A Realist Conception of Universal Basic Income
By Aleksander MasternakSupervisor: Dr Paul A. Raekstad Second Reader: Dr Enzo Rossi June 2017
Master Thesis Political Science – Political Theory University of Amsterdam
Abstract
Realist political theory claims it is concerned with real politics. According to Galston, realism exists to correct for excessive ideal theorising, make political theorising once again relevant to real world politics, and catch the attention of real policy makers. Recently, concerns have been raised about predicted mass technological unemployment and the idea of a Universal Basic Income has been raised as a viable solution. UBI proposals in Finland were increasingly reframed as practical solutions to political problems and since January 2017 Finland is running the only unconditional UBI experiment. By drawing on Williams’ and Sleat’s work on legitimacy a realist method is outlined to theorising UBI. Predictions of technological unemployment are treated seriously and UBI is assessed as a reasonable solution that maintains the legitimacy of the state. Empirical research on basic income experiments and proposals raise concerns over the implementation and portrayal of a UBI policy.
Acknowledgements
Firstly, I would like to whole-heartedly thank my supervisor Dr Paul Raekstad for his continued advice and guidance throughout the process of writing this thesis. His distinct expertise in realist political theory was particularly helpful.
Secondly, I would like to thank my dear friend Tobias Arbogast for his critical assessments and our hearty discussions.
Table of Contents Abstract ... 1 Acknowledgements ... 2 Chapter 0: Introduction ... 4 Chapter 1: Realism ... 7 1.0 Introduction ... 7 1.1 Liberalism ... 7 1.2 Realism ... 8 1.3 Williams’ Realism ... 9
1.4 Sleat’s Liberal Realism ... 13
1.5 Realist Method ... 15
Chapter 2: Context ... 16
2.0 Introduction ... 16
2.1 Realist Critique of Neoclassical Economic Theory ... 17
2.2 Mini-Genealogy of State Obligations Towards its Citizens ... 20
2.3 Recent Unemployment Trends ... 23
Figure 1: Total nonfarm employment, seasonally adjusted, Jan 1939-Dec. 2015 (Ghanbari & McCall, 2016) ... 24
2.4 Technological Disruption ... 24
2.5 Technological Unemployment ... 25
Figure 2: The distribution of BLS 2010 occupational employment over the probability of computerisation. Area under the graph is equal to total US employment (Frey & Osbourne, 2013: 37) ... 26
2.6 Skills and Education ... 27
2.7 Limits to Adopting Disruptive Technologies ... 28
2.8 Methodological Debate – Computerisation of Work ... 28
2.9 Post-Keynesian Political Economy ... 28
2.10 The First Political Question ... 29
2.11 Conclusion ... 32
Chapter 3: Basic Income ... 33
3.0 Introduction ... 33
3.1 Meaning and Work ... 37
3.2 Economic Feasibility: Job Guarantee Strategy ... 46
3.3 UBI Implementation: Mincome Experiment ... 48
3.4 Political Costs: Finnish Basic Income Experiments ... 51
3.5 Response to the First Political Question ... 53
3.6 A UBI Proposal ... 54
Conclusion ... 56
Bibliography ... 57
Chapter 0: Introduction
Universal Basic Income (henceforth UBI) is an income granted universally to all members of a society as their right as citizens without means testing or conditions (Koistinen & Perkiö, 2014: 26). The idea of a UBI is not new but only recently it became a serious policy consideration. Interest in the policy relate to predictions in recent works, such as Ford’s ‘Rise of the Robots’ (2015), of an impending transformation of the world of work where the increasing pace of computerisation takes over jobs on an unprecedented scale and in domains previously thought as irrefutably human. Now technologists, entrepreneurs, social activists, and even incumbent politicians have taken concern with these predictions. Advocates of UBI have taken these predictions seriously and now campaign UBI as a solution for these concerns. Additionally, UBI holds broad-spectrum political support and as such is thought as a solution that would be met with the least resistance (Ford, 2015: 260).
The debate on UBI has advanced furthest in Finland where in January 2017 a UBI experiment was launched. Koistinen and Perkiö (2014) conducted a survey of all UBI proposals posed in Finland and concluded that proposals became framed increasingly as real solutions to real political problems, and that innovative social policy takes time to become viable in society. Additionally, there seems to be a sense of urgency in literature predicting mass technological unemployment, where concerns over the stability of society are raised. A serious treatment of these concerns yields further appeal to policies that clear show their capacity to solve political issues. The starting assumption of this thesis is, the portrayal of a UBI policy as a concrete solution to real political problems attracts more support than appeals to ideologically-contingent values. Realist political theory asserts its concern with real politics and argues the sources of political normativity are not to be found in pre-political moral commitments but rather exist in a normativity inherent to politics. For this reason, I argue that a realist conception of a UBI policy would bring forth a portrayal of UBI as a concrete solution to real political problems.
My motive for using a realist framework is twofold, first as an attempt to advance the academic project by exploring a practical application of realist theory. As a countermovement to high liberal theory, recent realist theory concerned itself primarily with how political theorists should theorise (Galston (2010), Hall (2015), Hall & Sleat (2017), Raekstad (2016)). Other recent realist works have taken the discipline beyond the Methodentstreit, for instance Sleat’s ‘Liberal Realism’ (2013) (Rossi, 2016: 410). The specific application of realist political theory to theorising about UBI is my original contribution to academic literature.
Second, a realist conception of a UBI would portray the policy as political solution and thus yield the policy recognition and more concrete support by incumbent real politicians in the West. Galston argues that realism exists for correcting the excessive ideal theorising by drawing attention to "an experience based concept of feasibility" to catch the attention of policy makers and make political theorising once again relevant to real world politics (2010: 400-401). Additionally, I draw on Williams’ work on legitimacy partially because it was primarily concerned with theorising about state policy.
For this thesis, I use Galston’s broad definition of realism (i.e. the political as necessarily conflictual, emphasis on political possibility, rejection of utopianism, political moralism, and the wrong kind of ideal theorising) which I elaborate on in Chapter 1: Realism. Then, I draw upon Williams’ work on legitimacy and amend it with Sleat’s criticisms and additions, to devise a realist method of thinking about political problems, policy prescriptions, and the legitimacy of the state. Williams’ theory of legitimacy is an explorative exercise into how one may criticise a set of political practices or institutions while remaining committed to evaluating them with standards internal to political practices themselves. For our purposes, realism can be thought of as an exercise in political understanding, the assertion of the main political problems, and then an assessment of the reasonableness of the state’s response to the problems. Considering my aim and approach, my research question is:
‘How do we apply the realist method to political theory, to theorising Universal Basic Income?’ First, we develop a serious treatment of context, then we outline the main political issues. I argue the main political issues where UBI is a relevant solution relate to predictions of increasing technological unemployment. The state is expected to respond to these political issues otherwise its legitimacy is brought into question, and at worst the stability of society is threatened. I argue UBI is a reasonable solution to predicted technological unemployment, and adjacent political problems and if the state responds with UBI in the right way, it can preserve its legitimacy. At the end of the thesis I raise concerns over policy implementation and portrayal
In Chapter 1 I review relevant realist political theory lay out the way I will be (broadly) thinking about politics and the political. I use Galston’s broad definition of realism, specifically emphasising theorising that generates real solutions to real political problems. Then I use William’s methodology to theorise about state legitimacy (i.e. context, the BLD, and the FPQ). Thereafter, I incorporate Sleat’s criticisms of Williams, specifically a rejection of the need for a state to be liberal to be legitimate, and Sleat’s structural conception of legitimacy. I conclude the chapter with a summary of the realist method.
