Article body

“Let Faith be substituted for Fear; let mutual consideration and confidence supplant suspicion, and constructive good-will replace resistance; let the parties to Industry recognize a mutuality, not a conflict of interest, in all that pertains to maximum production and equitable distribution of wealth; and what is the result? Immediately, fresh energies are released, a new freedom is given to effort in industry. Productivity is increased, as are also the respective rewards of all the parties.”

W. L. Mackenzie King (1918: 172)

Introduction

The employment relationship in Canada and the United States (US) became increasingly dysfunctional and conflict-prone from the early 1870s to the end of World War I, symbolized by general strikes in both Winnipeg and Seattle in 1919 (Kealey, 1984; Adams, 1966). The field of industrial relations (IR) was born in the 1910s to study and understand the dynamics of the employment relationship and methods by which conflict can be reduced, harmony increased, and democratic capitalism stabilized. The raison d’être of the field is clearly articulated in the charge given by the Canadian government to the investigative Royal Commission on Industrial Relations (1919, also known as the Mathers Commission). The Commission’s Report (1919) states that its members are: “charged with the duty of considering and making suggestions for establishing permanent improvement in the relations between employers and employees, whereby, through close contact and joint action, they can improve existing industrial conditions and devise means for their continual review and betterment.” (p. 5, emphasis added).

Several years earlier, the American government formed a similar investigative body, the Commission on Industrial Relations, and its charge was also to examine the causes of an increasingly dysfunctional employment relationship. It was directed to, “inquire into the general conditions of labour in the principal industries… [and] any methods which have been tried… for maintaining mutually satisfactory relations between employers and employees” (Final Report, 1916: 6).

The IR field is now on the cusp of its 100th anniversary. To mark the event, this paper re-examines the central issue both Commissions were charged with addressing—the state of the employment relationship (ER) across the principal industries and geographic regions of the nation. Their mission was a three-step process: collect representative nationwide data and evidence on the state of the ER; sort, examine, and evaluate the data for what they show and imply about the state of the ER; and work up, report, and summarize the findings in a clear, impartial, and informative manner. For data collection, the two IR Commissions travelled cross-country, visited factories, mills and mines, held public hearings, and collected testimony from a wide cross-section of people. For data evaluation, the commissioners and staff winnowed and sifted all the information they had seen and heard, highlighted key events and pieces of testimony as supporting evidence, and cited the scattered statistical data available from government and private sources. For the presentation of findings, the commissioners wrote lengthy reports summarizing their interpretations of the evidence, albeit in the US case split into partisan majority and dissenting reports.

In this paper, we follow the same three-step process to evaluate the state of the employment relationship a century on but with much improved data and methods. Firstly, our data come from a new professionally designed and nationally representative survey, State of the Workplace Employment Relations Survey (SWERS), containing answers from more than 2000 managers and employees in the US on a wide variety of relevant ER issues. Second, the data are organized, interpreted, and ranked into low-to-high numerical scores on a performance and health measure of the ER using a pluralistic version of the balanced scorecard instrument from strategic management. Third, to keep a balanced scorecard from degenerating into a laundry list of measures, its structure and diagnostics need to be based on an overarching theoretical framework. The framework used for this paper, remarkably fitting given the centenary theme, is the diagrammatic model of an industrial relations system presented by W. L. Mackenzie King in Industry and Humanity (1918). King’s book has the honour of being the first theoretical treatise written specifically on the then-new subject of industrial relations. Fourth, the ER performance and health measure is calculated for each of the 2000+ workplaces and graphed as a frequency distribution on a 1-7 scale (1 = lowest). Our interpretation is that this frequency distribution has the distinction of representing, in one simple diagram, the key statistical moments describing the nationwide state of the ER which both IR Commissions a century ago laboured to summarize in a hundred and more pages of prose. Also, this type of empirical ER frequency has never before been presented in the IR literature, with the exception of a complementary but different analysis for Australia (Wilkinson et al., 2018), and yet can well be argued to diagrammatically depict the field’s paradigm-level dependent variable—the state of the employment relationship. Fifth, the frequency distribution of scores are cumulated and averaged to provide a single numeric score for the overall state of the ER as separately evaluated by managers and employees.

Evaluation Instrument

When a person goes to a doctor for a physical health exam, the doctor follows a checklist of procedures, measurements, and tests that cover all the bodily processes and parts. One block of tests and diagnostics, for example, covers the respiratory system, another covers the circulatory system, and so on. Then, within each block, the exam checklist has a number of specific tests, such as blood pressure and hemoglobin count for the circulatory system. At the end, the doctor meets with the patient, points out the strong and weak areas from the tests, and, through some weighted averaging process, gives the patient an overall health assessment, such as on a 1-to-10 scale or poor, average, good, excellent.

This paper seeks to do much the same type of health measurement and diagnosis evaluation for the employment relationship. This kind of evaluative ranking of employment relations across companies has not previously been done in the academic literature. A number of consulting firms, business magazines, and advocacy groups, however, regularly publish lists of great companies to work for, such as “Canada’s Top 100 Employers” and “Fortune 100 Best Companies to Work For” and, occasionally, lists of the worst companies to work for. A contribution of this paper is to put this type of evaluative exercise on a stronger analytical and social-science foundation. The start point is to find or develop a suitable ER evaluation instrument. The Center for Disease Control (CDC), for example, provides on its website (www.cdc.gov) a “Worksite Health Scorecard” but no IR organization or association provides a similar worksite employment relations scorecard. We, therefore, conducted a broader academic and practitioner search, including organizations doing the “best workplace” competitions (e.g. Bush, 2018). From the search, we found a suitable evaluation instrument, and along with it an interesting historical sidebar.

The evaluation instrument chosen is a modified version of the Balanced Scorecard (BSC) from accounting and strategic management. The BSC originated in an introductory 1992 article in the Harvard Business Review (Kaplan and Norton, 1992) and several follow-up books (Kaplan and Norton, 1996, 2000). Their idea caught fire and twenty-five years later is the subject of more than one-thousand journal articles and numerous books, and has been adopted worldwide by many hundreds of companies (Cooper, Ezzamel, and Qu, 2017). Surprisingly, however, an electronic database search found only one article in a human resource management (HRM) field journal that focuses specifically on the BSC (Walker and MacDonald, 2001) and none in the industrial relations field.

The BSC is a system of operational and performance indicators that mirror a company’s business strategy model and internal operating system. Its purpose is to give executives real-time information on the company’s progress in attaining its strategic financial goals and alert them to under-performing business units and operational processes. A favorite analogy is that a BSC helps managers monitor progress toward strategic goal attainment and alert them to needed business corrections and adjustments in the same way that an airplane’s instrument dashboard informs pilots on the status of all flight systems and needed adjustments to the flight path in order to best reach the destination (Kaplan and Norton, 1996).

