Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Social Interaction, Envy, and the Basic Income: Do Remedies to Technological Unemployment Reduce Well-being?

Social Interaction, Envy, and the Basic Income: Do Remedies to Technological Unemployment Reduce... 1IntroductionThe debate on technological unemployment has a long history, which in the past seemed to have reached the overarching conclusion that the phenomenon is of limited empirical relevance. Matters began to change at the end of the last century, when the ICT revolution was found to have an impact on employment, so that in recent years some authors (see, e.g. Brynjolfsson & McAfee, 2011, 2014) have questioned that old conclusion. The debate has since gained momentum, but has not converged on any broad consensus.In fact, technological progress is generally recognized as one of the causes of increasing income inequality over the last few decades, a phenomenon known as skill-biased technical change. Different approaches focus on tasks rather than skills, i.e., task-biased (or routine-replacing) technical change, but the logic is quite similar. In any case, there is less agreement about the possibility that technical change could cause a significant increase in the overall unemployment rate in the long run, i.e., technological unemployment, although a growing number of empirical studies are starting to track the phenomenon (see, e.g. Acemoglu & Restrepo, 2017, 2019), and other contributors have highlighted the devastating effect that artificial intelligence may have on employment (see, e.g. Acemoglu & Restrepo, 2020; Brynjolfsson & McAfee, 2011, 2014; Ford, 2015).Indeed, skill-biased (or task-biased) technical change and contemporary technological unemployment are two profoundly different phenomena. The former is a consequence of 20th-century technological progress, which has been able to bring down labor demand and hence wages for low-skilled workers (or workers carrying out routine and repetitive tasks) but drive up labor demand and hence wages for high-skilled workers (or workers carrying out non-routine tasks). Income inequality is the result. Technological unemployment is considered a possible consequence of 21st-century technological progress, associated with artificial intelligence, which may well jeopardize employment in most sectors of the economy for all workers, all jobs and all tasks in the future. The fact that robots are supposed to be more productive and less expensive than workers could result in high levels of unemployment and, according to some scholars, even the risk of the system converging on a state of full unemployment cannot be a priori ruled out (see e.g. the benchmark model in Berg, Buffie, & Zanna, 2018).Summing up the different positions on the issue, there is a general consensus that technological progress contributed to increasing inequality in the closing decades of the 20th century; there is a majority consensus that it could increase unemployment in the coming years; and there is minority but growing consensus that it might drive the system to mass unemployment.True, the possibility of the development of artificial intelligence leading to high levels of unemployment does not amount to a certainty. However, the basic idea behind the present article is that developing thorough theoretical analysis of and discussing the possible solutions to a problem that does not as yet demand attention but may well in the near future is perfectly feasible, appropriate and useful. Thus, the present article is in strong opposition to the consensus view that it is possible to discuss an economic problem only after empirical proof that it has actually arisen. And since empirical evidence of future events is inevitably lacking, in discussing the possible dynamics of the new process we can only undertake a thought experiment using the tools of economic theory.In any case, if we consider the possibility that robotization and the diffusion of artificial intelligence could generate high levels of unemployment, ex-ante study of public intervention policies that can counteract the phenomenon appears mandatory. And indeed many policies have been proposed. Surveying these policies, we find various possibilities: providing an unconditional basic income, taxing robots, subsidizing the hiring of humans, nationalizing firms and/or assigning the property of robots and/or firms to workers, and setting minimum human employment quotas or maximum robot employment quotas. Providing an unconditional (or universal) basic income to all citizens, irrespective of whether they are employed or unemployed, is the most debated and best-known of these possible policies. The present article focuses on this policy and in particular on introducing a basic income high enough to guarantee subsistence for human beings, as indeed it should be in a context that sees technological unemployment looming as a concrete possibility.Although the basic income has been searchingly discussed in the literature (see, e.g. Bowles, 1992; Gamel, Balsan, & Vero, 2006; Haagh, 2019a; Marchant, Stevens, & Hennessy, 2014; Pech, 2010; Tanner, 2015; Van der Linden, 1997, 2002), few studies have tried to evaluate the impact on well-being of a universal basic income explicitly designed to counteract technological unemployment (see, e.g. Ford, 2009, chapter 4; Ford, 2015, chapter 10; Brynjolfsson & McAfee, 2014, chapter 14): most studies have simply discussed the basic income as an instrument to reduce poverty and inequality. This is interesting in view of the fact that the contemporary resurgence of interest in this measure rests mainly on the possibility that it could play a role in counteracting technological unemployment. And economic theory is now fully equipped with theoretical instruments to evaluate the well-being consequences of such a policy. In particular, insights from behavioral and happiness economics – such as loss aversion, hedonic adaptation, escalation, social interaction and positional effects – appear useful for the purpose.Building on the above theoretical constructs and on Keynes’ short 1930 essay (“Economic Possibilities for Our Grandchildren”), which can be considered the seminal contribution on the issue, the present article aims to discuss in detail and in depth the impact that an unconditional basic income, implemented to counteract technological unemployment, might have on people’s well-being.It is worth noting that the basic income could represent a solution to all three ways (past, present and future) in which technological progress impacts on employment and wages (i.e., past increasing inequality due to skill-biased or task-biased technical change, current increase in unemployment due to technological progress, and a future with high levels of unemployment due to robotization and application of artificial intelligence). However, the focus in the present article is not on the (past) increase in wage polarization but on the (future, possible) increase in unemployment and the possibility of reaching a situation of mass unemployment (as will be shown, providing a basic income could allow wages to fall below subsistence and hence help prevent the system from drifting into a situation of mass unemployment).The main conclusion of the article will be that an unconditional basic income that remains unchanging over time could aggravate inequality in well-being and, more in general, lower the general level of well-being. The insights offered by behavioral economics lead all too easily to these outcomes. In particular, hedonic adaptation and loss aversion can be applied to show that provision of a basic income cannot adequately compensate for the psychological costs of unemployment, which empirical studies consider one of the worst experiences people may face in life (see, e.g. Clark & Oswald, 1994; Frey & Stutzer, 2002; Clark, Diener, Georgellisand, & Lucas, 2004). Furthermore, unemployed recipients of the basic income, far from provoking the envy of other people, could only envy the employed, the owners of the robots and the entrepreneurs. They would suffer problems of self-esteem and social stigma. Finally, they would be unable to escalate to higher-grade consumption behaviors, compensating for hedonic adaptation and passive envy, or fulfill their aspirations (and demoralized by the awareness of being unable to do so) – yet more circumstances that lower well-being. Provision of a higher basic income might overcome the first problem, i.e., the psychological costs, but not the others. By contrast, employed earners of the basic income (as well as entrepreneurs and owners of robots) would enjoy a significant increase in well-being because of the possibility to both provoke the envy of the unemployed and escalate to higher-grade consumption behaviors. Consequently, a universal basic income might well see inequality in well-being increasing.A possible alternative solution could be based on a policy mix combining a basic income that increases over time with a reduction in per capita working hours and a policy that mimics Tietenberg’s (1990 and 2003 tradable permits approach by imposing mandatory quotas for hiring human workers on firms. A solution of the sort has a number of positive effects on well-being and guarantees lower costs for firms that face higher costs in increasing the human–robot ratio.The present article is structured as follows.In Section 2, the impact of technological progress on employment, i.e., technological unemployment, is discussed from both the historical and the theoretical viewpoint. The possible policy solutions for this phenomenon, focusing in particular on provision of an unconditional basic income, are addressed in Section 3. Based on Keynes’ 1930 contribution and insights offered by behavioral and happiness economics, such as hedonic adaptation, envy, and escalation, the theoretical tools that one can find in the economics literature to measure the impact of different policies on aggregate well-being are described in Section 4. In Section 5, the theoretical tools described in Section 4 are applied to evaluate the effects of implementing the unconditional basic income on well-being, arguing that, in most cases, it aggravates inequality in well-being and, more in general, lowers its level. In Section 6, some alternative policies are discussed and a policy mix that combines the basic income with other measures is proposed. Section 7 concludes the paper.2Technological UnemploymentAs technological progress has impacted employment in different ways during the last 250 years, not only has the debate on technological unemployment crucially evolved over time but the phenomenon seems to elude any single definition. Consequently, many puzzling differences of opinion emerging in the theoretical literature can be accounted for with the fact that different scholars have discussed different phenomena as if they were one and the same.In particular, we can single out three different historical and theoretical phases in the debate on the impact of technological progress on employment, and hence three different groups of causes, consequences, and possible intervention policies.The first phase, corresponding to the first wave of technological unemployment, can be historically placed in the years of the First (roughly 1765–1830) and Second (roughly 1870–1914) Industrial Revolutions, when it appears unquestionable that technological progress, and in particular process innovation,Process innovation and product innovation are the main drivers of technological progress. Product innovation is generally considered employment-friendly, whereas the contrary applies to process innovation.reduced staffing needs for some industries.However, fierce as the debate turned out to be in those years, according to most contributors (see, e.g. Campa, 2007, 2014, 2017; Carboni, 2015; Schettkat & Yocarini, 2003; Wladawsky-Berger, 2015) the long-term empirical impact of technological progress on employment has proved modest: process innovation mainly relocated workers from one sector (to begin with, from agriculture; later from manufacturing) to another (at the outset, to manufacturing; later to services), without increasing long-term unemployment. These empirical findings are consistent, at least partially, with the great majority of theoretical contributions on the transition from agriculture to manufacturing, and from manufacturing to services, well described with Colin Clark’s “law of the three sectors” (Clark, 1940; Fisher, 1935): indeed, according to Baumol (1967), Fuchs (1968), and other scholars (for a review of the literature on the issue see Schettkat & Yocarini, 2003; D’Agostino, Serafini, & Ward, 2006), not only was the process eventually destined to generate a significant rise in employment in the service sector, but it was also to lead to a major increase in employment in general (mainly because of the low productivity growth in the service sector). Moreover, the empirical findings are also consistent with the conclusions arrived at in compensation theory, according to which technological progress may actually generate unemployment in the short run, but compensative forces would bring the system back to full employment in the long run (mainly) because of wage flexibility and increased employment deriving from increased production (of machines, in particular).For thorough discussion of compensation forces, see Vivarelli (2007 and 2014, Campa (2017), Peters (2017), Blien and Ludewig (2017), D’Orlando (2020a and 2020b), Calvino and Virgillito (2018).We may say that in this (initial) phase machines (capital) cooperate with human workers in the productive process while technological progress in the form of process innovation boosts labor and capital productivity, so that employment problems can arise only if demand does not keep up with increase in productivityBy definition, Productivity=ProductionEmployment$\text{Productivity}=\frac{\text{Production}}{\text{Employment}}$so that Employment=ProductionProductivity$\text{Employment}=\frac{\text{Production}}{\text{Productivity}}$. In the presence of an increase in productivity, the reduction in employment can be offset by an increase in demand, and hence in production, equal to (or greater than) the increase in productivity.(and compensation theory explains why this is actually not the case – see on this point Bessen, 2018; Berg et al., 2018; Blien & Ludewig, 2017; Neisser, 1942).The second phase started with the shift from manufacturing to services, in the first half of the 20th century, but gained momentum with the second wave of technological unemployment, which can be historically placed in the years of the Third Industrial Revolution (roughly 1980 up to the present), when progress in information technologies can be considered one of the main reasonsSee, e.g. Aghion, Caroli, and Garcìa-Peñalosa (1999), Card and Di Nardo (2002), Autor, Levy, and Murnane (2003, 2006, 2008, Violante (2008), Dustmann, Ludsteck, & Schönberg (2009), Acemoglu and Autor (2011), Autor and Handel (2013), Freeman (2015), Acemoglu and Restrepo (2016).(although not the only oneA number of scholars consider the reduction in public intervention in the economy and globalization as the other major factor responsible for wage polarization (see e.g. D’Orlando & Ferrante, 2022; Santarelli & Figini, 2002; Piketty, 2014).) for wage polarization and increases in inequality.This has been so since the shift from manufacturing to services, which had a huge impact on production and employment in all the highly industrialized countries, did not exactly have the same positive effects as predicted with the theoretical models described above, because it was accompanied by a shrinking of the role, as well as wages, of the blue-collar workers (see e.g. Bell, 1976). For these workers, reduced employment possibilities in manufacturing did not meet with increased employment possibilities in the service sector, which required different (higher) skills, creating a mismatch between workers’ skills and the skills demanded by the service firms that could be resolved only with lower wages. According to D’Agostino et al. (2006, p. 7) “the presence of adjustment barriers associated with the shift from manufacturing to services may have hindered the ongoing process of sectoral reallocation of the workforce.” Thus, the decline in manufacturing was accompanied by a decline in wages and income for middle-class workers, unable to maintain their standards of living.From an alternative perspective, focusing more on the role of technological progress than on the role of the shift to services, the whole process has been called skill-biased technical change,According to Violante (2008, p. 2): “‘Skill-Biased Technical Change’ (SBTC thereafter) is a shift in the production technology that favors skilled (e.g. more educated, more able, more experienced) labor by increasing its relative productivity and, therefore, its relative demand. Ceteris paribus, SBTC induces a rise in the skill premium—the ratio of skilled to unskilled wages.”emphasizing the fact that the process reduced employment possibilities and wages for unskilled workers but raised them for skilled workers. A slightly different version of this approach focuses more on tasksOn the task approach, see, e.g. Autor, Katz, and Kearney (2006); Autor (2013); and Autor and Handel (2013).than on skills and is known as task-biased (or routine-replacing) technical change. The tasks most affected by technological progress are routine tasks, so that workers whose jobs consisted mainly in such tasks suffered reduction in their employment possibilities and wages, whereas workers engaged in predominantly non-routine tasks enjoyed a corresponding increase.In this second phase, the impact of technological progress on aggregate employment was also relatively mild: during the years of the ICT revolution, employment levels rose rather than falling, sectors in expansion more than reabsorbing the unemployment generated by sectors in contraction, albeit at the cost of lower wages for some, the consequence thus turning out to be increase in inequality rather than in unemployment.However, from a theoretical viewpoint, a major new turn in this phase was that, for some tasks, machines began to replace workers, no longer serving simply to boost workers’ productivity. Thus, we have tasks (and sectors) in which capital cooperates with labor in the productive process, with technological progress increasing labor (and capital) productivity, as occurred in the preceding phase, and tasks (or even sectors) in which machines replace humans in the productive process, with technological progress increasing only (or mostly) capital productivity.On the above-described substitution process, see e.g. Brynjolfsson and McAfee (2011 and 2014, Ford (2015), West (2015), Acemoglu and Restrepo (2019).In any case, up to the end of the 20th century, the majority of theoretical contributors appeared to agree that compensation forces are effective in counteracting technological unemployment, although a number of dissenting opinions did persist.Dissenting opinions range from affirmation that technological unemployment exists but is concealed by the historical reduction in per capita working hours (i.e., Vivarelli, 1996) to the position that “economic theory does not have a clear-cut answer about the final employment effect of R&D and innovation,” such that “attention should be turned to the empirical analyses” (Piva & Vivarelli, 2017, p. 14).Radical changes began with the Fourth Industrial Revolution, corresponding to the third wave of technological unemployment, which started roughly around the beginning of the 21st century. The Third and Fourth Industrial Revolutions were based on an apparently similar form of technological progress; however, behind this apparent similarity a crucial difference was concealed. This difference lays in the new type of capital that represented the main innovative (and symbolic) characteristic of the Fourth Industrial Revolution: robots,According to Freeman (2015, p. 2), “[t]he term ‘robots’ refers broadly to any sort of machinery, from computer to artificial intelligence programs, that provides a good substitute for work currently performed by humans […] it does not matter whether a robot/machinery has a humanoid appearance, as long as it can perform human functions.”and in particular robots endowed with artificial intelligence. If robots and artificial intelligence enter the productive process, they may well eventually replace human beings in most jobs and tasks.According to Campa (2017, p. 14): “Artificial Intelligence develops exponentially and not only promises to further reduce the workforce in manufacturing, but it will begin to erode the work of specialists in the service sector. In the near future, unemployment could concern economic actors who have attended higher education institutions and invested much time and money to acquire their professional skills, such as journalists, physicians, teachers, lawyers, consultants, managers, etc.”If such a scenario were to come about, capital would no longer cooperate with human workers in enhancing labor productivity but could totally replace workers. Technological unemployment, which hit only unskilled workers or workers specialized in particular tasks in the Third Industrial Revolution, could affect all workers, skilled and unskilled, regardless of whether they are performing routine tasks or not, in most sectors of the economy.The theoretical approach that can be used to systematize this third wave of technological unemployment is, inevitably, a development of the approach used for the second wave, but only for the part in which robots compete with humans for employment: among the productive factors, we now have robots that, like capital, are produced but that, unlike capital, can be produced by using robots alone.To avoid formal complications that are not particularly relevant to an analysis focused on study of the basic income, we may assume that robots are humanoid programmable machines that can replace humans in most activities, and not merely in manual and/or routine activities. In this scenario, these humanoid machines can carry out tasks identical to those performed by humans. The scenario can be further simplified by assuming that capitalists own robots and rent them out to entrepreneurs. Meanwhile, entrepreneurs have to decide whether to hire robots or humans for their productive processes. At the microeconomic level, the choice between hiring labor (workers) or capital (robots) represents a choice between techniques: given the quantities produced, entrepreneurs will substitute robots for human workers if the value of the marginal product of robots, divided by the cost of the service provided by robots (i.e. rents paid to robots’ owners), is greater than the value of the marginal product of labor, divided by the wage rate. Given the quantity to be produced, entrepreneurs will substitute robots for human workers if:(1)p‾∂Q∂Rw‾R>p‾∂Q∂Lw‾L.$$\frac{\overline{p}\frac{\partial \text{Q}}{\partial \text{R}}}{{\overline{\text{w}}}_{\text{R}}} > \frac{\overline{p}\frac{\partial \text{Q}}{\partial L}}{{\overline{\text{w}}}_{\text{L}}}\text{.