The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
REGIONAL ECONOMIC ISSUES Working Paper Series Com m unity D evelopm ent-Fiscal Interactions: Theory and Evidence from the Chicago A rea William H. Oakland and William A. Testa FEDERAL RESERVE BANK OF CHICAGO WP- 1995/7 Community Theoiy D e v e l o p m e n t - F i s c a l Interactions: a n d E v i d e n c e f r o m the C h i c a g o A r e a William H. Oakland and William A. Testa* This study assesses the relationship between residential tax burdens and the process of business growth in Chicago metropolitan area suburbs. Itcorrects for the methodological shortcomings of previous studies through the development of an explicit model of local community behavior and through the use of a more complete measure of local tax burdens. The fragmented character of local government in the Chicago area provides an excellent laboratory for studying the local fiscal impacts of growth. This study is not limited to this single, albeit important, issue. It also addresses the question of the mechanism through which business development impacts local taxes and expenditures. Is the fiscal impact of business development primarily through itseffects on the community tax base, or does italso have importantexpenditure effects? Are the fiscalbenefits ofeconomic development manifested in greater public services or private goods and services? Only in the former case would the fiscal benefits take the form of lower local tax rates. ♦The authors are, respectively, assistant vice president at the Federal Reserve Bank o f Chicago and professor o f econom ics, Tulane University. This paper is drawn from a research project supported by the Metropolitan Planning Council and the Federal Reserve Bank o f Chicago. The materials and findings presented here are attributable to the authors and not to their respective employers. At the Federal Reserve Bank o f Chicago, David R. Allardice provided both direct assistance and program support. The authors would like to thank the following people for research assistance: Virginia Carlson, Xiao Chen, Richard Kaglic, and Robert McNamara. Thanks also to Jeny Szatan for his work on project organization, and for helpful comments. The authors also received helpful comments from members o f the Regional Development Committee and the Technical Advisory Group o f the Metropolitan Planning Council. Guidance on publicly available data was given by the staff o f the Property Tax Division o f the Illinois Department o f Revenue and the Illinois Department o f Employment Services. FRB CHICAGO Working Paper August 1995, WP-1995-7 1 We also address the issue of whether favorable fiscal consequences of economic development come partly at the expense of neighboring communities. This can happen ifjob creation in one community leads to population increases in other communities which in turn give rise to higher service demands than increased tax revenues. Such an eventuality raises the specter that growth policies that look attractive to one land-use authority may lead to negative economic consequences for the region as a whole. The Theoretical M odel Budget constraint The most basic relationship influencing the relationship between property tax rates and business development stems from the budget identity. This states that (1) t(Br+Bn) + O = E, where: t = property tax rate, Br = assessed value of residential property, B" = assessed value of business property, 0 = other revenues, and E = expenditure. By the budget identity, ifotherfactorson the right-hand side of (1) remain the same, an increase in business property would allow the community to reduce its tax rate. These other factors, however, may not remain the same. Thus, it is necessary to incorporate expenditure behavior into the model. FRB CHICAGO Working Paper August 1995, WP-1995-7 2 Expenditure Behavior An increase in business activity may require a community to increase its outlays for business services, such as highways, police, fire, etc. If there remains any surplus revenue, the community may also choose to expand services to its residents. Spending for residential services may also increase if new population is attracted by the increased employment opportunities. Finally, an increase in business tax base may cause a reduction in the "price" of residential public services. In effect, a larger share of any additional tax levies would be borne by the business community. Hence, the public may support an expansion of public consumption. Thus, there are several avenues through which an increase in business activity can stimulate public expenditures. These remarks can be formalized as follows: = E" + Er, (2) E (3) E" = E"(W), (4) Er = Er(N,p,Y,C,H), E" = business public services, Er = residential public services, W = employment within the jurisdiction, N = population of the jurisdiction, Y = per capita income within the jurisdiction, C = other relevant community demographics, H = average house value within the jurisdiction, and P = residential share of property tax base, Br/ (Br+ Bn). where FRB CHICAGO Working Paper August 1995.WP-1995-7 3 Equations (3) and (4) correspond to the demand for services by the business and residential sectors, respectively. Business services depend upon employment within the jurisdiction. Residential services depend upon community population, "price” of public services, income, and wealth within the community. The latter is represented by average house price while the "price" of public services is measured by the residential share of the property tax base.1 Like the demand for private goods and services, a reduction in the price of public services is expected to lead to an expansion in the public's demand for such services. To sum up, the expansion ofbusiness activity can impact expenditures through several routes: (1) directly, by increasing the need for services to support the business; (2) indirectly, by increasing the local population; and (3) indirectly, by lowering the "price" of public expenditures. Population The population of a community will reflect the number of people who seek to live there, population supply, and the number of people existing residents would like to accommodate, that is population demand. The population seeking residence within a community will depend upon the job opportunities both within the community and in itsenvirons. The latter is a potential important source of fiscal inter-relationship among neighboring communities, particularly ifnew population fails to cover its full expenditure costs. Population supply may also respond the fiscal environment; high tax rates will discourage population while ample and high quality public services will attract population. Finally, environmental quality and access to the CBD will influence population supply levels.5 (5) FRB CHICAGO Working Paper August 1995, WP-1995-7 Ns = N(t,W,Wn,e,Q), 4 where Ns = population supply, W" = employment in surrounding jurisdictions, e = per capita public expenditures, and Q = community amenities (distance to CBD, crime rate, congestion, etc.). Population demand, on the other hand, ismore difficultto pin down, to a large measure because such behavior is associated with tastes, as opposed to objective factors. Some communities may seek to exclude new residents because of the fear of congestion or crime. They may also fear that the new residents will not cover their full fiscal costs. Or they simply may seek exclusivity. All of these factors are likely correlated with income and wealth. A community predominantly inhabited by high-income individuals would have to subsidize the public services ofnew entrants whose incomes fall short ofthe community norm. They would also lose exclusivity ifless fashionable housing were to accompany new migrants. To avoid such outcomes, the governing authority may enact large lot zoning and restrictions to rental property. Thus, (6) N d = Nd(H,Y). The actual population of a community (equilibrium) will balance the supply and demand factors. Unlike most markets, where price adjusts to equate demand and supply, no such mechanism exists. Rather, the outcome will be a weighted outcome of the two sides, with the weights reflecting the relative political strength of developers and existing residents. Thus, (7) N with 0 < a < 1. FRB CHICAGO Working Paper August 1995, WP-1995-7 = a*Ns+ (l-a)*Nd, 5 Business activity L ike population, there are separate forces controlling the dem and for b u sin ess location and the supply o f b u sin ess sites w ithin a com m unity. On the dem and side, the level o f b u sin ess activity w ithin a com m u n ity w ill depend, in part, upon the tax rate it faces. W hile local taxes m ay not be the m ost sign ifican t elem en t o f a firm's co sts, it is one over w h ich it m ay have considerable control through its ch o ice o f b u sin ess location. W h ile labor co sts m ay be m ore important to a firm's bottom line, there is little they can d o to control them ; w age le v e ls tend to be relatively uniform w ithin a m etropolitan eco n o m y . B u sin ess dem and m ay also be stronger i f there is an am ple labor p o o l from w h ich to draw e m p lo y ees, and i f there is g o o d a c cess to markets. E xam ples o f the latter are: distance to the C B D , proxim ity to freew ay interchanges, and am ount o f econ o m ic activity nearby. (8 ) Wd = W d(t,N ,N n,W n,A ), N" = neighboring population, and A = b u sin ess am enities. w here O n the supply side, b u sin esses w ill be attractive to com m u n ities b ecau se o f the tax revenue they provide as w ell as jo b opportunities for residents. H ow ever, com m u n ities m ay have to pay a price for th ese advantages. Greater b u sin ess a ctivity m ay be accom panied by increased co n gestion and environm ental degradation. The w illin g n ess to trade o f f such am en ities for the fiscal advantages w ill be in versely correlated w ith in com e w ealth, and e x istin g em p loym en t le v e ls (co n g estio n ) and p o sitiv ely w ith the prevailin g tax rate (m easuring the need for fiscal relief). (9 ) FRB CHICAGO Working Paper August 1995, WP-1995-7 W* = W s(t,Y ,H ). 