Kaplan and Norton call their performance-monitoring instrument a balanced scorecard because it includes a mix of financial/operational measures and tangible/intangible inputs and outputs. They emphasize that the success of a BSC critically depends on, first, building in a set of multi-level measures that accurately signal the operational and financial ‘pulse’ of the business and, second, deriving these measures from an articulated model of the business that accurately depicts its key structural properties and cause-effect relations with respect to inputs, processes, and outputs. As a first step to BSC implementation, executives meet as a group and work out a multi-tiered flow-chart diagram, called a strategy map, which visually represents the company’s strategic goals, the structure and operation of the business, and the performance levels each unit and process has to achieve to attain the strategic targets.

One reason we chose a balanced scorecard instrument to measure the performance and health of the ER is because, though surprisingly little used in the HRM field, there are two very large and influential exceptions in this literature which provide helpful guidance on both dos and don’ts for an IR scorecard. The two exceptions are the complementary books The HR Scorecard by Becker, Huselid, and Ulrich (2001) and The Workforce Scorecard by Huselid, Becker, and Beatty (2005).

The aforementioned authors open both books by highlighting the employment relationship paradox that many company mission statements assert employees are their most valuable asset yet, in practice, they often give the workforce and HR function relatively low strategic emphasis and investment. As a result, Huselid, Becker, and Beatty (2005: 1) tell readers that the workforce is “the most important and most underperforming asset in most businesses” and, therefore, “anything less than optimal workforce success is not just a missed opportunity; in many cases it is a direct threat to the survival of the firm.” The principle cause, they say (p. 2), is measurement error arising from the fact many of the contributions of employees and the HR function are intangible or non-quantifiable and, thus, get omitted or under-valued in managerial decision making.

A BSC helps remove the measurement error, but only if it is built on an accurate representation of the company’s business model. Kaplan and Norton (1996: 17, emphasis added) observe: “A properly constructed Balanced Scorecard articulates the theory of the business. The scorecard should be based on a series of cause-and-effect relationships derived from the strategy.” Similarly, Huselid, Becker, and Beatty (2005) emphasize: “without developing models that show ‘what causes what’ throughout the business—we’ll end up with a series of unrelated metrics” (p. 68). Based on this reasoning, the front part of this paper is used to outline a theoretical framework for an IR scorecard. Among its implications are that the mediocre ER performance revealed in this paper’s scorecard has deeper structural causes than measurement error.

Surprisingly, Huselid, Becker, and Beatty (2005: Fig. 1-2) provide only a heuristic, broad-brush diagrammatic representation of the cause-effect model that anchors their workforce scorecard. The central performance variable is workforce success, defined as: “how well the workforce has contributed to the execution of the firm’s strategy.”(p. 6). This specification, adopted by Kaplan and Norton (1996), and common to the HRM field, evaluates the performance of the employment relationship only with respect to firms’ interests and wellbeing—a unitarist shareholder model—rather than a pluralist stakeholder model, as in the IR field, where weight is also give to employees’ interests and wellbeing.

In a shareholder model, employees are human resource inputs managed and valued as instrumental means for maximizing a profit end. Illustratively, Huselid, Becker, and Beatty (2005) declare: “Strategically relevant workforce success needs to be diagnosed and exploited like any other strategic opportunity or any other asset.” (p. 16, emphasis added). Workforce success, the dependent (Y) variable in their model, is driven by three main categories of intermediate and independent (X) variables: 1-Leadership and Workforce Behaviours; 2-Workforce Competencies; and 3-Workforce Mind-set and Culture. An explicit theoretical rationale for these variables is not given, but the prose discussion points to a version of the high-performance work system (HPWS) model popular in the HRM field (Paauwe, Guest, and Wright, 2013).

No study in neoclassical-based labour economics (NLE), located on the opposite side of the IR field from HRM, has developed a scorecard-type instrument for evaluation of ER performance. Since NLE has well-developed microeconomic models of firms, production, and personnel practices (Lazear and Gibbs, 2009; Borjas, 2016), they could also be used as a BSC organizing framework. The scorecard, however, would likely be considerably different from that used in HRM or IR, again emphasizing the contingent relationship between the paradigm vision of the employment relationship, design and construction of a BSC, and the resulting evaluation scores.

The historical sidebar discovery made in the search for an evaluation instrument is that, as with many ballyhooed new ideas in management, the balanced scorecard concept has antecedents going back many decades in the employment/labour area—in fact, to the very birth of the IR field. One of the expert witnesses called to testify before the Commission on Industrial Relations was labour consultant Robert Valentine, one of the very first consultants in the field and a progressive who worked for both companies and unions. He described to the Commission his new labour audit instrument developed to measure and evaluate the tenor and performance of the employment relationship at client companies. He subsequently published a copy of the audit in a business periodical (Valentine, 1915). Soon his labour audit evolved into a personnel audit (Tead and Metcalf, 1920) and six decades later into an HR audit (Fombrun, Tichy, and Devanna, 1984: Ch. 15). Similarly, the labour audit also developed along the IR branch and became an industrial relations audit performed in the 1920s at major companies by the non-profit consulting/research organization, Industrial Relations Counselors (IRC), established by IR co-founder Rockefeller Jr. (Kaufman, Beaumont, and Helfgott, 2003; Kaufman, 2010a: Ch. 6). After World War II, Dale Yoder and colleagues at Minnesota’s IR Center expanded it into a triple-IR audit (Yoder, Heneman, and Cheit, 1951) and, later still, an employment relations audit (Yoder and Staudohar, 1982).

ERS Model and Scorecard Framework

The balanced scorecard is a management-control instrument and not appropriate for a pluralistic IR evaluation of the employment relationship without significant modification. One plausible way to proceed is to extract from the literature a set of generic IR principles (e.g. Kaufman and Barry, 2014) and incorporate them into a modified BSC. This strategy has two downsides, however. The first is that a list of IR principles helps identify important subject blocks and individual questions for an IR scorecard but does not provide the broader theoretical framework needed to serve as the design skeleton for the scorecard. The second drawback is that one person’s set of IR principles may not match those of other people, with the resulting danger of the ad hoc.

A search was done, therefore, to find a theory, model, or framework of the employment relationship that is at once broad enough to serve as the organizational skeleton for a scorecard, contains what most people in the field consider essential IR principles and considerations, and is sufficiently detailed to provide concrete guidance on specific topics and questions. No well-suited diagram model of the ER was located in the modern literature, although Kaufman (2004) perhaps comes closest. Some studies on IR theory do not provide a diagrammatic representation (e.g. Dunlop, 1958; Adams, 1993; Budd and Bhave, 2008), while others cover only a portion of the employment relationship (e.g. Budd, 2004), portray the ER at a too general, abstract, or taxonomic level (e.g. Barbash, 1984; Meltz, 1993; Marsden, 1999), focus on the union-management part of the field (e.g. Kochan, Katz, and McKersie, 1986), or take a high-level political economy perspective (Wilkinson, Wood, and Deeg, 2014).

