Normative Behavioral Economics

Abstract

In normative behavioral economics , we study how the public sector and private sector should behave. Based on libertarian paternalism , we can use the idea of nudging people in better directions without forcing them. In order to evaluate resource allocation to judge how this nudge tool and other public policies should be used when preferences are endogenous, relying exclusively on welfarism may not be satisfactory in many cases because some preferences may be thought to be better than others. Introducing one of the three major approaches in normative ethics, virtue ethics , into a formal analytical framework of normative economics seems a possible solution to this problem.

Appendices

Appendix 1: Government’s Budget Problem and Ricardian Equivalence

We review Ricardian equivalence theory to see the role behavioral economics can play in important policy evaluations. In this appendix, we use a simple two-generation, two-period model. Ricardian equivalence theory deals with the fiscal problem, i.e., when or on which generation the government should impose a tax.

In many countries, a lower birthrate and aging population are becoming serious problems, and pension payment for the elderly and increases in medical expenses are imposing a large fiscal burden. The problem will intensify in the future. This is especially relevant for countries with persistent fiscal deficit such as Japan, where gross government debt per GDP is higher than 200% (the highest in the world) against the backdrops of low birthrate and rapidly aging population.

First, we consider a pure altruistic model (the parent to the child). Parents divide their income into three: C₀ for their own consumption, C₁ for their children’s consumption, and B for bequests. t_i is a fixed tax for generation i, y_i is an exogenous labor income of generation i (i = 0, 1), r is the interest rate. Thus, the bequest would be expressed as:

$${B} = {y}_{0} {-}{t}_{0} - {C}_{0}$$

(11.22)

The child’s consumption would equal his own disposable income plus bequest and interest on the bequest, thus:

$${C}_{1} = {y}_{1} - {t}_{1} + \left( {1 + {r}} \right){B}$$

(11.23)

Hence, the combined budget constraint of two generations of this household would be:

$${C}_{0} + \frac{{C_{1} }}{{1 + {r }}} = {y}_{0} {-}{t}_{0} + \frac{{y_{1} - t_{1} }}{1 + r}$$

(11.24)

In other words, the present value of consumption is the explicit value of disposable income. The parent is altruistic, and chooses C₀ and C₁ in the manner that maximizes:

$${u}\left( {{C}_{0} } \right) +\uptheta\,{u}\left( {{C}_{1} } \right)$$

(11.25)

under the budget constraint (11.24). In order to sustain this economy, expenditure for public goods G is necessary, and a fixed tax satisfies the government’s budget constraint:

$${G}_{0} + \frac{{G_{1} }}{{1 + {r}}} = {t}_{0} + \frac{{t_{1} }}{1 + r}$$

Initially, the government’s budget is balanced, i.e., G_i =  t_i (i = 0,1). If the government reduces the tax by one unit, issue government bonds, and pays off the debt by increasing the tax of the child’s generation, t₁ needs to increase by (1 +  r) unit. However, the right-hand side of (11.24) indicates that such policy change does not change the budget constraint of the household. Therefore, the optimal consumption to maximize utility for the parent (C₀) and a child (C₁) do not change. The parent will transfer his or her disposable income to a bequest, so consumption of both generations do not change. This is the essence of Ricardian equivalence—i.e., the timing of taxation does not affect consumption.

If Ricardian equivalence holds in the real world, the large budget deficit as in Japan is not a problem, because the parent’s generation would expect a tax increase in the future and act accordingly to increase the bequest. Traditional economics’ standard macroeconomic model uses the infinite-horizon model (i.e., not the two-period model). The infinite-horizon model does not assume that an individual’s life is infinite; instead, it assumes that generations are connected by pure altruism. Under such an assumption, Ricardian equivalence holds.⁵

Next, the warm glow model in Chap. 8 builds on the assumption that a parent derives utility from the act of bequest itself. Thus, the parent’s utility function would be:

$${u}\left( {{C}_{0} ,{B}} \right) +\theta {u}\left( {{C}_{1} } \right)$$

(11.26)