6 T he actual ou tcom e w ill be a w eigh ted average o f the dem and and supply factors (1 0 ) W = b*W d + (l-b )* W s 0 < b < 1. It w ill be noted that the tax rate enters both the dem and and supply sid es, w ith o p p osite effects. It is because o f this that studies som etim es find little relationship b etw een fiscal p o licies and bu sin ess location. C om m u n ities m ay be w illin g to e sc h e w the fiscal w in d falls provided by b u sin ess, b ein g m ore concerned w ith environm ental matters. O ften such com m u n ities h ave lo w property tax rates, thereby confounding the em pirical relationship b etw een tax le v e ls and b u sin ess location. Recapitulation The four relationships described ab ove provide a general equilibrium fram ework w ithin w hich the issue o f the fiscal co n seq u en ces o f eco n o m ic d evelop m en t can be investigated. T his m odel m akes ex p licit the im portant interrelationship betw een econ om ic developm ent, public expenditures, tax rates, and population. It further enables us to identify and understand the various lin k ages that ex ist betw een these important m agnitudes. F inally, it dem onstrates the potential w eak n esses o f em pirical studies w hich fail to a llo w for im portant feedback effects. N ot on ly w ould this result in statistical bias, but it m ay lead to m isunderstanding o f cause and effect relationships. Some Complications The preceding theoretical m odel id en tifies those variables that im p in ge on the relationship b etw een the property tax rate and com m ercial and industrial d evelop m en t. T he theoretical m odel w ill have to be m o d ified to reflect data availability and the fact that property taxes w ithin a m u n icip ality are lev ied by num erous overlapping governm ents. B egin n in g w ith overlapping governm ent, an increase in b u sin ess property w ill in flu en ce the tax rate o f the m unicipality by the am ount indicated by the budget identity in equation (1). H ow ever, for those overlapping ju risd iction s, the potential reduction in the property tax rate depends upon the siz e o f the property tax base w ithin the m unicipality in com parison w ith that o f the FRB CHICAGO Working Paper August 1995. WP-1995-7 7 overlapping district. S u ppose, for exam p le, that n ew b u sin ess d evelop m en t w o u ld increase the siz e o f the m unicipal property tax base b y 50% . T h is co u ld p oten tially reduce the tax rate lev ied b y the m unicipality b y 50% . H ow ever, su p p ose that the n ew b u sin ess d evelop m en t o n ly am ounts to 10% o f the property tax base o f the sch ool districts w hich overlap the m unicipality. T his w o u ld o n ly perm it a reduction in the sch o o l tax rate o f o n ly 5%. T he reason for the discrepancy b etw een the tw o e ffects is that the tax base b en efits o f the n ew b u sin ess d evelop m en t is shared w ith residents ou tsid e o f the m u n icip ality for sch ool tax purposes but accrues entirely to m u n icip ality residents for the m unicipal lev y . In effect, the m u n icip ality creates an externality for surrounding residents. N ot o n ly d oes th is reduce the fiscal b en efits from b u sin ess d evelopm ent, but it w ill also in flu en ce its w illin g n e ss to supply b u sin ess sites as h yp oth esized in equation (9 ) o f the theoretical m od el. T o take into account the p o ssib le dissipation o f the fiscal b en efits o f b u sin ess d evelop m en t, a m easure o f the dispersal o f tax base b en efits n eed s to be d evelop ed . For this purpose, w e em p lo y ed a sim ilar land area im putation as u sed to obtain the com p osite m unicipal property tax burden. Here* h ow ever, w e com puted the fraction o f the area served b y the overlapping govern m en t w h ich is located w ithin the m unicipality. T h ese w ill be aggregated u sin g the overlapping governm ent tax rate as w eigh ts. T he resulting variable w ill mirror the degree to w h ich the tax b en efits o f b u sin ess d evelop m en t are captured by the m unicipality. In turn, this variable w ill in flu en ce equations (1 ) and (1 0 ). The demand for services provided by overlapping jurisdictions T he fact that there are num erous overlapping governm ents servin g the m u n icip ality w ill also influ en ce the dem and for public serv ices b y residents, as described by equation (4). T he share o f taxable property w h ich is residential m ay vary across type o f governm ent. T his is apt to be m ost im portant for sch ool districts. Thus, the tax price o f a m u n icip ality w ill b e a co m p o site w h ich reflects the resp ective tax prices o f the sch o o l district and m u n icip ality governm ents w hich serve m unicipal residents. Sim ilar considerations apply to incom e and the other dem ographic variables in flu en cin g residential dem and for public services. H ow ever, data lim itations preclude a correction for such potential diversity. FRB CHICAGO Working Paper August 1995, WP-1995-7 8 W h ile our data set is rich, it is un lik ely that w e w ill have en ou gh inform ation to fu lly account for inter-jurisdictional variations in fiscal behavior. It is p o ssib le, therefore, that w e m ight erroneously im pute to bu sin ess d evelop m en t, effects w hich are cau sed b y other variables w hich w e cannot observe, or to errors in data m easurem ent. Fortunately, the availability o f data at m ultiple tim e points a llo w s us to sidestep this issu e b y focu sin g upon ch an ges in fiscal behavior over tim e. M any o f those variables w e cannot observe w ill not change sign ifican tly over our sam ple period; for exam ple, relative locational advantages are slo w to change. Sim ilar errors in m easurem ent caused, say by our land area im putations, m ay be relatively constant, and thus not in flu en ce changing behavior over tim e. F ocu sin g upon ch an ges over tim e is also com patible w ith our fiscal effects o f b u sin ess grow th. C om m unities w hich c h o o se to le v e ls o f com m ercial d evelopm ent say because o f environm ental m ay n everth eless react sim ilarly to changes in business activity. interest in the have different consideration, M oreover, the im pacts o f develop m en t are lik ely to e v o lv e w ith tim e. Such adjustm ents are best captured by an approach w hich fo cu ses upon change as o p p o sed to differen ces in b ehavior am ong disparate com m unities. Does Business Development Raise Taxes: Methodology T he unit o f observation for the study is the m unicipality. Our sam ple c o n sists o f incorporated m unicipal governm ents in the six cou n ties w h ich share a com m on border w ith C ook C ounty, including C ook itself, w ith a population in e x c e ss o f 10,000. B ecau se o f its size and eco n o m ic maturity, the city C h icago is exclu d ed from the sam ple. The m unicipality is ch osen , rather than say, sch o o l districts, because significant control over land use is v ested w ith this governm ental entity. B ecau se the sam ple ex clu d es the m any sm aller m un icip alities in the six county region, the study's results m ay not apply to such areas. The period o f observation is roughly 1980-1990. The period is not precise b ecau se so m e data w ere available on ly for the C ensus years 1979 and 1989, or for the fiscal years 1981 and 1991. For each m unicipality, an aggregate property tax rate has been com p iled . The ex isten ce o f num erous overlapping local ju risd iction s m ake it n ecessary to estim ate an aggregate tax rate or tax burden, w hich w o u ld reflect the com b in ed FRB CHICAGO Working Paper August 1995,WP-1995-7 9 burden o f the separate property tax le v ie s im p osed w ithin the boundaries o f a m unicipality. T h is w as a ccom p lish ed b y overlayin g m aps o f each typ e o f ju risd iction upon that for the m unicipality in question. T he fraction o f a m unicipality's tax base w h ich w as subject to the property tax le v y o f an overlapping jurisd iction o f a particular typ e (for exam p le, sch o o l district) is a ssu m ed equal the fraction o f m unicipal land area accou n ted for b y that particular jurisdiction. T h ese aggregate tax rates burdens have been further refined to reflect d ifferen ces in a ssessm en t practices and differen ces in the in cid en ce o f m ajor tax exem p tion s, such as the hom eow ners exem ption. T his en ab les us to estim ate an effe c tiv e rate o f tax on residential property as w ell as an e ffe c tiv e tax rate relative to incom e. B u sin ess grow th w ithin a m unicipality