}$$where w‾L${\overline{\text{w}}}_{\text{L}}$is workers’ wage, w‾R${\overline{\text{w}}}_{\text{R}}$is the price of the service provided by robots, ∂Q∂L$\frac{\partial \text{Q}}{\partial \text{L}}$is the marginal productivity of labor, ∂Q∂R$\frac{\partial \text{Q}}{\partial \text{R}}$is the marginal productivity of robots, and p‾$\overline{\text{p}}$is the price of the commodity produced.The price of the commodity produced, the wage rate, and the price of the service of capital are considered as given under the (here implicitly assumed) hypothesis of perfect competition.The problem of unemployment arises in a scenario where robots can replace human workers in most productive processes (or tasks) and show in those productive processes a productivity–wage ratio higher than that of human workers for all tasks and wage levels above subsistence. These circumstances imply that compensation theory cannot hold: if commodities can be produced by means of robots alone, no increase in demand can raise employment, and wage flexibility is likely to meet the subsistence bound before human workers’ productivity–wage ratio becomes as high as that of robots. As a result, high levels of unemployment are not only possible but, indeed, quite probable.It is certainly true that the resulting unemployment can be considered voluntary if viewed in the light of traditional (i.e. neoclassical) labor market microeconomics: workers voluntarily refuse to work, the wage they would earn being below subsistence and hence, inevitably, below their reservation wage; and firms will not pay higher wages if hiring robots proves a preferable option. To put it another way, according to orthodox economics the labor market is perfectly functioning and capable of arriving at an equilibrium where involuntary unemployment does not exist. The problem is that defining people who refuse to work for a wage rate below subsistence as voluntarily unemployed may be theoretically correct but in practice highly misleading, if not meaningless. As we will see a little later on, the basic income can impact exactly on this point, by reconciling theory with reality, since it would make it acceptable for humans to work for a wage rate below subsistence (subsistence being guaranteed by the basic income) so that unemployment, as well employment, would actually be voluntary.In any case, the spread of technological unemployment would inevitably take time, would not necessarily be complete and would probably affect different countries in different ways. In particular, it would affect the developed and less developed countries in different ways mainly due to technological reshoring, i.e., the revision of previous offshoring choices of multinational firms with production relocated to the home country (Ocicka, 2016, p. 105).On reshoring, or back shoring, in general see e.g. Gray, Skowronski, Esenduran, and Rungtusanatham (2013); Fratocchi, Di Mauro, Barbieri, Nassimbeni, and Zanoni (2014); Ancarani, Di Mauro, Fratocchi, and Orzes (2015).Before robotization, multinational firms located labor-intensive production phases in the less developed countries, given their low labor costs. And in these countries local firms could also be competitive in labor-intensive productions on the international markets for the same reason. In a Smithian perspective, we can see low labor costs as an absolute advantage for the less developed countries. Robotization makes using robots in the developed countries less expensive (and/or more productive) than using unskilled/routinary workers in the less developed countries, destroying the absolute advantage of the latter. So it is that multinational firms re-shore, i.e. close plants and fire workers in less developed countries and bring their productions back home, while labor-intensive local firms lose their competitiveness on the international markets and are likely to go bankrupt, stopping production and firing workers. Production in the less developed countries falls, and with it employment, even if they do not robotize. Unemployment would therefore hit above all the countries where the phases of production highly intensive in unskilled labor were located. As a result, the diffusion of robots in the developed countries could also generate high levels of unemployment in the less developed countries.However, at the outset the process would probably hit the countries specialized in productions highly intensive in unskilled labor but, as artificial intelligence and robotization spread, it would later also hit countries specialized in production highly intensive in labor, regardless of whether it is skilled or unskilled, or engaged in routine or non-routine tasks, i.e. also the developed countries. Unemployment would thus spread through countries irrespective of their factor endowments and specializations.If the labor market is segmented, whether in a skill or task perspective, we can expect the technological substitution of workers with robots to come about through a succession of steps. In a skill-biased technical change scenario, the process would first affect unskilled workers, and skilled workers only later. In a task-biased (or routine-replacing) technical change scenario, the substitution would first affect workers performing routine tasks, and later workers performing non-routine tasks. In any case, we would have a transition taking a long time and generating successive waves of unemployment for different categories of workers at different times, as innovations were developed in specific sectors or for specific tasks. This dynamic process could generate successive waves of inequality among workers whose jobs required different skills, workers whose jobs required routine or non-routine tasks, and workers living in different countries or regional areas (specialized in different jobs/tasks).In this situation, it may be difficult for humans to hold enough bargaining power to obtain an adequate share of profits and, furthermore, for the democratic institutions to survive, for if humans are unnecessary for production their bargaining power risks falling to zero. But since it is also in the interest of the firms owners to have adequate demand for the goods they produce, inevitably it would be up to the government to devise the appropriate intervention policies.3Policies against Technological Unemployment, and the Basic IncomeThe most popular of the policies capable of counteracting technological unemployment are taxing the hiring of robots, subsidizing the hiring of humans, subsidizing skill accumulation on the part of human workers (so as to enhance their productivity–wage ratio), reducing per capita working hours, nationalizing firms, assigning the ownership of robots (or firms) to citizens, encouraging workers to buy shares in firms, or providing an unconditional basic income to all citizens, regardless of whether they are employed or not.For thorough analysis of all these policy intervention mechanisms, see D’Orlando (2020a, pp. 9–10), D’Orlando (2020b, pp. 607–612) and Marchant and Stevens (2017).In the present article, the focus is on the last of these, which has in fact been referred to in many different ways: unconditional basic income, universal basic income, basic guaranteed income, basic income guarantee, basic income grant, guaranteed national income, etc. (see, e.g. Marchant & Stevens, 2017; Pech, 2010; Tanner, 2015).The unconditional basic income considered in the present article is “a cash grant provided to every citizen […] without any other eligibility requirement” (Tanner, 2015, p. 3) and specifically aiming to counteract technological unemployment. In particular, in view of the prospect that technological progress could generate high levels of unemployment that could not easily be reabsorbed, but unlike most other contributions, here the basic income is assumed to be sufficiently high to guarantee subsistence.In economic literature there are few studies on the consequences of provision of an unconditional basic income (see, e.g. Bowles, 1992; Gamel et al., 2006; Marchant, Stevens, & Hennessy, 2014; Pech, 2010; Tanner, 2015; Van der Linden, 1997, 2002).The debate on the basic income dates back to Paine (1797) and in the course of time has seen the participation of various social scientists, from economic historians (e.g. Polanyi, 1944) to economists (e.g. Friedman, 1962; Hayek, 1987; Tobin, 1966) and political philosophers (e.g. Van Parjis 2004).Furthermore, although the basic income has recently been at the center of public debate on a possible remedy for technological unemployment,On this point, see, e.g. Zheng, Guerriero, Lopez, and Haverman (2017, p. 13) and Martinelli (2017, pp. 27–29).even fewer studies consider it as such,See, e.g. Ford, 2009, chapter 4; Ford, 2015, chapter 10; Brynjolfsson & McAfee, 2014, chapter 14.in most cases proposing it simply as a remedy for poverty or inequality in a low unemployment scenario, i.e., as a possible substitute for the existing social insurance and benefits.Of the positive consequences of introducing an unconditional basic income, some scholars (see, e.g. Haagh, 2011a, 2019a; Standing, 2003, 2005) stress that it guarantees a lifelong, unconditional income support, ensuring greater income security with positive impact on the general well-being. Furthermore, provision an unconditional basic income would remedy the poor performance of the traditional welfare instruments (Jordan, 1992, pp. 170 ff.), which entailed social exclusion and structural unemployment, and their contemporary versions based on punitive approaches aimed at promoting labor supply (Haagh, 2019b). Finally, the basic income “would be simpler and more transparent than the current welfare bureaucracy… would reduce paternalism and government involvement in the lives of poor people… would more effectively alleviate poverty… could provide better incentives—or at least fewer disincentives—for work” (Tanner, 2015, p. 7; see also Widerquist, 2017). Other scholars come to the opposite conclusion, arguing that the basic income “fails to provide an incentive to find work or engage in other meaningful and beneficial activities […], provides an automatic government handout to individuals while reducing their esteem in their own eyes and those of their neighbors and contacts [and] would be very expensive” (Marchant & Stevens, 2017, p. 1). Studies based on income and substitution effects were developed but proved inconclusive regarding the impact of a basic income on labor supply, the wage equilibrium, and work effort (see, e.g. Martinelli, 2017, p. 51 ff). However, the idea that the basic income reduces labor supply and work effort, inducing people to remain idle, is fairly widespread, at least among economists. Moving on from the solidly traditional to the less traditional economic approaches, we again find different opinions on the impact of the basic income on labor supply and work effort. Some scholars look to behavioral economics and evolutionary game theory to conclude that when altruistic behavior and behaviors based on habits or heuristics replace self-interest and maximizing behavior, “individuals may not shirk when given income guarantees” (Widerquist, Lewis, & Pressman, 2005, p. 590). Nevertheless, insights from the debate on the role of intrinsic motivations suggest that when subsistence is guaranteed, it might be difficult to motivate people to work (and/or to put real effort into their work), at least for certain kinds of jobs, if the incentive is merely monetary (or extrinsic).On the basic income and intrinsic motivations see also Haagh, 2011b.Furthermore, “for those tasks in which a person demonstrates a high intrinsic motivation to perform, the introduction of an extrinsic incentive (in the form of a monetary reward or a fine) undermines her intrinsic motivation, which may cause her to decrease the level of effort” (Pech, 2010, p. 8).Notably, some authors, like Jordan (1992), consider the basic income from the point of view of the community, as a policy tool that can guarantee individuals the possibility “of caring or participating in other socially meaningful ways in society, or contributing to gender equality, social cohesion and social trust… [so that the basic income] has also been advocated from a liberal perspective in setting all people free to make their own decisions in life” (Roosma & van Oorschot, 2019).In the presence of technological unemployment, broadly speaking we may begin with the observation that provision of an unconditional basic income can essentially be targeted to two different goals, so that two different and alternative scenarios are possible.First scenario: with the guarantee of an unconditional basic income the public authorities could tolerate high levels of unemployment or, in an extreme and (as we will see below) mainly theoretical case, even full unemployment. Consider, for the sake of simplicity (and an understanding of Figure 1 below) this latter case: all humans are unemployed, and production is entirely accomplished by firms using robots alone, while the survival of humans is guaranteed by provision of the basic income, which must necessarily at least meet the subsistence level.Figure 1:Only robots are employed.Second scenario: since one of the problems that prevent workers from being competitive with robots (and hence from being hired by firms) is the existence of a lower subsistence bound below which human wages cannot fall, with the guarantee of an unconditional basic income for all citizens workers could accept wages below the subsistence level. In this case, the basic income would not necessarily be above the subsistence level, since it is the sum of the basic income and the wage that guarantees the survival of humans.Notably, the basic income is not an unemployment benefit, as humans receive it whether they are employed or not.However, in the present article, as pointed out above, we limit discussion to the case of a basic income guaranteeing subsistence, as indeed it should in a context where high levels of technological unemployment are hard to reabsorb and the unemployed cannot otherwise survive. Therefore, in this scenario, once the subsistence basket is obtained thanks to the basic income, humans can decide whether to agree to work on a low wage to raise their consumption above the subsistence level or to stay voluntarily unemployed and make do with subsistence.Referring to a scenario in which only robots are employed (i.e., the first scenario), the schematic functioning of the system as a whole is depicted in Figure 1: production is accomplished by robots alone; robots are privately owned by capitalists’ families; capitalists rent robots out to firms; firms produce and sell goods to households, pay taxes to the Government and rents (∑w‾R)$\sum {\overline{w}}_{R})$to the capitalists’ families; firms are managed by entrepreneurs who organize production; all humans are unemployed and their families receive the basic income; entrepreneurs’ families receive the basic income plus profits; capitalists’ families receive the basic income plus rents coming from robots renting; if we assume a balanced government budget, the basic income is financed by taxes paid by firmsFor the sake of simplicity let us assume that capitalists do not pay taxes (nothing would change except understanding of the figure in the contrary case).(firms’ taxes are equal to the basic income distributed to all citizens, so that ultimately firms pay the basic income).More in general, alternatives may include: a corporate tax; a wealth tax; a tax on owners of robots as distinguished from other forms of wealth; and a VAT (which, in combination with a UBI, can be designed to have progressive incidence); otherwise, we can imagine deficit financing of the basic income.Referring to the scenario in which also humans are employed (i.e. the second scenario), Condition 1 still depicts the firms’ microeconomic behavior. However, the aggregate outcome is more complex; it is depicted in Figure 2 below. The firms’ surplus is distributed to: i) entrepreneurs (profits); ii) robot owners (rents ∑w‾R)$\sum {\overline{w}}_{R})$); iii) workers (wages); iv) the government (taxes). Now, since the taxes paid by firms are equal to the basic income paid to humans, firms as a whole will pay workers as a whole an amount equal to the sum of the taxes they pay and the wages they pay.Figure 2:Also humans are employed.Therefore, for the representative firm, it is in its interest to hire robots rather than humans if(2)∑w‾L+basic incomep‾∂Q∂L>w‾R+basic incomep‾∂Q∂R$$\sum \frac{{\overline{\text{w}}}_{\text{L}}+\text{basic\ income}}{\overline{\text{p}}\frac{\partial \text{Q}}{\partial \text{L}}} > \frac{{\overline{\text{w}}}_{\text{R}}+\text{basic\ income}}{\overline{\text{p}}\frac{\partial \text{Q}}{\partial \text{R}}}$$It is worth noting that nothing changes for the single firms: as Condition 1) still holds, the single firms will not hire workers if their wage–productivity ratio is above that of robots. But there is a crucial difference if the government provides an unconditional basic income: the lower subsistence bound for wages would no longer exist, since the basic income guarantees subsistence for the workers. Consequently, the workers’ reserve wage could easily fall, at least for a part of the workforce, to a level low enough to guarantee a human wage–productivity ratio below that of robots, thus making it advantageous for firms to hire workers. Provision of a universal basic income would make the first scenario inevitably collapse into the second, eliminating the possibility of full unemployment and of a world in which production is accomplished by robots alone, even if the final employed/unemployed ratio inevitably depends upon the extent to which the wage decreases.Figure 2 shows the simplified schematic functioning of the resulting scenario for the economy as a whole: production is accomplished by robots and humans; a certain number of humans are employed and their families receive the basic income plus a wage; entrepreneurs’ and the robot owners’ families receive the basic income plus profits/rents; the basic income is financed by taxes paid by firms.Besides the way it functions, a thorough analysis of the basic income necessarily implies evaluation of its impact on the general well-being.4Measuring the Impact of Policies on Well-being: from Keynes to Hedonic Adaptation, Envy, and EscalationActually, behavioral and happiness economics offer an ample range of theoretical tools that economists can use to measure the impact of different policies on well-being. This also applies to the number of behavioral principles that could be used to address the impact of introducing an unconditional basic income on well-being. However, limiting analysis to the major principles, or at least those that appear the most relevant, and inevitably making some subjective choices, the focus here is limited to a subset of them. In particular, for our specific topic (i.e., policies to counteract the impact of technological unemployment), it can be useful to depart slightly from behavioral economics “stricto sensu” and also consider Keynes’s 1930 contribution. But to begin with we must briefly consider what we mean by well-being and what are its causes.4.1Well-being and its CausesSubjective well-being, also often referred to by economists as “happiness”, is a multidimensional concept that encompasses a number of possible meanings, and is much more complex than the old (and today almost abandoned) concept of utility. According to Nikolova and Graham (2020, pp. 4–6, but see also Bentea, 2019, pp. 258–261):Subjective well-being has three separate but related dimensions – affective (hedonic), evaluative, and eudaimonic […]. Affective subjective well-being refers to temporary experiences of emotions – positive ones such as happiness or joy, or negative ones, such as stress, anger, and sadness. Such positive and negative feelings and emotions are usually short-term and are influenced by the immediate circumstances and states of being and doing. […] Evaluative well-being, in contrast, is a judgment about one’s overall life circumstances and requires reflecting on life as a whole. Evaluative well-being also relates to specific assessments of life domains, such as work, family, housing, income, and the living standard. […]. Eudaimonic well-being is broader than having meaning and purpose in life and refers to the process of living well and having aspects, such as competence, autonomy, personal growth, and relatedness.Although the analysis proposed here refers mainly to affective subjective well-being, also often defined simply as “pleasure attainment and pain avoidance” (Bentea, 2019, p. 258),According to Bentea, 2019, p. 261, “[h]appiness appears when positive emotions… are frequent and intense while negative emotions are minimal and rarely experienced”.it inevitably also takes into account elements belonging to the two other dimensions of happiness (i.e. evaluative well-being and eudaimonic well-being) which are related to affective well-being.Accordingly, and according to some of the economists who focus their theoretical interest on these themes (see, e.g. Pugno, 2017, 2019, but also Scitovsky, 1992), one of the key determinants of well-being is change: “the person exercises her cognitive, social and emotional skills, with the purpose to search for the activities and challenges that give the best opportunities to develop her talents and aptitudes” (Pugno, 2019, p. 168), so that one can “describe human development by going beyond the alleviation of people’s deprivations to take account also of the appreciation and exploration of novel experiences that challenge and develop their inner skills…” (Pugno, 2017, p. 1178). In particular, happiness is considered as linked to the choice of activities that are characterized “by novelty, i.e. by the pleasure of learning” (Pugno, 2017, p. 1178). Within such a perspective novelty, different new behaviors, creative goods in Scitovsky’s idea, and more in general change are the main elements that can generate happiness.To make change possible, and to extract well-being from it, it is essential to have the right skills, since “people can increase the amount of happiness in their pleasant life when they have the skills to grow and multiply their pleasures” (Bentea, 2019, p. 261), and often (even if not always) a growing income is also needed (see again Pugno, 2019, p. 168). In these respects, the role played by the government is crucial, both for supplying citizens with (and/or financing acquisition of) an adequate skill endowment through formal and informal education, and, as we will see in detail a little later on, for providing them with an adequate income (possibly rising over time).It is worth noting that changing is not easy: a number of antagonistic forces, which we can summarize in the psychological principle of “preference for the status quo”, can hamper change. Scitovsky (1992), albeit in a different context and with rather different implications, discussed these forces when distinguishing between the pursuit of “comfort” and “novelty” activities, and considered the pursuit of comfort as a less effective way of achieving well-being than novelty activities and also as an obstacle to these better practices (see, e.g. Pugno, 2017, p. 1188).4.2Keynes’s Short 1930 EssayNow, returning to technological unemployment, it is worth recalling that Keynes discussed some of the well-being problems that could derive from technological unemployment and the possible remedies in his short 1930 essay “Economics Possibilities for Our Grandchildren” (Keynes, 1930). He described the hypothetical situation that might come about through continuous technological progress, 100 years later, in 2030. According to Keynes, “assuming no important wars and no important increase in population, [the struggle for subsistence] may be solved, or be at least within sight of solution, within a hundred years” (Keynes, 1930, p. 4), when fewer workers will be needed to guarantee an improving standard of living for the whole population. According to Keynes, the reduction in labor demand would not be a problem, since the resulting technological unemployment could be counteracted, or at least postponed, thanks to a progressive reduction in per-capita working hours. However, he explicitly admits that some psychological issues might nonetheless arise if humans are deprived of the need to work to receive subsistence. According to Keynes, there are two main issues.First, a possible “nervous breakdown of the sort which is already common enough in England and the United States amongst the wives of the well-to-do classes […] who have been deprived by their wealth of their traditional tasks and occupation” (Keynes, 1930, p. 4). Second, the impossibility of pursuing relative needs, i.e., those needs “that are relative in the sense that we feel them only if their satisfaction lifts us above, makes us feel superior to, our fellows” (Keynes, 1930, p. 4). These needs arise from our desire for superiority and can be satisfied by consuming the goods now commonly known as positional goods.In his essay, albeit in a rather preliminary form, Keynes anticipated some of the behavioral economics constructs that can be used to discuss the consequences of technological unemployment (and, at least implicitly, of the provision of an unconditional basic income) on well-being.4.3Loss AversionThe first behavioral economics construct that can be used to address the issue is loss aversion.See, e.g. Tversky and