Given the 100th anniversary theme of this paper, as luck would have it, our search continued through the literature until we found at the very beginning of the IR field Mackenzie King’s diagrammatic model of an industrial relations system. The model, developed in nine charts on fold-out pages attached at the back of Industry and Humanity (1918/1973), is a hidden gem never to the best of our knowledge reproduced or featured in another IR study. King’s book quickly sank into obscurity after publication, partly due to his short, non-academic time in IR—in 1921 King became Canada’s prime minister—and, also, an overly-prolix and moralistic writing style. Nonetheless, Industry and Humanity in three-hundred pages provides a remarkably broad and synthetic portrait of the animating objectives, ideas, principles, and theoretical building blocks of the new IR field which was only just then beginning to form. Further, his Chart 8 representation of an employment relations system (ERS), even with evident influence from King’s astrological interests (Levine, 2011), is an extraordinarily comprehensive and detailed model which far surpasses those of Dunlop (1958) and others.

We adopt a modified and reconfigured version of King’s ERS framework diagram to serve as the organizational skeleton for the IR balanced scorecard instrument, later to be filled in with SWERS data. The model is depicted in Figure 1. The top part is a moderately simplified and rearranged version of King’s Chart 8, while the middle and bottom parts are our supplementary add-ons. A relatively detailed explanation is given of this model not only for scorecard purposes, but also because it represents, particularly in augmented form, perhaps the most articulated paradigm vision and structured cause-effect model of industrial relations anywhere available.

Figure 1

Early IR Model of the Employment Relationship

Early IR Model of the Employment Relationship

-> See the list of figures

King’s ERS Model: Overview

King’s model is anchored on the institution of the employment relationship, as created and structured by the legal contract of employment (Chart 1, p. 340). It embeds the individual employer-employee relationship in a multi-level, hierarchical representation of a national-level ER system. The system includes a political economy/governance dimension, contains both external market and internal organizational sides of the employment relationship, and includes employer associations and independent unions and a wide range of cultural, historical, social, and technological forces and contingencies. King uses the model to elucidate what he claims (p. 115) are universal laws and principles of industrial relations.

King’s representation of an ER system can be abstractly conceptualized as a field-level Y = f (X) system of cause-effect relations, starting at the individual workplace with micro-level ER performance (variously defined and measured) as the Y variable and scorecard outcome, and then successively expanding the model to higher institutional/governance levels (e.g. company, industry, nation) with associated independent, contextual, mediating, and moderating X variables.

King’s ERS has three major functional components and coordinating modes. They are, respectively, government coordination through the exercise of sovereign state power embedded in executive action, legislation, courts, and police function; price coordination though forces of competition and demand/supply in markets; and management coordination in firms through authoritative direction, administration, and control. King earned his PhD from Harvard in political economy (1909) and wrote Industry and Humanity in the tradition of institutional economics (Kaufman, 2017). King, for example, knew and corresponded with Veblen, Commons, and the Webbs. When Commons wrote his theoretical treatise Institutional Economics (1934), the government, market, and management coordination modes became rationing, bargaining, and managerial transactions (i.e. the three fundamental categories of X variables), which collectively determine the pattern of ER system outcomes (Y variable), represented by the statistical frequency distribution at the bottom of Figure 1.

ERS Model: Top Part

The smaller triangle embedded in the larger triangle in the top part of the diagram is where King begins the ER system, starting with Chart 1 entitled “Nature of Industrial Relations.” It depicts the employment relationship and the three parties directly involved in the contract: the owner/entrepreneur (O), managers (M), and wage/salary labourers (L). Managers are for legal purposes considered agents of the owner(s), so the category of employer subsumes both owner(s) and managers. King distinguishes a fourth non-contractual party to the employment relationship, the community (society and public interest), which diagrammatically encompasses the ER as represented by the inner of the two large encompassing circles.

The three contractual parties to the ER are represented by the three partially overlapping but hierarchically-arranged circles marked O, M and L. The arrangement connotes the idea that company and worker negotiate in the labour market as legal equals (a horizontal power relation) but, inside the firm where the ER is enacted, the law makes management the superior and employee the subordinate (vertical power relationship). The vertical authority relation and management’s rule-making power make the ERS, and the individual firms, unions, and other institutions within it, political governance institutions, per King’s Ch. 11 “Government in Industry” and Commons’ book Industrial Government (1921). The general principle of ER governance that King enunciates is: “That form of government in Industry is best which doth actuate and dispose every part and member to the common good.” (p. 270, emphasis in original, paraphrasing John Pym’s address to Parliament in 1614).

Given this governance perspective, King takes a micro-to-macro approach to employment relations theory, starting in Chapter 3 on “The Human Aspect”—shown in the diagram as the box Human Aspect at the top of the larger triangle. He distinguishes three facets: 1-principles of human biology; 2-psychology (individual and social); and 3-ethics/morals. King observes that the problem of industrial relations in neoclassical labour economics is largely an issue of mechanical efficiency, given the many reductionist assumptions about human agents and markets.

King several times cites Adam Smith at the beginning of the book (e.g. p. 17, 18) and develops his principles of human behaviour and morals much in line with Smith’s Theory of Moral Sentiments (1759). Human beings are a mix of rational calculation, social comparison, and irrational passion with proclivities toward conflict-producing aggression, domination, and exploitation. These destructive forces can be controlled and turned to cooperation, peace, and wealth-production by appropriately structured institutions, norms, and rules that constrain and incentivize people to shift from negative-sum domination and predation to positive-sum cooperation and fair treatment. Unrestrained self-interest and free market competition, contrary to the utopian invisible-hand story in mainstream economics, work to undermine cooperation-based institutions and moral codes, leading to what King (p. 53) identifies as a Gresham’s Law phenomenon of bad ERs drive out good. King, like Commons, the Webbs, and progressive employers (e.g. Rockefeller, Jr., who employed King in the 1910s as his personal IR advisor), therefore favoured creating a floor of minimum ER standards through external institutions and regulation, such as labour law, unions, employers’ associations, and social welfare programs (located in the outer circle) (Kaufman, 2003; Budd, 2004; Domhoff and Weber, 2011).

In King’s view, the two superordinate end goals of an ER system are creation of wealth and wellbeing, with the latter taking precedence (p. 67). In the modern literature, by contrast, Budd (2004) proposes an alternative of efficiency, equity, and voice. These IR goals—per the epigraph—are promoted by an ERS designed to call forth the constructive side of people, serve all stakeholders’ interests, provide channels for employee participation, representation and voice, and within this framework seek through good management and market incentives to minimize areas of conflict and opposition of interests and maximize areas of cooperation and unity of interests.