In Chapter 2, I devise a serious treatment of context. The most relevant concerns stem from empirical research claiming technological unemployment is increasing in scale and scope. I start the chapter with a realist critique of the dominant belief that market corrective mechanisms solve technological unemployment. Then, I argue there is a historical precedent for the expansion of the obligations of the state towards its citizens stemming from previous responses to mass unemployment and economic downturns. Following, I assert that unemployment trends show increasing structural unemployment. Using Frey and Osbourne’s (2013) research I argue that the scope of computerisation is increasing and that predicted technological unemployment is of an unprecedented scale. Having
considered unemployment trends and future predictions, I argue education reform is necessary but insufficient. Thereafter, I argue the Post Keynesian political economy is more realist than a neoclassical one and adopt it for my policy prescriptions. Ultimately, I consider worst case-scenarios where
technological unemployment leads to state instability, and raise the need to implement future-proof policy.
In Chapter 3, I argue that UBI, in combination with education reform, is a reasonable response to the increasing scope and scale of technological unemployment. I argue, given that nor the market nor the state can provide meaningful employment in the face of high technological unemployment, and thus formal employment should not be considered an adequate default source of income and meaning. Instead UBI fulfils individual’s physiological needs, and frees individual’s intrinsic motivations to pursue their psychological needs. I reject the belief in the intrinsic value of work and instead should be allows to pursue their own conception of meaning. I argue intrinsically motivated pursuits produce higher quality work, which can yield taxable economic growth. Given a properly structured taxation policy a UBI could pay for itself. Then I consider specific concerns about policy implementation (i.e. timing, controversy) and its presentation (i.e. in pragmatic, non-stigmatising terms). Concluding I argue that a ‘low income insurance’ portrayal of a UBI has the highest probability of being considered a reasonable response by the state to real political problems, given a serious treatment of context. Adjacent concerns of
Chapter 1: Realism
1.0 Introduction
In this chapter I start with broad high liberal conceptions of the political, the purpose of politics, consent, and state legitimacy. The reason I start with outlining liberal political theory is twofold: first realism, in many ways, stems from critiques of liberal political theory; and second, given the current dominant political ideology is liberalism, and given my interest to avoid radical deviations from this status quo to minimise the political cost associated with introducing new policy, it is worth knowing what liberal theorising is about.
Then, I outline the realist critique of liberal political theory and outline the broad tenets of the discipline such as realist conceptions of politics and the political, the emphasis on power and its
legitimation, and serious treatments of context. Afterwards I introduce Williams’ version of realism which involves the process of state legitimating itself, acknowledge the theory’s main criticisms, and then motivate concessions to Williams’ classic realism with contemporary elaborations. Thereafter, building upon Williams’ liberal conclusions, I outline Sleat’s conception of Liberal Realism and outline how liberal realists should think about political conflict, the legitimacy of the state, and policy prescriptions. This section is closed with a summary of the main realist theoretical concepts to be employed in this thesis.
It is important to note that the realist political project is comprised of a series of at times diverging positions. When I refer to realism in latter chapters of this thesis I am primarily referring to Galston’s broad conception of realism defined in the author’s survey article ‘Realism in Political Theory’ (2010). The realist method is derived from Williams’ and Sleat’s work on legitimacy and liberalism.
1.1 Liberalism
“Liberal politics is grounded in the fundamental concern to respect all persons’ moral equality by protecting their liberty to pursue their conception of the good life free from state intervention” (Sleat, 2013: 136-137). In high liberal political theory, the assertion of moral equality of all persons originates from a neo-Kantian account of the autonomous individual. This account is then used to justify specifically liberal institutions and practices (Hall & Sleat, 2017: 280). The establishment of the
foundations of politics in some external non-political moral system that provides justification for a given political way of life is known as political moralism (Hall & Sleat, 2017: 280).
In liberal society institutions will reflect emphasis upon what constitute requirements for you to consider yourself an equal citizen (Sleat, 2011: 480). Individual rights act to restrict individual freedom such that it harmonises with the freedom of everyone else (Sleat, 2013: 23). Liberalism “attempts to show how people who disagree, fundamentally and seemingly intractably, about moral and religious issues can nevertheless live together under a common political authority” (Sleat, 2011: 471). This authority must be one that is consistent with and respects the freedom of all individuals subject to it. Liberals conceive of
this as a demand upon the state - that it must be one that enjoys the endorsement or consent of each and every citizen” (Sleat, 2011: 471). Consensus is possible if citizens adhere to a moral truth. Liberalism presents a consensus vision of the political where political relationships and actions are based on set principles (Sleat, 2013: 21).
From a liberal perspective, those who have not consented to the liberal regime would perceive the exercise of the regimes' coercive power as restrictions on their freedom (Sleat, 2011: 494). However, liberals assert that universal consent is not plausible in the context of reasonable pluralism. Liberals attempt to resolve the feasibility constraint over achieving universal consent on a set of fundamental values by asserting that reasonable individuals would accept these values. This results in hypothetical consent the conclusion of which thinks of those that disagree with this set of liberal values as simply unreasonable. By preferring compromise and toleration, over serious and potentially violent politics, liberalism’s politics is to eradicate politics by; imagining a hypothetical consensus, assuming liberalism is the default response to politics, and thus assume away deep-seated real disagreements that politics was originally meant to attempt to resolve (Sleat, 2013: 138, 142-143).
1.2 Realism
Realism is a countermovement to the recent 'high-liberalism' trend in political theory and presents itself as an alternative to ideal theory. Realism rejects utopianism, moralism, hypothetical consent, universal principles, and the priority of justice (Hall & Sleat, 2017: 278). Realists take issue with the lack of interest in liberal political thought with concepts like power and question liberal conceptions of human motivation (Sleat, 2011: 471).
Realists think of the political as deeply conflictual which “excludes the possibility of achieving a consensual basis for the use of political power, which is a necessary feature of liberal legitimacy” (Sleat, 2011: 471). Galston thought that the continued existence of a society without agreement upon first principles 'is often the only practical possibility' (2010: 385). Moreover, principles cannot serve as standards for political life unless they could be feasibly implemented (Galston, 2010: 395).
Realists critique liberal political thought as all important questions about liberty and distribution are decided pre-politically, by theoretical reasoning whose results are entrenched in constitutional law, and are not subsequently politically alterable (Sleat, 2011: 473). Instead realists aim to produce political theory without a reliance on understandings of morality, which realists found little intrinsic reason to endorse, and rather through a greater appreciation of "feasibility constraints or sensitivity to conditions of political possibility" (Hall & Sleat, 2017: 276-277). Context for the realist is important and it means to take seriously the historical and cultural circumstances, and social conditions of any given society (Sleat, 2013: 115). Context helps outline the set of viable policies that the state can reasonably assume it can legitimate.