w ill be m easured the grow th o f com m ercial and industrial a ssessed value and by aggregate em p loym en t. Measures of tax burden Three m easures o f tax burden are considered: (1 ) the statutory property tax rate that is applied to eq u alized a ssessed value, and hereafter d enoted as trate; (2 ) e ffe c tiv e tax rate on o w n er-occu p ied housing, denoted as rate; (3 ) e ffe c tiv e tax rate in term s o f in com e, denoted as burden. Each o f th ese m easures is interesting in its o w n right. Trate is o f interest b ecau se it is the m easure set b y local govern in g authorities; it is also the tax rate applied to com m ercial and industrial property. Rate is o f sig n ifica n ce b ecau se it reflects the degree to w h ich residential property is taxed and is relevant to d ecisio n s to build or im prove residential property. F inally, burden is o f sig n ifica n ce b ecau se it m easures the average sacrifice required o f hom eow n ers w ithin a com m unity; increases in this m easure im ply that le ss funds are availab le for private consum ption. M athem atically, th ese FRB CHICAGO Working Paper August 1995. WP-1995-7 can be represented as follow s: trate = R PT A X / E A V R E S , rate = R PT A X / F V R E S, burden = R P T A X / INC, 10 where R PT A X aggregate residential property tax paym ents, EAVRES aggregate equalized a ssessed value o f residential property, FV R E S aggregate market value o f residential property, and INC aggregate incom e. S im p le algebraic m anipulation reveals the fo llo w in g relationships am on g th ese m easures o f tax burden. trate = rate * (E A V R E S / FV R E S) burden = rate * (F V R E S / INC) The term (E A V R E S /F V R E S ) is com m on ly know n as the assessm en t ratio, w h ile (F V R E S /IN C ) is the average ratio o f house value to in com e. The assessm en t ratio differs am ong com m unities despite efforts by the State Board o f E qualization to keep them uniform . The ratio o f hou sin g valu e to in com e w ill d iffer becau se o f underlying d ifferen ces in land valu es and the average age o f h ou sin g stock. C on ceivab ly, m ovem en ts in these indicators o f tax burden m ay not necessarily agree. Empirical Model Specification T he em pirical m odel con sists o f four major sections; (1 ) tax burden; (2 ) population; (3 ) em ploym ent; (4 ) hou sin g appreciation. A ll but (4 ) have been d ev elo p ed in the theoretical m odel. The last, being static, is unable to treat dynam ic elem en ts such as ch an ges in h ousing price. Like the sim p le correlation analysis, focus w ill be upon explaining percentage ch an ges in tax burden m easures. C onsequently, m any o f the m o d el’s variables w ill th em selv es be expressed in term s o f percentage changes. FRB CHICAGO Working Paper August 1995, WP-1995-7 11 Tax burden T he tax burden section co n sists o f tw o m ajor relationships and a quasi-reduced form. The first relationship is predicated upon the com m unity's fiscal budget constraint. B y th is con d ition , Irate can be expressed as (1 1 ) trate = (Expenditure - Other R even u e) / E A V . Ignoring, for the m om ent, other revenue, trate is seen to d epend upon expenditure behavior and tax base grow th. In growth term s, w e have (1 2 ) chtrate = grexp - grE A V , w here ch or gr sig n ify percentage change. The tw o com p on en ts o f the tax burden section , then, are estim ating equations for expenditure and E A V . B y our p revious d iscu ssion in the theoretical section , grexp is a ssu m ed to depend upon the grow th o f population, grow th o f in co m e, grow th o f em p loym en t, tax price, and relevant dem ographics. (1 3 ) grexp = (grin c(+),grpop (+),grem p (+),taxp rice(+),d em o) + uexp, w here, for notational eco n o m y , w e have ch osen to suppress the c o e ffic ie n ts o f the right-hand side variables, and u is an error term. T he ex p ected sign o f the variable is indicated in parenthesis. It should be noted that w e have included the level o f tax price rather than its change as the m odel m ight oth erw ise su ggest. The le v e l variable is included to capture com p osition al e ffe c ts upon expenditure. T his w ill occur i f expenditure dem ands by ex istin g residents and b u sin ess increase at different rates. Such w ou ld be the ca se i f g ro w in g real in com e leads to higher residential dem ands for services, w ith b u sin ess dem ands h old in g relatively constant. In this ca se, com m u n ities w ith high tax prices w ill, other things constant, also exp erience higher tax rate grow th. H en ce, our ex p ected sign for tax price is p ositive. A change in tax price variable is deliberately om itted from the sp ecification o f (1 3 ) becau se its effe c ts w o u ld be FRB CHICAGO Working Paper August 1995, WP-1995-7 12 captured b y the grem p and grpop variables. T o include it as a separate factor w o u ld be to in vite serious m ulti-collinearity, i f not singularity. T he variable grE A V is assum ed to depend upon th ose factors in flu en cin g the grow th o f the tax base: grow th o f population, em ploym ent, in com e, and h ou se price. In addition, it w ill depend upon the lev el assrat and change in the assessm en t ratio (ch ass) for residential housing. F inally, sin ce C ook C ounty has a c la ssified assessm en t system under w h ich b u sin ess property is m ore h igh ly a ssessed , it w ill depend upon w hether the m unicipality is located in C ook C ounty, for w hich w e use a dum m y variable. (A 1) grE A V =(grinc(+),grpop(+),grem p(+),grhpric(+), cook (+),assrat(+),(grass(+ )) + uc,v. A ll variables in this equation are exp ected to have a p o sitiv e sign. B ecau se w e are not able to observe expenditures, w e cannot estim ate (1 3 ) directly. W e can on ly estim ate it indirectly through substitution into (1 2 ). A ccord in gly, our tw o "structural” equations in the tax burden section are ( A l ) and (A 2 ) grtrate=(grinc(+),grpop(+),grem p(?),taxprice(+),dem o,grE A V (-) trate(-),cook(?) + e,. The change in taxprice term is suppressed in (1 7 ) becau se, g iv en grE A V , it is h igh ly correlated w ith gremp. S p ecifically, chtaxprice = grbus - grE A V , B ecau se w e are unable to include the change in tax price as a separate variable, w e are unable to separate the taxprice and direct b u sin ess service dem and associated w ith gremp. S ince taxprice ch an ges should have a dam pening effect on expenditure (due to elasticity), w e are unable to p lace a prior on the sign o f grem p in (A 2). F inally, w e add tw o term s, the initial statutory tax rate and a dum m y for C ook C ounty. The form er is included b ecau se the expenditure equation is a relationship for "desired” increases in expenditure. I f a com m unity already has a high tax rate, it m ay fear lo sin g tax base to com p etin g jurisdiction. M oreover, som e com m u n ities operate at their tax rate lim its, beyon d w hich they m ust seek voter approval for tax rate increases. O fficia ls in high tax rate lo ca lities m ay hesitate to seek such approval and/or voters m ay refuse to grant such approval. T ax rates for m any FRB CHICAGO Working Paper August 1995, WP-I995-7 13 sp ecial districts are d efacto m andated b y state law and thus cannot be changed. T he C ook C ounty dum m y variable is introduced b ecau se o f its cla ssifica tio n sy stem for b u sin ess property. H ere the statutory tax rate understates b u sin ess tax le v e ls and overstates residential tax lev els. T herefore, resistance to higher tax rates w ill be lo w er in com m u n ities w ith a relatively large non-residential base. T hus, it is not p o ssib le to place a prior sign on the C ook C ounty dum m y variable in (A 2 ). In addition to the tw o structural equations, w e a lso have the sem i-red u ced form for th is sector w h ich w e obtain b y com b in in g ( A l ) and (A 2 ). (A ) grtrate=(grinc(?),grpop(?),grem p(?),grhpric(-),taxprice(?),dem o, chass(-),trate(-),cook (?)) + uA. E m pirically, ch ass is h igh ly correlated w ith c o o k and assrat and is as w e ll correlated w ith other elem en ts o f (A ). H en ce, it is not e x p licitly inclu d ed in the estim ation process. Sim ilarly, c o o k and assrat are very h ig h ly correlated, lead in g us to include one or the other, but not both w hen estim atin g (A ). It appears that eco n o m ic theory p rovides us w ith few priors on the sig n s o f th is sem i-red u ced form. A s a result, w e m ay w ell find the a b sen ce o f statistically sign ifican t relationships b etw een the grow th in tax burden and such variables. H ow ever, theory d oes predict that the sign o f h ou se price appreciation w ill be p o sitiv e and that the sign o f the initial tax rate w ill be negative. O therw ise, the data w ill have to speak for th em selves. Alternative tax burden measures T he sp ecification o f the tax burden sector for the alternative m easures o f tax burden fo llo w directly from the preceding m odel. For the e ffe c tiv e tax rate on h ou se value, w e have the relationship chrate = chtrate + grass. In other w ords w e need o n ly add a term to reflect the grow th o f the a ssessm en t ratio to the right hand side o f (A ) above. But, sin ce, it already e x ists, w e have b asically the sam e equation, w ith the ex cep tio n o f the initial tax rate term, w hich w