The workplace is the site of production and labour process, located at the bottom of the small triangle. It includes the direct contributions of all three parties to the employment relationship. Neoclassical labour economics telescopes all factor inputs into capital K and labour L and assumes each yields a deterministic marginal product/factor demand schedule. The effect is to reduce the employment relationship to an institutional datum and the process of coordinating and managing the dozens to millions of people working each day at airlines, banks, government agencies, retail stores, and software companies to the optimization of a technological input-output equation (Borjas, 2016). The assumption is that a certain stock of paid labour hours L translates into a predictable flow of labour services (effort/work) and rate of output (Q), typically obtained by either assuming L is a homogeneous commodity or work effort is fully specified in a complete labour contract.

King and the IR field, on the other hand, argue that labour’s marginal product is an indeterminate quantity because intervening between paid labour hours L and output Q is the employee’s volitionally supplied amount of work effort, called labour power by Marx (1867) and efficiency labour by Marshall (1890). Since employment contracts are incomplete and legally unenforceable, the amount of effective labour—broadly conceived as productive behaviours supplied by employees, such as engagement, cooperation, and customer service—can vary from a minimum of 0% (e.g. didn’t come to work, sleeping on the job) to a maximum of 100% (working at full-out potential). The amount of effective labour a company gets from its workforce is affected by a host of employment relationship considerations, starting at the most immediate level, such as interest/goal alignment, trust/confidence, fair/respectful treatment, and space for tacit effort bargaining, and extending outward to higher-level internal factors, such as shareholder vs stakeholder governance, command vs commitment styles of management, and breadth, depth and sophistication of HRM practices, and then to external factors, such as market conditions and labour law/social welfare provisions.

An implication is that even if the physical/tangible amounts of O, M, and L are held constant, or are identical across two firms, the production function(s) may yield very different Qs because differences in ER climates (Edwards, Bélanger, and Wright, 2002; Dastmalchian, 2008) lead to potentially large changes in effective O, M and L. With this idea in mind, King observes:

Fundamentally beyond all other considerations is the attitude of the parties to Industry toward each other. If the relationship be one of antagonism or hostility, of a regard for opposed as contrasted common interests, it matters little what the policies or methods governing production may be [e.g. advanced HRM practices], the foundations of economic and social development will be insecure.”

p. 99-100, emphasis in original

With respect to the traditional neoclassical production function, King further observes: “Assuming the existence of a right attitude toward each other, the rest is so simple as to be almost a matter of mathematical calculation” (p. 101, emphasis added).

The larger triangle in the centre of the diagram represents the hierarchically structured business organization, perhaps containing within it multiple plants and workplaces. King places seven circles above the tip of the triangle to indicate important internal organizational structure, governance, and human resource practice areas that the employer and management cadre have significant—but not complete—control over. Examples include amount/form of remuneration, training/development practices, and amount/form of employee voice and influence, possibly in the form of high-performance work practices (HPWPs). The larger triangle encompasses the boundary domain of the cause-effect workforce management model used by Becker, Huselid, and Ulrich (2001) and Huselid, Becker, and Beatty (2005) to build their scorecards and the seven small circles include many of the HRM/work design practices used as explanatory variables in the HPWS model.

Within the employment relationship are contradictions ignored by both NLE and HRM but captured within an institutionalist model such as King’s (and labour process, and radical and Marxist models, such as Edwards, 1986, Kelly, 1998, and Thompson and Newsome, 2004). These contradictions, far more than the measurement error cited by Huselid, Becker, and Beatty (2005), are responsible for ubiquitous shortfalls in workforce performance.

One contradiction, for example, is identified by Veblen (1904) as the conflict in the firm between goals of “making goods” and “making money” (also Commons, 1934; Edwards, 2003; Kaufman, 2017). On the one hand, the logic of making goods, for example, places priority on a stable economic environment, team cooperation, above-market wages as a motivational device, low turnover, and longer-term employment relationships. The logic of making money (profit), on the other hand, places priority on short-term gains from trade, individualistic and self-seeking competition, treating labour as a commodity to be bought for as little as possible and discharged when no longer needed, and high turnover and spot-market employment relationships. The two imperatives pull in opposite directions, one tugged by the visible hand of management and the other by the invisible hand of the market, and the ER has no built-in mechanism to reconcile and equilibrate them.

Another contradiction is an inherent tendency for a capitalist ER to degenerate into a low-trust/low-cooperation outcome, thus raising conflict and lowering productivity. As illustrated by the prisoner’s dilemma game, both management and workers gain from positive-sum cooperation but during day-to-day enactment of the ER both sides, and particularly management (with more power and discretion), have an incentive to opportunistically cheat on the deal (e.g. a mutual-gain promise of job security is later reneged on by unilateral layoffs). Likewise, even if management in good faith tries to honour the mutual-gain compact with workers it may still be forced by external events, such as market boom/bust or fending off a corporate raider, to take actions that break the psychological contract and embitter employment relations (Fox, 1974; Miller, 1991; Thompson, 2003; Bélanger and Edwards, 2007; Kaufman and Gall, 2015).

Outside the large triangle is the external environment of the firm. This part of the employment relationship is entirely omitted in the strategy map and scorecards built by Becker, Huselid, and Ulrich (2001) and Huselid, Becker, and Beatty (2005) (e.g. no labour or product markets in the workforce scorecard strategy map in 2005, Fig. 1-2). The external environment extends upward and outward from the firm in multiple layers of institutions, forces and levels of analysis, with government sovereign power, laws and rules as the highest-level ERS determinant. At a more immediate level, the most important ER external influences are three dimensions of the economic-socio environment, represented by the three boxes and arrows arrayed on each side of the large triangle. They are, respectively, the firm’s product market where output price and sales quantity are determined (P, Q), the firm’s labour market where market-level wage and employment are set (W, L), and the local-to-national set of living and working standards (e.g. median family income, average weekly hours) that serve as benchmarks for relative comparisons and minimum labour standards.

Important parts of the external market environment are also macro-economic factors, such as the business cycle and unemployment rate in the lower part of the outer circle. In fact, both King (1918) and Commons (1921) identify business cycle volatility and high unemployment as the single most important cause of deteriorated employment relations and, also, the explanation for the relatively small adoption and sustainability of unitarist/HPWP work systems. Business cycles, for example, lead to employee layoffs, wage cuts, and work intensification that sour employment relations and, likewise, firms become reluctant to invest in internal labour markets (ILMs) and HPWPs because they raise and rigidify short-run labour costs.

Within the two concentric circles in Figure 1 are a wide range of other external, contextual, and mediating forces, factors, and institutions that affect the firm and employment relationship. Many of these factors, such as unions and collective bargaining, have direct and indirect effects on ER outcomes. Unions have a direct effect as they deteriorate or improve the workplace relations between the parties through aspects such as more opposed vs common interests, more distrust/hostility vs trust/harmony, and positive vs negative effects on productivity. They also exert indirect effects at numerous points, such as changing one or more of the seven workplace circles (e.g. workplace governance, pay rules, dispute resolution), the operation and outcomes of the firm’s labour market, and their mediating influence on and synergistic interaction with other external environment variables, such as immigration, political governance, labour laws, and unemployment rate. At the bottom of the circles, within the dashed line channel leading to the middle-lower parts of the diagram, are the four levels of an ERS system, identified by King as Local, State/Industry, National, and International.