Galston reminded leaders and citizens that they should never think that the good once assumed is secured for good but rather that they should deliberate on political issues and cement agreements through binding rules, i.e. rules enforceable by the state (2010: 393-395, 400). These agreements then become embodied in political practices and institutions. The perceived temporary nature of these agreements point to a political environment of constant (potential) institutional change.
Realists think of politics as a distinct sphere of human life that concerns itself with collective arrangements. As such realist detest the appropriateness of applying individual morality to political goals (Hall & Sleat, 2017: 277). Moreover, Hall and Sleat argued that the issue with thinking about politics as applied ethics is that moral concerns are only going to replicate themselves in politics (Hall & Sleat, 2017: 280). For the realist the purpose of politics is to create and maintain political frameworks where
individuals can live together despite persistent and deep disagreement (Sleat, 2013: 47).
Realists do not deny that political arguments originate from normative commitments but they reject they need to be set out in an 'ideal theory' (Hall, 2015 (B): 4). "Political philosophy must use distinctively political concepts, such as power, and its normative relative, legitimation" (Galston, 2010: 388). Hall and Sleat assert that realist thought does not reject all ethical content but rather it aims to think about politics from an ethical standpoint that accounts for the realist's conception of the political and politics as a distinct domain of life (Hall & Sleat, 2017: 278).1
1.3 Williams’ Realism
Williams starts by taking issue with political moralism (or politics as applied ethics) because as Hall and Sleat assert that modern moral philosophy "fails to make sense of the essential untidiness and complexities of our lived ethical experiences" (2017: 279). Williams set out a methodology for states to assert political legitimacy based on a serious treatment of context; ‘the here and now.’ William's theory of legitimacy is an explorative exercise into how one may criticise a set of political practices or institutions while remaining committed to evaluating them with standards internal to political practices themselves (Rossi and Prinz, 2017: 5).
The initial step for a state to assert its legitimacy is to provide an adequate response to the First Political Question (henceforth FPQ) which asks whether the state has secured order, protection, safety, trust, and the conditions of cooperation (Williams, 2005: 3). The reasonableness of the FPQ originates from an assumption that individuals agree that the core challenge of politics is to "overcome anarchy without embracing tyranny" (Galston, 2010: 391). The set of criteria for state legitimacy is assessed through the Basic Legitimation Demand (henceforth BLD) which asserts whether a state is legitimate (LEG) or illegitimate (ILLEG) (Williams, 2005: 4). As part of the state’s response to a BLD, the state needs to provide a reasonable response to the FPQ (Williams, 2005: 4-5). Sleat argued the FPQ is only
part of BLD, thus creating order out of chaos is a necessary but insufficient condition for fulfilling a BLD (2013: 114). Hall asserts that BLD is central to Williams' realism because it avoids the moralism of much of contemporary political thought (Hall, 2015 (A): 466). The BLD, though potentially a thinly masked moral concept, is a political construct as it necessitates politics, which is inherent in the FPQ (Galston, 2010: 389; Hall, 2015 (A): 467).
Citizens make (often tacit) legitimation demands upon the state which the state is expected to respond to with a legitimation story that Makes Sense (henceforth MS) (Williams, 2005: 10-11). For Williams, a legitimation story MS when “we can comprehend a political regime as an example of a legitimate political authoritative order or not” (Sleat, 2013: 115). "The appropriate criterion for making sense is not 'Does this capture my favoured conception of justice?' but (the inherently political question) 'Is there an acceptable order of authority given that I must coexist with other citizens who have
conflicting interests and different moral beliefs?' " (Hall, 2015 (A): 473). Additionally, a legitimation story MS only when the realism constraint is upheld. Hall defined the realism constraint as the idea that for clarifications of various political values to be meaningful they require a historical and political
contextualisation that asserts what the elaboration of these political values means 'here and now' (Hall, 2015 (B): 6). That which MS should have the capacity of guiding political reform (Hall, 2015 (B): 10).
"A theory is ideal in the right way if its assumptions and expectations are challenging but feasible in its specified domain of application" (Galston, 2010: 407). Williams endorses the realism constraint but receives criticism from scholars who assert that restricting political theorising with considerations of how the world currently is "comes at the unacceptable cost of failing to consider how it ought to be" (Hall, 2015 (B): 6). Sleat thought more generally that realism’s focus on political realities came at a sacrifice for the distance necessary to evaluate (radical) political alternatives (Sleat, 2013: 5). Such a framework is appropriate for this thesis as the aim to prescribe policy that minimises radical deviations from what is thought acceptable.
Williams thought the relationship between the state and the citizen was either a political one or one of successful domination (Williams, 2005: 4-5). This idea is embodied in the critical theory principle that "holds that if the disadvantaged in society accept a justification of power purely as the exercise of power itself, the fact that they accept the story does not legitimate it" (Williams, 2005: 5-6). Internal warfare, in this sense, is thought of as successful domination. Williams asks us to consider the example of the Helot population of Sparta who he asserts are radically disadvantaged (i.e. no better than the enemies of the state) (2005: 5). In this instance, the radically disadvantaged are not in a political situation with the state because despite the state holding coercive power over the Helots there is no legitimation story the state could offer to the Helots that would MS such that they would be convinced not to revolt (Hall, 2015 (A): 467). "The point is that even though political power is coercive not all coercion is political, in much the same way that even though war might be diplomacy by other means, war is not politics by other means" (Hall, 2015 (A): 478).
Significantly Williams asserted that any coercive relationship that is sustained purely through the use of power is illegitimate and hence non-political (2005: 5-6). Williams’ distinction of state-citizen political relationships and ones of mere domination though interesting are not as relevant to the kind of liberal regimes we look at but do serve as a springboard for Sleat’s liberalism-realism reconciliation. Sleat takes issue with Williams’ assertion stating that “rule, appropriately legitimated, will necessitate instances of successful domination” and that “coercion necessarily plays a role in the establishment of political unity and stability (2013: 114). Political realists assert that coordination will require some form of coercion or the threat of coercion and that the presence of coercion is not a sufficient condition to deem a state illegitimate (Galston, 2010: 390).
Williams defines the state as a structure of power and authority distinct from any group or individual, which successfully enforces a monopoly of legitimate violence within an area. Williams’ classic realism focuses on the ability of the state to provide a compelling legitimation story to a substantive part of the population. For Williams, the state is necessary for there to be politics (Raekstad, 2016: 3). Raekstad attempted to save Williams’ assertion2 of the necessity of the state for politics by of the
persistence of the state as a feature of modernity (2016: 14). The debate whether politics necessitates the state can be left on the side-lines of this thesis because the kind of realist theorising I attempt is about state policy.