Kahneman (1991); Kahneman et al. (1991); D’Orlando and Ferrante (2009 and 2022; and D’Orlando, Ferrante, and Ruiu (2011).According to Kahneman, Knetsch, and Thaler (1991, p. 199), “[a] central conclusion of the study of risky choice has been that […] changes that make things worse (losses) loom larger than improvement or gains.” The idea that bad events bear more weight than good ones is confirmed by a number of contributions on similar behavioral principles (see, e.g. Knetsch, 1989; Knetsch & Sinden, 1984; Kahneman, Knetsch, & Thaler, 1990; Samuelson & Zeckhauser, 1988). These psychological principles hold important consequences for calculation of the true costs of unemployment, suggesting that unemployment episodes are events whose negative impact on well-being is not even fully offset by positive events such as rehiring. Although this does not mean that the psychological loss deriving from unemployment cannot be offset by an unemployment benefit, it does however imply that the benefit would have to outweigh the income loss net of the disutility of work, the gross damage being the sum of both the monetary and psychological costs.4.4Hedonic AdaptationHedonic adaptation is another behavioral economics construct essential for the purposes of the present article and fully compatible with loss aversion.See, e.g. Brickman and Campbell (1971), Brickman, Coates, and Janoff-Bulman (1978), Clark and Oswald (1994), Clark (1999), Diener, Suh, Lucas, and Smith (1999), Frederick and Loewenstein (1999), Di Tella, MacCulloch, and Oswald (2003), Clark et al. (2004), Stutzer (2004), Layard (2005), Oswald and Powdthavee (2006), and D’Orlando and Ferrante (2008, 2009 and 2022.Hedonic adaptation is founded on the empirical finding that people adapt to life events: “[l]ife events such as marriage, loss of a job, and serious injury may deflect a person above or below [her/his] setpoint, but in time hedonic adaptation will return an individual to the initial setpoint” (Easterlin, 2003, p. 1). Similarly, consuming a good or undertaking a consumption behavior increases well-being in the first phase, but later the consumer suffers habituation and her/his well-being returns to the baseline level. It is still debated whether adaptation is complete or incomplete, i.e., whether life shocks have a permanent effect on subjects’ well-being in the long run: some authors maintain that they can approach but never return to their baseline, so that an irreversible loss would persist for negative life events and an irreversible gain for positive life events.For discussion of complete and incomplete adaptation and the setpoint hypothesis, see Easterlin (2003) and Lucas, Clark, Diener, and Georgellis (2003).In particular, episodes of unemployment are considered among the most drastic cases of irreversible loss (see, e.g. Clark & Oswald, 1994; Clark et al., 2004; Frey & Stutzer, 2002), with well-being remaining below the baseline, even if the subject, after being fired, obtains a new job at the same wage as the old one (D’Orlando & Ferrante, 2018, p. 1264) – a finding perfectly consistent with the implications of loss aversion discussed above.4.5Positional Effects and EnvyWhen discussing the possibility that social position or, more in general, social interaction, plays a role in determining subjective well-being, theoretical interest inevitably focuses on envy.Reviews of the economics literature on envy can be found in Chaudhuri (1985), Hammond (1989), Mui (1995), and Kolm (1995).According to Parrott and Smith (1993, p. 906), envy “occurs when a person lacks another’s superior quality, achievement, or possession and either desires it or wishes that the other lacked it.” According to D’Orlando and Ricciotti (2021, p. 121), “[i]n these cases, subjects’ well-being depends upon their position in a somewhat defined economic hierarchy, and hence on others’ behaviors, income or wealth.” It is worth emphasizing that envy is not solely linked to material possessions, since people can envy others not only for what they have, but also for what they are: according to Goel and Thakor (2005, p. 2263) “we may envy another person’s success, happiness, intelligence, health, good looks, material possessions, power, title, job, or status”. Furthermore, theoretical contributions (see again Goel & Thakor, 2005, p. 2257–2258) often strongly maintain that envy is not inequality aversion, since envious people do not behave symmetrically: subjects experience a reduction in well-being when they are worse off than others, but they experience an increase in well-being when they are better off than others. Say it another way, they do not dislike inequality in general, but dislike being the losers in the inequality game.Envy is generally considered a negative emotion; in particular, Frank (2005) has discussed the negative externalities caused by this and similar positional effects, with conspicuous consumption playing a major role. Indeed, we can attribute to envy at least four possible negative consequences: (i) negative externalities that generate well-being losses for envious people (i.e. for people suffering from passive envy); (ii) the risk that envious people set about trying to make the others lose their possessions (disruptive or malicious or black envy); (iii) the possibility that envious people, by imitating the behaviors of their reference group, replicate and spread socially harmful behaviors through society (e.g. conspicuous consumption); (iv) the possibility that failing in imitating the consumption behaviors of the reference group negatively affects subjects’ self-esteem.However, some authors attribute a crucially positive role to the desire to be envied: for example, Veblen (1899 1970, p. 32) considered the “desire to excel in pecuniary standing and so gain the esteem and envy of one’s fellowmen” one of the main “incentives to acquisition and accumulation.” Furthermore, envy can also be seen as a broader concept that encompasses positive emotions such as admiration or esteem for people who accomplish socially valuable actions (such as environmental protection behavior, altruistic behavior, excellence in education or sports).In general, we can attribute to this broader concept of envy at least six possible positive consequences: (i) gains in well-being for envied people (i.e. for those who enjoy active envy); (ii) the possibility that envy will encourage subjects to strive to reach the others’ positions (constructive or benign or white envy); (iii) the possibility that envious people, by imitating the behaviors of their reference group, replicate and spread socially valuable behaviors through society (if, of course, the behaviors of the reference group are socially valuable); (iv) the possibility that succeeding in imitating the consumption behaviors of the reference group may positively affect subjects’ self-esteem; (v) the possibility that envy may stimulate change in consumption behaviors and so raise well-being (consistently with the observations on the main determinants of happiness proposed in Section 4.1 above); (vi) finally, the possibility that reaching the others’ position, or generating envy in others, may slow down, stop, or even reverse the habituation process.The last point in particular has been discussed by D’Orlando and Ricciotti (2019, p. 19): “until the time when others succeed [in imitating a] subject’s higher-level consumption behavior, the subject will continue undertaking a consumption behavior at a higher-than-average level, i.e., continue generating envy in others. This circumstance increases well-being by preventing (at least partially) the reduction caused by hedonic adaptation: my car is still the fastest on the highway, I pass all other cars and I am always, every day, happier and happier by generating envy in their drivers who unsuccessfully try to resist.”It is worth recalling that what improves a subject’s well-being is not envying others, i.e., suffering passive envy, which actually reduces well-being, but the possibility of lowering or eliminating passive envy by succeeding in imitating the reference group’s behaviors; and/or the capacity of increasing well-being by succeeding in changing behaviors to generate active envy, i.e., succeeding in being envied (esteemed, admired) by others.4.6EscalationFinally, a crucially useful behavioral economics construct is escalation. According to D’Orlando and Ricciotti (2021, p. 109):“[e]scalation occurs when over time people show a growing interest in stronger, harder, faster, riskier goods, substances, or behaviors. The key concept here is ‘more intense sensations’. Escalation is not simply a desire for novelty, that is, a desire for consuming a different good or engaging in different behavior. Instead, it requires a shift from goods, substances, and behaviors that give the consumer ‘less intense sensations’ to those which provide ‘more intense sensations’. People begin engaging in basic consumption behavior within a certain consumption type and are satisfied with this behavior. [But] after a while, they become less satisfied with it, which, as time passes, leaves them with less well-being. Thus, they become more interested in other consumption behaviors, which are stronger, harder, riskier, and faster.”Examples of escalating behaviors can be observed in drug consumers who begin with soft drugs but later escalate to harder ones; in wine consumers who progressively escalate to better wines; in skiers who initially prefer green slopes but, on becoming more expert, choose more difficult black slopes; and in swinging couples who start with exhibitionism and later escalate to intercourse with others (see, e.g. D’Orlando & Ricciotti, 2021, p. 109). It is worth noting that in all these cases, and more in general in the entire analysis presented in this article, we will refer to the most general definition of consumption to be found in theoretical literature, i.e. “the satisfaction of our wants by using goods, services, and even time” (D’Orlando & Ricciotti, 2021, p. 109), thus separating consumption from monetary expenditure.Escalation is founded on two different constructs: the reduction over time of the well-being derived from currently undertaken consumption behaviors, a phenomenon consistent with hedonic adaptation; and the increase over time in well-being derived from alternative consumption behaviors, a phenomenon consistent with the idea that change raises well-being. However, escalation is not simply pursuit of novelty since, in the case of escalation, the alternative consumption behaviors capable of increasing well-being are “stronger, harder, riskier, faster,” i.e., more intense, not simply different. And to be undertaken and enjoyed these “more intense” consumption behaviors often call for an adequate endowment of skills, and thus processes of skill accumulation. These processes are usually informal, or can rely on learning by doing, but in some cases formal education is necessary, at least to provide the subjects with the basic skills on which skill accumulation can be built. And if formal education is necessary, public intervention may have a role, at least as a supplier of basic skills.Although escalating to superior consumption activities or behaviors is not necessarily a consequence of increasing income (as in the case of escalating from skiing on green slopes to skiing on red slopes), sometimes a higher income is needed (as in the case of escalating from driving a slow car to driving a sports car) because more intense behaviors may necessitate a higher income.Actually, escalation can be hampered by the action of some counteracting forces, mainly relating to the behavioral principles of preference for the status quo and endowment effect, and most often these counteracting forces affect the less skilled and less wealthy subjects (on these points, see D’Orlando & Ferrante, 2018, 2022). There are, however, various other counteracting elements: more intense acts of consumption may be associated with higher monetary costs; there may be higher psychological costs associated with the same activities (due to social stigma, preference for the status quo, conformism, risk aversion, aversion to ostentation, frugality, etc.); subjects may expect to obtain low levels of increase in well-being through more intense consumption due to the difficulty of acquiring the skills necessary for making changes; etc. (for a detailed analysis of these forces, see D’Orlando & Ricciotti, 2021). Generally, these counteracting forces grow stronger as consumption behaviors become more intense.Ultimately, the actual realization of escalation depends on the relative strength of the forces driving change versus the forces hampering it.5Behavioral Economics and the Basic IncomePech (2010) is one of the few scholars who use behavioral economics to discuss the consequences of providing a basic income. In particular, Pech focuses on four major points: the preferable way to finance the basic income (i.e., the level of progressivity in taxation); the impact of the basic income on the wages of “good” and “bad” jobs; the impact of the basic income on cooperation and reciprocity; and the impact of the basic income on conspicuous consumption. However, Pech does not use behavioral economics to conduct analysis of the impact of the basic income on well-being, nor does he refer to the basic income as a mean to counteract technological unemployment. Unlike Pech’s contribution, in the present article constructs of behavioral and happiness economics (particularly, those described in the preceding section) are used to discuss the impact of the basic income on well-being in a scenario of high levels of (technological) unemployment.As discussed above, provision of an unconditional basic income high enough to ensure subsistence could play two different and (seemingly) alternative roles in counteracting technological unemployment: income integration for workers to keep their incomes above subsistence level, even if their wages are below it; or the only income for unemployed humans if robots alone are producing goods. Consequently, the policy purpose of the basic income can be either to increase humans’ productivity–wage ratio, making workers competitive with robots, so that it is in the firms’ interest to hire them, or guarantee subsistence for humans once robots produce (almost) all the goods. However, in Section 3, we showed that such a distinction is more philosophical than practical, since, if the government provides an unconditional basic income to all citizens, the citizens can in any case decide whether to work or to shirk, regardless of the original purpose of the policy. In the general case, a certain number of workers will therefore choose to remain employed. For this reason, here a scenario is depicted in which the government provides an unconditional basic income to all citizens disregarding the government’s reasons for doing so, and this scenario inevitably coincides with scenario 2 (and Figure 2) of Section 3.The above scenario will result in four categories of human beings: (i) those who receive the basic income and voluntarily choose to stay unemployed; (ii) those who receive the basic income and voluntarily choose to work on a low wage; (iii) entrepreneurs who manage firms; and (iv) capitalists who own robots.The supply of an unconditional basic income affects the well-being of these categories of subjects in different ways.To discuss these ways, let us begin by summarizing the circumstances that, in the light of the behavioral and happiness economics constructs described in Section 3, increase or reduce people’s well-being, and then go on to discuss whether provision of a basic income affects these circumstances.Let us start with hedonic adaptation. According to this approach, subjects’ well-being tends to return to the baseline after a (positive or negative) shock, but in the case of unemployment episodes, as discussed in Section 4.4 above, there is a general consensus that habituation may prove incomplete: each unemployment episode generates an irreversible loss that tends to cumulate – even if the irreversible loss is progressively smaller with each episode due to long-run hedonic adaptation. Furthermore, workers remain below the set-point even if they are re-hired. Therefore, to remain steadily above the baseline, subjects have not only to avoid negative shocks, particularly unemployment episodes, but also continuously to enjoy positive shocks to compensate for habituation. These positive shocks are mainly associated with the pursuit of novelty in consumption, i.e., changes in consumption behaviors that can halt and reverse the habituation process.In most circumstances, a different and better strategy (a strategy that guarantees slower habituation and greater increase in well-being) consistent also with the escalation approach is about not simply continuously changing consumption behaviors but escalating to higher-grade consumption behaviors. These higher-grade consumption behaviors do not necessarily call for greater expense, but, in most cases, a higher income makes thing easier. Furthermore, obtaining a higher income represents, in itself, an escalating behavior.However, neither changing to different consumption behaviors nor escalating to higher grade consumption behaviors are viable strategies for all subjects and in all circumstances: in most cases changing and escalating require an adequate skill endowment, and/or a process of skill accumulation, and/or adequate education, to make the new choices practicable and capable of positively impacting on subjective well-being, even in the cases in which a rising income is unnecessary. For low-skilled and low-income workers the forces hampering change can be hard to be overcome, so that preference for the status quo may lead them to pursuit of income security and stability in their consumption patterns – behavior that prevents them from enhancing their well-being.Extending analysis to social interaction, a crucial role is also played by positional externalities. In Section 4.5, we discussed in general some of the possible negative and positive consequences of envy. Here, we focus on the role that envy may play in driving change in consumption behaviors in general, and change towards higher-grade consumption behaviors in particular.It is true that in some cases envy generates negative social externalities that lower well-being. However – and even over and above the possible role of envy in the diffusion of socially valuable behaviors (see again Section 4.5 above) – given the loss in well-being, the circumstance that many subjects succeed in escalating to superior consumption behaviors, reducing/eliminating passive envy, enjoying active envy, raising their self-esteem, etc., can reduce the loss and, in some cases, might even fully compensate for it (and also for hedonic adaptation). Indeed, generally speaking, we have no reason to assume that the social loss deriving from negative envy externalities must inevitably be greater than the subjects’ gain. Furthermore, any policy aiming to eliminate envy, for example by endowing all subjects with the same income and consumption behaviors, would eliminate passive (and active …) envy, but not hedonic adaptation, with the consequence that subjects would not be able to remain steadily above the baseline. Thus a situation in which nobody can escalate is worse than one in which everybody can escalate (and possibly worse than one in which many can escalate).Here we must stress the role played by social interaction in the above-described dynamics and in particular in driving change. Since well-being increases (or at least does not return to the baseline) when subjects change their consumption behaviors, and the more often they change, the longer they stay above the baseline, positional concerns can enhance well-being by driving change. By dint of imitating others’ behaviors, suffering passive envy and observing the behaviors of the reference group, subjects are forced to change their behaviors. Social interaction might therefore be seen as one of the possible means that can be used to stimulate change and so to increase well-being (or reduce well-being losses): whether or not the change can fully compensate for the negative (external) social consequences of envy, changing at least reduces these negative consequences for the subject who changes, enhancing her/his situation; by contrast, if the subject does not change, she/he will endure the full impact of passive envy without being able to attenuate or compensate for it. These considerations may also have consequences for one of the main limits of the basic income, namely, its asserted negative impact on the incentive to find work and on work effort: positional concerns may generate the necessary non-monetary incentives to work when monetary incentives fail.We can now go on to discuss the way in which the behavioral economics constructs and dynamics discussed above affect the impact of the basic income on well-being and, conversely, whether better remedies exist. Note that while it is certainly true that receiving the basic income is better than receiving nothing if people are faced with technological unemployment, to evaluate whether the basic income is in fact the best policy to counteract technological unemployment we must look into all its drawbacks.Let us assume that, as soon as technological unemployment arises, the government begins to provide all citizens with an unconditional basic income, financing it by taxing firms. As discussed earlier, we will have: people who have lost their jobs remaining unemployed, and receiving the basic income; those who remain employed and continue working, receiving their wages plus the basic income; entrepreneurs, who manage firms and gain profits as well as receiving the basic income; capitalists, who own robots and receive the rents plus the basic income. In this situation, when workers lose their jobs they suffer a shock that drives them below the baseline level of well-being, and we know that unemployment is a kind of shock that may leave an irreversible long-term loss in well-being. So they will remain steadily below the baseline. And there is no telling whether the basic income will be high enough to compensate for the losses in well-being deriving from unemployment shocks, i.e. there is no certainty that the basic income will be high enough to enable unemployed workers to regain the baseline. Nevertheless, regardless of whether the basic income is high enough to restore the unemployed to the baseline, from then on they will never be able to rise above the baseline. This is so since they cannot escalate to higher-grade consumption behaviors if these behaviors necessitate a higher income; they suffer passive envy toward the employed who earn wages as well as receiving the basic income if replicating the consumption behaviors of the reference group requires a higher income; they suffer the same passive envy toward entrepreneurs and capitalists; they will not succeed in being envied by others; they suffer a loss of self-esteem for having lost their jobs and being unemployed; they suffer for not being able to realize their aspirations, if this entails a higher income; they might suffer social stigma for being unemployed; they suffer on recognizing that they have no role in society; and they are fully aware that they will never be able to escalate or realize their aspirations. The only chance they have of keeping above the baseline is by engaging in new consumption behaviors that do not call for a higher income but can nevertheless generate esteem, admiration or, more in general, envy: this is certainly possible, but, generally, a higher income would help, and a higher income would also help if change (or escalation) requires costly investment in skill (which could actually be facilitated by public intervention with training and education schemes). By contrast, the employed will have ample opportunities to remain steadily above the baseline, since they do not suffer loss in well-being for having lost their jobs, while obtaining an increase in income (and this is a positive shock that drives them above the baseline); do not suffer problems of self-esteem, while generating envy in others; and indeed they know that they can escalate. They might only suffer passive envy with respect to entrepreneurs and