ou ld substitute rate for trate. FRB CHICAGO Working Paper August 1995, WP-1995-7 14 Sim ilarly, for the effe c tiv e tax rate on in com e, burden, w e have the fo llo w in g chburden = chrate + chpric, w here h p iic is the ratio o f average house price to per capita incom e. B ut in chhprinc = grhpric - grinc, each o f the right-hand term s is already included in (A ). N ote, how ever, that the prior sign on grinc w ill be changed i f the c o efficien t in (A ) assu m es a value le ss than one. In this case, the sign o f in com e w ill be negative. Population F o llo w in g the theoretical m odel, the fo llo w in g equation can be d ev elo p ed for population. (B ) grpop=(grem p(+),grem pn(+),grhpric(-),chtrate(-),trate(-),schexp(+), h p rice(-),renters(+),density(-),annex(+)) + uB, w here grem pn = grow th o f nearby em ploym ent, grhpric = grow th o f hom e prices, trate = statutory tax rate 1981, sch exp = per pupil current expenditure, 1981, hpric = average hou sin g price 1981, renters = percentage renters, 1980, density = population density, 1980, and annex = change in land area. T he renters and density variables are included to proxy for com m unity attitudes tow ards n ew residents. L ow density m ay be a signal that a com m u n ity ad vocates large lot zon in g, w h ile a lo w rental share m ay identify com m u n ities seek in g to k eep out high -cost populations. Growth in land area is included for o b v io u s reasons. O therw ise, the specification o f (B ) fo llo w s directly from the theoretical argument. FRB CHICAGO Working Paper August 1995, WP-1995-7 15 Employment T he theoretical m odel su ggests the fo llo w in g equation for the grow th o f com m u n ity em ploym ent. (C ) grem p=(grpop(+),grpopn(+ ),grem p n,trate(?),chtrate(?), c o o k (-),lo ca t(+ ),in c(-),d issip (-)) + uc, where: gropn = grow th o f nearby population, co o k = dum m y variable for co o k county, locat = locational advantages, inc = average in com e, 1980, and d issip = dissipation index. T he population term s are included to reflect a ccess to labor m arkets, w h ile the em p loym en t variable attempts to m easure inter-industry supply relationships. C ook C ounty is included to reflect its cla ssified a ssessm en t practice w h ich discrim inates against b u sin ess and industry. The in com e variable represents dem and for environm ent, and the d issipation index reflects the degree o f leak age o f b u sin ess tax base to surrounding ju risd iction s. T he latter variables attem pt to m easure the w illin g n ess to supply sites to n ew b u sin esses. V arious m easures o f locational advantage are used, including the standard distance from C B D and h igh w ay a ccess variables. H ou sin g price, sin ce it is h ea v ily in flu en ced by land valu es, w ill be u sed as a proxy for locational advantages. F in ally, n o prior sign is placed on the tax variables b ecau se high taxes m ay increase supply o f sites in addition to raising co sts o f d oin g bu sin ess. Price appreciation U p to this point w e have not sp ecifica lly addressed the determ inants o f price appreciation. L ike m ost things, h ou se prices reflect supply and dem and considerations. T his, together w ith data availability, led to the fo llo w in g sp ecification . (D ) griipri=(grinc(+),chtrate(-),grem p(-),trate(-),inc(+),hprice(?) em p p o p (-),d en sity (+ ),d ista n ce(+ ),h ig h w a y (+ )) + uD, FRB CHICAGO Working Paper August 1995, WP-1995-7 16 where: em ppop = ratio o f em p loym en t to population, 1980, density = population per sq m ile, 1980, inc = per capita incom e, 1980, hprice = average h ousing price, 1980, distance = distance to the L oop, and h igh w ay = a ccess to expressw ay. H ou sin g appreciation should be more rapid the greater the grow th in resident in com e, but low er the greater the grow th o f tax rates (capitalization effect). It should also be low er in com m unities w ith rapid eco n o m ic d evelop m en t, reflecting con gestion grow th and environm ental degradation. A lso included are a group o f am enity variables: including the ex istin g tax rate, in com e, em p loym en t saturation, population density, density, and h igh w ay a ccess. T h ese variables cou ld be exp ected to in fluence the lev el o f house prices, and also their rate o f grow th i f prices adjust gradually. Statistical Results Each o f the m od els A -D above, as w ell as the structural equations A1 and A 2 , w as estim ated using ordinary least squares (O L S) and tw o-stage least squares (2 S ). Tax burden sector-structural relationships ( A l ) grow th o f assessed value (see Table A l ) T he O L S estim ates all carry the exp ected sign and are all but one sign ifican t at the 99% lev el o f con fid en ce, and the last, co o k , is sign ifican t at the 95% lev el. M oreover, the overall fit is quite go o d , w ith 82% o f the variance in E A V h avin g been accounted for. T his su ggests that the proxies for tax base grow th are quite adequate. C o efficien ts can be interpreted as partial elasticities. T hus, a 1% increase in population is associated w ith a .66% increase in E A V . Sim ilarly, a 1% increase in em p loym en t is leads to a .17% increase in the taxable base. T hus eco n o m ic d evelop m en t has clear tax base im pacts. T he discrepancy in siz e o f elasticity w ith population should not be overstated b ecau se the tw o m ovem en ts are lik ely to g o hand in hand. C om m ercial d evelop m en t often fo llo w s FRB CHICAGO Working Paper August 1995, WP-1995-7 17 residential develop m en t, w h ile industrial d evelop m en t m ay g iv e rise to residential develop m en t. The c o efficien t on c o o k su g g ests that grow th o f taxable base is 4 .3 percentage points higher than otherw ise. T he 2 S estim ates are not quite as accurate, nor are the c o e ffic ie n ts o f grhpri,chass, or c o o k sign ifican t at the sam e le v e ls o f co n fid en ce as for O L S. H ow ever, tw o o f the three are still sign ifican t at the 90% lev el. T he fact is that there is considerable colinearity am ong th ese three variables. T he point estim ates for population grow th and em p loym en t grow th are c lo se to their O L S counterparts, su g g estin g the sp ecification is robust. The upshot is that the pre con d ition s necessary for favorable tax co n seq u en ces o f eco n o m ic d evelop m en t are m e t-a n increase in the com m unity's tax base. (A 2 ) Statutoiy tax rates (se e Table A 2 ) T he O L S estim ates reveal a g o o d fit w ith nearly 60% o f the variance accou n ted for.2 A s exp ected , the grow th o f the tax base, G R E A V , exerts a strong and n egative in flu en ce on statutory tax rates, w hen expenditure effect are h eld constant. The initial tax rate, trate81 also perform s as exp ected , dem onstrating the p ow er o f fiscal constraint and tax rate lim itations. T he co effic ie n t poin ts to strong regression to the m ean. T he C ook dum m y is strongly p o sitiv e and significant, im p lyin g a p o sitiv e tax rate grow th differential o f nearly 15 percentage points. T h is su g g ests that sp en d in g caps in C ook county m ay be less binding than in the "collar" cou n ties. A s for expenditure determ inants, population grow th clearly p la ces a strong, p o sitiv e dem and upon local public services. T he situation for em p loym en t grow th, h ow ever, is le ss certain. W h ile the c o efficien t is p o sitiv e, it is not sign ifican t u sin g normal standards for statistical sig n ifica n ce. M oreover, its absolute size is less than on e tenth o f the population figure, su g g estin g o n ly m od est service dem ands by bu sin ess. A s m entioned ab ove, h ow ever, the c o e ffic ie n t on em p loym en t m ay reflect tax price effects as w ell as b u sin ess service effects. A s for tax price, it contributes to the grow th o f taxation, as exp ected . T h ose com m u n ities w ith lo w b u sin ess shares suffered h igher tax increases over the period. Thus, there m ay be a dynam ic b enefit from b u sin ess grow th not anticipated by static m od els. Initial average hom e valu e is n eg a tiv ely correlated w ith expenditure grow th. T his is le ss surprising than for in com e, sin ce it m ay sim p ly reflect reduced n eed for public serv ices in bedroom com m u n ities w ith large, e x p en siv e h om es. Finally, a renter variable, m easuring the share o f h ou sin g units rented, is d esign ed to test w hether renters are m ore lik ely to support public serv ices than hom eow n ers. FRB CHICAGO Working Paper August 1995, WP-1995-7 T he O L S 18 estim ates su g g est so m e support for this hypothesis, though sig n ifica n ce falls m arginally short o f the 90% level. A ll in all, the O L S equation is con sisten t w ith prior exp ectation s and fails to sh o w sign ifican t expenditure burdens im p osed b y b u sin ess grow th, w h ile pointing to sizeab le expenditure effects o f population grow th. A s w ith the E A V equation, the 2 S estim ates are less p recise than for the O L S. B y and large, how ever, the qualitative findings are unchanged. Certain o f the co efficien ts, how ever, are seen to be sen sitive to the estim ation technique. T he c o efficien t on E A V is considerably increased in absolute value. T his should not b e surprising sin ce the O L S estim ates are in flu