Economic Studies 112. Che-Yuan Liang Essays in Political Economics and Public Finance

Economc Studes 112 Che-Yuan Lang Essays n Poltcal Economcs and Publc Fnance Che-Yuan Lang Essays n Poltcal Economcs and Publc Fnance Department of Economcs, Uppsala Unversty Vstng address: Kyrkogårdsgatan
of 27
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Economc Studes 112 Che-Yuan Lang Essays n Poltcal Economcs and Publc Fnance Che-Yuan Lang Essays n Poltcal Economcs and Publc Fnance Department of Economcs, Uppsala Unversty Vstng address: Kyrkogårdsgatan 10, Uppsala, Sweden Postal address: Box 513, SE Uppsala, Sweden Telephone: Telefax: Internet: ECONOMICS AT UPPSALA UNIVERSITY The Department of Economcs at Uppsala Unversty has a long hstory. The frst char n Economcs n the Nordc countres was nsttuted at Uppsala Unversty n The man focus of research at the department has vared over the years but has typcally been orented towards polcy-relevant appled economcs, ncludng both theoretcal and emprcal studes. The currently most actve areas of research can be grouped nto sx categores: * Labour economcs * Publc economcs * Macroeconomcs * Mcroeconometrcs * Envronmental economcs * Housng and urban economcs Addtonal nformaton about research n progress and publshed reports s gven n our proect catalogue. The catalogue can be ordered drectly from the Department of Economcs. Department of Economcs, Uppsala Unversty ISBN ISSN Doctoral dssertaton presented to the Faculty of Socal Scences 2008 Abstract Dssertaton at Uppsala Unversty to be publcly examned n Hörsal 1, Ekonomkum, Frday December 5, 2008, at 1:15 p.m., for the degree of Doctor of Phlosophy. The examnaton wll be conducted n Englsh. LIANG, Che-Yuan, 2008, Essays n Poltcal Economcs and Publc Fnance, Department of Economcs, Uppsala Unversty, Economc Studes 112, 125 pp., ISBN , ISSN , urn:nbn:se:uu:dva-9340, Ths thess conssts of four self-contaned essays. Essay 1: Ths paper frst extends a recently developed nonparametrc approach to estmate labor supply (Blomqust and Newey 2002) to handle cases when there are ndvduals that do not work. The method s used to estmate the labor supply of marred women n Sweden. These estmates are then, together wth exstng estmates for marred men, used to smulate the earned ncome tax credt program ntroduced n Sweden n 2007 wth respect to marred couples. I fnd that the reform ncreases labor supply wth 3.6 percent and gross labor ncome wth 3.1 percent. The cost s a net revenue loss of 0.5 percent for the government. Essay 2: Ths paper nvestgates emprcally the effects of poltcal representaton on the electoral outcome at the party and coalton levels n proportonal electon systems. There are two notons of representaton: to hold seats and to belong to the rulng coalton. I refer to the effect of the former as the ncumbency effect and the effect of the latter as the effect of rulng. I fnd that ncumbency determnes the dstrbuton of 12 percent of the votes, whch s smlar to the advantage found n maortaran systems. Further, I fnd no effect of rulng, contrary to the commonly found cost of rulng n proportonal systems. Essay 3: Ths paper frst presents a model of how local government poltcans lobby to rase ntergovernmental grants to ther local governments. The model dentfes an effect of councl sze on grants receved. Ths relatonshp s then studed emprcally. I fnd a large negatve effect of councl sze on grants receved producng a group sze paradox. The effect could be explaned by free-rdng ncentves n ndvdual lobbyng effort contrbuton among local government poltcans whch arse due to a collectve acton problem. Ths pattern provdes ndrect evdence for the occurrence of ntergovernmental lobbyng. Essay 4 (co-authored wth Henrk Jordahl): Ths paper uses the Swedsh muncpal amalgamaton reform of 1952 to study the common pool problem n poltcs. The amalgams are common pools and the muncpaltes have ncentves to free-rde on ther amalgam partners by ncreasng debt pror to amalgamaton. We fnd that muncpaltes that merged n 1952 ncreased ther debt between 1948 and 1952 when the reform could