capitalists. Consequently, contrary to the general expectations, provision of a basic income to unemployed humans prevents them from changing their consumption behaviors, i.e., prevents them from escalating, and hence prevents them from reducing inequality in well-being. In other words, the introduction of a basic income may, far from reducing inequality in well-being, actually aggravate it.Furthermore, a universal basic income that remains unchanging over time can hardly generate constructive (or benign, or white) envy since a fixed income may prevent subjects from straining to reach the others’ positions; on the contrary, it can easily generate disruptive (or malicious, or black) envy since the impossibility to imitate others’ behaviors (if these behaviors require a higher income) may drive envious people to try to make the others lose their possessions. In these circumstances, a universal basic income that remains unchanging over time may also generate social instability.The role played by envy, and more generally by all the positional concerns that impact on well-being, is fully consistent with Keynes’ (1930) conclusions on the impossibility of pursuing relative needs. However, again in line with Keynes’ conclusions, boredom, too, can play a role in lowering well-being for unemployed humans. This idea is inconsistent with the traditional analysis of labor supply, which assumes that workers choose whether to work or not by referring to labor marginal disutility and hence implicitly assuming that working reduces well-being. However, according to Keynes, boredom, i.e., not working, could also negatively impact well-being; thus, the analysis to be developed is slightly more complex than generally assumed. In terms of traditional (neoclassical) microeconomics, we might say that working may generate both disutility and utility, or at least that not working may generate disutility.6Alternative (and Complementary) PoliciesThe weaknesses of the basic income as a tool to counteract technological unemployment leaves room for alternative solutions. Here, only two of the possible alternatives will be discussed: modifications to be made to the unconditional basic income policy and policy measures that can be implemented together with the basic income. The policy measures alternative to the basic income will not be discussed.For a thorough discussion of these possible alternatives see, e.g. D’Orlando (2020a and 2020b) and Marchant and Stevens (2017).6.1Implementing a Rising Basic incomeAs pointed out above, subjects can achieve increase in their well-being and remain steadily above the baseline by continuously changing their behaviors (pursuit of novelty), by choosing higher-level behaviors (escalation), by attempting to realize their aspirations (being aware that this is a feasible goal), by provoking the envy of others, by engaging in the consumption behaviors of the reference group, etc. In such a context, passive envy can be a strong stimulus to change and so may exert a positive impact on well-being. This pressure to change may be hampered by provision of a fixed basic income when changing requires a higher income. Therefore, avoiding a fixed basic income could be one possible solution to remedy some of the weaknesses in this policy. The basic income could, for example, correspond to the subsistence level (and so it should vary with regional patterns of consumption and the cost of living) but could rise over time, or rise as subjects engage in valuable activities or accumulate human capital, or whatever goal society sees as socially significant.The idea that the amount of the basic income to some extent depends upon the participation of individuals “in activities that enhance community, civic and cultural development” is to be found in Ford (2009, pp. 175 ff.) and is consistent with the badge system proposed by Marchant and Stevens (2017).The basic income could also be designed so as to allow people to use it to accumulate wealth, which could be bequeathed to heirs, or to acquire shares in firms, thus acting as a strong stimulus to change and escalation. In other words, the basic income could be designed in such a way as to allow for change by eliminating budget constraints or letting subjects reduce them, whereas the stimulus to change is generated by positional concerns.Apart from its consequences on individual well-being, a rising basic income would also bear another (aggregate) positive consequence, i.e. the rise of white envy over black envy: we have seen that a universal basic income that remains unchanging over time may diffuse a disruptive attitude among envious people and generate social instability; on the contrary a rising universal basic income may diffuse a constructive attitude among envious people that can enhance social stability.Two alternatives measures that could be implemented together with the basic income are reduction in per-capita working hours (which was also Keynes’ original 1930 proposal) and the imposition of mandatory quotas on firms regarding the number of workers to hire.6.2Reducing per-capita Working HoursReducing per-capita working hours has been, and to some extent still is, the most popular measure that economists have proposed to counteract technological unemployment. According to Vivarelli (1996), the “end of work literature”, based on the idea that technological progress might be able to completely displace human employment, is faulty since it fails to consider the historical reduction in per-capita working hours and its likely future evolution: “the weak point of this analysis is the lack of any consideration of the historical tendency towards a continuous decrease in per-capita annual working time; for instance, it has been shown that a manual worker used to work about 3000 h at the beginning of this century, whereas the average annual working time ranges nowadays between 1500 and 1900 h according to different national and institutional settings” (Vivarelli, 1996, p. 26).Albeit being historically robust, Vivarelli’s idea appears sound insofar as it takes into account what has happened in the past, but far less persuasive with regard to what might happen in the future, in a scenario seeing a proliferation of robots and artificial intelligence. In this respect, past experience cannot be considered a valid basis to forecast what will happen in the future. This is also true of a policy designed to reinforce, or accelerate, the historical tendency towards a decrease in per-capita working hours. Indeed, such a policy might be useful in coping with the increasing unemployment which could characterize the transitory phase, when robots begin to enter the productive process; but it would be completely useless in terms of effectively counteracting the third wave of technological unemployment, when robots endowed with artificial intelligence enter into most productive processes with the potential capacity to carry out all the tasks that humans do in these processes, but with a higher productivity–wage ratio. In such a context, unless the lower subsistence bound that prevents decrease in wages is removed, firms will in any case prefer to hire only robots rather than humans. And reducing per-capita working hours does not remove the bound. This policy, then, does not in itself counteract technological unemployment, even though it may be useful when combined with other measures, as we will see a little later on.6.3Imposing Mandatory Quotas for Workers to Be HiredHowever, both introducing a basic income policy (with the aim of lowering wages and boosting employment) and implementing a policy of reduction of per-capita working hours, relies excessively on the functioning of market mechanisms, and thus leaves the level of employment that can be achieved uncertain. Imposing mandatory quotas on firms for workers to be hired might therefore appear a better strategy, since it enables policymakers to achieve their employment goals with a fair degree of certainty. The problem with such a policy is that it jeopardizes the firms’ efficiency, and even their competitiveness, as indeed most intervention policies do. It might therefore be preferable to design such a policy in a way that mimics Tietenberg’s traditional tradable permits approach to environmental economics.On the tradable permits approach, see Baumol and Oates (1971), Montgomery (1972), and Tietenberg (1990, 2003.In his approach, pollution quotas are allocated among firms, which can either comply with these quotas or sell/buy them.The same theoretical framework and the same policy tools can be used to cope with technological unemployment problems.Tietenberg’s original approach works as follows. With the most widespread of the tradable permits programs, i.e., cap-and-trade programs,The others being credit programs. See Tietenberg, 2003, p. 408.the policymaker decides the maximum resource access limit (for example, the maximum level of air pollution that firms can produce) and allocates (free of charge or based on auctions in which firms pay more for greater quotas) permits to access the resource (for example, the quantity of CO2 that each firm is allowed to produce) among users. The firms can then trade quotas. With this procedure it would be possible to “minimize the cost of reaching a predefined environmental target” (Tietenberg, 2003, p. 401).“A principal theorem of environmental economics demonstrates that, under specific conditions, an appropriately defined tradable-permit system can minimize the cost of reaching a predefined environmental target (Baumol & Oates, 1971). In a perfectly competitive market, permits will flow toward their highest-valued use. Those that would receive lower value from using the permits (owing to lower abatement costs, for example) have an incentive to trade them to someone who would value them more. The trade benefits both parties. The seller reaps more from the sale than s/he could from using the permit and the buyer gets more value from the permit than s/he pays for it” (Tietenberg, 2003, p. 401).In the case of technological unemployment, things might go in a slightly different way. The policymaker would not set an upper bound to robot use, as would be the case in a cap-and-trade framework, but would instead set a lower bound to the number of workers to be hired by each firm. Such a procedure would be necessary if the government’s goal were to increase human employment, and not to reduce robot employment, considering also that it is really difficult to decide what exactly a robot is and so identify an objective standard for measuring the number of robots a firm is employing (this looks rather like an updated version of the traditional “capital measurement critique”). Therefore, the government first decides its employment goals and then allocates quotas of workers to be hired among the firms (free of charge or based on auctions in which firms pay proportionately more for smaller quotas).Just as in the case of environmental policies, the main problem here is how to determine the level of employment starting from which each firm is to increase it (with the pollution-reduction cap-and-trade approach, it was the level of pollution starting from which each firm had to reduce its pollution). In the traditional environmental cap-and-trade approach, the method that has been used most often is historical, i.e., based on the level of the use of the resource (for example, the level of air pollution) generated by a firm in the course of the last so many years. The same procedure might be used in the case of technological unemployment: each firm starts from its past level of labor hiring (or, better, the average of its labor hiring over a couple of years, to avoid opportunistic behaviors), which represents its initial allocation of rights, and must raise the level of its employment by a given percentage, representing its quota. Firms then have the option to trade their quotas. In this case, sellers pay for selling quotas of workers to be hired, and buyers are paid for buying quotas of workers to be hired. The possibility of trading quotas means increased efficiency for the firms, and indeed for the economic system as a whole:Firms or industries, for which the hiring of human workers is less profitable, would sell quotas to firms or industries for which the employment of human workers is less disadvantageous. Paraphrasing Tietenberg, quotas will flow toward their lowest-loss use. Those that would suffer a higher loss from being forced to hire humans have an incentive to trade them with those who suffer lower losses. Such trade benefits both (D’Orlando, 2020b, p. 22).Adopting a policy based on tradable quotas, firms as a whole would have to hire humans in the amount decided by the government, but human employment would be concentrated where it had the least negative effect on productivity.Furthermore, employed workers would have the opportunity to increase their income and hence enjoy pursuit of novelties, escalate, suffer/enjoy passive/active envy, realize their aspirations, etc.Another appreciable positive consequence of the policy described above concerns investments. If firms are through some mechanism or another forced to hire humans, as in the case of a Tietenberg-inspired policy, when they revamp their factories (or build new ones), they will create plants compatible with human employment, and not plants that can only be used by robots. On the other hand, if firms are free not to hire human workers, as would be the case with different intervention policies, they might concentrate investment on plants compatible with robot rather than human use. Thus, the long-term result of Tietenberg-like policies is the presence of factories in which humans can work, whereas, if other policies are implemented, there is a real risk that in the long-run there would only be robot-specific factories, with no room for human labor.However, the policy described above is exposed to the same weaknesses that mar environmental protection policies: firms risk losing their competitiveness. This is inevitable, since they have either to hire humans paying wages (in relation to productivity) higher than the cost of renting robots or pay a sum on top of the cost of renting robots to sell their quotas. And if they are exposed to international competition they may well find themselves unable to survive on the market due to excessive production costs. Therefore, as has turned out to be the case with environmental policies, this is not a policy for a single country: it has to be implemented by all countries, negotiating international treaties committing single countries to enforcement of the quotas. The problems that these types of international treaties have come up against over time (take the case of the Kyoto Protocol) would probably replicate with the application of employment quotas. There could, however, be one minor advantage if the international treaties, as in the case of environmental treaties, were to regulate the international trade in quotas, enabling the whole mechanism to work better.6.4A Comprehensive Policy MixThe basic income could be successfully combined with either of the policies described above.Let us first discuss the possibility of combining a reduction in per-capita working hours with provision of an unconditional basic income. In this case, as extensively discussed above, given the security of an unconditional basic income, wages could fall below subsistence, while nevertheless leaving humans the possibility to accept to be hired (since the basic income ensures the workers’ subsistence in any case). We also know that if the wages fell to the extent that the humans’ productivity–wage ratio rose above, or at least equal to, that of robots, it would be in the interest of firms to hire humans. Moreover, reduction in per-capita working hours, given the amount of humans’ worked hours demanded by firms, could allow firms to hire a greater number of humans than they would have if working hours were not reduced. Thus many human workers would be able to increase their income by adding the basic income to the wage, so as to be able to enjoy pursuit of novelties, escalate, etc., that is, to increase their well-being and reduce inequality in well-being.The second possibility is to combine quotas with provision of an unconditional basic income. Again, as in the previous case, the basic income could benefit both the single firms and the human workers: the former could hire workers instead of robots, paying wages low enough to make the productive–wage ratio of humans higher, or at least equal, to that of robots, while the latter could accept such wages since the basic income ensures subsistence in any case. But what if the wage does not fall low enough, or there is only limited increase in employment? The policymaker might force firms to hire humans by setting quotas of human workers to be hired. And firms could decide whether it was in their interest to hire workers paying wages below subsistence or hire robots alone, selling quotas and paying lower robot remuneration (since the reduction of human wages boosts competitive forces, and so reduces robot remuneration): this circumstance would be particularly important for firms operating in sectors where the cost of mandatory labor hiring was too high with respect to labor productivity. In any case the actual well-being-enhancing effect of this process depends upon how high the wage offered by firms actually turns out to be: it has to be high enough to allow wage-earners to undertake different (higher ranking) consumption behaviors.In any case, an alternative policy could be based on a combination of all the (three) measures discussed in this section: a basic income rising over time could guarantee that people remained steadily above the baseline and allow wages to fall below subsistence level, so that firms could profitably hire workers; a reduction in per-capita working hours could guarantee (nearly) full employment; setting quotas and making provision for trade in them could drive the system to the desired employment level and help reduce the cost of hiring humans for firms that had most to lose with the obligation to hire humans.However, in addition to implementing the three policies described in this section, the government should also provide citizens with an adequate endowment of basic skills, through basic education, and when possible also supply (or promote) programs for further formal and informal lifelong education. In fact, without the support of the basic skills and lifelong skill accumulation programs subjects would have difficulty in changing and/or escalating, due to the high and possibly rising levels of skill endowment that change/escalation requires, together with the possibly rising cost of skill accumulation processes and the rapid skill obsolescence caused by today’s impetuous technological progress.7ConclusionsAs discussed above, the third wave of technological unemployment has characteristics and implications radically different from those of the first and second wave. The first two waves reduced human employment in specific sectors (agriculture to start with, manufacturing later on), or lowered wages and/or employment for unskilled workers and/or workers carrying out routine tasks, thus increasing income inequality, with no substantial effect on long-term unemployment; the third might be capable of destroying all jobs for humans (both skilled and unskilled and both routine and non-routine) in most sectors of the economy, generating high levels of (technological) unemployment. This is due to the fact that the first two waves of technological unemployment were caused by the development of more productive machines, while the third would result from the introduction of robots endowed with artificial intelligence in the productive process. Unlike machines, robots do not cooperate with workers but can replace them in many, if not all, of their tasks. If such a scenario does indeed come about, the traditional compensation forces which proved effective in preventing the impact of technical progress on employment in the past will be ineffective due to the relatively high productivity and low cost of robots. This implies that wage reduction should be significant enough to counteract technological unemployment and guarantee human employment, but it will not pass the limit represented by the lower subsistence bound of the wage. Additionally, an increase in demand might lead to an increase in the production of commodities produced by robots alone and hence an increase in employment for robots alone without increasing human employment.The idea of an unconditional basic income is the solution to cope with the resulting high level of unemployment that has come under most discussion. However, despite the ever-growing debate, the impact of this solution on well-being has never been thoroughly thrashed out. Furthermore, when behavioral economics constructs (such as hedonic adaptation, pursuit of novelty, escalation, positional effects, and envy) are used to evaluate the well-being consequences of the basic income, a number of drawbacks emerge. In particular, this measure alone, without any accompanying measures, appears incapable of guaranteeing the process of continuously changing consumption behaviors that allows subjects to remain constantly above the baseline level of well-being, mainly due to the fact that the subjects’ income is fixed and the positive effects of positional concerns and social interaction as a stimulus for change cannot operate, and there can therefore be no escaping from habituation and passive envy. In other words, provision of a fixed basic income favors social immobility and thereby risks increasing rather than reducing inequality in well-being.At this point, let us consider the possible solutions to be implemented together with the basic income. Of these solutions, the most promising are reduction in per-capita working hours and the possible implementation of a “cap and trade” solution (which mimics Tietenberg’s traditional tradable permits approach to environmental issues), which sets quotas of workers to be hired on firms, but with the option to trade these quotas. Combining these two solutions with provision of a basic income which rises over time could guarantee the best results in terms of employment, well-being and firm efficiency.More in general, public policies should be designed to encourage change and escalation in subjects’ consumption behaviors. This could result in policies that stimulate prosocial behavior, policies designed to give people the opportunity to accumulate wealth and/or human capital (for themselves or for their heirs) by obtaining the ownership of firms (or robots), but also policies that allow people to engage in enviable, extreme or innovative consumption behaviors, so as to be envied and imitated by others. Any attempt to stimulate change should be promoted. But change (and escalation) also require an adequate endowment of basic skills on the part of citizens and the possibility to make use of lifelong programs of skill accumulation, also due to today’s rapid skill obsolescence caused by technological progress. Thus, public policies should also aim at guaranteeing adequate education and training programs, while promoting the spread of informal skill accumulation processes.A further conclusion of the analysis presented here is that, in order to minimize losses in well-being, an unconditional basic income (together with its companion policies) should have the main goal of ensuring the maximum possible level of employment, and not of financing the possibility to remain unemployed. The basic income can play this role by allowing wages to fall below subsistence, since workers can accept to be hired on such a low level of wages if subsistence is guaranteed by the basic income. And if wages fall below subsistence the possibility remains that the human wage/productivity ratio may fall below robot rent price/productivity ratio, allowing firms to hire human workers rather than concentrating solely on robots. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Basic Income Studies de Gruyter

Social Interaction, Envy, and the Basic Income: Do Remedies to Technological Unemployment Reduce Well-being?

D’Orlando, Fabio
Basic Income Studies , Volume 17 (1): 41 – Jun 1, 2022
Download PDF
41 pages

Loading next page...
