AN INTEGRATED VIEW regardless of the level of t(since tax payments and benefit receipts would be offsetting in their incidence). Or the revenue might finance public goods worth disproportionately more to the rich, in which case a higher level of the flat tax would increase inequality Or the revenue might target the poor, producing even more redistribution than when the revenue finances a uniform grant. Thus, in the abstract it is precarious to associate any distributive assessment with the use of purely proportional taxation, even if the level of the tax is known. o More broadly, particular forms of taxation--ranging from taxes on dividends and capital gains or on estates and gifts to taxes on consumption-as well as expenditure policies are often assessed in terms of distributive effects in a vacuum and thus with out regard to how the system as a whole might be brought into balance Such problems will be identified and addressed throughout this book. B Comprehensiveness of Instruments Considered The most pertinent tools for the present investigation include the main mechanisms by which governments do or could raise revenue and the various ways in which such funds are spent. Accordingly, it is necessary Nevertheless, in policy debates it is generally believed that a flat tax would be less progressive. The Economic Report of the President (1996, p. 91)bluntly states that "It]he prototypical flat tax would be less progressive than the current income tax. "A major rationale for the USA tax proposal in the mid-1990s, a form of graduated consumption tax, was to overcome what was understood to be the inherent limitations of the flat tax with regard to redistribution(see, for example, Seidman 1997). This(mis)understanding of proportional versus graduated taxes has a distinguished lineage, including Blum and Kalven's(1952, 1953) famous discussion of"The Uneasy Case for Progressive Taxation, " which argues as though undermining justifications for redistribution implies that a proportional income tax is supe- rior to a graduated one Critiques of Blum and Kalven include Bankman and Griffith(1987) Some in the debate may believe that the rates in most flat tax proposals are sufficiently low that they would raise significantly less revenue than the graduated tax schemes they would replace, which would make the proposals less redistributive if one adds particular assur tions about the incidence of the resulting expenditure cuts(on which see section 8.E) This statement is incomplete, for there exist other tools, such as wage regulation, that are potential means of redistribution. See also section 8G on the distributive effects of regu- lations more generally. 03_Kaplow_Ch02_p011-p034 ndd 19 1016/20071121:28AN
an integrated view 19 regardless of the level of t (since tax payments and benefi t receipts would be offsetting in their incidence). Or the revenue might fi nance public goods worth disproportionately more to the rich, in which case a higher level of the fl at tax would increase inequality. Or the revenue might target the poor, producing even more redistribution than when the revenue fi nances a uniform grant. Thus, in the abstract it is precarious to associate any distributive assessment with the use of purely proportional taxation, even if the level of the tax is known.10 More broadly, particular forms of taxation—ranging from taxes on dividends and capital gains or on estates and gifts to taxes on consumption—as well as expenditure policies are often assessed, in terms of distributive effects, in a vacuum and thus without regard to how the system as a whole might be brought into balance. Such problems will be identifi ed and addressed throughout this book. B. Comprehensiveness of Instruments Considered The most pertinent tools for the present investigation include the main mechanisms by which governments do or could raise revenue and the various ways in which such funds are spent.11 Accordingly, it is necessary 10 Nevertheless, in policy debates it is generally believed that a fl at tax would be less progressive. The Economic Report of the President (1996, p. 91) bluntly states that “[t]he prototypical fl at tax would be less progressive than the current income tax.” A major rationale for the USA tax proposal in the mid-1990s, a form of graduated consumption tax, was to overcome what was understood to be the inherent limitations of the fl at tax with regard to redistribution (see, for example, Seidman 1997). This (mis)understanding of proportional versus graduated taxes has a distinguished lineage, including Blum and Kalven’s (1952, 1953) famous discussion of “The Uneasy Case for Progressive Taxation,” which argues as though undermining justifi cations for redistribution implies that a proportional income tax is superior to a graduated one. Critiques of Blum and Kalven include Bankman and Griffi th (1987). Some in the debate may believe that the rates in most fl at tax proposals are suffi ciently low that they would raise signifi cantly less revenue than the graduated tax schemes they would replace, which would make the proposals less redistributive if one adds particular assumptions about the incidence of the resulting expenditure cuts (on which see section 8.E). 11 This statement is incomplete, for there exist other tools, such as wage regulation, that are potential means of redistribution. See also section 8.G on the distributive effects of regulations more generally. 03_Kaplow_Ch02_p011-p034.indd 19 3_Kaplow_Ch02_p011-p034.indd 19 10/16/2007 11:21:28 AM 0/16/2007 11:21:28 AM