be antcpated. The ncrease amounted to two thrds of new debt ssued or two percent of total revenues n the merged muncpaltes. But free-rdng dd not ncrease wth the common pool sze contrary to the predcton of the law of 1/n. Acknowledgement My chldhood dream was to become a researcher. Now that I know what research s about, I realze that t s even better than n the dream. Conductng research s, n my opnon, the most nterestng employment one could have: t s fun, challengng, and rewardng. Prerequstes for good research are good deas, n-depth analyss, and hard work. I could not have conducted all these parts by myself. I have reached ths stage thanks to many people. Most mportantly, my man supervsor Sören Blomqust has provded me wth excellent gudance n thnkng research, generous support wth the scarce resource that tme s, and nsghtful comments on endless drafts and redrafts. My assstant supervsor Henrk Jordahl has also made a maor contrbuton to ths thess, to my professonal and personal development, and has gven me plenty of good advce. Further, I want to acknowledge Davd Strömberg and Per Johansson dscussants at my Lcentate and fnal semnars, for comments and suggestons of hgh qualty. The Department of Economcs at Uppsala Unversty has been a perfect educatonal, research, and socal envronment. I have learned and benefted from, and enoyed beng around ntellgent, helpful, and frendly professors, researchers, fellow students, and admnstratve staff. These people nclude n partcular Per Engström, Ann-Sofe Kolm, Eva Mörk, Mkael Elnder, Håkan Seln, Qan Lu, Lars Lndvall, Hans Grönqvst, Carolne Hall, Jakob Wnstrand, Johan Söderberg, Nklas Bengtsson, and Katarna Grönvall. Specal thanks go to my wonderful and enthusastc offce mate Erk Glans. Wrtng a thess s absorbng, especally for an addctve personalty lke mne. I would not have survved t wthout a lfe outsde. Thank you, all my frends, for long dnners, deep conversatons, floorball and badmnton exercses, vdeo and board game sessons, and long crazy nghts. The team mates n the coolest floorball team ever BK Forever, deserve a specal menton. The rest of you, not mentoned by name but nevertheless not forgotten: you are mportant to me and enlghten my lfe. Needless to say, I would not be here wthout my mother Gnger Lang. She rased me durng the harshest tme n her lfe, but always put my best frst. My eternal grattude also goes to Per, Kerstn, and Smon Brandell for beng there durng the darkest hour. Fnally, Annette Östrand my soul mate and greatest love: I want to spend so much more tme wth you. Che-Yuan Lang Uppsala, October 2008 Content Introducton...1 Poltcal Economcs and Publc Fnance...1 Estmatng Causal Relatonshps...2 References...3 Essay 1. Nonparametrc Estmaton of Female Labor Supply and the 2007 Swedsh Earned Income Tax Credt Reform Introducton Lterature Revew Labor Supply Functons Emprcal Specfcaton and Estmaton Data and Summary Statstcs Estmaton Results Labor Supply Estmates Tax Systems and the Rse n Female Labor Supply Smulatng the 2007 Swedsh EITC Reform The 2007 Swedsh EITC Reform The Effects on Labor Supply The Effects on Incomes Concludng Dscusson...35 Acknowledgement...36 Appendx...37 Appendx A...37 Appendx B...40 Appendx C...41 References...43 Essay 2. Is There an Incumbency Advantage or a Cost of Rulng n Proportonal Electon Systems? Introducton Insttutonal Background and Data Summary Statstcs The Incumbency Effect The Effect of Rulng Concluson...67 Acknowledgement...67 References...68 Essay 3. Collectve Lobbyng n Poltcs: Theory and Emprcal Evdence from Sweden Introducton Theory The Model Equlbrum Propertes Other Theoretcal Issues Insttutonal Background Emprcal Strategy Descrpton of Varables Emprcal Specfcaton Descrptve Analyss Summary Statstcs Graphcal Analyss Regresson Results Man Results Threshold Specfc Estmates Concludng Dscusson...96 Acknowledgement...98 Appendx...98 References...99 Essay 4. Merged Muncpaltes, Hgher Debt: On Free-Rdng and the Common Pool Problem n Poltcs Introducton Insttutonal Background Identfcaton Strategy and Data Emprcal Specfcaton Results Concluson