 
/lp/de-gruyter/social-interaction-envy-and-the-basic-income-do-remedies-to-70est1UMy3
Publisher
de Gruyter
Copyright
© 2022 Walter de Gruyter GmbH, Berlin/Boston
ISSN
1932-0183
eISSN
1932-0183
DOI
10.1515/bis-2020-0001
Publisher site
See Article on Publisher Site

Abstract

1IntroductionThe debate on technological unemployment has a long history, which in the past seemed to have reached the overarching conclusion that the phenomenon is of limited empirical relevance. Matters began to change at the end of the last century, when the ICT revolution was found to have an impact on employment, so that in recent years some authors (see, e.g. Brynjolfsson & McAfee, 2011, 2014) have questioned that old conclusion. The debate has since gained momentum, but has not converged on any broad consensus.In fact, technological progress is generally recognized as one of the causes of increasing income inequality over the last few decades, a phenomenon known as skill-biased technical change. Different approaches focus on tasks rather than skills, i.e., task-biased (or routine-replacing) technical change, but the logic is quite similar. In any case, there is less agreement about the possibility that technical change could cause a significant increase in the overall unemployment rate in the long run, i.e., technological unemployment, although a growing number of empirical studies are starting to track the phenomenon (see, e.g. Acemoglu & Restrepo, 2017, 2019), and other contributors have highlighted the devastating effect that artificial intelligence may have on employment (see, e.g. Acemoglu & Restrepo, 2020; Brynjolfsson & McAfee, 2011, 2014; Ford, 2015).Indeed, skill-biased (or task-biased) technical change and contemporary technological unemployment are two profoundly different phenomena. The former is a consequence of 20th-century technological progress, which has been able to bring down labor demand and hence wages for low-skilled workers (or workers carrying out routine and repetitive tasks) but drive up labor demand and hence wages for high-skilled workers (or workers carrying out non-routine tasks). Income inequality is the result. Technological unemployment is considered a possible consequence of 21st-century technological progress, associated with artificial intelligence, which may well jeopardize employment in most sectors of the economy for all workers, all jobs and all tasks in the future. The fact that robots are supposed to be more productive and less expensive than workers could result in high levels of unemployment and, according to some scholars, even the risk of the system converging on a state of full unemployment cannot be a priori ruled out (see e.g. the benchmark model in Berg, Buffie, & Zanna, 2018).Summing up the different positions on the issue, there is a general consensus that technological progress contributed to increasing inequality in the closing decades of the 20th century; there is a majority consensus that it could increase unemployment in the coming years; and there is minority but growing consensus that it might drive the system to mass unemployment.True, the possibility of the development of artificial intelligence leading to high levels of unemployment does not amount to a certainty. However, the basic idea behind the present article is that developing thorough theoretical analysis of and discussing the possible solutions to a problem that does not as yet demand attention but may well in the near future is perfectly feasible, appropriate and useful. Thus, the present article is in strong opposition to the consensus view that it is possible to discuss an economic problem only after empirical proof that it has actually arisen. And since empirical evidence of future events is inevitably lacking, in discussing the possible dynamics of the new process we can only undertake a thought experiment using the tools of economic theory.In any case, if we consider the possibility that robotization and the diffusion of artificial intelligence could generate high levels of unemployment, ex-ante study of public intervention policies that can counteract the phenomenon appears mandatory. And indeed many policies have been proposed. Surveying these policies, we find various possibilities: providing an unconditional basic income, taxing robots, subsidizing the hiring of humans, nationalizing firms and/or assigning the property of robots and/or firms to workers, and setting minimum human employment quotas or maximum robot employment quotas. Providing an unconditional (or universal) basic income to all citizens, irrespective of whether they are employed or unemployed, is the most debated and best-known of these possible policies. The present article focuses on this policy and in particular on introducing a basic income high enough to guarantee subsistence for human beings, as indeed it should be in a context that sees technological unemployment looming as a concrete possibility.Although the basic income has been searchingly discussed in the literature (see, e.g. Bowles, 1992; Gamel, Balsan, & Vero, 2006; Haagh, 2019a; Marchant, Stevens, & Hennessy, 2014; Pech, 2010; Tanner, 2015; Van der Linden, 1997, 2002), few studies have tried to evaluate the impact on well-being of a universal basic income explicitly designed to counteract technological unemployment (see, e.g. Ford, 2009, chapter 4; Ford, 2015, chapter 10; Brynjolfsson & McAfee, 2014, chapter 14): most studies have simply discussed the basic income as an instrument to reduce poverty and inequality. This is interesting in view of the fact that the contemporary resurgence of interest in this measure rests mainly on the possibility that it could play a role in counteracting technological unemployment. And economic theory is now fully equipped with theoretical instruments to evaluate the well-being consequences of such a policy. In particular, insights from behavioral and happiness economics – such as loss aversion, hedonic adaptation, escalation, social interaction and positional effects – appear useful for the purpose.Building on the above theoretical constructs and on Keynes’ short 1930 essay (“Economic Possibilities for Our Grandchildren”), which can be considered the seminal contribution on the issue, the present article aims to discuss in detail and in depth the impact that an unconditional basic income, implemented to counteract technological unemployment, might have on people’s well-being.It is worth noting that the basic income could represent a solution to all three ways (past, present and future) in which technological progress impacts on employment and wages (i.e., past increasing inequality due to skill-biased or task-biased technical change, current increase in unemployment due to technological progress, and a future with high levels of unemployment due to robotization and application of artificial intelligence). However, the focus in the present article is not on the (past) increase in wage polarization but on the (future, possible) increase in unemployment and the possibility of reaching a situation of mass unemployment (as will be shown, providing a basic income could allow wages to fall below subsistence and hence help prevent the system from drifting into a situation of mass unemployment).The main conclusion of the article will be that an unconditional basic income that remains unchanging over time could aggravate inequality in well-being and, more in general, lower the general level of well-being. The insights offered by behavioral economics lead all too easily to these outcomes. In particular, hedonic adaptation and loss aversion can be applied to show that provision of a basic income cannot adequately compensate for the psychological costs of unemployment, which empirical studies consider one of the worst experiences people may face in life (see, e.g. Clark & Oswald, 1994; Frey & Stutzer, 2002; Clark, Diener, Georgellisand, & Lucas, 2004). Furthermore, unemployed recipients of the basic income, far from provoking the envy of other people, could only envy the employed, the owners of the robots and the entrepreneurs. They would suffer problems of self-esteem and social stigma. Finally, they would be unable to escalate to higher-grade consumption behaviors, compensating for hedonic adaptation and passive envy, or fulfill their aspirations (and demoralized by the awareness of being unable to do so) – yet more circumstances that lower well-being. Provision of a higher basic income might overcome the first problem, i.e., the psychological costs, but not the others. By contrast, employed earners of the basic income (as well as entrepreneurs and owners of robots) would enjoy a significant increase in well-being because of the possibility to both provoke the envy of the unemployed and escalate to higher-grade consumption behaviors. Consequently, a universal basic income might well see inequality in well-being increasing.A possible alternative solution could be based on a policy mix combining a basic income that increases over time with a reduction in per capita working hours and a policy that mimics Tietenberg’s (1990 and 2003 tradable permits approach by imposing mandatory quotas for hiring human workers on firms. A solution of the sort has a number of positive effects on well-being and guarantees lower costs for firms that face higher costs in increasing the human–robot ratio.The present article is structured as follows.In Section 2, the impact of technological progress on employment, i.e., technological unemployment, is discussed from both the historical and the theoretical viewpoint. The possible policy solutions for this phenomenon, focusing in particular on provision of an unconditional basic income, are addressed in Section 3. Based on Keynes’ 1930 contribution and insights offered by behavioral and happiness economics, such as hedonic adaptation, envy, and escalation, the theoretical tools that one can find in the economics literature to measure the impact of different policies on aggregate well-being are described in Section 4. In Section 5, the theoretical tools described in Section 4 are applied to evaluate the effects of implementing the unconditional basic income on well-being, arguing that, in most cases, it aggravates inequality in well-being and, more in general, lowers its level. In Section 6, some alternative policies are discussed and a policy mix that combines the basic income with other measures is proposed. Section 7 concludes the paper.2Technological UnemploymentAs technological progress has impacted employment in different ways during the last 250 years, not only has the debate on technological unemployment crucially evolved over time but the phenomenon seems to elude any single definition. Consequently, many puzzling differences of opinion emerging in the theoretical literature can be accounted for with the fact that different scholars have discussed different phenomena as if they were one and the same.In particular, we can single out three different historical and theoretical phases in the debate on the impact of technological progress on employment, and hence three different groups of causes, consequences, and possible intervention policies.The first phase, corresponding to the first wave of technological unemployment, can be historically placed in the years of the First (roughly 1765–1830) and Second (roughly 1870–1914) Industrial Revolutions, when it appears unquestionable that technological progress, and in particular process innovation,Process innovation and product innovation are the main drivers of technological progress. Product innovation is generally considered employment-friendly, whereas the contrary applies to process innovation.reduced staffing needs for some industries.However, fierce as the debate turned out to be in those years, according to most contributors (see, e.g. Campa, 2007, 2014, 2017; Carboni, 2015; Schettkat & Yocarini, 2003; Wladawsky-Berger, 2015) the long-term empirical impact of technological progress on employment has proved modest: process innovation mainly relocated workers from one sector (to begin with, from agriculture; later from manufacturing) to another (at the outset, to manufacturing; later to services), without increasing long-term unemployment. These empirical findings are consistent, at least partially, with the great majority of theoretical contributions on the transition from agriculture to manufacturing, and from manufacturing to services, well described with Colin Clark’s “law of the three sectors” (Clark, 1940; Fisher, 1935): indeed, according to Baumol (1967), Fuchs (1968), and other scholars (for a review of the literature on the issue see Schettkat & Yocarini, 2003; D’Agostino, Serafini, & Ward, 2006), not only was the process eventually destined to generate a significant rise in employment in the service sector, but it was also to lead to a major increase in employment in general (mainly because of the low productivity growth in the service sector). Moreover, the empirical findings are also consistent with the conclusions arrived at in compensation theory, according to which technological progress may actually generate unemployment in the short run, but compensative forces would bring the system back to full employment in the long run (mainly) because of wage flexibility and increased employment deriving from increased production (of machines, in particular).For thorough discussion of compensation forces, see Vivarelli (2007 and 2014, Campa (2017), Peters (2017), Blien and Ludewig (2017), D’Orlando (2020a and 2020b), Calvino and Virgillito (2018).We may say that in this (initial) phase machines (capital) cooperate with human workers in the productive process while technological progress in the form of process innovation boosts labor and capital productivity, so that employment problems can arise only if demand does not keep up with increase in productivityBy definition, Productivity=ProductionEmployment$\text{Productivity}=\frac{\text{Production}}{\text{Employment}}$so that Employment=ProductionProductivity$\text{Employment}=\frac{\text{Production}}{\text{Productivity}}$. In the presence of an increase in productivity, the reduction in employment can be offset by an increase in demand, and hence in production, equal to (or greater than) the increase in productivity.(and compensation theory explains why this is actually not the case – see on this point Bessen, 2018; Berg et al., 2018; Blien & Ludewig, 2017; Neisser, 1942).The second phase started with the shift from manufacturing to services, in the first half of the 20th century, but gained momentum with the second wave of technological unemployment, which can be historically placed in the years of the Third Industrial Revolution (roughly 1980 up to the present), when progress in information technologies can be considered one of the main reasonsSee, e.g. Aghion, Caroli, and Garcìa-Peñalosa (1999), Card and Di Nardo (2002), Autor, Levy, and Murnane (2003, 2006, 2008, Violante (2008), Dustmann, Ludsteck, & Schönberg (2009), Acemoglu and Autor (2011), Autor and Handel (2013), Freeman (2015), Acemoglu and Restrepo (2016).(although not the only oneA number of scholars consider the reduction in public intervention in the economy and globalization as the other major factor responsible for wage polarization (see e.g. D’Orlando & Ferrante, 2022; Santarelli & Figini, 2002; Piketty, 2014).) for wage polarization and increases in inequality.This has been so since the shift from manufacturing to services, which had a huge impact on production and employment in all the highly industrialized countries, did not exactly have the same positive effects as predicted with the theoretical models described above, because it was accompanied by a shrinking of the role, as well as wages, of the blue-collar workers (see e.g. Bell, 1976). For these workers, reduced employment possibilities in manufacturing did not meet with increased employment possibilities in the service sector, which required different (higher) skills, creating a mismatch between workers’ skills and the skills demanded by the service firms that could be resolved only with lower wages. According to D’Agostino et al. (2006, p. 7) “the presence of adjustment barriers associated with the shift from manufacturing to services may have hindered the ongoing process of sectoral reallocation of the workforce.” Thus, the decline in manufacturing was accompanied by a decline in wages and income for middle-class workers, unable to maintain their standards of living.From an alternative perspective, focusing more on the role of technological progress than on the role of the shift to services, the whole process has been called skill-biased technical change,According to Violante (2008, p. 2): “‘Skill-Biased Technical Change’ (SBTC thereafter) is a shift in the production technology that favors skilled (e.g. more educated, more able, more experienced) labor by increasing its relative productivity and, therefore, its relative demand. Ceteris paribus, SBTC induces a rise in the skill premium—the ratio of skilled to unskilled wages.”emphasizing the fact that the process reduced employment possibilities and wages for unskilled workers but raised them for skilled workers. A slightly different version of this approach focuses more on tasksOn the task approach, see, e.g. Autor, Katz, and Kearney (2006); Autor (2013); and Autor and Handel (2013).than on skills and is known as task-biased (or routine-replacing) technical change. The tasks most affected by technological progress are routine tasks, so that workers whose jobs consisted mainly in such tasks suffered reduction in their employment possibilities and wages, whereas workers engaged in predominantly non-routine tasks enjoyed a corresponding increase.In this second phase, the impact of technological progress on aggregate employment was also relatively mild: during the years of the ICT revolution, employment levels rose rather than falling, sectors in expansion more than reabsorbing the unemployment generated by sectors in contraction, albeit at the cost of lower wages for some, the consequence thus turning out to be increase in inequality rather than in unemployment.However, from a theoretical viewpoint, a major new turn in this phase was that, for some tasks, machines began to replace workers, no longer serving simply to boost workers’ productivity. Thus, we have tasks (and sectors) in which capital cooperates with labor in the productive process, with technological progress increasing labor (and capital) productivity, as occurred in the preceding phase, and tasks (or even sectors) in which machines replace humans in the productive process, with technological progress increasing only (or mostly) capital productivity.On the above-described substitution process, see e.g. Brynjolfsson and McAfee (2011 and 2014, Ford (2015), West (2015), Acemoglu and Restrepo (2019).In any case, up to the end of the 20th century, the majority of theoretical contributors appeared to agree that compensation forces are effective in counteracting technological unemployment, although a number of dissenting opinions did persist.Dissenting opinions range from affirmation that technological unemployment exists but is concealed by the historical reduction in per capita working hours (i.e., Vivarelli, 1996) to the position that “economic theory does not have a clear-cut answer about the final employment effect of R&D and innovation,” such that “attention should be turned to the empirical analyses” (Piva & Vivarelli, 2017, p. 14).Radical changes began with the Fourth Industrial Revolution, corresponding to the third wave of technological unemployment, which started roughly around the beginning of the 21st century. The Third and Fourth Industrial Revolutions were based on an apparently similar form of technological progress; however, behind this apparent similarity a crucial difference was concealed. This difference lays in the new type of capital that represented the main innovative (and symbolic) characteristic of the Fourth Industrial Revolution: robots,According to Freeman (2015, p. 2), “[t]he term ‘robots’ refers broadly to any sort of machinery, from computer to artificial intelligence programs, that provides a good substitute for work currently performed by humans […] it does not matter whether a robot/machinery has a humanoid appearance, as long as it can perform human functions.”and in particular robots endowed with artificial intelligence. If robots and artificial intelligence enter the productive process, they may well eventually replace human beings in most jobs and tasks.According to Campa (2017, p. 14): “Artificial Intelligence develops exponentially and not only promises to further reduce the workforce in manufacturing, but it will begin to erode the work of specialists in the service sector. In the near future, unemployment could concern economic actors who have attended higher education institutions and invested much time and money to acquire their professional skills, such as journalists, physicians, teachers, lawyers, consultants, managers, etc.”If such a scenario were to come about, capital would no longer cooperate with human workers in enhancing labor productivity but could totally replace workers. Technological unemployment, which hit only unskilled workers or workers specialized in particular tasks in the Third Industrial Revolution, could affect all workers, skilled and unskilled, regardless of whether they are performing routine tasks or not, in most sectors of the economy.The theoretical approach that can be used to systematize this third wave of technological unemployment is, inevitably, a development of the approach used for the second wave, but only for the part in which robots compete with humans for employment: among the productive factors, we now have robots that, like capital, are produced but that, unlike capital, can be produced by using robots alone.To avoid formal complications that are not particularly relevant to an analysis focused on study of the basic income, we may assume that robots are humanoid programmable machines that can replace humans in most activities, and not merely in manual and/or routine activities. In this scenario, these humanoid machines can carry out tasks identical to those performed by humans. The scenario can be further simplified by assuming that capitalists own robots and rent them out to entrepreneurs. Meanwhile, entrepreneurs have to decide whether to hire robots or humans for their productive processes. At the microeconomic level, the choice between hiring labor (workers) or capital (robots) represents a choice between techniques: given the quantities produced, entrepreneurs will substitute robots for human workers if the value of the marginal product of robots, divided by the cost of the service provided by robots (i.e. rents paid to robots’ owners), is greater than the value of the marginal product of labor, divided by the wage rate. Given the quantity to be produced, entrepreneurs will substitute robots for human workers if:(1)p‾∂Q∂Rw‾R>p‾∂Q∂Lw‾L.$$\frac{\overline{p}\frac{\partial \text{Q}}{\partial \text{R}}}{{\overline{\text{w}}}_{\text{R}}} > \frac{\overline{p}\frac{\partial \text{Q}}{\partial L}}{{\overline{\text{w}}}_{\text{L}}}\text{.