Moreover, Williams thought that “the legitimations appropriate to a modern state are essentially connected with the nature of modernity as the social thought of the past century, particularly that of Weber” (2005: 9). Weber defined modernity to include: organisation features (pluralism & bureaucratic forms of control), individualism, and cognitive aspects of authority (entzaubert) (Williams, 2005: 9). According to Williams’ conception of Modernity the only viable political framework that would allow a state to be legitimate today (provide a response to BLD that MS) is a liberal one. This was captured by Williams’ equation: LEG = Modernity + Liberalism (2005: 9). For a state to be LEG it must satisfy (more demanding) requirements of modernity: “avoidance of certain strong individual and collective
disadvantages, securing basic rights and freedoms (e.g. free speech), the avoidance of rationalisations of disadvantage in terms of race and gender, and a denial that hierarchical structures that generate
disadvantage are self-legitimating” (Raekstad, 2016: 6-7).3
Critics of Williams praise many aspects of his work on legitimacy in realist thought and have questioned Williams’ liberal conclusions. The result is one of the central criticisms of Williams’
conception of (classic) realism (2005). Hall argues that Williams' endorsement of a liberal state "is not a judgement about the ideal moral optimality of actually existing liberalism but a claim about the merits of realistically achievable competing ways of ordering our political institutions" (Hall, 2015 (A): 474). Sleat
2 An attempt Raekstad himself thought uncompelling. 3 Elaboration of (Williams, 2005: 7)
thinks that while “now and around here liberal regimes have compelling claims on legitimacy”, he asserts that realism still must accept that there can be non-liberal and still legitimate states (2013: 116, 122).
For legitimation to work the people need to agree what the main tenets of Modernity are (Sleat, 2013: 120). "The (often tacit) legitimation story must MS - and when we ask what MS to us here and now we are posing a normative question because 'what MS here and now is a structure of authority which we think we should accept'" (Hall & Sleat, 2017: 284). An interesting point Sleat raises is that the choice of the main tenets of Modernity rests upon an individual’s conception of the political good (Sleat, 2013: 119-120). For instance, Sleat thought that for Nietzscheans and existentialists ‘God is dead’ would constitute a tenet of Modernity. While Marxists could argue that a tenet of Modernity is the assertion that the current political framework and predominant institutions are construed in such a way to benefit a privileged minority – the capitalist class (Sleat, 2013: 121). To minimise the political costs of UBI I avoid deviations from Williams’ conception of Modernity.
Admittedly, prescriptive realist political arguments will inevitably be ethically laden (Hall & Sleat, 2017: 286). Hall and Sleat conceptualise the realist spirit as the ethic of truthfulness where there is a "willingness to see our political reality as it really is, to not succumb to illusions of wishful thinking or to imagine a greater fit between our needs, interests and values and the world that actually exists" (Hall & Sleat, 2017: 281). The way to embrace the ethic of truthfulness is by "endorsing a more historical perspective" and reflect on these historical accounts critically rather than simply accept them (Hall & Sleat, 2017: 281-282).
Per Williams (2005, 136), “Who has to be satisfied that the BLD has been met by a given formulation at one given time is a good question, and it depends on circumstances.” Broadly, the
standard practice for a state to assert legitimacy in Western liberal states relied on the ability of the state to justify a political framework by appealing to the beliefs of a substantial portion (i.e. majority or higher for more fundamental reforms) of those over whom this applies (Sleat, 2013: 51). Williams does necessitate a unanimous acceptance of the legitimation story for the state to be legitimate and moreover insists that 'considerations that support LEG are scalar' (2005: 10) (Hall, 2015 (A): 473).
The need to address BLD by the state is a continuous process and becomes an issue that the state needs to address when a substantial portion of the population invokes it. This can happen when the state implements a policy that affects a substantial portion of the population. This is coherent with the realist view that the status quo is temporary and consequently that institutional arrangements are impermanent.
1.4 Sleat’s Liberal Realism
For the purposes of this thesis, I am not interested in implementations of UBI that require drastic changes to the current political framework. I am not interested to argue against Williams’ need for a state for there to be politics and can assume as is. Without intentionally doing injustice to Sleat’s extensive project on Liberal Realism (2013) I have attempted to contract his main, relevant points into a brief and functional form for the purposes of my thesis. Additionally, this should also serve as a response to some of the main criticisms against Williams and his liberal conclusion.
Williams’ liberal conclusions suggest that he may have thought that a liberal regime is able to pass the critical theory principle. While liberalism MS to us in “conditions of modernity as a sufficient
response to BLD […] it is not clear [...] that Williams' account of politics as non-domination is consistent with the conditions of legitimacy that are intended to determine when a relationship of coercion is political or not" (Sleat, 2013: 120, 126). Sleat thinks the application of the critical theory test was practically difficult arguing that the mere presence of coercion was insufficient to assert that the state legitimation story came from there (2013: 119). A liberal regime has potential to fail Williams’ critical theory test precisely because it uses coercive forces to hegemonise its citizens, for instance by ensuring a liberal conception of freedom is universally valued (Sleat, 2013: 158-159). Liberal politics seems to be oppressive in precisely the way it wants to avoid.
Sleat attempts to reconcile liberalism and realism such that liberalism does not fall under the standard realist critique – this he terms liberal realism. Expanding on his critique of Williams, Sleat asserts that liberal realism must incorporate two claims: that politics can be instances of successful domination (and allow the state to remain legitimate), and that the absence of coercive influences cannot be a condition for a legitimation story (2013: 128).
Sleat proposes rather than think of the liberal state as a non-oppressive regime to instead think of it as a state that deliberately attempts to minimise the political cost endured by any one of its citizens (costs of living under the state). Admittedly, to maintain a reasonable response to FPQ, the liberal regime may increase these political costs upon its enemies (or those attempting to undermine the liberal regime) and may impose extreme costs on active enemies (Sleat, 2013: 161).
While BLD can still be met even if dissenters think otherwise, for “the mere voicing of a
complaint cannot be sufficient to generate a genuine BLD” and legitimation requires a substantial portion of the population affected by the policy to accept the state’s response to a BLD (Sleat, 2013: 116123-125). As such rule over dissenters can be allowed, they will inevitably be subjected to the state’s coercive powers, and the state’s use of coercive power over dissenters whose activities invoked the FPQ become increasingly legitimate as stability and order for a substantial portion of the population is prioritised (Sleat, 2013: 127-128).
Hall underlines the distinction between Sleat's structural and Williams' relational conception of legitimacy (2015 (A): 470). Sleat thought that Williams' insistence that the state needs to offer a
legitimation story to each and every subject relied on a moral premise that each individual is of equal moral worth (Hall, 2015 (A): 470). Instead Sleat's conception of legitimacy embraces Williams' assertion that a 'substantial' portion (defined circumstantially) of the population needs to accept the legitimation story of the state to be legitimate without the need for the state to offer the legitimacy story to each of its subjects (Hall, 2015 (A): 470). Williams thus uses more restrictive conditions on whom the state is expected to legitimate itself to in comparison to Sleat (Hall, 2015 (A): 472). Rather Hall asserts the legitimation story should be offered to citizens or 'political subjects' (Hall, 2015 (A): 470-471). For this thesis I adopt Sleat’s (less demanding) structural conception of legitimacy.
In liberal states the rule of law is a set of institutional arrangements that keep rulers from repressing the ruled (this yields to the idea that liberal rulers are ‘restrained masters’) (Sleat, 2013: 166-167). Citizens in liberal states have an important role to play in ensuring their rulers maintain their commitment to self-restraint by holding them to account if they transgress them (by making BLD, for instance) (Sleat, 2013: 169). “The realist’s vision of the political assumes that persons are at least
sufficiently committed to their causes such that they are willing to engage in the struggle for power, and in the face of deep disagreement with the ends and values they pursue” (Sleat, 2013: 139).
