Adaptive taxation stabilizes demographic waves---nuke war AND turns every impact.

For a given supply, the more inelastic the demand for a good, the smaller is the deadweight loss of imposing a tax. The ‘Ramsey rule’ states that more necessary goods should be taxed more heavily than goods with greater demand elasticities.11 This way the total excess burden caused by the introduction of the tax will be minimised: this tax structure would be economically efficient. ‘And extremely regressive!’, you may decry because the price-demand elasticity of a good is higher the more necessary the good, and it is the poor who tend to allocate a higher proportion of their income to ‘basic necessities’. Well, I said ‘under certain assumptions’, one of which is that the economic actor being model is a representative agent: all agents are assumed to be identical, so there are no distributional problems. If we assume there are, say, two groups of individuals—namely, the poor and the rich—and that the former have more inelastic demands because they tend to allocate a higher proportion of their lower income on necessities compared to the other group, Ramsey’s model highlights the tension between efficiency and equity that runs through much of the discussions in public finance and economic policy in general. As Mirrlees et al. (2011, p. 36) put it: ‘Optimal tax theory is all about the choice of a system of taxation that balances efficiency losses against the government’s desire for redistribution and the need to raise revenue.’

When it comes to efficient tax structures, a tailor-made tax on consumption that were able to distinguish between agents would be theoretically optimal: any two agents with different demand elasticities would be taxed differently. Butwhy restrict the analysis to the consumption of goods and services?Ramsey did not concentrate only on the demand side—after all, demand and supply are but the two blades of the same pair of scissors (Marshall 2013, V.III.7, p. 164). In fact, his basic insight was that—assuming linear and separable demand and supply schedules and the absence of income effects—the ratio between the unit tax on a good and its unit price must be equal to the sum of the inverse of the elasticity of demand and the inverse of the elasticity of supply; see Stiglitz (2015).

One key service is labour, so a tax on labour income could also be structured along Ramsey’s rule. Labour income is the product of the number of hours worked and the hourly wages, so the tax rate would be higher for agents with higher labour supply elasticity. Another important source of income, particularly for well-off individuals, is capital income; hence, in theory a tax on interest income could follow similar guidelines. Here we come to one intersection between ageing and taxation: age-adjusted optimal taxes.

2.2.2 Age-Adjusted Optimal Taxation

Ramsey was concerned about efficiency. The levying of taxes (other than lump-sum) introduces deadweight losses or inefficiencies. If the governments could observe things closely associated with efficiency such as effort or skills, they could estimate the elasticities of effort or skills and consequently levy taxes optimally. However, effort or skills cannot be observed directly—this is a key assumption: that the government (or social planner) cannot observe each individual’s labour ability or effort; hence taxes cannot be based on this characteristic, which is assumed to remain ‘private’ information.

We further assume that the government cannot even observe unit wages. Instead, only total income is observable, which conditions the fiscal instruments the government can use.More generally, time is assumed to be available solely either for labour or leisure.12 And leisure cannot be taxed directly, which, according to Erosa and Gervais (2002, p. 339), constitutes the ‘fundamental problem in setting optimal fiscal policy’. If we agree that labour income is positively and closely related to effort and skills, or inversely associated with leisure, then it could constitute an acceptable base for Ramsey-type taxation. Therefore, we need to look into the wage elasticity of labour supply.

Crucially, the elasticity of labour supply varies with age (Peterman 2016; Whalen and Reichling 2017), and the age profiles of capital income show a positive gradient, as the life-cycle hypothesis suggests—see, for example, Brugiavini and Padula (2003), Burtless (2005), and Garbinti et al. (2017). Consequently, the Ramsey model would recommend that income taxes should be adjusted by age. As we mentioned above—see also Chap. 1 in Volume III— older workers tend to exhibit a higher elasticity of labour supply with respect to changes in labour income, particularly along the extensive margin.