}$$where w‾L${\overline{\text{w}}}_{\text{L}}$is workers’ wage, w‾R${\overline{\text{w}}}_{\text{R}}$is the price of the service provided by robots, ∂Q∂L$\frac{\partial \text{Q}}{\partial \text{L}}$is the marginal productivity of labor, ∂Q∂R$\frac{\partial \text{Q}}{\partial \text{R}}$is the marginal productivity of robots, and p‾$\overline{\text{p}}$is the price of the commodity produced.The price of the commodity produced, the wage rate, and the price of the service of capital are considered as given under the (here implicitly assumed) hypothesis of perfect competition.The problem of unemployment arises in a scenario where robots can replace human workers in most productive processes (or tasks) and show in those productive processes a productivity–wage ratio higher than that of human workers for all tasks and wage levels above subsistence. These circumstances imply that compensation theory cannot hold: if commodities can be produced by means of robots alone, no increase in demand can raise employment, and wage flexibility is likely to meet the subsistence bound before human workers’ productivity–wage ratio becomes as high as that of robots. As a result, high levels of unemployment are not only possible but, indeed, quite probable.It is certainly true that the resulting unemployment can be considered voluntary if viewed in the light of traditional (i.e. neoclassical) labor market microeconomics: workers voluntarily refuse to work, the wage they would earn being below subsistence and hence, inevitably, below their reservation wage; and firms will not pay higher wages if hiring robots proves a preferable option. To put it another way, according to orthodox economics the labor market is perfectly functioning and capable of arriving at an equilibrium where involuntary unemployment does not exist. The problem is that defining people who refuse to work for a wage rate below subsistence as voluntarily unemployed may be theoretically correct but in practice highly misleading, if not meaningless. As we will see a little later on, the basic income can impact exactly on this point, by reconciling theory with reality, since it would make it acceptable for humans to work for a wage rate below subsistence (subsistence being guaranteed by the basic income) so that unemployment, as well employment, would actually be voluntary.In any case, the spread of technological unemployment would inevitably take time, would not necessarily be complete and would probably affect different countries in different ways. In particular, it would affect the developed and less developed countries in different ways mainly due to technological reshoring, i.e., the revision of previous offshoring choices of multinational firms with production relocated to the home country (Ocicka, 2016, p. 105).On reshoring, or back shoring, in general see e.g. Gray, Skowronski, Esenduran, and Rungtusanatham (2013); Fratocchi, Di Mauro, Barbieri, Nassimbeni, and Zanoni (2014); Ancarani, Di Mauro, Fratocchi, and Orzes (2015).Before robotization, multinational firms located labor-intensive production phases in the less developed countries, given their low labor costs. And in these countries local firms could also be competitive in labor-intensive productions on the international markets for the same reason. In a Smithian perspective, we can see low labor costs as an absolute advantage for the less developed countries. Robotization makes using robots in the developed countries less expensive (and/or more productive) than using unskilled/routinary workers in the less developed countries, destroying the absolute advantage of the latter. So it is that multinational firms re-shore, i.e. close plants and fire workers in less developed countries and bring their productions back home, while labor-intensive local firms lose their competitiveness on the international markets and are likely to go bankrupt, stopping production and firing workers. Production in the less developed countries falls, and with it employment, even if they do not robotize. Unemployment would therefore hit above all the countries where the phases of production highly intensive in unskilled labor were located. As a result, the diffusion of robots in the developed countries could also generate high levels of unemployment in the less developed countries.However, at the outset the process would probably hit the countries specialized in productions highly intensive in unskilled labor but, as artificial intelligence and robotization spread, it would later also hit countries specialized in production highly intensive in labor, regardless of whether it is skilled or unskilled, or engaged in routine or non-routine tasks, i.e. also the developed countries. Unemployment would thus spread through countries irrespective of their factor endowments and specializations.If the labor market is segmented, whether in a skill or task perspective, we can expect the technological substitution of workers with robots to come about through a succession of steps. In a skill-biased technical change scenario, the process would first affect unskilled workers, and skilled workers only later. In a task-biased (or routine-replacing) technical change scenario, the substitution would first affect workers performing routine tasks, and later workers performing non-routine tasks. In any case, we would have a transition taking a long time and generating successive waves of unemployment for different categories of workers at different times, as innovations were developed in specific sectors or for specific tasks. This dynamic process could generate successive waves of inequality among workers whose jobs required different skills, workers whose jobs required routine or non-routine tasks, and workers living in different countries or regional areas (specialized in different jobs/tasks).In this situation, it may be difficult for humans to hold enough bargaining power to obtain an adequate share of profits and, furthermore, for the democratic institutions to survive, for if humans are unnecessary for production their bargaining power risks falling to zero. But since it is also in the interest of the firms owners to have adequate demand for the goods they produce, inevitably it would be up to the government to devise the appropriate intervention policies.3Policies against Technological Unemployment, and the Basic IncomeThe most popular of the policies capable of counteracting technological unemployment are taxing the hiring of robots, subsidizing the hiring of humans, subsidizing skill accumulation on the part of human workers (so as to enhance their productivity–wage ratio), reducing per capita working hours, nationalizing firms, assigning the ownership of robots (or firms) to citizens, encouraging workers to buy shares in firms, or providing an unconditional basic income to all citizens, regardless of whether they are employed or not.For thorough analysis of all these policy intervention mechanisms, see D’Orlando (2020a, pp. 9–10), D’Orlando (2020b, pp. 607–612) and Marchant and Stevens (2017).In the present article, the focus is on the last of these, which has in fact been referred to in many different ways: unconditional basic income, universal basic income, basic guaranteed income, basic income guarantee, basic income grant, guaranteed national income, etc. (see, e.g. Marchant & Stevens, 2017; Pech, 2010; Tanner, 2015).The unconditional basic income considered in the present article is “a cash grant provided to every citizen […] without any other eligibility requirement” (Tanner, 2015, p. 3) and specifically aiming to counteract technological unemployment. In particular, in view of the prospect that technological progress could generate high levels of unemployment that could not easily be reabsorbed, but unlike most other contributions, here the basic income is assumed to be sufficiently high to guarantee subsistence.In economic literature there are few studies on the consequences of provision of an unconditional basic income (see, e.g. Bowles, 1992; Gamel et al., 2006; Marchant, Stevens, & Hennessy, 2014; Pech, 2010; Tanner, 2015; Van der Linden, 1997, 2002).The debate on the basic income dates back to Paine (1797) and in the course of time has seen the participation of various social scientists, from economic historians (e.g. Polanyi, 1944) to economists (e.g. Friedman, 1962; Hayek, 1987; Tobin, 1966) and political philosophers (e.g. Van Parjis 2004).Furthermore, although the basic income has recently been at the center of public debate on a possible remedy for technological unemployment,On this point, see, e.g. Zheng, Guerriero, Lopez, and Haverman (2017, p. 13) and Martinelli (2017, pp. 27–29).even fewer studies consider it as such,See, e.g. Ford, 2009, chapter 4; Ford, 2015, chapter 10; Brynjolfsson & McAfee, 2014, chapter 14.in most cases proposing it simply as a remedy for poverty or inequality in a low unemployment scenario, i.e., as a possible substitute for the existing social insurance and benefits.Of the positive consequences of introducing an unconditional basic income, some scholars (see, e.g. Haagh, 2011a, 2019a; Standing, 2003, 2005) stress that it guarantees a lifelong, unconditional income support, ensuring greater income security with positive impact on the general well-being. Furthermore, provision an unconditional basic income would remedy the poor performance of the traditional welfare instruments (Jordan, 1992, pp. 170 ff.), which entailed social exclusion and structural unemployment, and their contemporary versions based on punitive approaches aimed at promoting labor supply (Haagh, 2019b). Finally, the basic income “would be simpler and more transparent than the current welfare bureaucracy… would reduce paternalism and government involvement in the lives of poor people… would more effectively alleviate poverty… could provide better incentives—or at least fewer disincentives—for work” (Tanner, 2015, p. 7; see also Widerquist, 2017). Other scholars come to the opposite conclusion, arguing that the basic income “fails to provide an incentive to find work or engage in other meaningful and beneficial activities […], provides an automatic government handout to individuals while reducing their esteem in their own eyes and those of their neighbors and contacts [and] would be very expensive” (Marchant & Stevens, 2017, p. 1). Studies based on income and substitution effects were developed but proved inconclusive regarding the impact of a basic income on labor supply, the wage equilibrium, and work effort (see, e.g. Martinelli, 2017, p. 51 ff). However, the idea that the basic income reduces labor supply and work effort, inducing people to remain idle, is fairly widespread, at least among economists. Moving on from the solidly traditional to the less traditional economic approaches, we again find different opinions on the impact of the basic income on labor supply and work effort. Some scholars look to behavioral economics and evolutionary game theory to conclude that when altruistic behavior and behaviors based on habits or heuristics replace self-interest and maximizing behavior, “individuals may not shirk when given income guarantees” (Widerquist, Lewis, & Pressman, 2005, p. 590). Nevertheless, insights from the debate on the role of intrinsic motivations suggest that when subsistence is guaranteed, it might be difficult to motivate people to work (and/or to put real effort into their work), at least for certain kinds of jobs, if the incentive is merely monetary (or extrinsic).On the basic income and intrinsic motivations see also Haagh, 2011b.Furthermore, “for those tasks in which a person demonstrates a high intrinsic motivation to perform, the introduction of an extrinsic incentive (in the form of a monetary reward or a fine) undermines her intrinsic motivation, which may cause her to decrease the level of effort” (Pech, 2010, p. 8).Notably, some authors, like Jordan (1992), consider the basic income from the point of view of the community, as a policy tool that can guarantee individuals the possibility “of caring or participating in other socially meaningful ways in society, or contributing to gender equality, social cohesion and social trust… [so that the basic income] has also been advocated from a liberal perspective in setting all people free to make their own decisions in life” (Roosma & van Oorschot, 2019).In the presence of technological unemployment, broadly speaking we may begin with the observation that provision of an unconditional basic income can essentially be targeted to two different goals, so that two different and alternative scenarios are possible.First scenario: with the guarantee of an unconditional basic income the public authorities could tolerate high levels of unemployment or, in an extreme and (as we will see below) mainly theoretical case, even full unemployment. Consider, for the sake of simplicity (and an understanding of Figure 1 below) this latter case: all humans are unemployed, and production is entirely accomplished by firms using robots alone, while the survival of humans is guaranteed by provision of the basic income, which must necessarily at least meet the subsistence level.Figure 1:Only robots are employed.Second scenario: since one of the problems that prevent workers from being competitive with robots (and hence from being hired by firms) is the existence of a lower subsistence bound below which human wages cannot fall, with the guarantee of an unconditional basic income for all citizens workers could accept wages below the subsistence level. In this case, the basic income would not necessarily be above the subsistence level, since it is the sum of the basic income and the wage that guarantees the survival of humans.Notably, the basic income is not an unemployment benefit, as humans receive it whether they are employed or not.However, in the present article, as pointed out above, we limit discussion to the case of a basic income guaranteeing subsistence, as indeed it should in a context where high levels of technological unemployment are hard to reabsorb and the unemployed cannot otherwise survive. Therefore, in this scenario, once the subsistence basket is obtained thanks to the basic income, humans can decide whether to agree to work on a low wage to raise their consumption above the subsistence level or to stay voluntarily unemployed and make do with subsistence.Referring to a scenario in which only robots are employed (i.e., the first scenario), the schematic functioning of the system as a whole is depicted in Figure 1: production is accomplished by robots alone; robots are privately owned by capitalists’ families; capitalists rent robots out to firms; firms produce and sell goods to households, pay taxes to the Government and rents (∑w‾R)$\sum {\overline{w}}_{R})$to the capitalists’ families; firms are managed by entrepreneurs who organize production; all humans are unemployed and their families receive the basic income; entrepreneurs’ families receive the basic income plus profits; capitalists’ families receive the basic income plus rents coming from robots renting; if we assume a balanced government budget, the basic income is financed by taxes paid by firmsFor the sake of simplicity let us assume that capitalists do not pay taxes (nothing would change except understanding of the figure in the contrary case).(firms’ taxes are equal to the basic income distributed to all citizens, so that ultimately firms pay the basic income).More in general, alternatives may include: a corporate tax; a wealth tax; a tax on owners of robots as distinguished from other forms of wealth; and a VAT (which, in combination with a UBI, can be designed to have progressive incidence); otherwise, we can imagine deficit financing of the basic income.Referring to the scenario in which also humans are employed (i.e. the second scenario), Condition 1 still depicts the firms’ microeconomic behavior. However, the aggregate outcome is more complex; it is depicted in Figure 2 below. The firms’ surplus is distributed to: i) entrepreneurs (profits); ii) robot owners (rents ∑w‾R)$\sum {\overline{w}}_{R})$); iii) workers (wages); iv) the government (taxes). Now, since the taxes paid by firms are equal to the basic income paid to humans, firms as a whole will pay workers as a whole an amount equal to the sum of the taxes they pay and the wages they pay.Figure 2:Also humans are employed.Therefore, for the representative firm, it is in its interest to hire robots rather than humans if(2)∑w‾L+basic incomep‾∂Q∂L>w‾R+basic incomep‾∂Q∂R$$\sum \frac{{\overline{\text{w}}}_{\text{L}}+\text{basic\ income}}{\overline{\text{p}}\frac{\partial \text{Q}}{\partial \text{L}}} > \frac{{\overline{\text{w}}}_{\text{R}}+\text{basic\ income}}{\overline{\text{p}}\frac{\partial \text{Q}}{\partial \text{R}}}$$It is worth noting that nothing changes for the single firms: as Condition 1) still holds, the single firms will not hire workers if their wage–productivity ratio is above that of robots. But there is a crucial difference if the government provides an unconditional basic income: the lower subsistence bound for wages would no longer exist, since the basic income guarantees subsistence for the workers. Consequently, the workers’ reserve wage could easily fall, at least for a part of the workforce, to a level low enough to guarantee a human wage–productivity ratio below that of robots, thus making it advantageous for firms to hire workers. Provision of a universal basic income would make the first scenario inevitably collapse into the second, eliminating the possibility of full unemployment and of a world in which production is accomplished by robots alone, even if the final employed/unemployed ratio inevitably depends upon the extent to which the wage decreases.Figure 2 shows the simplified schematic functioning of the resulting scenario for the economy as a whole: production is accomplished by robots and humans; a certain number of humans are employed and their families receive the basic income plus a wage; entrepreneurs’ and the robot owners’ families receive the basic income plus profits/rents; the basic income is financed by taxes paid by firms.Besides the way it functions, a thorough analysis of the basic income necessarily implies evaluation of its impact on the general well-being.4Measuring the Impact of Policies on Well-being: from Keynes to Hedonic Adaptation, Envy, and EscalationActually, behavioral and happiness economics offer an ample range of theoretical tools that economists can use to measure the impact of different policies on well-being. This also applies to the number of behavioral principles that could be used to address the impact of introducing an unconditional basic income on well-being. However, limiting analysis to the major principles, or at least those that appear the most relevant, and inevitably making some subjective choices, the focus here is limited to a subset of them. In particular, for our specific topic (i.e., policies to counteract the impact of technological unemployment), it can be useful to depart slightly from behavioral economics “stricto sensu” and also consider Keynes’s 1930 contribution. But to begin with we must briefly consider what we mean by well-being and what are its causes.4.1Well-being and its CausesSubjective well-being, also often referred to by economists as “happiness”, is a multidimensional concept that encompasses a number of possible meanings, and is much more complex than the old (and today almost abandoned) concept of utility. According to Nikolova and Graham (2020, pp. 4–6, but see also Bentea, 2019, pp. 258–261):Subjective well-being has three separate but related dimensions – affective (hedonic), evaluative, and eudaimonic […]. Affective subjective well-being refers to temporary experiences of emotions – positive ones such as happiness or joy, or negative ones, such as stress, anger, and sadness. Such positive and negative feelings and emotions are usually short-term and are influenced by the immediate circumstances and states of being and doing. […] Evaluative well-being, in contrast, is a judgment about one’s overall life circumstances and requires reflecting on life as a whole. Evaluative well-being also relates to specific assessments of life domains, such as work, family, housing, income, and the living standard. […]. Eudaimonic well-being is broader than having meaning and purpose in life and refers to the process of living well and having aspects, such as competence, autonomy, personal growth, and relatedness.Although the analysis proposed here refers mainly to affective subjective well-being, also often defined simply as “pleasure attainment and pain avoidance” (Bentea, 2019, p. 258),According to Bentea, 2019, p. 261, “[h]appiness appears when positive emotions… are frequent and intense while negative emotions are minimal and rarely experienced”.it inevitably also takes into account elements belonging to the two other dimensions of happiness (i.e. evaluative well-being and eudaimonic well-being) which are related to affective well-being.Accordingly, and according to some of the economists who focus their theoretical interest on these themes (see, e.g. Pugno, 2017, 2019, but also Scitovsky, 1992), one of the key determinants of well-being is change: “the person exercises her cognitive, social and emotional skills, with the purpose to search for the activities and challenges that give the best opportunities to develop her talents and aptitudes” (Pugno, 2019, p. 168), so that one can “describe human development by going beyond the alleviation of people’s deprivations to take account also of the appreciation and exploration of novel experiences that challenge and develop their inner skills…” (Pugno, 2017, p. 1178). In particular, happiness is considered as linked to the choice of activities that are characterized “by novelty, i.e. by the pleasure of learning” (Pugno, 2017, p. 1178). Within such a perspective novelty, different new behaviors, creative goods in Scitovsky’s idea, and more in general change are the main elements that can generate happiness.To make change possible, and to extract well-being from it, it is essential to have the right skills, since “people can increase the amount of happiness in their pleasant life when they have the skills to grow and multiply their pleasures” (Bentea, 2019, p. 261), and often (even if not always) a growing income is also needed (see again Pugno, 2019, p. 168). In these respects, the role played by the government is crucial, both for supplying citizens with (and/or financing acquisition of) an adequate skill endowment through formal and informal education, and, as we will see in detail a little later on, for providing them with an adequate income (possibly rising over time).It is worth noting that changing is not easy: a number of antagonistic forces, which we can summarize in the psychological principle of “preference for the status quo”, can hamper change. Scitovsky (1992), albeit in a different context and with rather different implications, discussed these forces when distinguishing between the pursuit of “comfort” and “novelty” activities, and considered the pursuit of comfort as a less effective way of achieving well-being than novelty activities and also as an obstacle to these better practices (see, e.g. Pugno, 2017, p. 1188).4.2Keynes’s Short 1930 EssayNow, returning to technological unemployment, it is worth recalling that Keynes discussed some of the well-being problems that could derive from technological unemployment and the possible remedies in his short 1930 essay “Economics Possibilities for Our Grandchildren” (Keynes, 1930). He described the hypothetical situation that might come about through continuous technological progress, 100 years later, in 2030. According to Keynes, “assuming no important wars and no important increase in population, [the struggle for subsistence] may be solved, or be at least within sight of solution, within a hundred years” (Keynes, 1930, p. 4), when fewer workers will be needed to guarantee an improving standard of living for the whole population. According to Keynes, the reduction in labor demand would not be a problem, since the resulting technological unemployment could be counteracted, or at least postponed, thanks to a progressive reduction in per-capita working hours. However, he explicitly admits that some psychological issues might nonetheless arise if humans are deprived of the need to work to receive subsistence. According to Keynes, there are two main issues.First, a possible “nervous breakdown of the sort which is already common enough in England and the United States amongst the wives of the well-to-do classes […] who have been deprived by their wealth of their traditional tasks and occupation” (Keynes, 1930, p. 4). Second, the impossibility of pursuing relative needs, i.e., those needs “that are relative in the sense that we feel them only if their satisfaction lifts us above, makes us feel superior to, our fellows” (Keynes, 1930, p. 4). These needs arise from our desire for superiority and can be satisfied by consuming the goods now commonly known as positional goods.In his essay, albeit in a rather preliminary form, Keynes anticipated some of the behavioral economics constructs that can be used to discuss the consequences of technological unemployment (and, at least implicitly, of the provision of an unconditional basic income) on well-being.4.3Loss AversionThe first behavioral economics construct that can be used to address the issue is loss