In terms of legitimation stories in liberal realism the state is expected to provide two; one towards its enemies (those who reject the political framework) and one towards friends and adversaries (those who broadly accept the political framework but may have substantial disagreements over it) (Sleat, 2011: 152-153). This can be thought of as a means of motivating a political relationship with those that think all state coercive force is illegitimate. I argue this motivation is compatible with Sleat’s structural conception of legitimacy. Legitimate rule rejects the need to consistently use force to maintain order, rather it is the threat of force that maintains order against minorities that have rejected the legitimation story (Sleat, 2013: 62).
Liberalism should not be thought of as the default response to the political question (Sleat, 2013: 138). Liberal realists must embrace their partisanship and defend their values against both enemies and simply dissenters (Sleat, 2011: 140). Though the realist conception of the political as deeply conflictual, simply asserting the nature of the political was insufficient to reconcile liberalism and realism (Sleat, 2013: 129). Politics is a contest of power and not merely an exchange of opinions.4 Williams thought we tend to
misrepresent political enemies as fanatics (referencing the portrayal of Islam in contemporary times) (2005: 137). "Realist conceptions of political disagreement help us recognise that a political decision does not 'announce that the other party was morally wrong, or indeed, wrong at all. What it immediately announces is that they have lost' " (Williams, 2005: 13).
Sleat thinks the way to solve this is for political participants to embrace agonism. Mouffe’s ‘agonism’ assumes the political as necessarily conflictual but certain kinds of conflict as potentially order-creating. For this to be achieved political opponents embrace agonism and thus accept the legitimacy of their opponents’ irreconcilable views (Sleat, 2013: 145-146). Liberals realists must accept the endorsement of anti-liberal alternatives as reasonable. The realist liberal not only accepts the reasonability of their enemies’ positions, asserts mastery over their enemies (dominates them), but accepts that the nature of their relationship with their enemies may be a temporary one (Sleat, 2013: 172). Democratic politics then aims not to eradicate conflict (or politics) but rather to channel and direct it such a way so that it becomes compatible with pluralist democracy (Sleat, 2013: 147).
"Now and around here we only permit liberal solutions because 'other supposed legitimations are now seen to be false and in particular ideological'" (Hall, 2015 (A): 469). Even though a liberal state may be the most reasonable response to BLD (given Williams’ conception of modernity), a realist liberal must accept the potential for legitimate regimes that are not necessarily liberal in character. A regime that came to power by asserting that its normative framework is X (and has commitments towards certain
substantial policies) is legitimate if it pursues policies that aim to achieve an institutional arrangement X. These policies are decided upon through adversarial interactions among several interpretations of various specific normative values (Sleat, 2013: 155).
Liberal realism attempts to provide a specifically liberal response to FPQ while maintaining the reasonableness of (at times radical) opposition. Liberal realists should allow for not only the voicing of the arguments of the opposition but being open to the potential that political disagreement leads to new legislature and drastic changes to the political framework. Currently within the liberal state, Sleat argued that through political constitutionalism individual rights and their protection can be conceived of as political constructs which have the capacity to change through legislative processes (2013: 171). This constitutes an ethics approach to politics that is itself derived from the political.
1.5 Realist Method
First, we develop a serious treatment of context, then we outline the main political issues. These political issues can be thought of as the content of a BLD made upon the state. The state is expected to respond to the BLD with a legitimation story that MS. The response to the BLD should MS to a substantial portion of the population. In our case part of the state’s response is UBI. We consider how does UBI address the main political concerns for it be considered a reasonable response to the BLD. In the end, I raise real concerns over implementation and portrayal of UBI. In the next chapter I develop a serious treatment of context and outline the most relevant political concerns.
Chapter 2: Context
2.0 Introduction
The purpose of this chapter is to develop a serious treatment of context that sets-up an argument in favour of a Universal Basic Income policy in the next chapter. I start with a realist critique the capacity of the market to correct for technological unemployment. Then, I argue for a historical precedent for the expanding role of the state in response to mass unemployment and economic downturns. Thereafter I concentrate on the content of the context which include: theoretical and empirical predictions about the trends of technological innovation, the effects technological innovation is predicted to have on
employment (i.e. technological unemployment), and the changes to skill and education premiums. The main claims that make up the context are the following:
1. The corrective capacity of the market to resolve technological unemployment (i.e. theory of compensation), rests on overly ideal foundations, and is, at least limited, and at worst non-existent. Alternative, more realist theoretical explanations should be considered.
2. Today the consensus is that the state is obliged to help bring about full-employment and to intervene in an economic downturn. Empirical evidence shows expansionary macroeconomic policy has been effective in dealing with recessions and unemployment
3. The technological disruption is predicted to yield mass unemployment first affecting the lowest-skilled jobs and then moving up the skills ladder. The upcoming technological disruption will uniquely benefit the high-skilled most (as opposed to previous historical episodes of
technological unemployment.
4. There has been an inflation in the value of formal accreditation (degree and skills) such that a university degree no longer guarantees employment (to the same extents as before).
5. Educational reform is necessary but insufficient to tackle the problems the upcoming technological disruption is expected to bring.
6. There may be historical precedent to thinking long term technological unemployment does not happen. The critique of neoclassical theory, and the scale and scope of the predicted
technological disruption suggest long term unemployment should be expected. My aim is to argue that this context yields a series of BLDs that:
1. Necessitate government action
2. Requires a change in government policy so that the response to the BLD is deemed reasonable 3. This change may involve the introduction of new policy that could be presented as a natural
potential expansion of the obligation of the state towards its citizens, and subsequently be also a reasonable response to the BLD
The state respond to BLDs stemming from an updating context (e.g. technological
unemployment trends). To achieve an accurate description of the context I consult several relevant and recent empirical studies. The last section in this chapter asks whether the technological disruption lead to state instability such that fundamental policy changes become necessary?
2.1 Realist Critique of Neoclassical Economic Theory
Historically, fears of technological unemployment tended to arise in times of radical technological changes. Since the beginning of the technological unemployment debate economic theory pointed to the existence of corrective economic forces (Vivarelli, 2013: 1). In the late 1920s several economists argued that workers recently displaced by technological innovation were being absorbed into new sectors not currently covered by statistics (Woirol, 2006: 483). Technological innovation was thought to create new jobs elsewhere after the automatable ones were destroyed. The market would respond to the supply of unemployed workers by providing an adequate amount of (reasonably meaningful) work. Broadly, this effect has been termed the ‘theory of compensation’ or the ‘compensation mechanism’ (Vivarelli, 2013: 2). Woirol argued that empirical evidence from the times did not provide clear evidence for either side of the technological unemployment debate (Woirol, 2006: 485). Until 1931-1932 the theory of compensation argument formed the core of the consensus of the economic forces at work on unemployment.
Broadly, the theory of compensation asserted that technological innovation brought about a reduction in the per unit production cost (assuming perfectly competitive markets). In 1931-1932 the new economic theory consensus emerged (and lasted until 1939) was based in the neoclassical tradition. Any lasting unemployment was argued to be a malfunction of the price system (Woirol, 2006: 489).