In labour economics, there is a distinction between the extensive and the intensive margin of labour supply: the extensive margin refers to the decision of whether to work in paid employment or not; the intensive margin refers to the decision of how many hours to allocate to paid employment, having already decided to work. For men and women the elasticity of labour supply, both at the extensive and the intensive margins, varies along their life cycle (Blundell et al. 2013). As Hemel (2010, p. 1888) explained:

If individual workers cannot make marginal adjustments to their hours, then we would expect the price elasticity of labor supply to be greater for individuals who are considering exit from the workforce than for individuals who might desire an incremental reduction in hours. In other words, we would expect the disincentive effects of taxation to be greater around the retirement decision than at midcareer.

This changing elasticity over the life cycle constitutes the main theoretical pillar of the position that proposes the adoption of age-dependent taxation: ‘The need for age-dependent taxes is a natural implication of life-cycle behavior’ (Erosa and Gervais 2002, p. 341).

Akerlof (1978) advanced the idea that if poorer groups of the population could be identified or ‘tagged’ costlessly according to relevant characteristics they could not modify, transferring subsidies to them would increase social welfare. Chronological age is an observable and non-manipulable trait and, given its association with income, a good candidate for tagging (Bastani et al. 2013).

Furthermore, unlike other observable traits which are also associated with earnings and economic behaviour in general, such as height or gender, tagging based on chronological age may be more acceptable politically given that most people are expected to reach a given chronological age at some time during their lifetimes (Banks and Diamond 2010). For example, in a US context, Hemel (2010, p. 1896) noted that the endorsement by interest groups representing older people of age-dependent tax rates, which made the policy recommendations stemming from optimal taxation theory easier to implement, was ‘a rare (and perhaps unique) congruence’ between economic theory and interest group politics.