en ced by feedback from tax rate grow th to E A V growth. H ence, the 2S c o efficien t m ay be m ore representative. A sim ilar observation applies to population grow th, though to a lesser degree. The 2S estim ates su ggest an even larger expenditure e ffe c t o f population grow th. Furthermore, any marginal in flu en ce o f em p loym en t grow th disappears altogether. Tax burden sector—semi-reduced forms (see Table A3) Statutoiy tax rates O n ce the E A V variable is elim inated from the estim ating equation, the precision o f the estim ates falls. N everth eless, the O L S equation con tin u es to exp lain m ore than 50% o f the variance o f the dependent variable. U nfortunately, the co efficien ts o f a number o f k ey variables, such as population and em ploym ent b eco m e statistically insignificant. T h is is not surprising g iv en the offsettin g effects o f these variables through expenditure and tax base effects. A s predicted, housing appreciation has led to a decrease in statutory tax rates, although it barely fails the 90% test o f sign ifican ce in the 2 S m od el. Initial h ou sin g price continues to depress tax rate grow th, but w ith h a lf o f the force as in the structural equation. The C ook dum m y and initial tax rate have sim ilar effe c ts as in the structural equation. The co effic ie n t o f population grow th is p o sitiv e, though not significant. The k ey variable, grow th o f em p loym en t is approxim ately sign ifican t in the O L S estim ate at the 90% co n fid en ce lev el. Its co efficien t is n egative, im p lyin g som e tax rate b en efits from em p loym en t grow th. H ow ever, it fails the test for sig n ifica n ce in the tw o -sta g e version. Thus, these results are too marginal to serve as d efin itive ev id en ce for im m ediate favorable tax rate effects o f eco n o m ic d evelop m en t. FRB CHICAGO Working Paper August 1995, WP-1995-7 19 H ow ever, the p o sitiv e and sign ifican t co effic ie n t on taxprice su g g ests lo n g term tax rate b en efits o f eco n o m ic d evelop m en t. T ogether w ith the n egative co effic ie n t on em p loym en t grow th, this p la ces the burden o f p r o o f o f n eg a tiv e fiscal co n seq u en ces o f eco n o m ic d evelop m en t squarely on th ose w h o p o sit it. E ffectiv e tax rate on h ousing W hen account is taken o f assessm en t variations, there is a m uch clearer con n ection b etw een em p loym en t grow th and tax burden. In both the O L S and 2S version s, the c o efficien t o f g ie m p is n egative and sign ifican t at the 95% le v e l o f con fid en ce. The 2 S estim ate is larger that the O L S, im p lyin g that a 1% change in em p loym en t grow th is accom p an ied b y a .12% reduction in e ffe c tiv e tax rate grow th. W hile this effect is m odest, it is by no m eans trivial. The c o efficien t on population grow th, another variable o f great interest, is p o sitiv e but fails the sig n ifica n ce test. Thus, it is not p o ssib le to infer that population grow th w ill have im m ediate n egative e ffectiv e property tax rate co n seq u en ces. H ow ever, the p o sitiv e c o efficien t on taxprice poin ts to lo n g term c o sts o f population grow th (w ithout accom p an yin g em p loym en t grow th). T hus, the w eig h t o f ev id en ce w o u ld seem to b e against the proposition that population grow th typ ically has p o sitiv e effe c tiv e tax rate co n seq u en ces. H ou sin g appreciation has n egative tax e ffects in the O L S version, provid in g so m e that assessm en t lag hypothesis. H ow ever, in the 2S version , the co effic ie n t is insignificant. F in ally, the rental variable is strongly sign ifican t for both the O L S and 2 S m ethods. T his adds credence to the contention that renters are m ore prone to support public expenditure, perhaps b ecau se they b eliev e their landlord p ays the tax. E ffectiv e rate on in com e For the in com e m easure, there is apparently little con n ection b etw een em p loym en t grow th and tax burden. For neither estim ation technique is grem p statistically sign ifican t. Indeed, in the 2 S version, the co effic ie n t is positive! Here, h ow ever, there is ev id en ce o f a p o sitiv e tax burden im pact o f population grow th. A 1% increase in population grow th appears to result in a .25% increase in tax burden grow th. That in com e grow th and house appreciation have opposite, sign ifican t im pacts on tax burden co m e s as no surprise, b ecau se o f the dep en d en ce o f ch b u n len on the (h ou se p rice)/in com e ratio. T hus, in com e grow th, h old in g price FRB CHICAGO Working Paper August 1995, WP-1995-7 20 appreciation constant, leads to an decrease in this k ey ratio, and h en ce a decrease in tax burden. G iven that the absolute value o f the c o efficien t is le ss than on e, it su g g ests p o sitiv e expenditure effects o f in com e grow th. For this m easure o f tax burden, both the C ook dum m y and the 1980 assessm en t ratio are included in the equation. W hile the n egative sign o f cook w as anticipated, the negative sign on assrat co m es as a surprise. Perhaps the answ er lies w ith som e interaction w ith price appreciation. Summaiy—tax burden effects Our em pirical analysis, w h ile largely supporting the underlying m od el, has provided m ixed ev id en ce concerning the im pact o f eco n o m ic d evelop m en t upon property tax burdens. I f w e hold constant the tax base co n seq u en ces, em p loym en t grow th appears to have no tax rate im pact. If w e cou p le this w ith the fact that em p loym en t growth has significant tax base effects, and the fact that the expansion o f tax base, holding expenditure determ inants constant, w ill tend to reduce tax rates, w e arrive at the co n clu sion that the em p loym en t grow th leads to a reduction in statutory tax rates. If w e test this hyp oth esis directly, through a sem i-reduced form, w e also find a n egative association. H ow ever, at best, the statistical sig n ifica n ce o f this relationship is marginal. Thus, the errors in the relationship b etw een em p loym en t grow th and base grow th and that betw een base grow th and tax rate grow th are su fficien t to w eaken the ultim ate relation betw een em p loym en t grow th and tax rate grow th. N everth eless, there appears to be sign ifican t lo n g run b en efits o f em p loym en t grow th through its im pact on taxprice. T h ese b en efits do not accrue im m ediately, but are felt as taxprices are gradually low ered, reducing tax rate burden o f in com e growth. For the effe c tiv e property tax rate, how ever, there is no am biguity. There appears to be a statistically significant negative relationship b etw een em p loym en t grow th and tax rate growth. For m any purposes, e ffe c tiv e property tax rates are a better m easure o f burden than statutory rates. H ence, w e can advance the proposition that eco n o m ic d evelop m en t has led to a reduction in property tax burdens in the five county C h icago area. The tax co n seq u en ces o f the other major type o f develop m en t, that arising from population grow th, has sim ilar am biguities. T he ev id en ce is clear that population grow th both increases expenditures and increases property tax revenues. H ence, the im pact on property tax rates w ill depend upon the relative strength o f these forces. W ith respect to statutory rates, there is so m e FRB CHICAGO Working Paper August 1995, WP-1995-7 21 e v id en ce o f an adverse im pact on tax rate, but the effe c t fa ils the test o f statistical sign ifican ce. W e cannot reject the h yp oth esis that there is no im m ediate p o sitiv e effect on tax rates. H ow ever, w e can reject the h y p oth esis that there are m ajor im m ediate tax rate b enefits. M oreover, lik e em p lo y m en t grow th, population grow th has lo n g run im plications via the taxprice effect. H ere, how ever, the result is to increase long-run tax rates. T he sam e remarks apply w ith regard to the e ffectiv e tax rate on ow n ero ccu p ied property—the c o efficien ts o f the sem i-reduced form are p o sitiv e but not statistically sign ifican t from zero. It is on ly i f w e m easure tax burden w ith regard to in com e can w e u n eq u ivocally argue that population grow th increases tax rates. Ironically, w e are unable to find p o sitiv e tax co n seq u en ces under th is m easure for em p loym en t grow th. Population sector (see Table B) T he m ost striking aspect o f our estim ates o f population grow th con cern s its relationship to em p loym en t grow th. W hether w e use O L S or 2 S , em p loym en t grow th, both lo ca lly and nearby, appears to have a strong p o sitiv e in flu en ce on population grow th. T hus, population fo llo w s jo b s. (W hether the reverse is a lso true w ill be taken up in the next sectio n .) It is a lso apparent that neighboring jo b grow th has a larger effect than local jo b g r o w th -n e a r ly four tim es as strong. Thus, eco n o m ic d evelop m en t activities o f one's neighbors p ow erful have im plications for one's ow n residential d evelop m en t. T h is in turn m ay have n egative fiscal con seq u en ces depending upon the m easure o f tax burden adopted. Thus, even i f it is true that em p loym en t grow th p rovid es lo ca l fiscal b en efits, this d o es not insure that