Economic theory, from both neoclassical and modern schools, explained the compensation mechanism through several causal effects. Vivarelli (2012) provides a survey of the main arguments for the theory of compensation and offers rebuttals. The rebuttals originate primarily from criticisms of the fundamental axioms of neoclassical economic theory. "A theory is ideal in the right way if its assumptions and expectations are challenging but feasible in its specified domain of application" (Galston, 2010: 407). Three of main arguments in support of compensation theory go as follows5:
1) New technologies generate new sectors and thus new employment opportunities in these sectors 2) A decrease in unit production costs, through competitive market pressures, cause a drop in
prices. The price drop causes an increase in demand for products and thus for new employment. 3) New investments generate new jobs. The funds (savings) for the investments originate in the lag
between the initial drop in production costs and the subsequent drop in prices due to competitive pressures.
Relevant neoclassical economics assumptions include: A) perfect means-ends rationality
B) full information C) perfect competition
Galston thought neoclassical economic theory engaged in the wrong kind of ‘ideal’ theorising by basing its core theories on idealised, contrary to fact, assumptions about the real world (2010: 403). More comprehensive attempts at criticising and (partially) rejecting these assumptions has been done elsewhere (Arnsperger & Varoufakis (2006), Kahneman (2011)). Assumptions A and B are manifested in the economic conception of the rational agent which Kahneman (2011) thought unrealistic. Admittedly, Arnsperger and Varoufakis (2006: 7-8) argued that the contemporary conception of the
homo-oeconomicus “has evolved to resemble us more,” however, Galston argued that economic theory should not be “derived from minimal deviations of Adam Smith but rather from deviations from real
occurrences and observations” (2010: 403).
The full information assumption relates to the capacity of the market to signal information about products and trends through prices. Even if the assumption about the market’s capacity to provide a full set of information to economic agents were true, the capacities of economic agents to fully, absorb, evaluate, and appropriately act upon this information has been disputed (Kaheneman, 2011). For neoclassical analysis, the labour market is thought of in similar terms as the product market. Assumption C is not a reasonable definition of the persisting real world market (Gallant, 2014).
These assumptions stood as the basis of economic theorising in the USA for the past half century. “These assumptions permitted the construction of elegant theoretical structures whose
connection with the world became increasingly remote. Knowing what good policy would be if economic actors were completely informed and fully rational said almost nothing about good policy in the real world” (Galston, 2010: 403).
Following, largely intuitive, criticism of relevant neoclassical assumptions here are responses to arguments supporting the theory of compensation:
1) The assumption of perfect information applied to the labour market supports the idea that the market has capacities to respond to changes in the supply of labour with reasonably meaningful employment opportunities. In reality the (labour) market need not react with an adequate amount of employment opportunities. Moreover, employment opportunities are not homogenous, may be meaningful for only specific individuals (not non-differentiable rational agents), and require certain skills (or skill level) that the current skill make-up of the labour supply may not be able to satisfy. For example, a substantial group of recently unemployed retail workers need not have transferable skills necessary to work in the IT industry where new employment opportunities were recently created. If we assume labour-replacing technologies are employed to minimise
costs then a machine is usually only employed if it produces more than the number of workers it replaces.
2) The idea that prices would drop when production costs would decrease in an industry due to the introduction of a new technology relies on the assumption that the market is perfectly
competitive. As several companies in an industry introduce this cost-saving technology individual companies reduce prices to maximise demand for their product and under-cut competitors. Other companies, to retain demand for their product, follow suit. The first effect following the adoption of labour-replacing technologies is a drop in aggregate demand as workers become unemployed and lose their incomes (Vivarelli, 2013: 8).
3) This relies on the assumption that economic agents are rational agents. In reality the,
expectations of economic agents play a significant role. For instance, during downturns, where unemployment tends to increase, economic agents become pessimistic about the future and are thus less likely to invest. All profits are not necessarily translated into investments and relies heavily upon ‘animal spirits.’6
Considering these criticisms credence in the capacity of market forces to correct technological unemployment should at least be questioned, and at worst entirely rejected. Working with the theoretical consensus of the 1930s Frederick Mills' empirical research showed that technological change played a significant role in the unemployment that emerged in the 1930s (Woirol, 2006: 490). Mills argued that the central economic issue of the 1920s and 1930s was that productivity gains were distributed predominantly towards the producers rather than towards lower prices or wages (Woirol, 2006: 490). The coming of WW2 and the resultant employment increases geared towards the war effort silenced the technological unemployment debate.
Many contemporary neoclassical economists point to history and argue that technological innovation has created new jobs displaced workers could take on. The predicted effects of the upcoming technological revolution on employment were thought palatable overall; companies enter industries where productivity is relatively high, invest in those industries (i.e. capitalisation effect), and thus provide
expanded employment opportunities. For instance, Atkinson has argued that in the future
“unemployment and labour force participation rates […] should be quite similar to the rates of today” (2016). Atkinson labelled proponents of the idea that recent computerisation will lead to mass structural unemployment as ‘techno-distopians’ and asserted that historically fears of mass unemployment due to technological disruptions were often unfounded (2016). Skidelsky rejects the optimist’s view that technological innovation will create many new types of jobs (2013).
Vivarelli argues that while ‘the end of work’ theories7 underestimate the opportunities for
compensation, economists often forget that compensation is partial, reliant upon “historical and institutional circumstances” (2013: 11). Cesaratto, Serrano, and Stirati argue that expansionary
macroeconomic effects of technical change depend crucially on the macroeconomic policies of the state (2003: 49). Additionally, Cesaratto et al. argue that compensation effects are largely absent (2003: 33-34).
According to Woirol, apart from the technological unemployment debates of the 1930s and 1960s, the consensus (both public and professional economists) has been that technological innovation does not cause long-term unemployment (1996: 3). Cesaratto et al. argue that technological
unemployment could lead to long-term unemployment if expansionary macroeconomic policy is not pursued (2003: 50-51). Admittedly, the abandonment of expansionary macroeconomic policy has been the case since the mid-1970s (Cesaratto et al., 2003: 51).
2.2 Mini-Genealogy of State Obligations Towards its Citizens
We have explored the ways in which unemployment is thought of in mainstream economic theory, provided realist rebuttals to prescribed policy measures stemming from such thinking, and additionally provided supportive historical evidence that rejected, or at least questioned, the capacity of the market to correct technological unemployment. The realist critique of neoclassical models of the labour market suggest real world instances of failures of such economic policy and the potential need for (emergency) state intervention.
Following from the realist critique of labour market economics I present a mini-genealogy of the expanding obligations of the state towards its citizens in response to instances of substantial and
persistent national unemployment that market mechanisms failed to resolve within a reasonable timeframe. “A genealogy is obviously not a critique in the everyday sense […], it does not automatically imply the rejection of what is subjected to genealogical analysis” (Geuss, 2002: 212). The ultimate purpose of this part is to motivate a potential future legitimation of state intervention against predicted mass technological unemployment that is broadly similar to instances where state intervention was necessary in historical episodes of mass unemployment.
Woirol conducted a survey of several letters US citizens sent to President Roosevelt in the 1930s. These letters contained avid descriptions of the hardships and struggles millions of American unemployed endured. Most letters sent requested aid in some form, be it immediate relief, employment, or loans for their homes or farms, while a smaller category of these proposed economic plans of some form. “The plans rarely suggest[ed] radical schemes, but the acceptance of government action as a solution to the depression was acceptance of government action as a solution to the depression was accepted almost as
7 I.e. theories that assert that automation is growing at an increasing rate and viable work opportunities are ever
an article of faith” (Woirol, 2012: 572). These letters can be thought of as explicit BLDs a substantial portion of the population (max unemployment rate: 25% (1933), and widespread strife) made upon the state.