CHAPTER 2 to consider income taxes, commodity taxes, corporate taxes, estate and ft taxes, and payroll taxes as well as transfer programs, income security programs for retirement, disability, and unemployment, and direct pub- lic expenditures, whether on education, police, roads, or any other gov ermen ntal program. The advantages of taking a comprehensive view are illustrated by the application of economic analysis to the control of externalities, notably, environmental pollution. Economics provides a useful framework for de- fining the problem-the very concept of an externality constituting a sig- nificant contribution--and for assessing interventions-through the ap- plication of cost-benefit analysis. Additionally, and most relevant at this oint, economics has offered significant insights into the choice of instru ments for controlling externalities. In addition to command-and-control gulations, Pigou(1920)introduced taxes and subsidies, Coase(1960) emphasized the possibility of private bargaining, and Dales(1968)sug- gested the potential virtues of tradeable permits. arguably the greatest modern contribution of economics in this field has been to broaden the focus with regard to the available tool set and to offer incisive analysis that indicates which tool or combination of tools is best depending on the particular characteristics of a given setting, such as the quality of the governments information The benefits that flow from a comprehensive view of instruments however, have not been fully realized in the analysis of taxation andn nd an explicit attempt to determine their comparative advantag lated subjects in public economics. For example, it is common to assess the redistributive effects of specific forms of taxation and types of gov ernment expenditure without simultaneously contemplating the alter natives and examining which policies are superior under various condi tions. Th ider what adjust VAT--perhaps exemptions for food-would best meet some distribu tive target without inquiring whether some combination of food stamps (vouchers), cash transfers, and income tax adjustments might better achieve the objective and involve less distortion Or a study may assess modifications to transfer(estate and gift )taxation without compar income tax adjustments that would have similar effects on revenue and 03_Kaplow_Ch02_po11-p034 ndd 1016/20071121:28AN
20 chapter 2 to consider income taxes, commodity taxes, corporate taxes, estate and gift taxes, and payroll taxes as well as transfer programs, income security programs for retirement, disability, and unemployment, and direct public expenditures, whether on education, police, roads, or any other governmental program. The advantages of taking a comprehensive view are illustrated by the application of economic analysis to the control of externalities, notably, environmental pollution. Economics provides a useful framework for de- fi ning the problem—the very concept of an externality constituting a signifi cant contribution—and for assessing interventions—through the application of cost-benefi t analysis. Additionally, and most relevant at this point, economics has offered signifi cant insights into the choice of instruments for controlling externalities. In addition to command-and-control regulations, Pigou (1920) introduced taxes and subsidies, Coase (1960) emphasized the possibility of private bargaining, and Dales (1968) suggested the potential virtues of tradeable permits. Arguably the greatest modern contribution of economics in this fi eld has been to broaden the focus with regard to the available tool set and to offer incisive analysis that indicates which tool or combination of tools is best depending on the particular characteristics of a given setting, such as the quality of the government’s information. The benefi ts that fl ow from a comprehensive view of instruments and an explicit attempt to determine their comparative advantages, however, have not been fully realized in the analysis of taxation and related subjects in public economics. For example, it is common to assess the redistributive effects of specifi c forms of taxation and types of government expenditure without simultaneously contemplating the alternatives and examining which policies are superior under various conditions. Thus, a study may consider what adjustments to a sales tax or VAT—perhaps exemptions for food—would best meet some distributive target without inquiring whether some combination of food stamps (vouchers), cash transfers, and income tax adjustments might better achieve the objective and involve less distortion. Or a study may assess modifi cations to transfer (estate and gift) taxation without comparing income tax adjustments that would have similar effects on revenue and distribution. 03_Kaplow_Ch02_p011-p034.indd 20 3_Kaplow_Ch02_p011-p034.indd 20 10/16/2007 11:21:28 AM 0/16/2007 11:21:28 AM
