中心城市的工资税或混合工资税对就业的影响

中心城市的工资税或混合工资税对就业的影响

作者：RalphM.Braid

国籍：美国

出处：区域科学与城市经济学，2009年2月17日

【摘要】本文讨论中心城市的工资税对就业的影响。税收一般会减少中心城市的就业率，并增加郊区就业率。影响的大小依赖于中心城市征收的确切的工资税类型（对工作在中心城市的中心城市居民以及在中心城市工作的郊区居民征收的税率）、生产函数、通勤成本参数及多项式评定选择参数。

【关键词】当地工资税地方所得税中心城市就业

1、引言

本文研究的中心城市工资税的固定额对中心城市及其郊区之间的就业分布上的影响。在美国至少有14个州的地方政府使用当地工资或所得税。我一些早期的文献

(Braid,1996,2000,2002,2003,2005)用多种模型来研究这类问题，但所有的模型都基于单一来源的工资税的假设。此种类型的工资税也被一些国家的地方政府采用。本文的模型与我早期的文献有显著的不同，美国各种不同类型的中心城市的工资税的影响将被用以检验。

首先，其他理论和实证文献是值得讨论的。通过地方公共财政的理论模型对整个大都市区或“地区”的物业税或工资税变化的影响进行研究。然而了解大城市区域的不同地区之间的税率差异影响也是很重要的。首先，当地有代表性的财产税税率在大城市区域不同的司法管辖区之间通常会有差异。其次，一些中心城市有工资税或所得税，然而周围的郊区没有。

令人惊讶的是，很少有理论文章考虑到这种税差的影响。许多实证论文验证了大城市区域的当地税收差异在就业或企业场所方面的影响。值得注意的是

Grieson(1980),Haughwout等人（2004），Inman(1987)和Luce（1994）估计了费城所得税对就业的影响，其中非居民通勤的所得税率比美国任何一个其他城市的居民都要高。

我以前的一些论文，旨在补充这些基础数据分析，包括Braid(1996,2005)和Guo(2009)的理论文献，可用相关的方程来解决不同的（关于税收竞争）问题。商业用地是固定的，劳动力在大都市区的当地司法管辖区之间自由移动的。这些论文，都假设通勤是不花钱的，从而忽略了通勤成本。

一些论文假设了一维连续位置模型，即假设大都市区域是一个线段和消费者的通勤成

本的比例距离。Braid(2000)和Peralta(2007）假设在每两个地方管辖就有一个单一的就业中心。Braid（2002，,203）假设中心城市是被郊区所包围，并且可以在任何地方就业。

上文提到在过去的七篇文献中，工资税是一种纯粹的基于来源的工资税（有时也称为工资单税）。也就是说，中心城市的工资税税率同样适用于在中心城市工作的城市居民和郊区居民，但不适用于在郊区工作的城市和郊区居民。美国的几个大城市包括旧金山，洛杉矶，纽瓦克（新泽西州），伯明翰（阿拉巴马州）有这种工资税，而他们周围的郊区却没有。

在美国的几个城市有一种以纯居民为主的所得税，此外还有更多的“混合型”的所得税，所有中心城市的居民以同一种税率征税（无论他们在哪里工作），而在中心城市工作的郊区居民以同一税率或较低税率征税。一些大都市包括威尔明顿，堪萨斯城，圣路易斯，底特律和密歇根州，纽约市和费城等城市有此种税，而周围的郊区却没有（或拥有较低的税率）。

许多除美国之外的国家的地方政府也采用工资税或所得税，这些国家包括瑞典，丹麦。奥地利，法国，德国，日本和西班牙（见Braid,2005,第1790页）。我不能肯定的是，这些国家的当地工资税或所得说确切类型，但我相信在某些情况下，税收是纯粹以居民为主的，而Peralta(2007)指出，澳大利亚，奥地利，法国，希腊和墨西哥“在州或地方一级有工资税”，并在韩国，“所得税税收收入是地方一级的经济来源”。

在我的文献中这两种假设是不同的。首先，居住在一个区域但去另一个区域工作的任何人都会产生固定通勤成本（无成本的通勤或与距离成正比的通勤费用对比）。其次，在某个区域会有多少居民将在此地工作可用多种回归方式进行检测，该方法依赖于一个地区每个工人能赚取的净收入。

我的理论文章补充了Haughwout 等人（2004年）的实证结果，并提供了一个基本的理论模型（见第二节），同时提出对中心城市工资税影响的感觉和见解（见第三节，包括第1-6论点）。各种实证结果与我的计算结果之间的相似之处和差异在第五节被指出（见表1、

第四节和第五节）。此外，他们也估计纽约市居民基本税的影响（接近纯粹理论），以及“标准的混合工资税”在费城的影响，但不考虑一个城市以纯粹来源为基础的工资税（见我的命题3和表1）或”密歇根混合工资税“的影响（见命题4和表1）。

The employment effects of a central city’ssource-based

wage tax or hybrid wage tax

Author ：Ralph M.Braid

Nationality ：USA

Provenance ：Regional Science and Urban Economics,17February,2009

【Abstract 】This paper considers the employment effects of a central city’swage tax.The tax will normally decrease employment in the central city and increase employ ment in the suburbs.The magnitude of the effects depends on the exact type of wage tax that the central city imposes (thetax rates on central-city residents who work in the ce ntral city,on central-city residents who work in the suburbs,and on suburbanites who wor k in the central city),on a production function,on a commuting cost parameter,and on a multinomial-logit-choice parameter.