Acknowledgement References v Introducton The essays n ths thess study rather dvergent ssues n economcs. Common features are ther focus on postve and appled emprcal analyss usng Swedsh mcrodata. In ths ntroducton, I wll try to llumnate some other ponts of contact and put the essays n a broader perspectve, both wth respect to the topcs studed and the methods used. Poltcal Economcs and Publc Fnance The lowest common denomnator of the essays s publc economcs, whch s the study of the government n the economy, n contrast to e.g. the study of households and frms. More narrowly, the essays deal wth ssues n the felds of poltcal economcs (Essay 2, 3, and 4) and publc fnance (Essay 1, 3, and 4). Poltcal economcs s the study of poltcal decson makng and the mpact of ths process on economc polcy. 1 The phlosophcal pont of departure s that actors n poltcs, be they voters, nterest groups, or poltcans, have ndvdual preferences and act ratonally upon them n the same way as actors n a market. Arrow s mpossblty theorem 2 (1951) had a maor mpact on the drecton of development toward postve analyss at the expense of normatve analyss. 3 In representatve democraces, poltcs s characterzed by voters electng poltcans, who decde upon polces, whch n turn feed back on voters. Essay 2 looks at the lnk between voters and poltcans by studyng whether poltcans gan popularty whle n offce. In contrast, Essay 3 and 4 look at the lnk between poltcans and polcy by studyng whether poltcans act n the nterest of certan groups of voters more than others due to allegance or re-electon concerns. The knd of problem studed both n Essay 3 and 4 s the common pool problem, whch arses when the costs of an actvty for a smaller group of people s shared among a larger group. An everyday example would be a dnner where two famles have decded to splt the bll: both famles then have ncentves to order too much. Such stuatons could, as Buchanan and Tullock (1959) ponted out, also arse n poltcs. Publc fnance s, n a narrow sense, the study of the fnance of governmental actvtes. 4 In contrast to poltcal economcs, the focus les exclusvely on polcy, and not on the decson makng process. There s also a more wdespread nterest n normatve analyss. 1 Poltcal economy s perhaps a more commonly used term for ths feld. However, that term was hstorcally used to descrbe what s now smply referred to as economcs. 2 The theorem states that t s mpossble to convert several ndvdual preferences nto a sngle socetal preference wthout volatng reasonable prncples of how ths should be done. 3 See Persson and Tabelln (2000) for a treatment of ths feld from a unfyng framework. 4 In a broader sense, publc fnance s synonymous wth publc economcs. In the narrower sense, publc fnance does not nclude non-budgetary ssues such as monetary polcy. 1 Taxaton s the foundaton of governmental actvtes. However, taxaton often dstorts the allocaton of resources, labor, and consumpton n a way that creates neffcences. Optmal taxaton s a core area n publc fnance that deals wth the optmal desgn of the tax system. Ramsey (1927) lad the basc theoretcal foundaton, and a maor nsght of hs s that the answer partly depends on the responsveness of behavor to taxaton. The responsveness of labor supply to taxaton s examned n Essay 1. Governng power s partly delegated to local governments n most modern democraces. Such decentralzaton could brng governance closer to the people and allow polcy dfferentaton across regons. Despte some ndependence, local governments are typcally ted to the central government to some degree. Fscal federalsm s an area wthn publc fnance that deals wth the fnancal relatonshp between dfferent levels of government. Although the man focus s on poltcal decson makng, the polcy outcomes studed n Essay 3 and 4 are also of mportance for understandng the operaton and fnances of local governments. Estmatng Causal Relatonshps Economcs could be descrbed as the applcaton of natural scence methods to socal scence topcs. Quanttatve methods are almost exclusvely used n emprcal analyss. In appled work, the ambton s often