ERS Model: Middle and Bottom Parts

We have added to King’s ERS diagram three additional components in the middle and bottom parts of Figure 1 that strengthen its analytic content and empirical scorecard connections.

The bottom-most diagram depicts a frequency distribution of workplaces/companies ordered along the horizontal axis by a summary measure of the state of the ER, referred to here with school report card and physical exam analogies in mind as its performance and health (P&H). The P&H measure is constructed as a weighted average or index number of various individual data items from SWERS that measure the performance and health dimensions—or the pulse—of the employment relationship relevant to the survival, prosperity, and wellbeing of the stakeholders. The P&H distribution is bounded on the left-hand side of the continuum by the worst existing state of ER and, on the right, by the best state. In turn, variation in ERS independent-contextual variables in the top part of Figure 1, either cross-section or over time, cause the ER distribution to shift to the left or right (worse or better employment relations).

A frequency distribution has recently been used to represent ERS lower-tier dependent variables, such as breadth/depth of workplace employee voice (Kaufman, 2014). The idea also extends to the ER system level, per the observation in the Final Report of the Commission on Industrial Relations that: “Considering the whole field of American industry, there are almost infinite variations of relationship between employers and employees…[with] many bright spots [and] diseased spots in industry” (noting the health analogy, p. 25, emphasis added). This quotation suggests an ER system has a dispersed distribution of employment relations, such as depicted at the bottom of Figure 1.

The statistical moments of the frequency distribution, such as mean, median, variance, standard deviation, and skewness, provide important descriptive features of the state of the employment relationship. The frequency distribution also represents the dependent Y variable, as alternatively defined and measured, in an employment relation-based model, such as King’s model of an ER system. Generalized to a paradigm level, the objective of the science part of industrial and employment relations, in this interpretation, is to theoretically and empirically explain the location and shape of the ER outcomes summarized in a frequency distribution and the position of individual workplaces, industries or countries in the distribution. The practice and policy part of the field, as intended by King (1918), is aimed at using the explanatory theory model and associated scorecard instrument to help identify, develop, and implement new institutions, practices, and public policies that improve ER performance and health at micro, meso, and macro levels. In terms of the diagram, the practice and policy goal is to move individual employer-employee relationships along a given frequency distribution from lower-performing left to higher-performing right and, for ER aggregates such as country or industry, to improve relationships across the board so the entire frequency distribution shifts rightward (the X variables = shift factors).

The photos under the top part of Figure 1 are included as a visible representation of King’s fundamental IR principle that the key to creating and sustaining high-performing employment relationships is an institutional infrastructure and governance system that promotes right relations and common interests in the workplace (i.e. more overlap among O, M and L input circles). The three photographs are from the Colorado Fuel and Iron Company (CFI) where Rockefeller Jr. sent King as his newly-hired industrial relations advisor.

The middle photo shows the steel mill complex that depicts the inanimate capital input variable (K) in CFI’s production function. In labour economics, and to a significant degree HRM (e.g., Huselid, Becker, and Beatty, 2005), the state of employment relations is not included as a productivity determinant and, hence, at CFI or some other company the capital stock K, with a given workforce input of L, produces the same output Q regardless of whether ER relations are great or terrible. A universal principle of industrial relations, elucidated by King, is that the state of employment relations is, in fact, a critical variable affecting the amount of Q that comes from a given K-L combination. The three photos illustrate this idea, beginning with the left-most one. It shows CFI workers marching out of town on strike (Rees, 2010). They were marching out of town because CFI management evicted the strikers and families from the company’s housing (company housing is an HRM practice) and they were forced to set up tent villages on the desolate high prairie and live through the freezing winter conditions of 1913-1914.

On April 20, 1914, one of the tent communities was attacked and burned by state militia with two dozen people killed, an event quickly dubbed the Ludlow Massacre (Gitelman, 1988). Testimony to the Commission on Industrial Relations revealed a hugely deteriorated employment relationship at CFI with arbitrary/autocratic management, harsh and inhumane treatment of workers, and constant hire/fire insecurity. Rockefeller Jr., the absentee owner in New York City who had never been to CFI, sent King to investigate and make recommendations for transforming the operations from a disastrously low-performing ERS to a much improved higher-performing ERS. King, in effect, was asked to develop a new workforce management strategy for CFI that would move it from the left-hand tail of the ER frequency distribution to as far on the right-hand side as conditions permitted (e.g. excess capacity and cutthroat competition in the coal market).

King did an in-depth investigation of conditions at CFI and worked-out a scorecard-like evaluation. Had King relied on a 1910s version of the HPWS/shareholder model dominant in today’s HRM field, his mental scorecard would resemble the HR/workforce scorecards developed by Becker, Huselid, and Ulrich (2001) and Huselid, Becker, and Beatty (2005). On the one hand, guided by the cause-effect logic in these scorecards, he would be led to a managerialist strategy in which the CFI executive team implements new governance, work, and HR systems designed to create levels of common interests, commitment, and cooperation that optimize the profits and ROI (return on investment) for the Rockefeller family. On the other, had King worked from a labour economics-type scorecard, he would be led to a different strategy—again with the goal of maximum profit/ROI, but using an external labour market (ELM) hire/fire system, a self-interest form of motivation based on carrot/stick incentives and threats, and terms/conditions set at market levels and varied with demand/supply conditions.

King rejected both HRM and NLE strategies and adopted a distinctive IR strategy grounded on the real driver of ER/workforce success: a team spirit of energized trust and common purpose created by demonstrated management commitment to a stakeholder governance system emphasizing mutual gain, fair-dealing, respect and voice; replacement of short-term ELM instability, insecurity, and minimalist skill-training with longer-term ILM stability, security, and human capital investment; and substantive employee involvement and consent. King’s book Industry and Humanity (1918) describes the theoretical principles underlying this strategy and the right-hand photo in Figure 1 illustrates the (intended) stakeholder/partnership spirit with Rockefeller (front row, 4th from left) seated before a CFI building with a group of employee representatives elected from across the company to serve on the joint labour-management works council—the new ER practice that King considered the linchpin innovation (MacDowell, 2000; Gollan et al., 2015: ch. 8; Taras, 2003).

From an IR perspective, essential to a high-performance employment relationship is a meaningful stakeholder model and shift in the psychological contract from employee equals exploitable resource to employee equals organizational team member. In shareholder-driven HRM/NLE models, employees know they are expendable resources, like trucks or computers, and the company’s objective is to efficiently exploit (utilize) them for maximum profit, perhaps with good conditions and HPWPs if it pays. Employees then reciprocate with a similar calculated, self-interested ‘get the most/give the least’ strategy which, in prisoner’s dilemma fashion, creates a dynamic toward a less energized, more transactional employment relationship—illustrated by a leftward movement on the ER frequency distribution. To reach and stay at a higher performance level on the distribution, King made a stakeholder governance system the cornerstone principle in his ERS strategy model.