【Keywords 】

1、Introduction

In this paper,I examine the effects of central city’swage tax on the distribution of a fixed amount of employment between the central city and its suburbs.Local wage or inc ome taxes are used by at least some local governments in 14states in the United States. Some of my earlier papers (Braid,1996,2000,2002,2003,2005)canbe used to examine this sort of question in a variety of models,but all assume a pure source-based wage tax.This type of wage tax is used by local governments in some states.The model in this paper differs significantly from those in all of my earlier papers,and the effects of a wide varie ty of different types of central-city wage taxes that are actually used in the United State s can be examined.

First,it is worth discussing other theoretical and empirical literature.The effects of a change in a property tax or wage tax by a whole metropolitan area or a "region"are exa mined in theoretical models of local public finance.However,it is also important to unders tand the effects of tax rate differentials between different parts of a metropolitan area.Fir st,local property tax rates typically differ between the different jurisdictions of a metropol itan area .Second,some central cities have wage taxes or income taxes,whereas the surrou

Local wage taxes;Local income taxes;Central city;Employment

nding suburbs do not.

Surprisingly,few theoretical papers consider the effects of such tax differentials.Many empirical papers have examined the effects of localized tax differentials within a metrop olitan area on employment or business location.Notably,Grieson(1980),haughwoutet

al(2004),Inman(1987),andLuce (1994)estimatethe employment effects of the Phil

adelphia income tax,which has a much higher tax rate on non-residents who commute in than any other city in the United States.

Some of my previous papers are intended to complement these data-based analyses.B raid(1996,2005)andGuo(2009)aretheoretical papers that include relevant equations but use them to adddress different questions (abouttax competition).Business land is immobile,a nd labor is freely mobile among the local jurisdictions of the metropolitan area.These pa pers ignore commuting costs by assuming that commuting is costless.

Several papers assume one-dimensional-continuous-location models in which the metr opolitan area is a line segment and consumers have commuting costs proportional to dist ance traveled.Braid(2000)andPeralta(2007)assumea single employment center in each of two local jurisdictions.Braid(2002,2003)assumesthat a central city is surrounded by subur ban areas,and that employment can be anywhere.

The last seven papers mentioned above assume that a wage tax is a pure source-ba sed wage tax (sometimesknown as a payroll tax).That is,a central city’swage tax applie s at the same rate to central-city and suburban residents working in the central city,but n ot to central-city and suburban residents working in the suburbs.Several large cities in th e United States have this sort of wage tax,while surrounding suburbs do not,including Sa n Francisco ,Los Angeles ,Newark (NewJersey),and Birmingham(Alabama);

A couple of cities in the United States have a pure residence-based income tax.Ma ny more have a “hybrid”incometax ,with all central-city residents taxed at one rate (wherever they work),and with suburbanites who work in the central city taxed at the same or a lower rate .Large cities with this sort of tax,while surrounding suburbs do not have them (orhave them at a much lower rate),include Kansas City ,St,Louis,Wilmington,Det roit and other cities in Michigan,New York City,and Philadelphia.

Local government use wage or income taxes in a number of countaies beside the U.S,including(seeBraid,2005,p.1790)Sweden,Denmark,Austria,France,Germany,Japan and Sp ain.I am not sure of the exact type of local wage or income taxes in these countries,but I believe that in some cases the taxes are purely residence-based,whereas Peralta (2007)points out that Australia,Austria,France,Greece and Mexico“havepayroll taxes at the state or local level”,andthat in Korea,“incometaxation at the local level is source-based”.

This paper presents a long-run model that is a significant variation of Section 1of Braid(1996).5I assume that there are several discrete zones in the metropolitan area,one of which is the central city.Each zone has a fixed amount of business land ,which is a t hird input into production (alongwith labor and business capital).This is similar to Braid (1996,2005)butcontrasts with the absence of land in Braid(2000,2002)andPeralta (200

7).

Two assumptions are different than in any of my other papers.Frist,there are fixed c ommuting costs for anyone who lives in one zone and commutes to work in another zo ne （in contrast to costless commuting or commuting costs proportional to distance).Seco nd,there is a multinomial logit determination of how many workers living in each zone will work in that zone and in each other zone,depending on the net wages a worker can earn in each zone.

My theoretical paper complements the empirical results of Haughwout et al. （2004），by providing an underlying theoretical model (seeSection 2),and by providing intuition a nd insights about the effects of central-city wage taxes (seeSection 3including Propositi ons 1-6).Various similarities and differences between their empirical results and my nume rical results (seeTable 1and Section 4and 5)are pointed out in Section 5.Also,they estim ate the effects of the (close-to-pure)residence-basedwage tax in New York City,and the “standardhybrid wage tax”inPhiladelphia,but do not consider a city with a pure source-based wage tax(seemy Proposition 3and Table1)or “Michiganhybrid wage tax”(seePr oposition 4and Table 1).