to reveal causal relatonshps, as opposed to mere correlatons. In natural scence, the causal effect of varable A on varable B s usually establshed by conductng an experment where B s studed as A s vared, ceters parbus,.e. keepng everythng else equal. Economsts can seldom conduct experments and generally only have access to observatonal data. The problem s typcally that observed covaraton between A and B may be caused by systematc varaton n other confoundng factors,.e. the ceters parbus clause does not hold. There are two dfferent approaches n tryng to solve ths problem: the structural approach (used n Essay 1) and the reduced-form approach (used n Essay 2, 3, and 4). The structural approach s to develop a theoretcal model that descrbes how B s determned by A and confoundng factors, and then to derve an emprcal specfcaton that s tested from ths model. The results can be nterpreted n the lght of the model whch reveals channels of causaton. It s, however, dffcult to account for all relevant confoundng factors n the theoretcal model, whch easly becomes too complcated to work wth. A structural approach s used to nvestgate the mpact of the earned ncome tax credt program ntroduced n Sweden n 2007 on labor supply n Essay 1. The reduced-form approach tres to solate random varaton n A. Ths could be done by accountng for confoundng factors n the emprcal specfcaton (generally n crude ways wthout theoretcal modelng), or more elegantly through the use of natural experments, mostly by studyng nsttutonal reforms. Reduced-form approaches usually dentfy causal effects 2 more convncngly than structural approaches. The results are, on the other hand, more dffcult to nterpret and less useful for understandng the economcs behnd the causal relatonshp. One reduced-form desgn s dfference-n-dfferences (DD). Suppose A could take two values, treatment and no treatment, and that we are nterested n the causal treatment effect on B. The problem wth the total dfference n B between treated and untreated unts s that t could contan a dfference due to varous confoundng factors n addton to the causal treatment effect dfference. The dea of DD s to approxmate the confoundng dfference by measurng the dfference n a tme perod where none of the unts were treated and remove t from the total dfference. A DD desgn s used to nvestgate the mpact of expectatons of local government amalgamaton on debt n Essay 4. Another, more recently developed, reduced-form desgn s regresson dscontnuty (RD). RD makes use of stuatons wth abrupt dscontnuous varaton n A that s completely determned by smooth contnuous varaton n another assgnment varable C. Suppose that treatment s gven only f C s larger than a cutoff value. The dea of RD s to compare treated unts ust above the threshold wth untreated unts ust below. The only systematc dfference between the groups s then treatment status and a neglgble dfference n C. 5 RD desgns are used to nvestgate the mpact of poltcal representaton on the electoral outcome n Essay 2, and the mpact of local government councl sze on ntergovernmental grants receved n Essay 3. Emprcal methods could also be classfed along a parametrc to nonparametrc dmenson dependng on the assumptons made on the exact way varables n and outsde the analyss affect the outcome varable. Nonparametrc methods allow weaker assumptons but also requre more data and computatonal power, whch are resources that have become ncreasngly avalable lately. Nonparametrc methods are used n Essay 1, but also to some extent n Essay 2 and 3. References Arrow, K. J., 1951, Socal Choce and Indvdual Values, Wley, New York. Buchanan, J., and Tullock, G., 1962, The Calculus of Consent: Logcal Foundatons of Consttutonal Democracy, Unversty of Mchgan Press, Ann Arbor. Imbens, G. W., and Lemeux, T., 2008, Regresson Dscontnuty Desgns: A Gude to Practce, Journal of Econometrcs 142, Persson, T., and Tabelln, G., 2000, Poltcal Economcs: Explanng Economc Polcy, MIT Press, Cambrdge, Massachusetts. Ramsey, F. P., 1927, A Contrbuton to the Theory of Taxaton, Economc Journal 37, See Imbens and Lemeux (2008) for an ntroducton to RD. 3 Essay 1 Nonparametrc Estmaton of Female Labor Supply and the 2007 Swedsh Earned Income Tax Credt Reform 1. Introducton The am behnd many tax reforms s to ncrease labor supply. The earned ncome tax credt (EITC) s a program that recently has been ntroduced n a number of countres. An EITC program has been n place snce 1986 n the Unted States. Other countres wth a smlar program nclude Great Brtan, Canada, Ireland, New Zealand, Austra, Belgum, Denmark, Fnland, France, and the Netherlands. In 2007, Sweden became the latest country to ntroduce such a program. An EITC program s supposed to enhance work ncentves, especally at low hours of work. One reason for ts popularty may be that t s wdely thought that labor supply s more senstve on the partcpaton margn than on the hour margn. Snce the partcpaton rate s lower for women, knowledge about the responsveness of female labor supply s crucal for evaluatons of the mpact of EITC programs. It s, however, econometrcally challengng to estmate labor supply. A maor dffculty arses because the tax and transfer system creates nonlnear budget sets wth many segments. Indvduals can then affect the margnal net wage rate they face by selectng whch segment to locate on. Hours of work and the margnal wage are therefore smultaneously determned. Addtonally, there are knks n the budget set around whch the behavor could be fundamentally dfferent than on the segments. Another ssue s that the functonal form of the labor supply functon s unknown. Further, some methods depend on strong dstrbutonal assumptons that may drve the estmaton results. In the case of female labor supply, there are also a consderable number of ndvduals that choose the corner soluton of zero hours of work, whch makes censorng an mportant factor to account for. Blomqust and Newey (2002) recently developed a nonparametrc approach that ncely handles the mentoned dffcultes when there s no censorng. Blomqust (2001), Blomqust and Newey (2002), Wu (2005), and Kumar (2006, 2007) are the only other papers adoptng ther approach. I extend the approach to the case wth censorng. In addton to estmaton of expected hours of work, the extenson enables estmaton of the partcpaton probablty wthn the same framework, whch makes t possble to sort out responses on the partcpaton and hour margns. The method s then used to 5 estmate the labor supply of marred or cohabtng women n Sweden wth data from four years between 1973 and Labor supply estmates are usually summarzed n terms of elastctes. Snce budget sets are nonlnear there are many net wage varables. It s therefore artfcal to speak of the effect of changng the net wage. Ths problem s partcularly severe wth the present method as varables characterzng the whole budget set enter the estmaton. These ssues (dscussed at length n secton four) lead me to conclude that the estmates from the present method are most useful for predctng labor supply under dfferent budget sets. Taxes could be evaluated through smulatons of behavor under dfferent tax systems. Such smulatons n ths paper reveal that the evoluton of the tax system does explan part of the rse n female labor supply durng the sample perod. The fnal part of ths paper evaluates the EITC program ntroduced n Sweden n 2007 through smulatons usng the labor supply estmates. Ths s a controversal reform, and there are dscussons of further extensons. The proponents argue that t mproves work ncentves n partcular partcpaton ncentves, and that ths fnances large parts of the reform. They also beleve that t ncreases net household ncomes and ncome equalty. The opponents argue that t s an expensve reform that deterorates publc fnances, and they fear that eventual defcts could be fnanced n a way that ncreases ncome nequalty. The reform conssts of two components, one that ncreases basc deductons n the lowest labor ncome bracket, and one that reduces margnal tax rates more moderately n a subsequent larger bracket. There are no practcal dfferences between tax cr
Similar documents
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks