations, with a focus on favoritism, would lead to identifying four distinctive modes: Coordinated-balanced-power, coordinated-state-capture, liberal-balanced-power, and liberal-state-capture. Table 1 shows which countries belong to each of these modes. To design this table and easily identify the four modes, the scale of the used indicators, coordination and favoritism, are reconstructed to run from +50 to −50. The geometric and arithmetic means of the (absolute) favoritism and coordination scores of the studied countries are also reported. Geometric means should be given more credit while ranking the countries within each SBLR mode. This measurement pushes down the rank of countries having a low scores in either of the constituent variables (coordination and favoritism) and ensures that higher mean values reflect higher scores in both variables. Figure 1 visualizes the categorization of the studied countries using coordination and favoritism graph coordinates. The more countries are placed at the corners of the graph’s frame, the more they exhibit characteristics of their mode; the more central they are placed the less they show these characteristics. The characteristics of the four modes are discussed below.

3. Methodology

In this section, regression analysis is used to investigate the effect of various policies implemented in different SBLR modes on the growth paths of the studied countries. The assumption made here is that SBLR modes could strengthen, mitigate, or reverse the effects of these policies. The regressions are run for the studied 22 developed countries and the period between 2007 and 2018. A summary statistics of the used variables is reported in Table A1 in the Appendix.

The conducted regressions have the following general equation:

The dependent variable (Y) is labor productivity in the first, number of patents in the second, and inequality in the third regression. The proxies used for the different considered variables are calculated as explained in the previous section and reported in Table 3.

Dummy variables are constructed for the four SBLR modes. For instance, Co-Balanced SBLR has a positive level of coordination (coordination>0) and a negative level of favoritism (favoritism<0). For each of the studied countries, categorization differs from one year to the other according to each country’s score in coordination and favoritism. To avoid multicollinearity, the Lib-Balanced SBLR dummy is excluded and used as the benchmark. Gross domestic product (GDP) per capita, coordination, and favoritism are used as control variables. The latter two variables are rescaled into an ascending percentage scale (from 0 to 100). Both variables are used to test the effect of different levels of coordination and favoritism on growth paths regardless to the four SBLR modes (see Table 3).

The two-stage least squares (2SLS) model is used to run the different regressions [4]. 2SLS is used because of concerns on endogeneity in terms of reverse causality. Running the Hausmann test on the conducted regressions confirmed that all regressions are better run by 2SLS. As instruments, a number of legal (British, French, and German legal origins), cultural (Protestant and Catholic percentages and ethnic fractionalization), and social (size in terms of population) variables are used. Instrumented variables are the policy variables and their interaction with SBLR dummies (as well as the GDP per capita indicator). SBLR dummies, however, are treated as exogenous, and the same is true for the coordination and favoritism indicators. It is reasonable to think that labor productivity, patent creation, and Gini index have little (if any) causality effect on the SBLR modes or their two constituting indicators which are rather deeper entrenched institutions that are largely the outcome of path dependent historical developments, especially in the developed world.

To avoid multicollinearity, each of the studied five policies are introduced individually (together with its interaction terms with SBLR dummies) in a sperate regression. Finally, sequential elimination of statistically insignificant variables is done automatically using the Gretl program’s function “sequential elimination at 10%” which guarantees, through an F-Test, that removing insignificant variables did not considerably alter the consistency of the original regressions. The process led to maintaining only statistically significant variables and these were mostly highly significant variables (at the 1% of significance).

The results of the conducted regressions are reported in Tables 5–7, one table for each dependent variable.

4. Results and discussion

Table 5, reporting the regressions that have labor productivity as the dependent variable, show the following. The control variables of favoritism and coordination both tend to have statistically significant positive effects on labor productivity. Given that the Lib-Balanced mode is the benchmark, venture capital availability increases productivity in both Lib-Capture and Co-Balanced modes, where a 1% increase in venture capital leads to a comparable increase in productivity in both modes (3.3% and 3.2% respectively given the use of log for the dependent variable). Government technological procurement leads to a slightly negative effect on labor productivity only in the Co-Balanced mode (1% leading to −0.5%). It is to be noted that the level of government technological procurement in this mode is the highest in comparison to the other modes, and thus it might be rather too excessive. Longer starting business durations decrease labor productivity in both the Co-Balanced and Lib-Capture modes, with a higher effect on the former. Stated differently, more easiness of starting business has a stronger effect on productivity in Co-Balanced followed by Lib-Captured modes. Higher labor taxes decrease productivity in the Lib-Capture followed by the Co-Capture mode. This provides some evidence supporting the arguments made by He and Liu (2008) about how higher labor taxes could discourage skill formation, especially in the Co-Capture mode where it is the highest among the four modes. Profit taxes, however, have varying effects. They increase labor productivity in Co-Balanced followed by Lib-Capture modes and decrease it in the Co-Capture mode. The positive effect of profit taxes in the Co-Balanced mode supports the arguments made by Coccia (2018) that, at low levels, raising profit taxes increases productivity (where the tax has the second lowest level in the Co-Balanced mode). Generally, it could be said that concerning policies in which the Co-Balanced mode has one of the two highest levels among the four modes, productivity is fostered by venture capital availability and easiness of starting business. For the Lib-Capture mode, this is rather the case for its comparatively high levels of profit tax, while it would benefit from raising both venture capital availability and easiness of starting business for which the mode has comparatively low levels. On the other hand, the remarkable implemented policy that benefits labor productivity in the Co-Capture mode is rather having low profit taxes (the lowest among the four modes), where lowering it further would increase productivity.

Table 6 reports the regressions having patents per million inhabitants as the dependent variable. In two of the conducted five regressions, favoritism has a negative effect on patents while coordination has a positive effect in most of the regressions, both matching the findings of Sabry (2019) that favoritism generally harms innovation while manifestations of high coordination such as broadly-participated PPD have a general positive effect, especially in more open political systems. Venture capital availability increases patents in the Co-Balanced mode but decreases it in the Lib-Capture mode. Government technological procurement sharply decreases patents (a 1% increase in procurement decreases patents by about 27 patents per million inhabitants), again suggesting that the excessiveness of this policy in this mode might be responsible for unfavorable economic outcomes. Longer durations for starting business sharply decreases patents in the Lib-Capture followed by Co-Capture mode (1% causing a fall of 27 and 21 patents, respectively). The two modes have on average the highest duration for starting business among the SBLR modes, suggesting that innovation is likely better supported by a moderate level of market entry and competition (see Aghion & Griffith, 2005; Goto, 2009), in a way that balances the entrance of entrepreneurs’ innovative firms and protecting the market share of tycoons. Higher labor taxes decrease patents in the three considered modes, with the highest effect being on the Lib-Capture, followed by the Co-Balanced, then the Co-Capture mode (an increase of 1% decreasing patents by 23, 16, and 9 per million inhabitants, respectively). Similarly, profit taxes reduce patents but only in the Co-Balanced followed by the Co-Capture mode (1% increase causing a fall of 27 and 12 patents, respectively). Generally, for the Co-Balanced mode, the most effective implemented policy in increasing patents is venture capital availability. On the other hand, for both Co-Balanced and Co-Capture modes low levels of profit taxes seem to be the most beneficial policy for patents creation.

Finally, Table 7 reports the results of the regressions that have the Gini index as the dependent variable. Favoritism has a statistically significant positive effect on the Gini index (lead to higher inequality) in most of the conducted regressions while the opposite is true for coordination. Venture capital availability decreases the Gini index, and thus inequality, only in the Co-Balanced mode but raises it in both the Co-Capture followed by the Lib-Capture mode. The effect is the strongest in the Co-Balanced mode (a 1% increase in venture capital availability leading to a decrease of 0.436% in inequality). Government technological procurement increases inequality in the Co-Capture mode, suggesting a negative impact of favoritism on allocating procurement offers, even when the mode has comparatively the lowest level of the policy among the SBLR modes. Longer duration for starting business expectedly increases the Gini index in the three considered modes, with the biggest effect being on the Co-Balanced mode. Labor taxes increase the Gini index in the Co-Capture mode (the mode has and by a wide margin the highest level of the policy). Finally, profit taxes decrease the Gini index and inequality in both the Co-Balanced and Co-Capture modes. Generally, venture capital availability again proves to be the best policy for the Co-Balanced mode where it decreases inequality as it increases labor productivity and patents. The second policy for which high levels are beneficial for the mode is the easiness of starting business as it decreases inequality the same as it increases labor productivity. For the Co-Capture mode, lower government technological procurement is a beneficial implemented policy for this mode with regard to equality.

5. Conclusion

This research argues that state-society relations, and especially SBLR, strongly affect the growth paths of developed countries, whether in terms of long-term industrially-led economic growth and/or equality of distribution. The differentiation between SBLR modes goes beyond the usual coordination dichotomy between LMEs and CMEs, present in the VoC literature, and integrates a very important dimension which is power relations between the various studied actors. This presents coordination and favoritism as the two important determining dimensions for differentiating between four main SBLR modes in developed countries.