Given that labour supply elasticity increases with chronological age, the fiscal policy conclusion (based upon a number of simplifying assumptions) is that marginal tax rates should decline with advancing chronological age (Hemel 2010). The literature on this topic, as on any other in economics, has developed from relatively simple models13 to more complex specifications, where initial assumptions are gradually relaxed. We are going to use Weinzierl (2011a,b) as a guide to reflect on the relationship between taxation and individual ageing, without going into the mathematical details. Weinzierl discussed there alternative tax policies: a labour income tax that is a function of the income but unrelated to the chronological age of the taxpayer; a labour income tax that depends on the income and the chronological age of the taxpayer; and a labour income tax that depends on the income, the chronological age, and the lifetime path of incomes of the taxpayer—termed, respectively, ‘Static Mirrlees’, ‘Partial Reform’, and ‘Full Optimum’ policies. The starting point is an individual’s utility function, which is generally assumed to depend on consumption and leisure. Instead of leisure, what in this case enters the utility function is the disutility of the labour effort. Labour effort is equal to the ratio between the labour income and the wage per unit of labour. Labour income, in turn, is the product of the wage and the labour effort. Ability or effort is not readily observable and is assumed to be distributed among the population. Economic life starts in period 1 when an individual enters the labour market and finishes in period T , when she retires. Each individual lives and works for the same number of periods: in this model there is no retirement and chronological age and period coincide. The utility of an individual of ability i and age t results from the difference between the utility derived from consumption and the disutility from labour: Each individual maximises her utility subject to the budget constraint after the taxes whose specification depends on the policy. For example, under a Static Mirrlees regime where income tax only depends on income, we have T = T (y), whereas under a Partial Reformpolicy, we have T = T (y, t) as the income tax depends on income and age; finally, in the Full Optimum case, the tax becomes T y(.)T t=1, t because it depends on the lifetime path of income and the current chronological age of the taxpayer. Finally, the model assumes the existence of a social welfare function that a benevolent policy maker or social planner seeks to maximise. The standard assumption, adopted by Weinzierl, is that this social welfare function can be represented by a weighted utilitarian function of the individual utilities—that is, as the weighted sum of the individual utilities, where the weights, known as Pareto weights (Saez and Stantcheva 2016), reflect social preferences (d’Aspremont and Gevers 2002; Sen 1986).Weinzierl (2011b) calibrated the model with data for over 10,000 people in paid employment aged 25–55 in the USA between 1968 and 2001 and obtained that age-dependent taxes generated efficiency (i.e. welfare) gains equivalent to between 0.6 per cent and 1.5 per cent of aggregate annual consumption. These efficiency gains resulted from a reduction in the marginal taxes on high-income young workers (because higher taxes on these individuals would introduce substantial the deadweight costs withmuch lower tax revenue fromthis same group of people) and from a reduction, on average, on younger workers relative to older workers under the assumption of imperfections in the capital markets such that private saving and borrowing were restricted. The conclusion was that the welfare gains under age dependence would exceed those under an age-independent tax regime. Similarly, based on US data between 1982 and 2008, Bastani et al. (2013) estimated that switching from a non-linear age-independent income tax to a non-linear age-dependent income tax would generate welfare gains equalling about four per cent of total output. The same approach has been extended to taxes on consumption and interest income: if elasticities vary with chronological age, under the same assumptions as above—particularly, the separability of the utility function—it would be optimal to levy these age-dependent taxes. However, would it be optimal to introduce age-dependent taxes on wages, consumption, and interest income at the same time? Bymeans of a similarmodel as the one described above, Alvarez et al. (1992) showed that it would not. Under the separability assumption between consumption and leisure, these authors presented the following ranking of tax (or subsidy) configurations in terms of the decreasing distortions or deadweight losses each configuration would create (Alvarez et al. 1992, Table 1, p. 119): 1. Age-conditioned wage taxes (or subsidies) and age-conditioned consumption taxes (or subsidies), or a uniform, non-aged-conditioned consumption tax (or subsidy) and age-conditioned wage taxes (or subsidies), or ageconditioned wage taxes (or subsidies) 2. Age-conditioned consumption taxes (or subsidies) and a non-agedconditioned interest income tax (or subsidy) 3. Uniform, non-aged-conditioned consumption tax (or subsidy) and a nonaged- conditioned interest income tax (or subsidy) 4. Age-conditioned consumption taxes (or subsidies) or uniform, non-agedconditioned consumption taxes (or subsidies) This ranking is not based on redistributive arguments or on consideration of the social insurance repercussions of imposing age-dependent taxes, but purely on efficiency grounds. Nevertheless, Gervais (2012) showed that agedependent tax rates on labour income and on capital income would be progressive: under age-dependent labour income tax, the tax rates would increase with labour supply and vice versa, whilst the tax rates on capital income would be negative when labour supply increased (and vice versa). The models in the optimal taxation literature have become more and more complex as authors have relaxed assumptions and explored particular conditions and situations. Different specifications have included the interplay between labour decisions and retirement—for example, the elasticity of labour supply to changes in the pension system, which we will touch upon in Volume III, Part II—or between investment in human capital and wages and productivity along the life cycle; see Volume III, Part I. Lehmann et al. (2013), using data for France between 2003 and 2006, presented the interesting finding that the elasticity of labour supply (at the extensive margin) with respect to the marginal net-of-income-tax rate would be higher than that with respect to the marginal net-of-payroll (or social security contribution) tax rates (which would not be significantly different from zero) despite both taxes affect after-tax income in exactly the same amount. Furthermore, these authors also reported that for workers aged 50 or over, the elasticity was negative, whilst it was positive for younger workers, which adds empirical clout to the recommendations that marginal income tax rates should be lower for older workers. A study looking into Swedish data for 2007 (Laun 2017) concluded that the introduction of two income tax credits—an earned income tax credit which increases for workers aged 65 or over and a reduction in social security contributions (a payroll tax credit) for workers aged 65 or above—had positive short-term effects at the extensive margin: the elasticity of labour supply to these age-related tax credits was 0.22, meaning workers aged 65 years old extended their labourmarket participation as a result of these tax credits, which contributed to a gain in welfare per dollar spent of about 1.15. Other extensions have considered the effects on capital accumulation, the uncertainty around the extension of the lifespan, decisions about planned bequests, and the stochasticity of the relationship between labour productivity and chronological age. Woodland (2016, p. 742) surveyed the literature and concluded that there seems to be an agreement that age-dependent taxes produce net welfare gains, which could ‘partially relieve’ the pressure that population ageing creates on government budgets. 