The letters were uniformly serious and almost exclusively proposed reform over revolution. Out of the sample set of individual policy proposals substantially more aimed at the demand side of labour markets: 1) Increase consumer spending and purchasing power, 2) Direct job creation plans, 3) Stop job displacement - in this order of popularity (Woirol, 2012: 573). The second type of policy proposed aimed at the supply side of job markets and only the third at improving the functioning of markets. The state’s response to the BLD should include policy proposals that are both believed to work in theory and that a substantial portion of the population is bound to accept. Almost all letters surveyed made sense from an economic point of view and a “large portion made suggestions that had the possibility of having a positive effect on economic recovery” (Woirol, 2012: 571). Therefore, policies aimed primarily at the demand side of the labour market are bound to have a higher probability of being accepted.
Economists in the 1930s strongly believed in the self-correcting power of the markets to reduce technological unemployment, even in the short term. Admittedly, several classical economists believed even though compensation effects would remedy technological unemployment they understood that market adjustments were not immediate and advocated for government aid to help displaced workers gain new employment (Blaug, 1997: 182). The New Deal implemented policies that offered relief
(through transfer payments) to the third of the American population that were worst off and put millions of people to work through public work programmes (Krugman, 2008). The New Deal policies seemed to reflect the policy recommendations discontented Americans had made upon the state.
In terms of transfer payments, the New Deal was developed as a two-track welfare system. The first track was designed for the white male population and aimed to partially replace the workingman’s family wage, whilst the second was deliberately aimed at women and minorities. The first was designed with the appearance that “beneficiaries merely got back what they put in” and was thus portrayed as a reward for workingman status, whilst the second came with a means-testing requirement for eligibility aimed to find the neediest in the ’second-track’ social group (Fraser & Gordon, 1994: 321-322).
According to Fraser and Gordon the social stigma associated with receiving welfare payments originates in the deliberate moral opprobrium inherent in welfare payment eligibility testing (1994: 322).
Additionally, Fraser and Gordon argue that the social stigma surrounding welfare payments is a post-industrial age phenomenon related to social connotations of dependency (1994: 324). Whilst during the industrial times dependency may have been thought of as necessary, in the post-industrial age all forms of dependency were thought of as avoidable and blameworthy (1994: 323). For neoclassical economists, the moral/psychological meanings of dependency are present but uninterrogated, assumed to be undesirable” (Fraser & Gordon, 1994: 328). “Liberals and conservatives […] rarely situate the notion
of dependency in its historical or economic context; nor do they interrogate the presuppositions” (Fraser & Gordon, 1994: 328).
The New Deal had several long-term achievements which set a precedent on what constitute as legitimate state actions. It established “durable and essential” institutions (e.g. bank deposit insurance and social security) that remain “the bedrock of [the USA’s] economic stability” (Krugman, 2008). Admittedly, Krugman argued that FDR’s New Deal policies proved of limited success in the short run because the policies were economically cautious (2008). “What saved the economy, and the New Deal, was the enormous public works project known as World War II, which finally provided a fiscal stimulus adequate to the economy’s needs” (Krugman, 2008).
During the 1940s the previous consensus on policy against technological unemployment was rejected and instead Keynes’ macroeconomic arguments in favour of full employment became the consensus (Woirol, 2006: 491). By the 1960s it became accepted that short-term technological
unemployment should be corrected through government intervention (i.e. job creation and spending). The expanded governmental role can be conceived of as an expanded set of obligations the government has upon its citizens.8
Obama’s American Recovery and Reinvestment Act was a policy package aimed at reigniting the American economy following the 2008 financial crisis and subsequent recession. The ‘stimulus’ package aimed at inducing public spending by temporarily boosting direct government spending, cutting taxes, and increasing public aid (to increase family incomes) (Krugman, 2014). Now there is strong empirical evidence, and a broad consensus among economists, that the package had a strong impact on
employment and output (Krugman, 2014). After 2010 concerns over budget arose again and austerity measures took priority over job creation programmes (Krugman, 2014).
The New Deal expanded the set of obligations of the state upon its citizens. Nowadays, the institutions established then were maintained and transformed such that several government welfare provisions are taken for granted. A precedent for breadth of the role of the state in the economic lives of its citizens had been set. The Western state now often provide unemployment benefits, aids in the job search, and the government creates jobs in times of economic turmoil to boost the economy. Through the New Deal the US Department of Defence became, and remains, the biggest single employer in the world today.
In this previous section I argued that there is a historical precedent for the expansion of the role of the state in response to instances of mass employment and economic downturns. In the next few sections I broadly review recent unemployment trends, assess Frey and Osbourne’s (2013) predictions of the technological disruption, and consider the need for educational reform.
8 Or, admittedly, the pursuit of already agreed upon obligations (e.g. Universal Declaration of Human Rights
2.3 Recent Unemployment Trends
The figure below shows employment trends coincide with business cycles – drops during recessions (grey areas on the timeline) and increases during recoveries. Up to the last two recessions (in 2001 and in 2009) recoveries tended to coincide well with increases in employment, or at the very least ends to job losses. Within 3 months following the trough of a business cycle employment tended to rise. In 2001 employment continued to fall nineteen months before recovering (Economist, 2009). The Economist attributed the employment trends lag following the 2001 recession to increases in structural unemployment (Economist, 2009). The 2009 recession brought about a 6.3% drop in non-farming employment (Ghanbari & McCall, 2016). The decrease in employment took 25 months and recovery to pre-crisis levels took 51 months (Ghanbari & McCall, 2016). Skidelsky argued in the past 25 years there has been an upward trend in structural unemployment and it has become increasingly difficult to keep unemployment down (2013).
According to Gorz, industrial productivity has increased 5-6% every year since 1978 while the economy grew an average of 3-4% and production of commercial goods and service has risen by 2% a year since (2001: 3.1). This means that every year despite a growing economy, the amount of labour it requires is reduced by 2%. This means that without accounting for increasing rate of computerisation in the next fifteen years the amount of labour the economy requires would decrease by a third (Gorz, 2001: 3.1).
In 1998, the US business sector put in 194 billion hours of labour. By 2013 the output of the business sector increased by 42% but the total number of hours of labour necessary to produce this remained 194 billion. Despite this 40 million people were added to the workforce (Ford, 2015: 281). Ford argued US economy needs to produce around 1mill jobs a year to keep up with the growth of the
workforce. Between 2000 and 2010 no new jobs were created on the labour market (2015: xi). Moreover, a recent McKinsey report noted that 30-45% of the working-age population in the world is underutilised (i.e. unemployed, inactive, or underemployed (Manyika, 2017). According to the report while most attention is placed on unemployment, too little attention is placed on
Figure 1: Total nonfarm employment, seasonally adjusted, Jan 1939-Dec. 2015 (Ghanbari & McCall, 2016)
2.4 Technological Disruption
In a 2013 article “The Future of Employment: How Susceptible are Jobs to Computerisation” Frey and Osbourne present a compelling case, based on empirical data, for what the world of work might look like given our current and predicted pace of technological development in relevant fields. The authors categorise jobs in the USA based on their susceptibility to computerisation into low, medium, and high-risk categories. According to this study 47% of US employment is at high-risk of potentially being automatable "perhaps [in] a decade or two"9 (Frey & Osbourne, 2013: 38).