INTEGRATED VIEW 21 It is usually best to use instruments that directly related to the matter in question. In the case of redistribution, the income tax system (including cash transfers) is that instrument. Furthermore, in identify- ing the core role of the income tax system, the revenue-raising function of taxation should also be included because were it not for differences among individuals that give rise to distributive concerns, revenue would optimally be raised instead through uniform lump-sum taxation, with- ut causing any distortion. 2 Policy tools other than the income tax sys- tem tend to be advantageous in pursuing distributive and revenue ob jectives only when they are able to address particular shortcomings of a more direct approach, such as by mitigating evasion or reducing the labor-leisure distortion in various subtle ways. Accordingly, throughor this book, the income tax will always be kept in focus, and other form of taxation as well as government expenditure policies will be related Although administrative and political considerations often require substantial deviations from what would otherwise be ideal. it is difficult to reach a sensible accommodation without first obtaining a reasonably comprehensive view of the relevant options. An additional problem is that a single instrument or a particular reform package-such as park improvements financed by a gasoline tax increase--often affects multi ple margins simultaneously, which complicates the task of determining what combination of instruments is best when addressing multiple ob- jectives. This challenge brings us to the third component of the pro posed integrated view That is, the distributive objection to uniform lump-sum taxation when individuals income-earning abilities vary substantially is the primary justification for reliance on distor tionary taxation, such as an income tax, which allows tax obligations to reflect individuals circumstances. As a consequence, much analysis of the efficiency of taxation is intimately lated to distribution, even if this connection often remains unrecognized or implicit. For further related discussion, see section 6 D on how the presence of an income tax and the problem of income distribution bears on the Ramsey tax problem and chapter 8(especially subsection 8 D3)on the relationship between redistribution and the need for revenue to finance government expenditures on goods and services as part of the optimal income tax 03_Kaplow_Ch02_po11-p034 ndd 1016/20071121:28AN
an integrated view 21 It is usually best to use instruments that are most directly related to the matter in question. In the case of redistribution, the income tax system (including cash transfers) is that instrument. Furthermore, in identifying the core role of the income tax system, the revenue-raising function of taxation should also be included because, were it not for differences among individuals that give rise to distributive concerns, revenue would optimally be raised instead through uniform lump-sum taxation, without causing any distortion.12 Policy tools other than the income tax system tend to be advantageous in pursuing distributive and revenue objectives only when they are able to address particular shortcomings of a more direct approach, such as by mitigating evasion or reducing the labor-leisure distortion in various subtle ways. Accordingly, throughout this book, the income tax will always be kept in focus, and other forms of taxation as well as government expenditure policies will be related explicitly to it. Although administrative and political considerations often require substantial deviations from what would otherwise be ideal, it is diffi cult to reach a sensible accommodation without fi rst obtaining a reasonably comprehensive view of the relevant options. An additional problem is that a single instrument or a particular reform package—such as park improvements fi nanced by a gasoline tax increase—often affects multiple margins simultaneously, which complicates the task of determining what combination of instruments is best when addressing multiple objectives. This challenge brings us to the third component of the proposed integrated view. 12 That is, the distributive objection to uniform lump-sum taxation when individuals’ income-earning abilities vary substantially is the primary justifi cation for reliance on distortionary taxation, such as an income tax, which allows tax obligations to refl ect individuals’ circumstances. As a consequence, much analysis of the effi ciency of taxation is intimately related to distribution, even if this connection often remains unrecognized or implicit. For further related discussion, see section 6.D on how the presence of an income tax and the problem of income distribution bears on the Ramsey tax problem and chapter 8 (especially subsection 8.D.3) on the relationship between redistribution and the need for revenue to fi nance government expenditures on goods and services as part of the optimal income tax problem. 03_Kaplow_Ch02_p011-p034.indd 21 3_Kaplow_Ch02_p011-p034.indd 21 10/16/2007 11:21:28 AM 0/16/2007 11:21:28 AM