aversion.See, e.g. Tversky and Kahneman (1991); Kahneman et al. (1991); D’Orlando and Ferrante (2009 and 2022; and D’Orlando, Ferrante, and Ruiu (2011).According to Kahneman, Knetsch, and Thaler (1991, p. 199), “[a] central conclusion of the study of risky choice has been that […] changes that make things worse (losses) loom larger than improvement or gains.” The idea that bad events bear more weight than good ones is confirmed by a number of contributions on similar behavioral principles (see, e.g. Knetsch, 1989; Knetsch & Sinden, 1984; Kahneman, Knetsch, & Thaler, 1990; Samuelson & Zeckhauser, 1988). These psychological principles hold important consequences for calculation of the true costs of unemployment, suggesting that unemployment episodes are events whose negative impact on well-being is not even fully offset by positive events such as rehiring. Although this does not mean that the psychological loss deriving from unemployment cannot be offset by an unemployment benefit, it does however imply that the benefit would have to outweigh the income loss net of the disutility of work, the gross damage being the sum of both the monetary and psychological costs.4.4Hedonic AdaptationHedonic adaptation is another behavioral economics construct essential for the purposes of the present article and fully compatible with loss aversion.See, e.g. Brickman and Campbell (1971), Brickman, Coates, and Janoff-Bulman (1978), Clark and Oswald (1994), Clark (1999), Diener, Suh, Lucas, and Smith (1999), Frederick and Loewenstein (1999), Di Tella, MacCulloch, and Oswald (2003), Clark et al. (2004), Stutzer (2004), Layard (2005), Oswald and Powdthavee (2006), and D’Orlando and Ferrante (2008, 2009 and 2022.Hedonic adaptation is founded on the empirical finding that people adapt to life events: “[l]ife events such as marriage, loss of a job, and serious injury may deflect a person above or below [her/his] setpoint, but in time hedonic adaptation will return an individual to the initial setpoint” (Easterlin, 2003, p. 1). Similarly, consuming a good or undertaking a consumption behavior increases well-being in the first phase, but later the consumer suffers habituation and her/his well-being returns to the baseline level. It is still debated whether adaptation is complete or incomplete, i.e., whether life shocks have a permanent effect on subjects’ well-being in the long run: some authors maintain that they can approach but never return to their baseline, so that an irreversible loss would persist for negative life events and an irreversible gain for positive life events.For discussion of complete and incomplete adaptation and the setpoint hypothesis, see Easterlin (2003) and Lucas, Clark, Diener, and Georgellis (2003).In particular, episodes of unemployment are considered among the most drastic cases of irreversible loss (see, e.g. Clark & Oswald, 1994; Clark et al., 2004; Frey & Stutzer, 2002), with well-being remaining below the baseline, even if the subject, after being fired, obtains a new job at the same wage as the old one (D’Orlando & Ferrante, 2018, p. 1264) – a finding perfectly consistent with the implications of loss aversion discussed above.4.5Positional Effects and EnvyWhen discussing the possibility that social position or, more in general, social interaction, plays a role in determining subjective well-being, theoretical interest inevitably focuses on envy.Reviews of the economics literature on envy can be found in Chaudhuri (1985), Hammond (1989), Mui (1995), and Kolm (1995).According to Parrott and Smith (1993, p. 906), envy “occurs when a person lacks another’s superior quality, achievement, or possession and either desires it or wishes that the other lacked it.” According to D’Orlando and Ricciotti (2021, p. 121), “[i]n these cases, subjects’ well-being depends upon their position in a somewhat defined economic hierarchy, and hence on others’ behaviors, income or wealth.” It is worth emphasizing that envy is not solely linked to material possessions, since people can envy others not only for what they have, but also for what they are: according to Goel and Thakor (2005, p. 2263) “we may envy another person’s success, happiness, intelligence, health, good looks, material possessions, power, title, job, or status”. Furthermore, theoretical contributions (see again Goel & Thakor, 2005, p. 2257–2258) often strongly maintain that envy is not inequality aversion, since envious people do not behave symmetrically: subjects experience a reduction in well-being when they are worse off than others, but they experience an increase in well-being when they are better off than others. Say it another way, they do not dislike inequality in general, but dislike being the losers in the inequality game.Envy is generally considered a negative emotion; in particular, Frank (2005) has discussed the negative externalities caused by this and similar positional effects, with conspicuous consumption playing a major role. Indeed, we can attribute to envy at least four possible negative consequences: (i) negative externalities that generate well-being losses for envious people (i.e. for people suffering from passive envy); (ii) the risk that envious people set about trying to make the others lose their possessions (disruptive or malicious or black envy); (iii) the possibility that envious people, by imitating the behaviors of their reference group, replicate and spread socially harmful behaviors through society (e.g. conspicuous consumption); (iv) the possibility that failing in imitating the consumption behaviors of the reference group negatively affects subjects’ self-esteem.However, some authors attribute a crucially positive role to the desire to be envied: for example, Veblen (1899 1970, p. 32) considered the “desire to excel in pecuniary standing and so gain the esteem and envy of one’s fellowmen” one of the main “incentives to acquisition and accumulation.” Furthermore, envy can also be seen as a broader concept that encompasses positive emotions such as admiration or esteem for people who accomplish socially valuable actions (such as environmental protection behavior, altruistic behavior, excellence in education or sports).In general, we can attribute to this broader concept of envy at least six possible positive consequences: (i) gains in well-being for envied people (i.e. for those who enjoy active envy); (ii) the possibility that envy will encourage subjects to strive to reach the others’ positions (constructive or benign or white envy); (iii) the possibility that envious people, by imitating the behaviors of their reference group, replicate and spread socially valuable behaviors through society (if, of course, the behaviors of the reference group are socially valuable); (iv) the possibility that succeeding in imitating the consumption behaviors of the reference group may positively affect subjects’ self-esteem; (v) the possibility that envy may stimulate change in consumption behaviors and so raise well-being (consistently with the observations on the main determinants of happiness proposed in Section 4.1 above); (vi) finally, the possibility that reaching the others’ position, or generating envy in others, may slow down, stop, or even reverse the habituation process.The last point in particular has been discussed by D’Orlando and Ricciotti (2019, p. 19): “until the time when others succeed [in imitating a] subject’s higher-level consumption behavior, the subject will continue undertaking a consumption behavior at a higher-than-average level, i.e., continue generating envy in others. This circumstance increases well-being by preventing (at least partially) the reduction caused by hedonic adaptation: my car is still the fastest on the highway, I pass all other cars and I am always, every day, happier and happier by generating envy in their drivers who unsuccessfully try to resist.”It is worth recalling that what improves a subject’s well-being is not envying others, i.e., suffering passive envy, which actually reduces well-being, but the possibility of lowering or eliminating passive envy by succeeding in imitating the reference group’s behaviors; and/or the capacity of increasing well-being by succeeding in changing behaviors to generate active envy, i.e., succeeding in being envied (esteemed, admired) by others.4.6EscalationFinally, a crucially useful behavioral economics construct is escalation. According to D’Orlando and Ricciotti (2021, p. 109):“[e]scalation occurs when over time people show a growing interest in stronger, harder, faster, riskier goods, substances, or behaviors. The key concept here is ‘more intense sensations’. Escalation is not simply a desire for novelty, that is, a desire for consuming a different good or engaging in different behavior. Instead, it requires a shift from goods, substances, and behaviors that give the consumer ‘less intense sensations’ to those which provide ‘more intense sensations’. People begin engaging in basic consumption behavior within a certain consumption type and are satisfied with this behavior. [But] after a while, they become less satisfied with it, which, as time passes, leaves them with less well-being. Thus, they become more interested in other consumption behaviors, which are stronger, harder, riskier, and faster.”Examples of escalating behaviors can be observed in drug consumers who begin with soft drugs but later escalate to harder ones; in wine consumers who progressively escalate to better wines; in skiers who initially prefer green slopes but, on becoming more expert, choose more difficult black slopes; and in swinging couples who start with exhibitionism and later escalate to intercourse with others (see, e.g. D’Orlando & Ricciotti, 2021, p. 109). It is worth noting that in all these cases, and more in general in the entire analysis presented in this article, we will refer to the most general definition of consumption to be found in theoretical literature, i.e. “the satisfaction of our wants by using goods, services, and even time” (D’Orlando & Ricciotti, 2021, p. 109), thus separating consumption from monetary expenditure.Escalation is founded on two different constructs: the reduction over time of the well-being derived from currently undertaken consumption behaviors, a phenomenon consistent with hedonic adaptation; and the increase over time in well-being derived from alternative consumption behaviors, a phenomenon consistent with the idea that change raises well-being. However, escalation is not simply pursuit of novelty since, in the case of escalation, the alternative consumption behaviors capable of increasing well-being are “stronger, harder, riskier, faster,” i.e., more intense, not simply different. And to be undertaken and enjoyed these “more intense” consumption behaviors often call for an adequate endowment of skills, and thus processes of skill accumulation. These processes are usually informal, or can rely on learning by doing, but in some cases formal education is necessary, at least to provide the subjects with the basic skills on which skill accumulation can be built. And if formal education is necessary, public intervention may have a role, at least as a supplier of basic skills.Although escalating to superior consumption activities or behaviors is not necessarily a consequence of increasing income (as in the case of escalating from skiing on green slopes to skiing on red slopes), sometimes a higher income is needed (as in the case of escalating from driving a slow car to driving a sports car) because more intense behaviors may necessitate a higher income.Actually, escalation can be hampered by the action of some counteracting forces, mainly relating to the behavioral principles of preference for the status quo and endowment effect, and most often these counteracting forces affect the less skilled and less wealthy subjects (on these points, see D’Orlando & Ferrante, 2018, 2022). There are, however, various other counteracting elements: more intense acts of consumption may be associated with higher monetary costs; there may be higher psychological costs associated with the same activities (due to social stigma, preference for the status quo, conformism, risk aversion, aversion to ostentation, frugality, etc.); subjects may expect to obtain low levels of increase in well-being through more intense consumption due to the difficulty of acquiring the skills necessary for making changes; etc. (for a detailed analysis of these forces, see D’Orlando & Ricciotti, 2021). Generally, these counteracting forces grow stronger as consumption behaviors become more intense.Ultimately, the actual realization of escalation depends on the relative strength of the forces driving change versus the forces hampering it.5Behavioral Economics and the Basic IncomePech (2010) is one of the few scholars who use behavioral economics to discuss the consequences of providing a basic income. In particular, Pech focuses on four major points: the preferable way to finance the basic income (i.e., the level of progressivity in taxation); the impact of the basic income on the wages of “good” and “bad” jobs; the impact of the basic income on cooperation and reciprocity; and the impact of the basic income on conspicuous consumption. However, Pech does not use behavioral economics to conduct analysis of the impact of the basic income on well-being, nor does he refer to the basic income as a mean to counteract technological unemployment. Unlike Pech’s contribution, in the present article constructs of behavioral and happiness economics (particularly, those described in the preceding section) are used to discuss the impact of the basic income on well-being in a scenario of high levels of (technological) unemployment.As discussed above, provision of an unconditional basic income high enough to ensure subsistence could play two different and (seemingly) alternative roles in counteracting technological unemployment: income integration for workers to keep their incomes above subsistence level, even if their wages are below it; or the only income for unemployed humans if robots alone are producing goods. Consequently, the policy purpose of the basic income can be either to increase humans’ productivity–wage ratio, making workers competitive with robots, so that it is in the firms’ interest to hire them, or guarantee subsistence for humans once robots produce (almost) all the goods. However, in Section 3, we showed that such a distinction is more philosophical than practical, since, if the government provides an unconditional basic income to all citizens, the citizens can in any case decide whether to work or to shirk, regardless of the original purpose of the policy. In the general case, a certain number of workers will therefore choose to remain employed. For this reason, here a scenario is depicted in which the government provides an unconditional basic income to all citizens disregarding the government’s reasons for doing so, and this scenario inevitably coincides with scenario 2 (and Figure 2) of Section 3.The above scenario will result in four categories of human beings: (i) those who receive the basic income and voluntarily choose to stay unemployed; (ii) those who receive the basic income and voluntarily choose to work on a low wage; (iii) entrepreneurs who manage firms; and (iv) capitalists who own robots.The supply of an unconditional basic income affects the well-being of these categories of subjects in different ways.To discuss these ways, let us begin by summarizing the circumstances that, in the light of the behavioral and happiness economics constructs described in Section 3, increase or reduce people’s well-being, and then go on to discuss whether provision of a basic income affects these circumstances.Let us start with hedonic adaptation. According to this approach, subjects’ well-being tends to return to the baseline after a (positive or negative) shock, but in the case of unemployment episodes, as discussed in Section 4.4 above, there is a general consensus that habituation may prove incomplete: each unemployment episode generates an irreversible loss that tends to cumulate – even if the irreversible loss is progressively smaller with each episode due to long-run hedonic adaptation. Furthermore, workers remain below the set-point even if they are re-hired. Therefore, to remain steadily above the baseline, subjects have not only to avoid negative shocks, particularly unemployment episodes, but also continuously to enjoy positive shocks to compensate for habituation. These positive shocks are mainly associated with the pursuit of novelty in consumption, i.e., changes in consumption behaviors that can halt and reverse the habituation process.In most circumstances, a different and better strategy (a strategy that guarantees slower habituation and greater increase in well-being) consistent also with the escalation approach is about not simply continuously changing consumption behaviors but escalating to higher-grade consumption behaviors. These higher-grade consumption behaviors do not necessarily call for greater expense, but, in most cases, a higher income makes thing easier. Furthermore, obtaining a higher income represents, in itself, an escalating behavior.However, neither changing to different consumption behaviors nor escalating to higher grade consumption behaviors are viable strategies for all subjects and in all circumstances: in most cases changing and escalating require an adequate skill endowment, and/or a process of skill accumulation, and/or adequate education, to make the new choices practicable and capable of positively impacting on subjective well-being, even in the cases in which a rising income is unnecessary. For low-skilled and low-income workers the forces hampering change can be hard to be overcome, so that preference for the status quo may lead them to pursuit of income security and stability in their consumption patterns – behavior that prevents them from enhancing their well-being.Extending analysis to social interaction, a crucial role is also played by positional externalities. In Section 4.5, we discussed in general some of the possible negative and positive consequences of envy. Here, we focus on the role that envy may play in driving change in consumption behaviors in general, and change towards higher-grade consumption behaviors in particular.It is true that in some cases envy generates negative social externalities that lower well-being. However – and even over and above the possible role of envy in the diffusion of socially valuable behaviors (see again Section 4.5 above) – given the loss in well-being, the circumstance that many subjects succeed in escalating to superior consumption behaviors, reducing/eliminating passive envy, enjoying active envy, raising their self-esteem, etc., can reduce the loss and, in some cases, might even fully compensate for it (and also for hedonic adaptation). Indeed, generally speaking, we have no reason to assume that the social loss deriving from negative envy externalities must inevitably be greater than the subjects’ gain. Furthermore, any policy aiming to eliminate envy, for example by endowing all subjects with the same income and consumption behaviors, would eliminate passive (and active …) envy, but not hedonic adaptation, with the consequence that subjects would not be able to remain steadily above the baseline. Thus a situation in which nobody can escalate is worse than one in which everybody can escalate (and possibly worse than one in which many can escalate).Here we must stress the role played by social interaction in the above-described dynamics and in particular in driving change. Since well-being increases (or at least does not return to the baseline) when subjects change their consumption behaviors, and the more often they change, the longer they stay above the baseline, positional concerns can enhance well-being by driving change. By dint of imitating others’ behaviors, suffering passive envy and observing the behaviors of the reference group, subjects are forced to change their behaviors. Social interaction might therefore be seen as one of the possible means that can be used to stimulate change and so to increase well-being (or reduce well-being losses): whether or not the change can fully compensate for the negative (external) social consequences of envy, changing at least reduces these negative consequences for the subject who changes, enhancing her/his situation; by contrast, if the subject does not change, she/he will endure the full impact of passive envy without being able to attenuate or compensate for it. These considerations may also have consequences for one of the main limits of the basic income, namely, its asserted negative impact on the incentive to find work and on work effort: positional concerns may generate the necessary non-monetary incentives to work when monetary incentives fail.We can now go on to discuss the way in which the behavioral economics constructs and dynamics discussed above affect the impact of the basic income on well-being and, conversely, whether better remedies exist. Note that while it is certainly true that receiving the basic income is better than receiving nothing if people are faced with technological unemployment, to evaluate whether the basic income is in fact the best policy to counteract technological unemployment we must look into all its drawbacks.Let us assume that, as soon as technological unemployment arises, the government begins to provide all citizens with an unconditional basic income, financing it by taxing firms. As discussed earlier, we will have: people who have lost their jobs remaining unemployed, and receiving the basic income; those who remain employed and continue working, receiving their wages plus the basic income; entrepreneurs, who manage firms and gain profits as well as receiving the basic income; capitalists, who own robots and receive the rents plus the basic income. In this situation, when workers lose their jobs they suffer a shock that drives them below the baseline level of well-being, and we know that unemployment is a kind of shock that may leave an irreversible long-term loss in well-being. So they will remain steadily below the baseline. And there is no telling whether the basic income will be high enough to compensate for the losses in well-being deriving from unemployment shocks, i.e. there is no certainty that the basic income will be high enough to enable unemployed workers to regain the baseline. Nevertheless, regardless of whether the basic income is high enough to restore the unemployed to the baseline, from then on they will never be able to rise above the baseline. This is so since they cannot escalate to higher-grade consumption behaviors if these behaviors necessitate a higher income; they suffer passive envy toward the employed who earn wages as well as receiving the basic income if replicating the consumption behaviors of the reference group requires a higher income; they suffer the same passive envy toward entrepreneurs and capitalists; they will not succeed in being envied by others; they suffer a loss of self-esteem for having lost their jobs and being unemployed; they suffer for not being able to realize their aspirations, if this entails a higher income; they might suffer social stigma for being unemployed; they suffer on recognizing that they have no role in society; and they are fully aware that they will never be able to escalate or realize their aspirations. The only chance they have of keeping above the baseline is by engaging in new consumption behaviors that do not call for a higher income but can nevertheless generate esteem, admiration or, more in general, envy: this is certainly possible, but, generally, a higher income would help, and a higher income would also help if change (or escalation) requires costly investment in skill (which could actually be facilitated by public intervention with training and education schemes). By contrast, the employed will have ample opportunities to remain steadily above the baseline, since they do not suffer loss in well-being for having lost their jobs, while obtaining an increase in income (and this is a positive shock that drives them above the baseline); do not suffer problems of self-esteem, while generating envy in others; and indeed they know that they can escalate. They might only suffer passive envy with