In this model, the parent derives utility from B, so they choose C₀ and B to maximize their utility based on (11.26) under the budget constraint of (11.22) and (11.24). The tax cut of the parent’s generation does not affect (11.24), but does affect budget constraint (11.22), so Ricardian equivalence does not hold.⁶

Lastly, in the bounded rationality model, there is a possibility that, even if the tax cut is done in the parent’s generation, he or she does not consider the tax increase in the future generation. In that case, Ricardian equivalence does not hold. Also, a person with hyperbolic discounting who is unable to fulfill the sophisticated individual commitment in the multiple-period model, disposable income and consumption would co-move, so an increase in disposable income by reducing the tax would simply lead to disposable income of the current generation, thus Ricardian equivalence does not hold.

Appendix 2: Conditional and Unconditional Preference Orderings in Models of Endogenous Preferences

In the habit-formation model, as in Prospect theory , utility function depends on the reference point. If the reference point is one’s past consumption, it is called the endogenous habit formation model . On the other hand, if the reference point is average consumption of the overall economy such as other people’s consumption, it is called the exogenous habit-formation model.⁷ For example, a person with nicotine addiction from smoking can be considered as having the past smoking as the reference point which impacts on today’s smoking pattern. The idea of dependency (habit) on the reference point is similar to the endogenous preference model in Prospect theory , so it is categorized as behavioral economics in this book. However, many studies assume that the consumption that is less than the reference point will not be selected (i.e., people will not reduce their consumption from the reference point), and habit formation was used in macroeconomics and the standard financial model in a way that is compatible with other definitions.

In order to explain the concept of conditional and unconditional preference ordering as in Pollak (1978), we consider a habit formation model as follows. In an economy, there are N consumers. Consumer i’s reference point is the average of everybody else’s consumption average. Then:

$${u}_{i} \left( {{C}_{i} - {C}_{ia} } \right)$$

(11.27)

where C_i is consumer i’s consumption, C_ia is the average of other people’s (number of people is N − 1) consumption, then consumer i maximizes his or her utility under the normal assumption of u_i(·) (monotonically increasing convex function) under the budget constraint. For consumer i, C_ia is given a state variable, and the preference ordering of C_i given the state variable is defined as conditional preference ordering. The utility function of this conditional preference ordering can be expressed by the utility function U_i(C_i; C_ia) =  u_i( C_i − C_ia). On the other hand, for a policymaker, the state variable is not given. Without the assumption that C_ia is given, the utility of consumer i (C₁, …, C_N) can be expressed as u_i(C_i − C_ia), and the preference ordering can be expressed by the utility function U_i(C₁, …, C_N) =  u_i(C_i − C_ia). This ordering is defined as unconditional preference ordering.⁸

Pollak (1978) proposed that policy evaluation should be done using fixed unconditional preference ordering instead of variable conditional preferential ordering. As long as we use the utility function of unconditional preferential ordering, U₁(C₁, …, C_N), …, U_N(C₁, …, C_N), analysis based on Pareto improvement and social welfare functions can be used even in the models with endogenous preferences.

Appendix 3: Tough Love Model with Bequest

Bhatt et al. (2017) extend the tough love model of cultural transmission of preferences to analyze bequest and bequest tax. In this model, the parent divides his income into three: (1) own consumption, (2) future consumption for the child during his or her childhood (T), and (3) bequest after the child reaches the working age (B). The government decides the bequest tax rate τ and collects the bequest tax τ B and gives a subsidy s to the child of working age. s is set to be fixed so that it will offset the decline in income from the bequest tax. This assumption is made so that we can analyze the effect of bequest tax with given income, by offsetting income decline from the bequest tax with a fixed amount of subsidy. Thus, the government’s budget constraint will satisfy τ B =  s. Because the amount of the subsidy is fixed, the bequest tax only affects the decision of the parent on whether they give the money to his child while the child is still under the working age (T), or gives the bequest (B). Under this assumption, the extended model of consumption of the child after his or her retirement would be⁹:

$${C}_{3} = \left( {1 + {r}} \right)({y}_{2} + (1{-}\tau){B} + {s}{-}{C}_{2} ).$$