2.2.3 Taxation in Dynastic Models With infinitely lived agents, life-cycle considerations cease to exist.14 However, the timing of saving decisions does matter because of the repercussions on capital accumulation, economic growth, and welfare. These distortions in inter-temporal choices lead to the conclusion that the optimal tax rate on capital or inheritance is zero (Chamley 1986; Judd 1985). Alas, it only takes the introduction of additional assumptions to, or the relaxation of, some assumptions in the most basic dynasticmodels to reach the conclusion that the optimal tax rate on wealth or assets is positive. Among other extensions that lead to tax rates on inheritance and capital income greater than zero, we can list: idiosyncratic labour income shocks—that is, shocks that affect individuals or their households rather than whole economic aggregates; accidental bequests; and the lack of government commitment over time (infinitely, in fact) to the fiscal policy (Piketty and Saez 2013). Furthermore, in their influential survey of direct taxation, Banks and Diamond concluded: The empirical evidence on the consumption patterns of parents and adult children alive at the same time is strongly contradictory of the idea that people typically behave as if there were a single dynastic utility function being jointly maximized.Moreover, taking this literally and recognizingmarriage (which links dynasties to each other) leads to absurdities. (Banks and Diamond 2010, p. 575) 2.3 Population Ageing and Fiscal Space Fiscal space has been defined as the availability of budgetary room that allows a government to provide resources for a desired purpose without any prejudice to the sustainability of a government’s financial position. (Heller 2005, p. 3) and as the financing that is available to government as a result of concrete policy actions for enhancing resource mobilization, and the reforms necessary to secure the enabling governance, institutional and economic environment for these policy actions to be effective, for a specified set of development objectives. (Roy et al. 2009, p. 33) Ostry et al. (2010, p. 17) proposed a different definition of fiscal space focused on fiscal sustainability: ‘the difference between the current level of public debt and the debt limit implied by the country’s historical record of fiscal adjustment’. These authors estimated, using data from 2015, that many developed countries needed to implement fiscal adjustment efforts that far exceeded their historical records. The concept of fiscal space is not without critics: according to Perotti (2007), it is nothing new as it merely re-states the notion of the inter-temporal government budget constraint. However, other analysts see it as a useful conceptual tool to look beyond budgetary straitjackets in the sense that fiscal space can be created and expanded. For instance, Ortiz et al. (2015, p. 1) listed eight policy options available even in the poorest countries to ‘make budgetary room’: • increasing tax revenues • expanding social security coverage and contributory revenues • lobbying for increased aid and transfers • eliminating illicit financial flows • borrowing or restructuring debt • adopting a more accommodative macroeconomic framework • reallocating public expenditures • tapping into fiscal and foreign exchange reserves Park (2012) looked into the impact of population ageing on fiscal space using yet another definition: ‘[the] distance between the current tax revenue level and the peak of Laffer curve’ (p. 3).15 Using data from 1995 to 2009 for the G-7 countries,16 this author estimated a Laffer curve under the then existing population structure and calibrated the Laffer curve for each of these countries up to 2050 using demographic projections. The idea behind the exercise was that population ageing would shrink labour supply and hence reduce revenue capacity (i.e. the tax base) and the fiscal space, but if a country was on the ‘right’ side of the Laffer curve (i.e. if the average tax rate was below its maximum revenue-generating level), the pressure from the change in the demographic structure could be offset, at least in part, by the excess room for fiscal manoeuvre. The results pointed to a smaller fiscal space for Germany, France, and Italy—the three countries in the G-7 with a relatively larger share of the public sector in the economy. These countries would be ‘more susceptible to aging shock’ (op. cit., p. 14). Bogetic et al. (2015) also looked into population ageing and fiscal space, but operationally defined the latter as debt per dependent person ‘as a measure of fiscal space to finance the aging related commitments that are no longer feasible to be financed by the contributions from the active population’ (p. 20). They defined a maximum threshold for the debt-to-GDP ratio and then estimated the policy adjustments needed to avoid exceeding that target. The authors explained that this cap showed ‘the extent to which the government can implement a transfer and expenditure system if it cannot be financed by the aging work force’ and elaborated: Typically, any age-related expenditures are financed by higher taxes or borrowing beyond a certain threshold. Because the threshold is determined by the relative size of the dependent population and the characteristics of the tax and transfer systems and of public spending, the extent to which a government can issue debt per dependent person without breaching a specified debt ceiling shows how much fiscal space it has available to accommodate the fiscal pressures of an aging society. A cap on the growth rate of debt per dependent person -a “speed rule” defined by the fiscal target- will be tighter if the dependent population grows faster than the working-age population. Similarly, an increase in the initial debtto- GDP ratio or a tightening in the fiscal target will suggest need for a downward adjustment to the speed rule. (Bogetic et al. 2015, p. 20) The authors analysed long-term fiscal policy options in 40 countries in Europe and Asia by means of one population ageing indicator—the oldage dependency ratio (see Volume I, Chap. 2)—and several fiscal variables as a percentage of gross domestic product, such as gross public expenditure, revenue, and debt, social contribution revenues, social benefit expenditure, and so on. From these simple bivariate relationships and the threshold defined as explained above, they concluded that the fiscal space in most of these countries is limited ‘for discretionary and growth-oriented public spending’ and ‘to accommodate increases in aging-related spending’ (Bogetic et al. 2015, p. 7). Fiscal space is a useful concept for macroeconomic analysis, but economists should be wary of the danger that lies within: an over-mechanistic numerical approach that does not take into account that: …decisions on financing and spending usually respond to the specific interests of stakeholders that have sufficient power to influence them. No intervention on financing or public spending is neutral in terms of which groups in society are winners or losers. Thus, the creation of fiscal space, whatever its purpose, unfolds in a context dominated by the relationships between the power groups in society. (Durán-Valverde and Pacheco 2012, p. 1) Or, in the words of (Estes et al. 2003, p. 19): …portrayals of the threatened bankruptcy of public treasuries to support the retirement of the elderly and the ‘appropriate’ policy response (such as the proposed privatisation of social security) are crisis constructions embedded in intense power struggles that are momentous in their social, political and economic consequences, including the sacrifices that are demanded.