respect to entrepreneurs and capitalists. Consequently, contrary to the general expectations, provision of a basic income to unemployed humans prevents them from changing their consumption behaviors, i.e., prevents them from escalating, and hence prevents them from reducing inequality in well-being. In other words, the introduction of a basic income may, far from reducing inequality in well-being, actually aggravate it.Furthermore, a universal basic income that remains unchanging over time can hardly generate constructive (or benign, or white) envy since a fixed income may prevent subjects from straining to reach the others’ positions; on the contrary, it can easily generate disruptive (or malicious, or black) envy since the impossibility to imitate others’ behaviors (if these behaviors require a higher income) may drive envious people to try to make the others lose their possessions. In these circumstances, a universal basic income that remains unchanging over time may also generate social instability.The role played by envy, and more generally by all the positional concerns that impact on well-being, is fully consistent with Keynes’ (1930) conclusions on the impossibility of pursuing relative needs. However, again in line with Keynes’ conclusions, boredom, too, can play a role in lowering well-being for unemployed humans. This idea is inconsistent with the traditional analysis of labor supply, which assumes that workers choose whether to work or not by referring to labor marginal disutility and hence implicitly assuming that working reduces well-being. However, according to Keynes, boredom, i.e., not working, could also negatively impact well-being; thus, the analysis to be developed is slightly more complex than generally assumed. In terms of traditional (neoclassical) microeconomics, we might say that working may generate both disutility and utility, or at least that not working may generate disutility.6Alternative (and Complementary) PoliciesThe weaknesses of the basic income as a tool to counteract technological unemployment leaves room for alternative solutions. Here, only two of the possible alternatives will be discussed: modifications to be made to the unconditional basic income policy and policy measures that can be implemented together with the basic income. The policy measures alternative to the basic income will not be discussed.For a thorough discussion of these possible alternatives see, e.g. D’Orlando (2020a and 2020b) and Marchant and Stevens (2017).6.1Implementing a Rising Basic incomeAs pointed out above, subjects can achieve increase in their well-being and remain steadily above the baseline by continuously changing their behaviors (pursuit of novelty), by choosing higher-level behaviors (escalation), by attempting to realize their aspirations (being aware that this is a feasible goal), by provoking the envy of others, by engaging in the consumption behaviors of the reference group, etc. In such a context, passive envy can be a strong stimulus to change and so may exert a positive impact on well-being. This pressure to change may be hampered by provision of a fixed basic income when changing requires a higher income. Therefore, avoiding a fixed basic income could be one possible solution to remedy some of the weaknesses in this policy. The basic income could, for example, correspond to the subsistence level (and so it should vary with regional patterns of consumption and the cost of living) but could rise over time, or rise as subjects engage in valuable activities or accumulate human capital, or whatever goal society sees as socially significant.The idea that the amount of the basic income to some extent depends upon the participation of individuals “in activities that enhance community, civic and cultural development” is to be found in Ford (2009, pp. 175 ff.) and is consistent with the badge system proposed by Marchant and Stevens (2017).The basic income could also be designed so as to allow people to use it to accumulate wealth, which could be bequeathed to heirs, or to acquire shares in firms, thus acting as a strong stimulus to change and escalation. In other words, the basic income could be designed in such a way as to allow for change by eliminating budget constraints or letting subjects reduce them, whereas the stimulus to change is generated by positional concerns.Apart from its consequences on individual well-being, a rising basic income would also bear another (aggregate) positive consequence, i.e. the rise of white envy over black envy: we have seen that a universal basic income that remains unchanging over time may diffuse a disruptive attitude among envious people and generate social instability; on the contrary a rising universal basic income may diffuse a constructive attitude among envious people that can enhance social stability.Two alternatives measures that could be implemented together with the basic income are reduction in per-capita working hours (which was also Keynes’ original 1930 proposal) and the imposition of mandatory quotas on firms regarding the number of workers to hire.6.2Reducing per-capita Working HoursReducing per-capita working hours has been, and to some extent still is, the most popular measure that economists have proposed to counteract technological unemployment. According to Vivarelli (1996), the “end of work literature”, based on the idea that technological progress might be able to completely displace human employment, is faulty since it fails to consider the historical reduction in per-capita working hours and its likely future evolution: “the weak point of this analysis is the lack of any consideration of the historical tendency towards a continuous decrease in per-capita annual working time; for instance, it has been shown that a manual worker used to work about 3000 h at the beginning of this century, whereas the average annual working time ranges nowadays between 1500 and 1900 h according to different national and institutional settings” (Vivarelli, 1996, p. 26).Albeit being historically robust, Vivarelli’s idea appears sound insofar as it takes into account what has happened in the past, but far less persuasive with regard to what might happen in the future, in a scenario seeing a proliferation of robots and artificial intelligence. In this respect, past experience cannot be considered a valid basis to forecast what will happen in the future. This is also true of a policy designed to reinforce, or accelerate, the historical tendency towards a decrease in per-capita working hours. Indeed, such a policy might be useful in coping with the increasing unemployment which could characterize the transitory phase, when robots begin to enter the productive process; but it would be completely useless in terms of effectively counteracting the third wave of technological unemployment, when robots endowed with artificial intelligence enter into most productive processes with the potential capacity to carry out all the tasks that humans do in these processes, but with a higher productivity–wage ratio. In such a context, unless the lower subsistence bound that prevents decrease in wages is removed, firms will in any case prefer to hire only robots rather than humans. And reducing per-capita working hours does not remove the bound. This policy, then, does not in itself counteract technological unemployment, even though it may be useful when combined with other measures, as we will see a little later on.6.3Imposing Mandatory Quotas for Workers to Be HiredHowever, both introducing a basic income policy (with the aim of lowering wages and boosting employment) and implementing a policy of reduction of per-capita working hours, relies excessively on the functioning of market mechanisms, and thus leaves the level of employment that can be achieved uncertain. Imposing mandatory quotas on firms for workers to be hired might therefore appear a better strategy, since it enables policymakers to achieve their employment goals with a fair degree of certainty. The problem with such a policy is that it jeopardizes the firms’ efficiency, and even their competitiveness, as indeed most intervention policies do. It might therefore be preferable to design such a policy in a way that mimics Tietenberg’s traditional tradable permits approach to environmental economics.On the tradable permits approach, see Baumol and Oates (1971), Montgomery (1972), and Tietenberg (1990, 2003.In his approach, pollution quotas are allocated among firms, which can either comply with these quotas or sell/buy them.The same theoretical framework and the same policy tools can be used to cope with technological unemployment problems.Tietenberg’s original approach works as follows. With the most widespread of the tradable permits programs, i.e., cap-and-trade programs,The others being credit programs. See Tietenberg, 2003, p. 408.the policymaker decides the maximum resource access limit (for example, the maximum level of air pollution that firms can produce) and allocates (free of charge or based on auctions in which firms pay more for greater quotas) permits to access the resource (for example, the quantity of CO2 that each firm is allowed to produce) among users. The firms can then trade quotas. With this procedure it would be possible to “minimize the cost of reaching a predefined environmental target” (Tietenberg, 2003, p. 401).“A principal theorem of environmental economics demonstrates that, under specific conditions, an appropriately defined tradable-permit system can minimize the cost of reaching a predefined environmental target (Baumol & Oates, 1971). In a perfectly competitive market, permits will flow toward their highest-valued use. Those that would receive lower value from using the permits (owing to lower abatement costs, for example) have an incentive to trade them to someone who would value them more. The trade benefits both parties. The seller reaps more from the sale than s/he could from using the permit and the buyer gets more value from the permit than s/he pays for it” (Tietenberg, 2003, p. 401).In the case of technological unemployment, things might go in a slightly different way. The policymaker would not set an upper bound to robot use, as would be the case in a cap-and-trade framework, but would instead set a lower bound to the number of workers to be hired by each firm. Such a procedure would be necessary if the government’s goal were to increase human employment, and not to reduce robot employment, considering also that it is really difficult to decide what exactly a robot is and so identify an objective standard for measuring the number of robots a firm is employing (this looks rather like an updated version of the traditional “capital measurement critique”). Therefore, the government first decides its employment goals and then allocates quotas of workers to be hired among the firms (free of charge or based on auctions in which firms pay proportionately more for smaller quotas).Just as in the case of environmental policies, the main problem here is how to determine the level of employment starting from which each firm is to increase it (with the pollution-reduction cap-and-trade approach, it was the level of pollution starting from which each firm had to reduce its pollution). In the traditional environmental cap-and-trade approach, the method that has been used most often is historical, i.e., based on the level of the use of the resource (for example, the level of air pollution) generated by a firm in the course of the last so many years. The same procedure might be used in the case of technological unemployment: each firm starts from its past level of labor hiring (or, better, the average of its labor hiring over a couple of years, to avoid opportunistic behaviors), which represents its initial allocation of rights, and must raise the level of its employment by a given percentage, representing its quota. Firms then have the option to trade their quotas. In this case, sellers pay for selling quotas of workers to be hired, and buyers are paid for buying quotas of workers to be hired. The possibility of trading quotas means increased efficiency for the firms, and indeed for the economic system as a whole:Firms or industries, for which the hiring of human workers is less profitable, would sell quotas to firms or industries for which the employment of human workers is less disadvantageous. Paraphrasing Tietenberg, quotas will flow toward their lowest-loss use. Those that would suffer a higher loss from being forced to hire humans have an incentive to trade them with those who suffer lower losses. Such trade benefits both (D’Orlando, 2020b, p. 22).Adopting a policy based on tradable quotas, firms as a whole would have to hire humans in the amount decided by the government, but human employment would be concentrated where it had the least negative effect on productivity.Furthermore, employed workers would have the opportunity to increase their income and hence enjoy pursuit of novelties, escalate, suffer/enjoy passive/active envy, realize their aspirations, etc.Another appreciable positive consequence of the policy described above concerns investments. If firms are through some mechanism or another forced to hire humans, as in the case of a Tietenberg-inspired policy, when they revamp their factories (or build new ones), they will create plants compatible with human employment, and not plants that can only be used by robots. On the other hand, if firms are free not to hire human workers, as would be the case with different intervention policies, they might concentrate investment on plants compatible with robot rather than human use. Thus, the long-term result of Tietenberg-like policies is the presence of factories in which humans can work, whereas, if other policies are implemented, there is a real risk that in the long-run there would only be robot-specific factories, with no room for human labor.However, the policy described above is exposed to the same weaknesses that mar environmental protection policies: firms risk losing their competitiveness. This is inevitable, since they have either to hire humans paying wages (in relation to productivity) higher than the cost of renting robots or pay a sum on top of the cost of renting robots to sell their quotas. And if they are exposed to international competition they may well find themselves unable to survive on the market due to excessive production costs. Therefore, as has turned out to be the case with environmental policies, this is not a policy for a single country: it has to be implemented by all countries, negotiating international treaties committing single countries to enforcement of the quotas. The problems that these types of international treaties have come up against over time (take the case of the Kyoto Protocol) would probably replicate with the application of employment quotas. There could, however, be one minor advantage if the international treaties, as in the case of environmental treaties, were to regulate the international trade in quotas, enabling the whole mechanism to work better.6.4A Comprehensive Policy MixThe basic income could be successfully combined with either of the policies described above.Let us first discuss the possibility of combining a reduction in per-capita working hours with provision of an unconditional basic income. In this case, as extensively discussed above, given the security of an unconditional basic income, wages could fall below subsistence, while nevertheless leaving humans the possibility to accept to be hired (since the basic income ensures the workers’ subsistence in any case). We also know that if the wages fell to the extent that the humans’ productivity–wage ratio rose above, or at least equal to, that of robots, it would be in the interest of firms to hire humans. Moreover, reduction in per-capita working hours, given the amount of humans’ worked hours demanded by firms, could allow firms to hire a greater number of humans than they would have if working hours were not reduced. Thus many human workers would be able to increase their income by adding the basic income to the wage, so as to be able to enjoy pursuit of novelties, escalate, etc., that is, to increase their well-being and reduce inequality in well-being.The second possibility is to combine quotas with provision of an unconditional basic income. Again, as in the previous case, the basic income could benefit both the single firms and the human workers: the former could hire workers instead of robots, paying wages low enough to make the productive–wage ratio of humans higher, or at least equal, to that of robots, while the latter could accept such wages since the basic income ensures subsistence in any case. But what if the wage does not fall low enough, or there is only limited increase in employment? The policymaker might force firms to hire humans by setting quotas of human workers to be hired. And firms could decide whether it was in their interest to hire workers paying wages below subsistence or hire robots alone, selling quotas and paying lower robot remuneration (since the reduction of human wages boosts competitive forces, and so reduces robot remuneration): this circumstance would be particularly important for firms operating in sectors where the cost of mandatory labor hiring was too high with respect to labor productivity. In any case the actual well-being-enhancing effect of this process depends upon how high the wage offered by firms actually turns out to be: it has to be high enough to allow wage-earners to undertake different (higher ranking) consumption behaviors.In any case, an alternative policy could be based on a combination of all the (three) measures discussed in this section: a basic income rising over time could guarantee that people remained steadily above the baseline and allow wages to fall below subsistence level, so that firms could profitably hire workers; a reduction in per-capita working hours could guarantee (nearly) full employment; setting quotas and making provision for trade in them could drive the system to the desired employment level and help reduce the cost of hiring humans for firms that had most to lose with the obligation to hire humans.However, in addition to implementing the three policies described in this section, the government should also provide citizens with an adequate endowment of basic skills, through basic education, and when possible also supply (or promote) programs for further formal and informal lifelong education. In fact, without the support of the basic skills and lifelong skill accumulation programs subjects would have difficulty in changing and/or escalating, due to the high and possibly rising levels of skill endowment that change/escalation requires, together with the possibly rising cost of skill accumulation processes and the rapid skill obsolescence caused by today’s impetuous technological progress.7ConclusionsAs discussed above, the third wave of technological unemployment has characteristics and implications radically different from those of the first and second wave. The first two waves reduced human employment in specific sectors (agriculture to start with, manufacturing later on), or lowered wages and/or employment for unskilled workers and/or workers carrying out routine tasks, thus increasing income inequality, with no substantial effect on long-term unemployment; the third might be capable of destroying all jobs for humans (both skilled and unskilled and both routine and non-routine) in most sectors of the economy, generating high levels of (technological) unemployment. This is due to the fact that the first two waves of technological unemployment were caused by the development of more productive machines, while the third would result from the introduction of robots endowed with artificial intelligence in the productive process. Unlike machines, robots do not cooperate with workers but can replace them in many, if not all, of their tasks. If such a scenario does indeed come about, the traditional compensation forces which proved effective in preventing the impact of technical progress on employment in the past will be ineffective due to the relatively high productivity and low cost of robots. This implies that wage reduction should be significant enough to counteract technological unemployment and guarantee human employment, but it will not pass the limit represented by the lower subsistence bound of the wage. Additionally, an increase in demand might lead to an increase in the production of commodities produced by robots alone and hence an increase in employment for robots alone without increasing human employment.The idea of an unconditional basic income is the solution to cope with the resulting high level of unemployment that has come under most discussion. However, despite the ever-growing debate, the impact of this solution on well-being has never been thoroughly thrashed out. Furthermore, when behavioral economics constructs (such as hedonic adaptation, pursuit of novelty, escalation, positional effects, and envy) are used to evaluate the well-being consequences of the basic income, a number of drawbacks emerge. In particular, this measure alone, without any accompanying measures, appears incapable of guaranteeing the process of continuously changing consumption behaviors that allows subjects to remain constantly above the baseline level of well-being, mainly due to the fact that the subjects’ income is fixed and the positive effects of positional concerns and social interaction as a stimulus for change cannot operate, and there can therefore be no escaping from habituation and passive envy. In other words, provision of a fixed basic income favors social immobility and thereby risks increasing rather than reducing inequality in well-being.At this point, let us consider the possible solutions to be implemented together with the basic income. Of these solutions, the most promising are reduction in per-capita working hours and the possible implementation of a “cap and trade” solution (which mimics Tietenberg’s traditional tradable permits approach to environmental issues), which sets quotas of workers to be hired on firms, but with the option to trade these quotas. Combining these two solutions with provision of a basic income which rises over time could guarantee the best results in terms of employment, well-being and firm efficiency.More in general, public policies should be designed to encourage change and escalation in subjects’ consumption behaviors. This could result in policies that stimulate prosocial behavior, policies designed to give people the opportunity to accumulate wealth and/or human capital (for themselves or for their heirs) by obtaining the ownership of firms (or robots), but also policies that allow people to engage in enviable, extreme or innovative consumption behaviors, so as to be envied and imitated by others. Any attempt to stimulate change should be promoted. But change (and escalation) also require an adequate endowment of basic skills on the part of citizens and the possibility to make use of lifelong programs of skill accumulation, also due to today’s rapid skill obsolescence caused by technological progress. Thus, public policies should also aim at guaranteeing adequate education and training programs, while promoting the spread of informal skill accumulation processes.A further conclusion of the analysis presented here is that, in order to minimize losses in well-being, an unconditional basic income (together with its companion policies) should have the main goal of ensuring the maximum possible level of employment, and not of financing the possibility to remain unemployed. The basic income can play this role by allowing wages to fall below subsistence, since workers can accept to be hired on such a low level of wages if subsistence is guaranteed by the basic income. And if wages fall below subsistence the possibility remains that the human wage/productivity ratio may fall below robot rent price/productivity ratio, allowing firms to hire human workers rather than concentrating solely on robots.

Journal

Basic Income Studies – de Gruyter

Published: Jun 1, 2022

Keywords: basic income; technological unemployment; well-being; behavioral economics; envy; escalation; hedonic adaptation; loss aversion; E24; I31; I38; J64

There are no references for this article.