it w ill provide glob al fiscal ben efits. T he rest o f the population equation g o e s ju st about as exp ected , w ith the ex cep tio n o f per pupil education spending (w o p ex o p ). T he sign o f this variable is the reverse o f w hat w as exp ected . It m ay w ell b e that th ose com m u n ities w h ich spend ab ove the average also en gage in population grow th control m easures to a void having to su b sid ize the children o f n ew m igrants. FRB CHICAGO Working Paper August 1995, WP-1995-7 22 Employment sector (see Table C) U n lik e the findings o f a recent study o f the Philadelphia area, there appears to be little im pact o f nearby population grow th upon em p loym en t exp an sion. W h ile the O L S version points to a relationship to o w n population grow th, this lik ely mirrors the finding in the population equation. W hen the 2 S procedure is em p loyed , the result disappears. Thus, it is apparently not true that em p loym en t fo llo w s p o p u la tio n -w h ic h contradicts a p ie c e o f contem porary public w isd om . Perhaps this is due to the fact that m uch o f the em p loym en t grow th over the period w as due to increased labor force participation. Such jo b s m ay thus have little to do w ith population expansion. It is interesting to note that high statutory tax rates appear to deter develop m en t. C uriously, the C ook dum m y, is p o sitiv e, contrary to expectations. A s expected, h igh -in com e com m u n ities tend to repel d evelop m en t. A t the sam e tim e, those com m u n ities w ith high property va lu es tend to attract n ew developm ent; high property va lu es m ay be a proxy for locational advantages. The m anufacturing density variable, w h ich w as d esign ed to represent com m unity tolerance for environm ental disam en ities, has the w rong sign . C om m unities rich in m anufacturing tend to have slo w er em p loym en t grow th. T his variable m ay be proxying for an older suburb, w ith little vacant space for n ew developm ent. Housing appreciation (see Table D) L astly, w e consider the grow th o f inflation adjusted h ou sin g prices. Our estim ates su ggest pow erful capitalization effects. Both high initial and rapidly grow in g statutory rates are associated w ith decreased price appreciation. Grow th o f in com e, a major dem and for hou sin g services also increases h ou sin g inflation. R icher com m unities, m easured by initial in com e b en efit disproportionately from h ousing price grow th. H ow ever, h old in g in co m e constant, th ose areas w ith higher initial h ousing prices en joyed le ss rapid appreciation. The environm ental variables, em p loym en t saturation (em p loym en t/p op u lation ) and population density both had the ex p ected sign. T he form er is a measure o f con gestion and other form s o f environm ental degradation associated w ith b u sin ess developm ent. F inally, it appears that the o u tlyin g suburbs b enefited m ost from hou sin g appreciation during the 1980s. FRB CHICAGO Working Paper August 1995, WP-1995-7 23 Study Conclusions Our em pirical an alysis, w h ile largely confirm ing the underlying m o d el, has provid ed m ix ed ev id en ce concerning the im pact o f eco n o m ic d ev elo p m en t on property tax burdens. I f w e hold constant the tax base co n seq u en ces, em p loym en t grow th appears to have no observable tax rate im pact through expenditure channels. I f w e cou p le this w ith the fact that em p loym en t grow th has sign ifican t p o sitiv e effects on the local property tax base, w e arrive at the co n clu sio n that the em p loym en t grow th leads to a reduction in statutory tax rates. I f w e test this h yp oth esis directly, w e also find a n egative association. H ow ever, at best, the statistical sig n ifica n ce o f this relationship is m arginal. A pparently, there is su fficien t n oise in the relationship b etw een em p loym en t grow th and base grow th and b etw een base grow th and tax rate grow th so as to w eaken the observable relationship betw een em p loym en t grow th and tax rate grow th. For the e ffe c tiv e property tax rate, how ever, this matter has reso lv ed itself. There appears to be a statistically sign ifican t n egative relationship b etw een em p loym en t grow th and tax rate grow th. For m any purposes, e ffe c tiv e property tax rates are a better m easure o f burden than statutory rates. H en ce, w e can advance the proposition that eco n o m ic d evelop m en t has led to a reduction in property tax burdens in our sam ple o f C hicago's suburbs. T he tax co n seq u en ces o f the other major type o f d evelop m en t, that arising from population grow th, has sim ilar am biguities. Population grow th in creases expenditures, but it also increases property tax revenues. H en ce, the im pact on property tax rates w ill depend upon the relative strength o f th ese forces. W ith respect to statutory rates, there is som e ev id en ce o f an net adverse im pact on tax rate, but the effect fails the test o f statistical sig n ifica n ce. W e cannot reject the h ypoth esis that there is no effect on tax rates. H ow ever, w e can reject the h yp oth esis that there are major tax rate benefits. A sim ilar result is found w ith regard to the e ffectiv e tax rate on ow n er-occu p ied property. It is o n ly i f w e m easure tax burden w ith regard to in com e can w e u n eq u iv o ca lly argue that population grow th increases tax rates. B oth local and nearby em p loym en t grow th are found to exert a strong p o sitiv e in flu en ce on population grow th, that is, p eop le fo llo w jo b s. M oreover, n eighboring jo b grow th is found to have a larger effect than local jo b grow th — nearly four tim es as strong. E con om ic d evelop m en t activ ities o f one's n eighbors FRB CHICAGO Working Paper August 1995, WP-1995-7 can have sign ifican t im p lication s for one's ow n residential 24 develop m en t. A nd i f it is the case that residential grow th is a ccom p an ied b y c o stly fiscal con seq u en ces, then b u sin ess d evelop m en t in a n eigh b orin g com m u n ity has been found to indirectly p lace added pressure on residential property tax rates. Thus, even i f it is true that em p loym en t grow th p rovid es local fiscal ben efits, this d o es not insure that it w ill provide b enefits. global fiscal Footnotes 'T his assu m es that residents b eliev e that, at the m argin, increases o f b u sin ess taxes are exported outside the jurisdiction. T his is a strong assum ption, but on e com m o n ly m ade in the literature. 2In percentage change m od els, fits tend to be notoriously low . References Illin ois D epartm ent o f E m ploym ent Security, Where Workers Work, annual, various issu es. Illin ois D epartm ent o f R evenue, Assessment-Sales Price Ratio Studies, annual, various issu es. O akland, W illiam H. and W illiam A. Testa, D o es B u sin ess D ev elo p m en t R aise T axes: A n Em pirical Appraisal, Economic Perspectives, M arch/April 1995, pp. 2 2 -3 2 . Oakland, W illiam H. and W illiam A. Testa, "Com m unity D ev elop m en t-F iscal Interactions: A R ev iew o f the Literature," Regional Economic Issues, Federal R eserve Bank o f C h icago, A ugust, 1995, W P -1995-6. U .S . D epartm ent o f C om m erce, Bureau o f the C ensus, Census ofPopulation, 1980. _______________________ , Bureau o f the C ensus, FRB CHICAGO Working Paper August 1995, WP-1995-7 Census of Population, 1990. 25 Table A1 Regression analysis of growth in equalized assessed values (OLS and two stage least squares) Growth in equalized assessed value (OLS) Intercept Growth in equalized assessed value (2SLS) 42.44 (19.71) 41.16 (11.87) 0.66 (9.17) 0.59 (4.30) 0.17 (6.59) (3.13) Growth in average home price 0.17 (4.63) (1.62) Growth in income 0.55 (6.50) 0.61 (4.45) Change inassessment ratio 0.30 (4.27) (1.75) Cook County dummy variable 4.30 (1.97) 4.38 (1.22) 0.80 0.65 Growth in population Growth inemployment R sq (adj) 0.21 0.1 0.213 NOTES: The variables "Growth Inpopulation", Growth Inemployment", "Growth inhome prices" and "Change inassessment ratio’are jointlydependent variables inthe two-stage least squares (2LS) model. Values are given in bold face type. Estimated t-values are given inparentheses FRB CHICAGO Working Paper August 1995, WP-1995-7 26 Table A2 Regression analysis of changes in tax rates Change in statutory tax rates (TRATE) O LS Intercept C hange in statutory tax rates (TRATE) 2SLS 47.19 (6.32) 57.90 (4.82) Growth in equalized assessed value -0.49 (-6.60) -0.74 (-3.47) Growth in population 0.42 (5.42) 0.59 (3.90) Growth in employment 0.03 (1.30) 0.04 (0.77) Growth in income -0.01 (-0.08) 0.19 (0.96) Statutory tax rate, 1981 -3.87 (-5.05) -4.38 (-4.50) -0.0010 (-2.96) -0.0010 Tax price (EAV res/EAV total) 0.84 (3.96) 1.05 (3.68) Percentage of renters in jurisdiction 0.11 (1.60) 0.13 (1.59) Cook County dummy variable 15.16 (6.46) 16.38 (5.84) 0.56 0.44 Average home price R sq (adj) (-2.07) NOTES: Jointly dependent variables are given in bold face type. Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995, WP-1995-7 27 Table A3 Regression analysis of changes in tax rates (2SLS) C hange in statutory tax rates (TRATE) Change in effective tax rates (RATE) C hange in effective tax rates relative to incom e (B U R D EN ) Intercept 33.00 (4.17) 58.29 (5.98) 92.31 (7.43) Growth in population 0.10 (1.36) 0.11 (1.29) 0.27 (3.46) Growth in employment -0.04 (-1.64) -0.07 (-2.33) -0.01 (-0.37) Growth in income -0.15 M .2 8 ) 0.15 (1.11) -0.52 (-4.04) Growth in average home price -0.13 (-3.38) -0.11 (-2.25) 0.62 (10.04) Statutory tax rates, 1981 -3.07 (-3.60) Effective tax rate, 1981 -0.29 (-6.40) -11.23 (-6.41) Tax payment as a share of income, 1981 Average home price, 1981 -0.001 (-3.24) -0.002 (-5.58) -0.0002 (-0.29) Tax price (EAV res/EAV total) 0.50 (2.16) 0.94 (3.28) 0.54 (2.30) -1.81 (-3.55) Assessment ratio Percentage of renters in jurisdiction, 1981 0.05 (0.63) 0.20 (2.17) 0.19 (2.13) Cook County dummy variable 9.36 (3.32) -13.33 (-3.90) -17.27 (-3.74) 0.43 0.49 0.88 R sq (adj) NOTES: Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995, WP-1995-7 28 Table A3 (continued) Regression analysis of changes in tax rates (OLS) Change in statutory tax rates (TRATE) Change in effective tax rates (RATE) Change in effective tax rates relative to Incom e (BURDEN ) Intercept 30.25 (3.52) 54.24 (5.31) 101.72 (5.89) G rowth In population 0.15 (1-37) 0.17 (1.35) 0.26 (1.78) G rowth in em ploym ent -0.06 (-1.24) -0.12 (-2.02) 0.09 (1.16) Growth in income -0.25 (-1.69) 0.05 (0.35) -0.55 (-2.54) G rowth In average home price -0.08 (-1.49) -0.04 (-0.68) 0.61 (4.63) Statutory tax rates, 1981 -3.10 (-3.35) -0.29 (-6.15) Effective tax rate, 1981 -12.34 (-4.01) Tax payment as a share of income, 1981 Average home price, 1981 -0.001 (-2.22) -0.002 (-4.37) -0.0002 (-0.23) Tax price (EAV res/EAV total) 0.53 (2.13) 0.99 (3.32) 0.51 (1.68) -2.10 (-2.71) Assessment ratio Percentage of renters in jurisdiction 0.06 (0.78) 0.25 (2.52) 0.19 (1.60) Cook County dummy variable 10.80 (3.40) -12.43 (-3.34) -18.31 (-2.98) 0.39 0.46 0.80 R sq (adj) NOTES: Jointly dependent variables are given by bold face type. Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995. WP-1995-7 29 Table B Regression analysis of growth in population (OLS and two-stage least squares) Growth of population (OLS) Growth of population (2LS) Intercept 2.06 (0.23) 5.95 (0.51) Growth of employment 0.10 (3.99) (2.50) 0.42 (4.20) 0.41 (3.70) -0.11 (-0.14) -0.23 (-0.25) Change in statutory tax rate 0.08 (0.80) - 0.014 (-0.08) Average home price, 1980 0.0003 (0.93) 0.0001 (0.25) Growth in average home price 0.06 (1.63) (0.47) Population density 1980 -0.001 (-1.81) -0.001 (-1.55) Change in land area of jurisdiction, 1980 to 1990 2.75 (4.18) 2.75 (3.97) Percentage of renters in jurisdiction 0.18 (2.36) 0.15 (1.88) -0.001 (-2.51) -0.001 (-2.28) 0.64 0.61 Growth of employment in nearby jurisdictions Statutory tax rate, 1981 Per pupil education spending R sq (adj) 0.123 0.023 NOTES: The variables "Growth in employment", "Growth in average home price" and "Change in statutory tax rates" are predicted values in the two-stage least squares (2LS) model and are given by bold face type. Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995, WP-1995-7 30 Table C Regression analysis of growth in employment (OLS and two-stage least squares) Growth In employment (OLS) Growth in employment (2LS) Intercept 77.95 (2.52) 111.56 (2.70) Growth in population 0.82 (2.79) (0.36) -0.60 (-1.16) -0.11 (-0.18) 0.39 (0.92) 0.44 (0.97) Manufacturing employment per square mile, 1980 -0.02 (-4.37) -0.02 (-4.50) Per capita income 1980 -0.02 (-3.60) -0.02 (-3.90) Statutory tax rates, 1981 -4.11 (-1.41) -6.34 (-1.87) Change in statutory tax rates -0.39 (-1.20) - 1.003 (-1.58) Average home price, 1981 0.01 (3.87) 0.01 (3.91) Access to highways Cook County dummy variable 0.58 (0.18) 15.76 (1.43) -0.51 (-0.15) 26.43 (1.93) Index of tax base dissipation -0.94 (-1.22) -0.88 (-1.08) R sq (adj) 0.44 0.41 Growth in nearby population Growth of employment in nearby jurisdictions 0.182 NOTES: The variables "Growth in population" and "Change in statutory tax rates" are jointly dependent variables in the two-stage least squares (2LS) model and are given by bold face type. Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995,WP-1995-7 31 Table D Regression analysis of housing appreciation (OLS and two-stage least squares) Appreciation of housing Appreciation of housing (OLS) (2LS) 20.67 (1.14) 21.06 (0.88) Per capita income 1980 0.02 (7.30) 0.02 (5.30) Population density 1980 0.003 (3.063) 0.003 (2.84) Intercept Employment to population ratio, 1981 -24.10 (-3.40) -20.78 (-2.09) Growth of employment -0.06 (-1.25) -0.01 (-0.05) 0.68 (3.64) 0.58 (2.59) Statutory tax rates, 1981 -7.46 (-5.21) -7.38 (4.55) Changes in statutory tax rates -0.64 (-4.04) -0.82 (-2.94) Average home price, 1981 -0.01 (-6.42) -0.01 (-4.55) Access to highways -0.33 (-0.20) -0.36 (-0.21) Distance to the loop 0.68 (3.51) 0.73 (3.46) 0.64 0.61 Growth in per capita income R sq (adj) NOTES: The variables "Growth in employment", "Change in statutory tax rates" are predicted values in the two-stage least squares (2LS) model and are given by bold face type. Estimated t-values are given in parentheses. FRB CHICAGO Working Paper August 1995, WP-1995-7 32 Appendix I County locations of municipalities in study Municipality name and government form Location: County(ies) Addison village Alsip village Arlington Heights village Aurora city Bartlett village Batavia city Bellwood village Bensenville village Berwyn city Bloomingdale village Blue Island city Bolingbrook village Bridgeview village Brookfield village Buffalo Grove village Burbank city Calumet City city Calumet Park village Carol Stream village Carpentersville village Chicago Heights city Chicago Ridge village Cicero town Country Club Hills city Crestwood village Crystal Lake city Darien city Deerfield village Des Plaines city Dolton village Downers Grove village Elgin city Elk Grove Village village DuPage Cook Cook, Lake DuPage, Kane Cook, DuPage DuPage, Kane Cook Cook, DuPage Cook DuPage Cook DuPage, Will Cook Cook Cook, Lake Cook Cook Cook DuPage Kane Cook Cook Cook Cook Cook McHenry DuPage Cook, Lake Cook Cook DuPage Cook, Kane Cook, DuPage FRB CHICAGO Working Paper August 1995, WP-1995-7 33 Appendix I (continued) County locations of municipalities in study Municipality name and government form Location: County(ies) Elmhurst city Elmwood Park village Evanston city Evergreen Park village Forest Park village Franklin Park village Geneva city Glencoe village Glendale Heights village Glen Ellyn village Glenview village Glenwood village Hanover Park village Harvey city Hazel Crest village Hickory Hills city Highland Park city Hinsdale village Hoffman Estates village Homewood village Joliet city Justice village La Grange village La Grange Park village Lake Forest city Lake Zurich village Lansing village Libertyville village Lincolnwood village Lisle village Lombard village Lyons village McHenry city DuPage Cook Cook Cook Cook Cook Kane Cook DuPage DuPage Cook Cook Cook, DuPage Cook Cook Cook Lake Cook, DuPage Cook, Kane Cook Will Cook Cook Cook Lake Lake Cook Lake Cook DuPage DuPage Cook McHenry FRB CHICAGO Working Paper August 1995. WP-I995-7 34 Appendix I (continued) County locations of municipalities in study Municipality name and government form Location: County(ies) Markham city Matteson village Maywood village Melrose Park village Midlothian village Morton Grove village Mount Prospect village Mundelein village Naperville city Niles village Norridge village Northbrook village North Chicago city Northlake city Oak Forest city Oak Lawn village Oak Park village Orland Park village Palatine village Palos Heights city Palos Hills city Park Forest village Park Ridge city Prospect Heights city Riverdale village River Forest village River Grove village Riverside village Rolling Meadows city Romeoville village Roselle village Round Lake village St. Charles city Cook Cook Cook Cook Cook Cook Cook Lake DuPage, Will Cook Cook Cook Lake Cook Cook Cook Cook Cook Cook Cook Cook Cook, Will Cook Cook Cook Cook Cook Cook Cook Will Cook, DuPage Lake DuPage, Kane FRB CHICAGO Working Paper August 1995.WP-I995-7 35 Appendix I (continued) County locations of municipalities in study Municipality name and government form Location: County(ies) Sauk Village village Schaumburg village Schiller Park village Skokie village South Holland village Streamwood village Summit village Tinley Park village Villa Park village Waukegan city Westchester village West Chicago city Western Springs village Westmont village Wheaton city Wheeling village Wilmette village Winnetka village Wood Dale city Woodridge village Woodstock city Worth village Zion city Cook, Will Cook, DuPage Cook Cook Cook Cook Cook Cook, Will DuPage Lake Cook DuPage Cook DuPage DuPage Cook, Lake Cook Cook DuPage DuPage, Will McHenry Cook Lake FRB CHICAGO Working Paper August 1995, WP-1995-7 36 Working Paper Series A series ofresearch studies on regional economic issues relating to the Seventh Federal Reserve District, and on financial and economic topics. REGIONAL ECONOMIC ISSUES Estimating Monthly Regional Value Added by Combining Regional Input With National Production Data Philip R. Israilevich and Kenneth N. Kuttner Local Impact of Foreign Trade Zone David D. Weiss Trends and Prospects for Rural Manufacturing William A. Testa State and Local Government Spending—The Balance Between Investment and Consumption Richard H. Mattoon Forecasting with Regional Input-Output Tables P.R. Israilevich, R. Mahidhara, and G.J.D. Hewings W P-92-8 W P-92-9 W P-92-12 