The financial situation of the public sector in a country is a relevant variable to assess its fiscal space. However, its growth prospects, debt reduction and consolidation policies, the political commitment by policy-makers, the relative clout and effectiveness of the different interest groups, institutional characteristics and dynamics, and—for developing countries—the use of development assistance from abroad are of equal importance.

Finally, economists should be wary of the prevalent mode of political competition (Roemer 2009) in the jurisdiction under study or where they work and provide their advice or make decisions, because as Lynch (2006, pp. 67–68) correctly asserted ‘…the dominant mode of political competition is crucial for the eventual age orientation of social spending’. 2.3.1 The Global Aging Preparedness Index The Global Aging Preparedness Index (or GAP Index), compiled by the Global Aging Initiative of the Center for Strategic and International Studies in Washington, D.C., USA, is an index that seeks to assess how countries are getting ready for ‘particularly the “old-age dependency” dimension of the challenge’ (Jackson et al. 2013, p. iii) posed by population ageing (the authors of the index define old age as 60 years or over). In part, it includes the notion of fiscal space, but it extends the concept as it brings together two sub-indices: a Fiscal Sustainability Index and an Income Adequacy Index. The Fiscal Sustainability Index is composed of three ‘categories’: • the Public Burden category, which measures the level of and projected (all projections for the 2013 edition were run until 2040) growth in total public benefits to older people and is composed of two indicators: – the Benefit Level (the projected public benefits to older people as a proportion of GDP) – the Benefit Growth (the projected growth in public benefits to older people as a proportion of GDP) • the Fiscal Room category, which measures the fiscal space of a country by combining three indicators: – the Tax Room (projected government revenue as a percent of GDP if all growth in benefits to older people are paid by taxes) – the Budget Room (projected proportion of benefits as a percentage of total public spending assuming cuts in all other items finance the projected growth in benefits to older people) – the Borrowing Room (the projected net debt as a percent of GDP assuming all projected growth in benefits to older people is funded by borrowing) • the Benefit Dependency category,which measures the degree of dependence of older people on public benefits and is composed of two indicators: – the Benefit Share (average of projected benefits to older people as a per cent of the cash income of the median-income older person) – the Benefit Cut (the percentage of older people that would fall into poverty—measured as earning an income below 50 per cent of the median income of the whole population—if public benefits were cut by 10 per cent) The indicators in the first two categories of the Fiscal Sustainability Index are weighted equally, but in the Benefit Dependency category, the share indicator receives two-thirds and the cut indicator receives one-third. The Public Burden category receives a weighting factor of 40 per cent, whilst the Fiscal Room and the Benefit Dependency categories are assigned 30 per cent each. The Income Adequacy Index is also composed of three ‘categories’: • the Total Income category, which measures the ratio between the level of income of older people and younger people, and the trends, and is composed of two indicators: – the Income Level (the projected ratio of average after-tax total income per capita of older people to younger people) – the Income Trend (the projected change in the ratio of average after-tax total income per capita of older people to younger people) • the Income Vulnerability category, which measures income adequacy for middle-income older people and the extent of poverty in old age, and is composed of three indicators: – the Median Income Level (the projected ratio of median after-tax cash income per capita of older people to younger people) – the Median Income Trend (the projected change in the ratio of median after-tax cash income per capita of older people to younger people) – the Poverty Level (the percentage of older people with incomes below 50 per cent of the median income for the whole population) • the Family Support category, which measures family support networks, and is composed of two indicators: – the Family Ties (the percentage of older people livingwith adult children) – the Family Size (the projected change in the number of surviving children of older people) The indicators in the first two categories of the Income Adequacy Index are weighted equally, but in the Family Support category, the Family Ties indicator receives two-thirds and the Family Size indicator receives one-third. The Total Income and the Income Vulnerability categories receive a weighting factor of 40 per cent each, whilst the Family Support category is assigned 20 per cent. Finally, the Global Preparedness Index is calculated by assigning the Fiscal Sustainability and the Income Adequacy indices the same weights. The latest (2013) GAP index report found that fiscal sustainability would not be an important policy concern: India, Mexico, and Chile, but that countries such as Brazil, Japan, France, the Netherlands, Germany, Italy, and above all Spain would be highly vulnerable. Looking into income adequacy, Netherlands, the United States, Brazil, Australia, and Germany stand out as highly prepared; amongst the most vulnerable countries are Mexico, Russia, South Korea, and Poland.