Computerisation means the replacement of human labour with machines equally (or more) capable of performing the same work for a lower cost. The replacement of human labour with machines originates from an incentive to continually minimise production costs. Industry-wide adoption of such a technology may lead to the obsolescence of skills necessary to perform the original task. Technological substitution of labour leads to a destructive effect where workers lose their jobs and are required to relocate their labour supply. Current technological advancement may prove qualitatively different as, the current pace of technological development and adoption is outrunning the pace at which we can find new uses for obsolete labour (Frey & Osbourne, 2013: 2).
The reason that human labour has consistently prevailed as a factor of production is due to the ability of human beings to use their cognitive abilities and their capacity to adapt to dynamic tasks. "While technological progress throughout economic history has been largely confined to the mechanisation of manual tasks, requiring physical labour, technological progress in the twenty-first century can be expected
9 It is important to note that this study only tested for the full automation of a given occupation and did not concern
to contribute to a wide range of cognitive tasks, which, until now, have largely remained a human
domain" (Frey & Osbourne, 2013: 19). Now computerisation is entering more cognitive domains and the ability of the market to provide low-skilled employment opportunities, even in newly expanding industries is diminishing (Frey & Osbourne, 2013: 13).
The pace of technological innovation is increasing and the adaptability of these technologies to various industries is increasingly less confined to routine manufacturing jobs. Jobs that were thought of as inherently irreplaceable (such as driving and handwriting recognition) because they relied on specifically human qualities are becoming increasingly susceptible to computerisation and will inevitably be replaced by machines. Frey and Osbourne's study shows that developments in machine learning "will reduce the aggregate demand for labour input in tasks that can be routinized by means of pattern recognition, while increasing the demand for labour performing tasks that are not susceptible to computerisation (2013: 36).
The computerisation of a given occupational task first relies on the ability of engineers to determine an adequately specific problem they are trying to solve. Even though technological innovation as a process is continual and ever accelerating, the adoption of technologies relies on more sporadic technological developments that manifest themselves once an adequately specified engineering problem has been solved (Frey & Osbourne, 2013: 4).
A comparative advantage of computers over human labour is scalability and this is exhibited clearly with Big Data (Frey & Osbourne, 2013: 16). Technologies become capable of replacing human labour when the size and complexity of the data set is both large enough to define the problem, a given algorithm is adequately sophisticated to find viable solutions, and feasible engineering solutions have been developed that turn out to be cheaper than human labour (Frey & Osbourne, 2013: 15-16).
The extent of computerisation will be defined by how well engineering bottlenecks can be overcome. These engineering bottlenecks correspond to three categories: perception and manipulation tasks, creative and intelligence tasks, and social intelligence tasks (Frey & Osbourne, 2013: 24-27).
2.5 Technological Unemployment
Frey and Osbourne predict two waves of computerisation separated by a 'technological plateau" (2013:38). In the first wave, most workers in the transportation and logistics occupations, together with the bulk of office workers and labour in production occupations are 'likely to be substituted by computer capital' (Frey & Osbourne, 2013: 38).
"Generalist occupations involving the development of novel ideas and artefacts, are the least susceptible to computerisation (Frey & Osbourne, 2013: 40). Frey and Osbourne "predict that most management, business, and finance occupations, which are intensive in generalist tasks requiring social intelligence" and "most occupations in education, healthcare, [..] fine arts, and media" are largely confined
to the low-risk category (Frey & Osbourne, 2013: 40). Admittedly, computerisation is coming for the high-wage, high-skill jobs as well but occupations that require great degrees of social and creative intelligence are least likely to be computerised in the nearest future (Ford, 2015: 27).
Figure 2: The distribution of BLS 2010 occupational employment over the probability of computerisation. Area under the graph is equal to total US employment (Frey & Osbourne, 2013: 37)
Recently jobs that had been previously off-shored to minimise production costs are coming back to the USA as computerisation technologies are becoming more efficient, and capable of competing with even the lowest off-shore workers (Ford, 2015: 9). The off-shoring trend has been reversing in recent years with China losing 15% of manufacturing jobs between 1995 and 2002 (Ford, 2015: 10). In 2014 106,000 contract workers lost their jobs as Nike decided it is cheaper to invest in computerisation technologies in the long-run than continuing to pay the world's lowest-cost labour in China (Young, 2014). If long-term cost analyses point in favour of investing into computerisation over human labour there should be concern over technological unemployment.
2.6 Skills and Education
Technological innovation is skill-biased, i.e. technology tends to more readily replace jobs that required low skilled labourers and create jobs that require high qualifications. Technological advancement is beneficial to the educated and high skilled (Vivarelli, 2013: 23). Frey and Osbourne argued that it was a 21st century phenomenon that contemporary technological advancement favours higher skilled workers
(Frey & Osbourne, 2013: 9). Current technological development will test the ability of human beings to overcome the capacities of machines through education and thus become demanded on the labour market once again. The increasing rate of technological progress and the seemingly limited (or delayed) adaptation of human societies to these disruptions points to a potential for a rise in the natural rate of unemployment10 (Frey & Osbourne, 2013: 14). There is growing employment in high and low income
occupations and the hollowing-out of middle-income jobs (Frey & Osbourne, 2013: 42).
Initially computerisation is expected to consume the “base of the jobs skills pyramid,” however advances in AI and adjacent technologies is expected to have capacities to replace jobs that require more higher level of skill to complete (Ford, 2015: 252-253). This assertion must however consider other obstacles to technological adoption such as economic feasibility and government policy. "The expansion in high-skilled employment can be explained by the falling price of carrying out routine tasks by means of computer, which complements more abstract and creative services" (Frey & Osbourne, 2013: 12).
Ford noted a credential inflation where the requirements are increasing but the complexity of work remains the same (2015: 252). Historically there seemed to be a reasonable match between job requirements and workforce capabilities. Additionally, where education is free substantial portions of the employed (e.g. 27% in Canada, 43% in China) tend to be overqualified for the work they are doing (Ford, 2015: 251). Beaudry et al. (2016) documents a decrease in the demand for skills, even as the supply of skilled labour increased. This constituted a reversal of demand for cognitive tasks around the year 2000. In response, high skilled labourers have moved down the occupational ladder (underemployed) and have thus displaced low skilled labourers, at times beyond the labour market (Beaudry et al., 2013:
199). Moreover, the McKinsey survey11 pointed out that 40% of employers said the main reason for
entry-level job vacancies was a lack of skills (Manyika, 2017). 60% of graduates were thought to be inadequately prepared to the world of work.
Ford argues against conventional knowledge that emphasises investment in education and training as a solution to the rapid computerisation of the workplace is false asserting that we cannot simply “cram everyone into that shrinking region at the very top” of the jobs skills pyramid (2015: 253-254). Krugman argued that if workers got college degrees there may still be no jobs available, or at the
10 Combination of frictional and structural unemployment that persists in efficient markets. 11 Sample: employers and young people in 9 countries.