2 C Comparability of Proposals under Assessment Section A explains the need to specify a policy completely in order to nalyze it properly. But completeness, notably the requirement of budget balance, can be achieved in many ways. Some means of doing so better facilitate comparability than others. Moreover, the artful choice of pol icy combinations can take advantage of the range of available tools, as suggested in section B, while also addressing the difficulty of achieving multiple objectives simultaneously. The most successful methods of meeting these demands will gener ally be ones that enable apples-to-apples comparisons, thereby permit ting separate examination of each dimension of a policy, which in turn allows identification of the policy' s intrinsic features. In the present set- ting, this result often is best achieved by postulating adjustments to the income tax and transfer system that yield distribution neutrality(and not merely revenue neutrality). Specifically, such an approach is useful whenever the policy in question is not purely redistributive(which it is in the case of a labor income tax or cash transfers). The result of this analytical technique is to separate redistributive effects of a policy, along with whatever distortion is caused by the act of redistribution itself, from other effects of the policy. The reasoning underlying this view will now be presented more fully, followed by further discussion of why this approach to comparability is so powerful in providing an integrated un derstanding of taxation, government expenditures, and redistribution. 5 1. The problen To make the analysis concrete, consider whether luxury taxes shoul be imposed Such taxation appears on its face to be redistributive, and that attribute is a likely motivation for implementing luxury taxes particular. As we know from section A, however, this characterization The approach presented in this section was initially presented in Kaplow(1996d)and is further elaborated in Kaplow(2004) Similarly, one could consider subsidies to necessities, such as are implicitly provi by the common practice of exempting expenditures on food from sales taxes and VATs. 03_Kaplow_Ch02_po11-p034 ndd ⑥ 1016/20071121:28AN
22 chapter 2 C. Comparability of Proposals under Assessment Section A explains the need to specify a policy completely in order to analyze it properly. But completeness, notably the requirement of budget balance, can be achieved in many ways. Some means of doing so better facilitate comparability than others. Moreover, the artful choice of policy combinations can take advantage of the range of available tools, as suggested in section B, while also addressing the diffi culty of achieving multiple objectives simultaneously. The most successful methods of meeting these demands will generally be ones that enable apples-to-apples comparisons, thereby permitting separate examination of each dimension of a policy, which in turn allows identifi cation of the policy’s intrinsic features. In the present setting, this result often is best achieved by postulating adjustments to the income tax and transfer system that yield distribution neutrality (and not merely revenue neutrality). Specifi cally, such an approach is useful whenever the policy in question is not purely redistributive (which it is in the case of a labor income tax or cash transfers). The result of this analytical technique is to separate redistributive effects of a policy, along with whatever distortion is caused by the act of redistribution itself, from other effects of the policy. The reasoning underlying this view will now be presented more fully, followed by further discussion of why this approach to comparability is so powerful in providing an integrated understanding of taxation, government expenditures, and redistribution.13 1. The Problem To make the analysis concrete, consider whether luxury taxes should be imposed. Such taxation appears on its face to be redistributive, and that attribute is a likely motivation for implementing luxury taxes in particular.14 As we know from section A, however, this characterization 13 The approach presented in this section was initially presented in Kaplow (1996d) and is further elaborated in Kaplow (2004). 14 Similarly, one could consider subsidies to necessities, such as are implicitly provided by the common practice of exempting expenditures on food from sales taxes and VATs. 03_Kaplow_Ch02_p011-p034.indd 22 3_Kaplow_Ch02_p011-p034.indd 22 10/16/2007 11:21:28 AM 0/16/2007 11:21:28 AM