2.4 The Fiscal Implications of Risk

Population ageing is one of the socially amplified risks (Kasperson et al. 1988, 2003) in contemporary societies, and what Taylor-Gooby (2004) identified as one of the four processes of ‘social risk’ in post-industrial societies. In fact, Taylor-Gooby understands that population ageing exerts the most important pressure on welfare states, which emerges as ‘new social risks’. These new risks translate, from the perspectives of citizens, into changes in employment relations, family life, social care, career, and so on. From the perspective of governments, the new social risks are manifested in changes in policy-making, institutional structures, power dynamics, and legitimisation.

We saw in Volume I, Chap. 1, examples of apocalyptic demography, with concerns that population ageing would bring about a ‘crisis’, a ‘time bomb’, an ‘agequake’, or a ‘silver tsunami’ and authors warning of the seismic consequences of ‘gray dawn’ and the ‘coming generational storm’. Some go even beyond this. For example, Richard Jackson, President of the Global Aging Institute based in Virginia, USA, suggested that global population ageing threatens to ‘overturn’ (Jackson 2006) or ‘shape’ (Howe and Jackson 2011) the world order. In the same vein, Heller (2003a,b) asserted that a number of major long-term risks and challenges beset almost every country and pose mounting threats to fiscal stability, including population ageing—which Heller17 placed on the same level with climate change, rapid technological change, the intensification of globalisation, increasing inequality, and bioterrorism and the proliferation of weapons of mass destruction. Such a statement may look hyperbolic, but it mirrors what Peterson (1999a,b) had previously written:

The list of major global hazards in the next century has grown long and familiar. It includes the proliferation of nuclear, biological, and chemical weapons, other types of high-tech terrorism, deadly super viruses, extreme climate change, the financial, economic, and political aftershocks of globalization, and the violent ethnic explosions waiting to be detonated in today’s unsteady new democracies. Yet there is a less-understood challenge -the graying of the developed world’s population- that may actually do more to reshape our collective future than any of the above.

(Peterson 1999b, p. 42)