The empirical results, depending on regression analysis and statistical data, offer many interesting findings. The Co-Balanced mode having higher levels of coordination and lower favoritism enjoys the best growth path among the four modes. Statistically, Co-Balanced countries have better long-run economic growth prospects and the lowest level of inequality. The lower levels of favoritism in this mode lead to more balanced power relations that encourage SMEs and their innovation; and this is reflected in having the highest level of venture capital availability and a comparatively shorter duration for starting a business. Higher levels of coordination added to low favoritism seem to make the best use of these two policies on growth path indicators as the empirical results have shown.

On the other hand, the Lib-Capture mode, characterized by lower coordination and higher favoritism, seems to have the worst growth path. Statistically, Lib-Capture countries have comparatively lower long-run economic growth prospects and the highest levels of inequality. Higher favoritism seems to provide a bias against SMEs, leading to a longer duration for starting a business and low availability of venture capital. The latter is suggested by the empirical results to affect equality in this mode. The only policy that the countries of this mode have on average high levels and that at the same time brings some benefits is high profit taxes which increase labor productivity.

Between the above-mentioned two poles of distinctive modes, two other SBLR modes with different policy profiles and growth paths exist. The most remarkable outcome in these two modes is the notably low level of patent creation (i.e.: innovation) in Co-Capture countries. The heavy impact of high favoritism in this mode on this outcome could be arguably be detected from the negative effect of strict market entry-the strictest in all four modes-on patent creation. The countries of this mode in which entrepreneurs are likely the main losing actor would benefit from policies and regulations that increase their relative economic power vis a vis tycoons. The policies that seem to bring benefits are actually that which involve less government intervention: lower profit taxes for labor productivity and patents and low government technological procurement for inequality. In comparison to the Co-Balanced mode, the results point to how government intervention and the benefits of high coordination are distorted by favoritism that renders government involvement less beneficial.

Hence, to realize the best industrial growth path, fighting favoritism should be an important objective. The negative impact of favoritism on innovation could not be disregarded in the eve of the fourth industrial revolution, where innovation is increasingly pivotal to future industrial development. Actively engaging societal groups in the policymaking process is important in addressing their concerns and balancing them at the same time. This should lead to the double benefit of formulating better policies that should foster growth as well as provide better distribution of this growth. High levels of coordination should help in realizing this objective. Yet, this could only be possible if societal groups are free to associate and aggregate their power and when there are means of preventing one actor from gaining more favorite treatment and exclusive influence over policymakers. As pointed out by Hall and Thelen (2008), capacities for coordination are much in doubt if labor lost the belief that the existing industrial relations provide them with “a just set of rewards”. The presence of both powerful and broadly represented business associations and labor unions and the existence of a government interested in coordinating their efforts-rather than letting itself be controlled by one group at the expense of the others-should help in the realization of the best growth path. Thus, institutional reform that empowers societal groups and enables them to defend their interests as well as fights all forms of corruption should lead to the realization of a more prosperous and equitable industrial development, with the “re-industrialization” of the developed world being no exception. The technological and social challenges of intensive automation and digitalization accompanying the fourth industrial revolution make the envisaged institutional reform more urgent.

The discussion is specifically relevant for the Lib-Capture mode, for which South Korea is the best representative, as discussed in the paper. The country’s economic development has been a textbook case for a successful economic development process, with the term developmental state being synonym to its experience (Evans, 1989; Lee & Han, 2006). Critical accounts pointed to this experience though as crony capitalism with high levels of favoritism to connected tycoons who were however subservient to the more dominant state (Enderwick, 2005; Kang, 2003). The democratization of the state seemed to flip the balance between the state and tycoons, to the latter’s favor, but favoritism survived and the country witnessed an institutional transformation into state capture (Sabry, 2023). The comparatively poor figures in the considered variables of the growth path of the Lib-Capture mode suggest that South Korea needs to yet move to the left of Figure 1 to follow the historical developmental path of its historical role model, Japan (Evans, 1989), which is in the Lib-Balanced mode; or it might even further move up as well towards the Co-Balanced mode. In either cases, curbing favoritism would be the key for this institutional change that would open more chances for further economic development. This is the case as well for other countries for which the characteristics of the Lib-Capture mode are relatively strong.

Despite the important findings that this research has reached, this paper is mainly meant to open a further investigation into this topic and open this dimension that the research on VoC and political economy have under-researched. A deeper investigation of SBLR typologies that could only be possible by having richer datasets with more data on coordination for the whole world, rather than only the advanced economies, would further our understanding of the dynamics that shape the growth paths of different countries of the world. While acknowledging the limitations of this paper, it is hoped that this research would motivate the collection of richer empirical data that would boost research in this field.