The last component of Figure 1 not yet discussed is the box marked Competing Social Interests, Values, and Welfare Objectives, placed between the CFI photos and bottom frequency distribution. The outcomes generated in the ERS above the box have no objectively-determined desirability or goodness that can be ordered along the horizontal axis of the frequency distribution. Criteria such as desirability, goodness, health, and performance, collectively measuring ER Success (analogous to Workforce Success in the HRM scorecard model), are intrinsically normative assessments which differ among people according to their individual and group affiliations, interests, values, and welfare objectives (Budd and Scoville, 2005; Heery, 2016). Labour radicals, for example, may regard strikes as good (harbinger of working class rebellion against capitalist exploitation) and employee representation plans as bad (chloroforming workers with false consciousness of partnership), while business conservatives look at the same two events and reach opposite good/bad conclusions. Thus, within this box each person normatively evaluates and ranks the ERS outcomes produced in the top part of Figure 1, the individual rankings are aggregated through some social choice mechanism, and the aggregate ranking yields a national-level ER frequency distribution, such as in the bottom diagram.

The shape and location of the ER frequency distribution is an empirical question but, for illustrative purposes, is shown in Figure 1 as a normal distribution with a left-hand tail ending at the lowest, worst-quality P&H score, the peak of the curve at the median quality level of P&H, and the right-hand tail ending at the highest, best-quality P&H workplace. The shape and location of the distribution, and evaluative statements about ER success, however, depend on which individuals and groups (e.g. only managers, only employees, or both?) are polled. Statements about the overall level of employment relations also depend on the relative weights given to different groups and interests, such that a scorecard based on a shareholder-only assessment likely yields a different ER distribution and conclusion about ‘ER success’ than one based on a joint stakeholder assessment. Thus, every scorecard evaluation is a function of the normative, governance-based question of whose interests count, and how much?

State of Workplace Employment Relations Survey (SWERS)

The ERS model in Figure 1 is used as an organizational skeleton to construct an employment relations (ER) scorecard. However, the scorecard cannot be operationalized without a suitable survey data source. Large-scale workplace surveys have been done in both Canada and Australia (see Godard, 2001; Lipset and Meltz, 2004; Fair Work Australia, 2015) but are now mostly rather dated with relative emphasis on either the union/collective bargaining or labour market/productivity sides of the field. Only the Workplace Employment Relations Survey (WERS) in Britain provides for recent years readily available, detailed workplace-level data on all aspects of the employment relationship (see Brown et al., 2009), although it gives more emphasis to institutional forms and practices (e.g. compensation forms, grievance resolution practices) than SWERS, and less to behavioural and relational dimensions of work (e.g. strength of employee engagement, control vs commitment management style). Also, greater effort was made with SWERS to connect to the HR practitioner/consultant literature, such as concepts of a ‘great place to work’ and ‘employee value proposition.’ An upside of SWERS, therefore, is that it contains many new measures not found in previous surveys; a downside is that these questions often ask for subjective evaluations, by being new are not psychometrically validated, and the data are not always comparable with other surveys.

A survey company helped design the questionnaire, create the panels of respondents, conduct the survey, and handle data collection and tabulation. To capture the pluralist IR stakeholder perspective, surveys were given to separately-selected panels of employees and employer representatives (from non-matched companies for reasons of cost). The employer respondents, obtained by the survey company from a representative cross-section of American companies, were screened to include only mid-to-upper level executives-managers and omit those who managed 10 or fewer people or reported a small knowledge of company employee relations and HR practices. Employees, separately assembled by the survey company, were screened so all respondents were 18 years or older, working at least 20 hours per week, and not performing managerial or supervisory duties. Minimum workplace size for both panels is 20 employees.

The employer questions were framed at the company, business unit, or facility level with respondents instructed to pick the highest level for which they had reasonable knowledge; also, managers were instructed to answer more detailed questions on employee attitudes/behaviours and employment/HR practices for the largest, most representative employee group. Employee questions were typically framed at the workplace level but with respondents instructed to answer for a smaller part of the workplace if needed for accurate answers, such as a department or work unit. Some element of non-comparability between employer and employee responses is, thus, introduced, since the answers from managers typically cover a larger and perhaps different workforce segment, but it was deemed more important for reliable results that employees answer for only the workgroup size for which they have reasonable knowledge. Also promoting comparability is that roughly 75 percent of the measures in the scorecards are from similarly worded questions used in both employer and employee surveys.

The employer and employee surveys were answered online by the respondents. The respondents were selected using a quota sampling method. Categories of important respondent characteristics are identified, such as age, gender, industry, occupation, and union status, and people in the overall panel with a specific characteristic are invited to participate until the required number within and across categories are obtained to ensure national representativeness (thus, effectively, a 100% response rate), subsequently checked against government-reported data.

The surveys were pre-tested in late 2015 and final results collected by late January 2016. The surveys were completed online by approximately 2000 employees and 500 managers. The sample sizes were determined by the survey company to ensure representativeness.

The answers to nearly all the questions in the survey were solicited on a 1-7 scale with 1= lowest/worst and 7 = highest/best. To facilitate interpretation, the numerical scores were converted into letter grades on an F to A scale, with grades above F also distinguished into minus and plus categories. The conversion scale from numeric to letter grade is made as symmetric as possible: 1.0 - 2.49 = F; 2.5 - 3.49 = D; 3.50 - 4.49 = C; 4.5 - 5.49 = B; and 5.5 - 7.0 = A. In-between letter grades are defined using the same symmetric approach (e.g. C- = 3.5 - 3.83; C = 3.84 - 4.16; and C+ = 4.17 - 4.49). Negatively-framed questions were inverted so they follow the positive 1-7 scale.

From ERS Model to IR Scorecard

In building an IR scorecard, it is important to start out with the purpose clearly in mind. The purpose is to provide a structured set of questions and measures covering all the important parts and processes of an organization’s workforce, labour/production process, and employment relationship—like the questions and measures in a physical exam checklist on different body parts and functions—which signal each area’s strengths/weaknesses and performance/health, and can be aggregated to provide an overall company-level P&H score and ER evaluation. A theory of the employment relationship, like a theory of the human body, is thus important for organizing the scorecard in terms of key components, functions, processes and outcomes lest it degenerate into an ad hoc and potentially misleading laundry list of data items. The King model, therefore, provides a helpful framework for working out a theoretically-congruent structure for an IR scorecard, and also distinguishing it from an HRM scorecard.