AN INTEGRATED VIEW 23 is premature because the policy description is incomplete For example, if the revenue were spent deepening a harbor so it could handle larger yachts, the net effect might even be favorable to the rich This section focuses on what manner of completing the system best facilitates comparability. A common approach to analyzing taxation is to keep expenditures fixed(in level and composition)so that attention is confined to different ways of raising a specified amount of revenue To focus analysis further, adjustments to the income tax schedule are often considered. Even with this amount of refinement however, there still are many degrees of freedom, because there exists an infinite variety of ways in which the income tax may be adjusted to meet a given reve- nue target, in our example, to reduce tax collections by the amount of funds raised by the proposed luxury taxes. Regarding redistribution, virtually any outcome--netting the effects of the luxury tax scheme and the as-yet-unspecified income tax reduction-is possible. At this point, certain seemingly neutral yet essentially arbitrary focal points are usually invoked. For example, it may be hypothesized that everyone's marginal income tax rate is to be reduced by the same amount or percentage or that everyone 's tax payments are to be reduced by the same amount or percentage. Even this incomplete list, however, allows for a great range of distributive effects. For example, reducing every- one's tax payments by the same amount would provide an equal dollar benefit to everyone; reducing payments by the same percentage would benefit higher-income individuals substantially more. Accordingly, two analysts of a proposed luxury tax scheme could provide very different assessments of the distortion, degree of redistribution, and effect on overall welfare that it would produce even if they agree on their empiri al assumptions and other relevant factors. And when the underlying source of all the disagreement-different income tax adjustments-is mentioned only in passing, perhaps amidst many other technical details This formulation is problematic because many sorts of expenditures, including trans- fers, can be made within some tax systems, notably as part of an income tax, such as is don in the United States through the Earned Income Tax Credit. More broadly, there is no clear heoretical distinction between taxes and expenditures(see Green and Kotlikoff 2007 and Shaviro 2004). The approach presented later, involving distribution-neutral finance, provides a determinate resolution to the issue raised in the text without relying on any such distinction. 03_Kaplow_Ch02_p011-p034 ndd 23 ⑥ 1016/20071121:28AN
an integrated view 23 is premature because the policy description is incomplete. For example, if the revenue were spent deepening a harbor so it could handle larger yachts, the net effect might even be favorable to the rich. This section focuses on what manner of completing the system best facilitates comparability. A common approach to analyzing taxation is to keep expenditures fi xed (in level and composition) so that attention is confi ned to different ways of raising a specifi ed amount of revenue.15 To focus analysis further, adjustments to the income tax schedule are often considered. Even with this amount of refi nement, however, there still are many degrees of freedom, because there exists an infi nite variety of ways in which the income tax may be adjusted to meet a given revenue target, in our example, to reduce tax collections by the amount of funds raised by the proposed luxury taxes. Regarding redistribution, virtually any outcome—netting the effects of the luxury tax scheme and the as-yet-unspecifi ed income tax reduction—is possible. At this point, certain seemingly neutral yet essentially arbitrary focal points are usually invoked. For example, it may be hypothesized that everyone’s marginal income tax rate is to be reduced by the same amount or percentage or that everyone’s tax payments are to be reduced by the same amount or percentage. Even this incomplete list, however, allows for a great range of distributive effects. For example, reducing everyone’s tax payments by the same amount would provide an equal dollar benefi t to everyone; reducing payments by the same percentage would benefi t higher-income individuals substantially more. Accordingly, two analysts of a proposed luxury tax scheme could provide very different assessments of the distortion, degree of redistribution, and effect on overall welfare that it would produce even if they agree on their empirical assumptions and other relevant factors. And when the underlying source of all the disagreement—different income tax adjustments—is mentioned only in passing, perhaps amidst many other technical details 15 This formulation is problematic because many sorts of expenditures, including transfers, can be made within some tax systems, notably as part of an income tax, such as is done in the United States through the Earned Income Tax Credit. More broadly, there is no clear theoretical distinction between taxes and expenditures (see Green and Kotlikoff 2007 and Shaviro 2004). The approach presented later, involving distribution-neutral fi nance, provides a determinate resolution to the issue raised in the text without relying on any such distinction. 03_Kaplow_Ch02_p011-p034.indd 23 3_Kaplow_Ch02_p011-p034.indd 23 10/16/2007 11:21:28 AM 0/16/2007 11:21:28 AM