W P-92-14 W P-92-20 A Primer on Global Auto Markets Paul D. Ballew and Robert H. Schnorbus W P-93-1 Industry Approaches to Environmental Policy in the Great Lakes Region David R. Allardice, Richard H. Mattoon and William A. Testa W P-93-8 The Midwest Stock Price Index-Leading Indicator ofRegional Economic Activity William A. Strauss W P-93-9 Lean Manufacturing and the Decision to Vertically Integrate Some Empirical Evidence From the U.S. Automobile Industry Thomas H. Klier W P-94-1 Domestic Consumption Patterns and the Midwest Economy Robert Schnorbus and Paul Ballew W P -94-4 i Workingpaperseriescontinued To Trade or Not to Trade: Who Participates in RECLAIM? Thomas H. Klier and Richard Mattoon W P -94-11 Restructuring & Worker Displacement in the Midwest Paul D. Ballew and Robert H. Schnorbus W P-94-18 Financing Elementary and Secondary Education in the 1990s: A Review of the Issues Richard H. Mattoon Community Development-Fiscal Interactions: A Review of the Literature William H. Oakland and William A . Testa Community Development-Fiscal Interactions: Theory and Evidence from the Chicago Area William H. Oakland and William A. Testa W P-95-2 W P -95-6 W P -95-7 ISSUES IN FINANCIAL REGULATION Incentive Conflict in Deposit-Institution Regulation: Evidence from Australia Edward J. Kane and George G. Kaufman W P -92-5 Capital Adequacy and the Growth of U.S. Banks Herbert Baer and John McElravey W P -9 2 -1 1 Bank Contagion: Theory and Evidence George G. Kaufman W P -92-13 Trading Activity, Progarm Trading and the Volatility of Stock Returns James T Moser W P -92-16 Preferred Sources of Market Discipline: Depositors vs. Subordinated Debt Holders Douglas D. Evanoff W P-92-21 An Investigation of Returns Conditional on Trading Performance James T Moser and Jacky C. So The Effect of Capital on Portfolio Risk atLife Insurance Companies Elijah Brewer I lf Thomas H. Mondschean, and Philip E. Strahan W P -92-24 WP-92-29 2 Workingpaperseriescontinued A Framework forEstimating the Value and InterestRate Risk of Retail Bank Deposits David E. Hutchison, George G. Pennacchi WP-92-30 Capital Shocks and Bank Growth-1973 to 1991 Herbert L Baer and John N. McElravey WP-92-31 The Impact of S&L Failures and Regulatory Changes on the CD Market 1987-1991 Elijah Brewer and Thomas H. Mondschean W P-92-33 Junk Bond Holdings, Premium Tax Offsets, and Risk Exposure atLife Insurance Companies Elijah Brewer III and Thomas H. Mondschean W P-93-3 Stock Margins and the Conditional Probability of Price Reversals Paul Kofman and James T. Moser W P-93-5 IsThere Lif(f)e After DTB? Competitive Aspects ofCross Listed Futures Contracts on Synchronous Markets Paul Kofman, Tony Bouwman and James T. Moser Opportunity Cost and Prudentiality: A RepresentativeAgent Model ofFutures Clearinghouse Behavior Herbert L Baert Virginia G. France and James T. Moser The Ownership Structure ofJapanese Financial Institutions Hesna Genay Origins of the Modem Exchange Clearinghouse: A History ofEarly Clearing and Settlement Methods atFutures Exchanges James T. Moser The Effect of Bank-Held Derivatives on Credit Accessibility Elijah Brewer III, Bernadette A. Minton and James T Moser Small Business Investment Companies: Financial Characteristics and Investments Elijah Brewer III and Hesna Genay W P -93-11 W P -93-18 W P -93-19 W P-94-3 W P-94-5 W P -94-10 3 Workingpaperseriescontinued Spreads, Information Flows and Transparency Across Trading System Paul Kofinan and James T. Moser WP-95-1 MACROECONOMIC ISSUES An Examination of Change in Energy Dependence and Efficiency in the Six Largest Energy Using Countries-1970-1988 Jack L Hervey W P-92-2 Does the Federal Reserve Affect Asset Prices? Vefa Tarhan W P -92-3 Investment and Market Imperfections in the U.S. Manufacturing Sector Paula R. Worthington W P -92-4 Business Cycle Durations and Postwar Stabilization of the U.S. Economy Mark W. Watson W P-92-6 A Procedure for Predicting Recessions with Leading Indicators: Econometric Issues and Recent Performance James H. Stock and Mark W. Watson Production and Inventory Control atthe General Motors Corporation During the 1920s and 1930s Anil K. Kashyap and David W. Wilcox Liquidity Effects, Monetary Policy and the Business Cycle Lawrence J. Christiano and Martin Eichenbaum Monetary Policy and External Finance: Interpreting the Behavior of Financial Flows and Interest Rate Spreads Kenneth N. Kuttner W P -92-7 W P -92-10 W P -92-15 W P -92-17 Testing Long Run Neutrality Robert G. King and Mark W. Watson W P -92-18 A Policymaker’s Guide to Indicators ofEconomic Activity Charles Evans, Steven Strongin, and Francesca Eugeni W P -92-19 Working paper series continued 4 Barriers to Trade and Union Wage Dynamics Ellen R. Rissman W P-92-22 Wage Growth and Sectoral Shifts: Phillips Curve Redux Ellen R. Rissman W P-92-23 Excess Volatility and The Smoothing ofInterestRates: An Application Using Money Announcements Steven Strongin Market Structure, Technology and the Cyclicality ofOutput Bruce Petersen and Steven Strongin The Identification ofMonetary Policy Disturbances: Explaining the Liquidity Puzzle Steven Strongin Earnings Losses and Displaced Workers Louis S. Jacobson, Robert J. LaLonde, and Daniel G. Sullivan Some Empirical Evidence of the Effects on Monetary Policy Shocks on Exchange Rates Martin Eichenbaum and Charles Evans W P-92-25 W P-92-26 W P -92-27 W P -92-28 W P-92-32 An Unobserved-Components Model of Constant-Inflation Potential Output Kenneth N. Kuttner W P -93-2 Investment, Cash Flow, and Sunk Costs Paula R. Worthington W P -93-4 Lessons from the Japanese Main Bank System for Financial System Reform in Poland Takeo Hoshi, Anil Kashyap, and Gary Loveman Credit Conditions and the Cyclical Behavior of Inventories Anil K. Kashyap, Owen A. Lamont and Jeremy C. Stein Labor Productivity During the Great Depression Michael D. Bordo and Charles L. Evans WP-93-6 W P -93-7 W P -93-10 Working paper series continued 5 Monetary Policy Shocks and Productivity Measures in the G-7 Countries Charles L Evans and Fernando Santos WP-P3-12 Consumer Confidence and Economic Fluctuations John G. Matsusaka and Argia M. Sbordone W P -93-13 Vector Autoregressions and Cointegration Mark W. Watson W P -93-14 Testing forCointegration When Some of the Cointegrating Vectors Are Known Michael T. K. Horvath and Mark W. Watson Technical Change, Diffusion, and Productivity Jeffrey R. Campbell Economic Activity and the Short-Term Credit Markets: An Analysis of Prices and Quantities Benjamin M. Friedman and Kenneth N. Kuttner Cyclical Productivity in a Model of Labor Hoarding Argia M. Sbordone W P-93-15 W P -93-16 W P -93-17 W P -93-20 The Effects ofMonetary Policy Shocks: Evidence from the Flow ofFunds Lawrence J. Christiano, Martin Eichenbaum and Charles Evans W P -94-2 Algorithms for Solving Dynamic Models with Occasionally Binding Constraints Lawrence J. Christiano and Jonas D.M. Fisher W P-94-6 Identification and the Effects of Monetary Policy Shocks Lawrence J. Christiano, Martin Eichenbaum and Charles L Evans W P -94-7 Small Sample Bias in G M M Estimation of Covariance Structures Joseph G. Altonji and Lewis M. Segal W P -94-8 Interpreting the Procyclical Productivity ofManufacturing Sectors: External Effects of Labor Hoarding? Argia M. Sbordone W P -94-9 Working paper series continued 6 Evidence on Structural Instability in Macroeconomic Time Series Relations James H. Stock and Mark W. Watson W P-94-13 The Post-War U.S. Phillips Curve: A Revisionist Econometric History Robert G. King and Mark W. Watson W P -94-14 The Post-War U.S. Phillips Curve: A Comment Charles L Evans W P-94-15 Identification of Inflation-Unemployment Bennett T. McCollum W P-94-16 The Post-War U.S. Phillips Curve: A Revisionist Econometric History Response to Evans and McCallum Robert G. King and Mark W. Watson W P -94-17 Estimating Deterministic Trends in the Presence of Serially Correlated Errors Eugene Canjels and Mark W. Watson W P -94-19 Solving Nonlinear Rational Expectations Models by Parameterized Expectations: Convergence to Stationary Solutions Albert Marcet and David A. Marshall W P -94-20 The Effect of Costly Consumption Adjustment on Asset Price Volatility David A. Marshall and Nayan G. Parekh W P-94-21 The Implications of First-Order Risk Aversion for Asset Market Risk Premiums Geert Bekaert, Robert J. Hodrick and David A. Marshall W P-94-22 Asset Return Volatility with Extremely Small Costs ofConsumption Adjustment David A. Marshall W P -94-23 Indicator Properties of the Paper-Bill Spread: Lessons From Recent Experience Benjamin M. Friedman and Kenneth N. Kuttner W P -94-24 7 Workingpaperseriescontinued Overtime, Effort and the Propagation of Business Cycle Shocks George J. Hall WP-94-25 Monetary policies in the early 1990s~reflections of the early 1930s Robert D. Laurent W P -94-26 The Returns from Classroom Training forDisplaced Workers Louis S. Jacobson, Robert J. LaLonde and Daniel G. Sullivan W P -94-27 Isthe Banking and Payments System Fragile? George J. Benston and George G. Kaujman W P -94-28 Small Sample Properties ofG M M for Business Cycle Analysis Lawrence J’ Christiano and Wouter den Haan W P -95-3 The Fed Funds Futures Rate as a Predictor of Federal Reserve Policy Joel T. Krueger and Kenneth N. Kuttner W P -95-4 Capital Utilization and Returns to Scale Craig Burnside, Martin Eichenbaum and Sergio Rebelo W P -95-5 8