Also important to keep in mind, a scorecard is an instrument to measure, calculate, and evaluate workforce/ER performance and health, which corresponds to the IR dependent variable in the frequency distribution at the bottom of Figure 1. Thus, a separate follow-on research project is to identify the causal explanatory factors accounting for the pattern of P&H scores in Figure 1 within and across companies (i.e. the shape/location of the ER frequency distribution). This type of analysis requires, however, more sophisticated multivariate regression techniques and an expanded set of explanatory external, contextual, and moderating variables. The organizational structure and measures for the scorecard presented below, therefore, are largely from inside the firm (the larger triangle) in the King model.

The theoretical model that informs the structure of the HRM workforce scorecard developed by Huselid, Becker, and Beatty (2005) is from the high-performance work practices (HPWP) literature. In general form, it posits HPWPs → Unique Employee Skills/Motivation → Employee Work Effort/Citizenship Behaviour → Firm Performance → Shareholder Wealth (Paauwe, Guest, and Wright, 2013). A drawback of this work is that these authors neither explicitly articulate this model nor use it as a conceptual skeleton for working out a generic example of their proposed scorecard. Nonetheless, the HPWP model clearly suggests a basic scorecard structure that starts with measures of breadth/depth of HR function and practices, with higher numerical scores a “plus” (as higher scores on a hearing test are a plus on a physical exam), followed with measures of employee skills and motivation (e.g. training hours, work engagement), measures of work effort and citizenship behaviour, and so on until the end set of measures on firm performance (e.g. return on assets, increase in stock price).

King’s ERS model offers points of critique of this type of HRM scorecard and also guidance for an alternative IR scorecard. The driving force of firm performance, according to King, is not breadth/depth of HPWPs, although in the right contexts and amounts they provide helpful support, but a positive, energized team spirit that induces from top to bottom effective cooperation, going the extra mile, and selfless dedication to the success of the enterprise. Energized team spirit, in turn, cannot be created and maintained unless employees feel a material and emotional stake in the enterprise, impossible to achieve without genuine stakeholder governance, sharing of profit, risk, power, voice and status, and trust in management. High performance, therefore, starts with owner/executive commitment to a mutual-gain, shareholder model that creates a positive employment relationship that ripples through all aspects of the business. Furthermore, when combined with high-quality management and employee resources, high-quality employment practices, and high-quality operational practices, it makes the firm—with cooperative external market/social conditions—a top-rated employer and top-rated financial performer. Schematically, the King model can be represented as: Committed/Trusted Management + Accommodative External Environment → Stakeholder Governance + Effective Organizational Management → Positive/Energized Team → Cooperation/Engagement/Commitment → High Productivity/Customer Satisfaction → High Firm Performance → Shareholder Mutual Gains.

Incorporating these features creates an IR workforce scorecard with nine sections, shown in Table 1. The stakeholder dimension is emphasized with company/workplace performance outcomes in Section I and employee performance outcomes in Section II. Sections III and IV, respectively, capture the quantity/quality dimension of the inputs/capabilities provided by management and employees. Section V captures the team spirit/positive relations dimension of the workplace. Section VI includes organization-level ingredients of positive relations/climate, such as management style and joint governance. Section VII includes more specific HR and work design features that influence mutuality and positive relations. Section VIII captures the community/public interest stakeholder dimension. And, finally, all the scores on individual measures across the eight sections are cumulated, averaged, and reported in Section IX.

Table 1

State of the Employment Relationship Report Card, Employee and Employer Surveys (United States)

State of the Employment Relationship Report Card, Employee and Employer Surveys (United States)

-> See the list of tables

Sections I-VIII are filled in with as many SWERS data items as available and relevant. The result is a total of fifty-three (53) measures. Each measure listed in Table 1 is worded to parallel the wording of that question in SWERS. The individual measures in each section are summed and averaged to give the overall section average score (on a 1-7 scale) and grade (a F-A+ scale). Median section scores were also calculated, but differed little from the mean scores and are, thus, not reported.

Section IX reports the overall average P&H score/grade for this nationally representative cross-section of American workplaces. It is obtained by summing the workplace score/grade for each of the 2000+ companies in the SWERS data set and calculating the average. The averaging process gives each scorecard item and company an equal weight, at least at this stage the most practical and non-arbitrary assumption. The overall workforce scores/grades in section IX represent a unique summary measure of the performance and health of the employment relationship in the USA, much as if a nationally representative set of adults in the USA were given a common physical exam with fifty-plus measures, and the individual scores summed and averaged to provide a national-level estimate of U.S. physical health and bodily performance.

Table 1 is divided into two halves, with the first six columns showing scores/grades calculated from the responses given by upper-level managers in the employer panel of SWERS and the last six columns calculated from responses by non-managerial workers in the employee panel. For both managers and employees, the first two columns report mean scores/grades for the entire samples and the last four columns report, respectively, scores/grades for the companies whose mean scores are in the bottom 20% and top 20% of the distributions.

Results and Findings

The fifty-three data items and overall scores shown in Table 1 provide the most comprehensive, detailed, physical exam-like portrait of the health and performance of the employment relationship and workforce management in North America since the two government-sponsored IR Commissions a century ago.

Scanning the first and fourth columns of data in sections I-VIII reveals that the mean scores from managers cluster in the B+ to A- range and in the C to B- range from employees. These indicators are then summed and averaged to get an overall scorecard score/grade for each company, which, in turn, are summed and averaged across all 2000+ companies in the respective SWERS panels to get the grand means, shown in section IX as Overall Score. Since the SWERS data are constructed to be nationally representative, Table 1 represents, in effect, a national-level ER evaluation instrument. The scores/grades are 5.25/B+ from managers (the bottom for B+ is 5.16) and 4.48/C+ from employees (close to B-, starting at 4.50). It is important to emphasize that these estimates are contingent on the structure and measures of the scorecard (hence the emphasis on a strong theoretical foundation) and the questions and sample of respondents in SWERS (hence the professional survey company).

These findings have a good news/bad news quality. The good news is that company managers, in the aggregate, rate most aspects of the workforce/employment relation in the B+ to A- range. On the bad-news side, the employees are noticeably less positive and most grades are in the C to B- range. On the one hand, given two to three decades of widespread organizational restructuring and downsizing, a hollowing out of the industrial sector and the loss of middle-class jobs, and the economic-social polarization mirrored in stagnant employee real earnings and the meteoric rise in executive pay, worse results than a high C+ grade from employees might have been expected. But, on the other hand, an aggregate employee grade of C+, along with the large number C to B- grades on individual measures, certainly raises questions whether the average American company, at least as seen from the employees’ perspective, is falling significantly short on competitiveness and quality of workforce management and employee performance.

It is also important to look at the distribution of scores/grades around the mean. As a first step, shown in Table 1 are the averages for workplaces located in the bottom 20% and top 20% of the employer and employee distributions. The bottom 20% of firms, as rated by managers, get a low C- grade while the top 20% get a high A+ grade; employees give respective grades of F and A. This analysis reveals that America has a top tier of workplaces that are ‘A range’ in workforce and employment relations performance, as rated by both employers and employees, but it also has a bottom tier of workplaces rated by managers and employees as bad-to-terrible (C-/F). The well-known IR ‘frames of reference’ model (Fox, 1974; Budd and Bhave, 2008; Heery, 2016) divides workplaces into critical/radical, pluralist, and unitarist groups based on the degree of conflict/harmony and opposed/common interests, and it is an appealing inference, but one also needing stronger statistical validation, to think of the bottom 20%, middle 60%, and top 20% of workplaces in Table 1 as ordering into these three theoretical categories.

More information on the range and dispersion of workforce/ER scores is provided in Figure 2. Each panel depicts two frequency distributions, one formed from the employer scores in Table 1 and the other from employee scores. The distributions are formed by rank ordering from lowest to highest the two sets of overall scores for all the companies in each SWERS panel. These distributions are empirical representations of the theoretical P&H frequency distribution at the bottom of the King model and summarize the state or pulse of the employment relationship in companies across the US. As such, the distributions can be thought of as a representation of the IR field’s paradigm-level dependent variable.

Figure 2

State of the employment relationship : Narrow, Medium and Broad Workplace Performance and Health Scores

State of the employment relationship : Narrow, Medium and Broad Workplace Performance and Health Scores

Note: The frequency distribution of employer (ER) and employee (EE) responses on the 7 point Likert survey scales are plotted for different sections of the State of the Employment Relationship Report Card. For items contained in each section see Table 1.

Source: State of the Workplace and Employment Relationship Survey (SWERS), 2016. Both the employer and employee survey data were used in Figure 1.

-> See the list of figures

However, as earlier pointed out, there is no ‘correct’ or ‘best way’ to measure ER performance because numerous subjective choices have to be made about whose interests count, how much they count, and specification and measurement of interests. Thus, the employer and employee frequency distributions are also calculated using three alternative specifications, called Narrow, Medium, and Broad. These labels connote alternative assumptions about the number of stakeholders and range of interests that go into calculating P&H; for example, narrow = employers and employees only care about their own final-end outcomes in the ER and broad = employers and employees both care about their joint interests and outcomes across all ER stages and dimensions. Specifically, for the two SWERS panels:

  • Narrow: Employer includes only Company Performance Outcomes (section I, column 1) and Employee includes only Employee Satisfaction/Performance Outcomes (section II, column 4).

  • Medium: Employer includes average of Company Outcomes (section I, col. 1) + Employee Outcomes (section II, col. 1); Employee includes average of Company Outcomes (section I, col. 4) + Employee Outcomes (section II, col. 4).

  • Broad: Employer includes all fifty-three manager-reported interests/measures (section IX, col. 1); Employee includes all fifty-three employee-reported items (section IX, col. 4).

Looking across the resulting three panels and six distributions, three common features are evident. The first is the higher mean scores given by employer respondents, the second is the employee distribution is more dispersed (larger variance), and the third is that both distributions have a moderately skewed left-hand tail. The mean scores (employer/employee) across the three panels are, respectively, 5.27/4.80, 5.40/4.70, and 5.24/4.48. These numbers reveal that the employers’ evaluation of the state of the ER is relatively uniform across the three P&H specifications, but the employees’ evaluation successively declines, particularly so when the employment relations climate and practices (sections VI-VII) are put into the broad specification.

These data reveal that employers and employees clearly have different perspectives on each other, their relationship, the work situation, and performance outcomes (see also, Liao et al., 2009). Scores are not entirely comparable because manager and employee respondents come from different organizations and work group units but, given this qualifier, differences in mean scores are large and pervasive. Of the thirty-eight individual indicators spanning employer and employee columns, mean letter grades are the same in zero cases. For twenty-seven indicators, the mean score given by employers is at least one letter grade higher than given by the employees and, for eleven, the gap is larger still. In only five cases are the employers’ mean grades lower than employees’ (company competitive position, low employee turnover, expensive to replace employee group, low conflict/infighting, employees interested in what management says/does). The SWERS data indicate, therefore, that employers, for more than 8 out of 10 indicators, have a more positive assessment than do employees.

A valuable service provided by a scorecard is to identify particularly weak or troublesome performance areas in the ER. Scanning sections I-VIII, the section with the relatively lowest manager scores is employer-employee relations and climate (sec. V) while the three lowest scoring areas for employees are management style (sec. VI), HRM functional practices (sec. VII), and employment practices for mutualism (sec. VIII). The individual indicators in the scorecard that employees rate as very low but managers rate as much higher include: quality of people management, (A-/C+), confidence/trust in management (A-/C+), family/partnership feeling (A-/C+), and workplace morale (A-/C+). They deserve attention because all are critical ingredients for a successful high-performance ER and workplace.

Summary and Conclusion

Doctors produce better health outcomes for patients if they first conduct a complete physical exam before prescribing medicines or performing surgery and, likewise, airplane pilots are more likely to get their passengers safely and on-time to the destination when guided by a full set of well-calibrated flight instruments. This paper applies the same reasoning to the employment relationship and management of the workforce. For this purpose, we constructed a balanced scorecard for the employment relationship based on the theoretical model of an employment relations system presented by W. L. Mackenzie King and filled it in with fifty-three measures of ER performance and health obtained from employer and employee panels of the new SWERS data set. Although the analysis is substantially descriptive and not in the standard mode of multivariate regression and hypothesis testing, the upside is the ability to address a paradigm-level research question, develop new concepts and tools, and provide empirical findings having transparency and use value for researchers, practitioners, and policy makers alike.

Our summary interpretation of the results is that a slice of American companies enjoys very positive, mutual-gain employment relationships. However, a largely equal-sized slice has very poor, low-performing relationships, and the majority of companies are centered in a middle-level C to B+ range of mediocre-to-relatively good ER performance. For future research, these findings invite deeper multivariate analyses to identify the workplace characteristics and practices that shape each firm’s low-to-high position in the ER performance frequency distribution; for practice and policy, these findings have ‘bad news’ in that a majority of companies, particularly as seen by employees, have under-performing employment relationships but also ‘good news’ in that managers are considerably more positive in their evaluations. In either case, the many companies with performance shortfalls provide an opportunity for mutual-gain improvement in competitiveness and profit for employers, and pay and quality of work life for employees.

The opportunity to realize these mutual gains is easier said than done, however, because of numerous inhibiting factors identified by King (1918) a century ago. Examples include inequality of power and authority in the workplace, macroeconomic cycles and competitive undercutting by business rivals, lack of employee voice and employment security, short-sighted greed and opportunism, and withholding of effort and cooperation due to ill-will and lack of trust. The enduring insight and value of the employment-industrial relations field is that it accepts these factors as inherent to employment relationships and, thus, to be studied through objective science and resolved through pluralist/stakeholder voice, participation, and negotiation at all levels of governance—rather than largely airbrushed out of the picture through selective assumptions and idealist abstractions as too often is the tendency in rival fields of human resource management and neoclassical labour economics.