Loading...
Ball State Study on TIF District RevenueCENTERS RIBa51NE55GNod EC,ONOM C"RESEARCH irinfactitt Michael°1 Hicks,` PhD is.thea0,e- o' r�'°gegand Franncs Banti ti guj ti ormal. sor of.economss andthe ,.y.�..-..a._.. t .a diwrecttoyyo1 the Center for, Busiest and Ecdnomic Research a[ Ball State Universiet;is researj jk ,;ntere�R m stalPpdy` 3.., cal public financezznd edstott oflpubhc;poiicye tithe locatioh compost,, s. ion and size of e'conomlct' c ittigHe has a'utAoredy� hree books andjmor' titin 50 seholarlyjpapets ndd his -Wolk Bass been"",ry' 4,, ighkght&d Inlnews outlets -a crosstthe natlonn;. ,:. s Fap DDaagneytorofk search Isdlre`ad' firesearch.m trftcentiLieusiness nd.Economic Research 11Bal, ltS11e'University.' Her research focuses an tate anddocaltla policy nddregio�nalleconomic evelopment issues?She has workedIolli merous 1 dia foc"u d polity ` s i udesles ad,aretynto_ p cs ciudmg regtoea tri t g.v e nment.taxesrandenpen- d tures, senior migration, a d localgovernment r form. .. Pamela1Qumn Howard served as s' rcti7 reassistznt in GIS fondle Centeerdor Business and EconomOResearch from trlay 2013`'thraugh Jul 2y1 014She' received he'mas`ter s degree ".. in urban `and*Tetional planning �, lin i, 2�O0114Yd issccurrently,4emppoyed ias' associate planner for the City of wititr�'d IN. s©201 ditch r Busfness` tI EEconomic Research°NBalirgile University' g .tjf '` `c: t " r4Sx rtif- `ra:X? r�"i tgv �,gt s't"'4.C"„a�_+w B'VaSTylAIE� NIMERSIIIL CENTER)F,ORhBUS.I,,yN.ESS AND ECO,tNNOMIC1RERSEgpRCH. t4 M .G7xiw' .t r ?!'".''4.tL ;r un, eV: tai ".aF1dv ,. sIt` 'iMi'fi.4>sJ.C�ut,'ai4v i‘r: Some Economic Effects of Tax Increment Financing in Indiana Michael J. Hicks, PhD Director, Center for Business and Economic Research and George & Frances Ball Distinguished Professor of Economics, Miller College of Business Dagney Faulk, PhD Director of Research, Center for Business and Economic Research Pam Quirin Graduate Research Assistant in GIS, Center for Business and Economic Research Ball State University Executive Summary This study evaluates the impact of tax increment financing (TIF) in Indiana counties from 2003-2012. Using a spatial panel model as part of a two-stage research strategy into the economic effects of TIFs, we find that larger TIF districts are associated with higher effective tax rate of non -TIF areas within a county. This motivates the second stage analysis, which estimates the net effect of TIF districts on TIF and non -TIF regions within a county Here we find that TIFs are associated with small but positive growth in assessed value. However, we find uniform negative impacts ofTIFs on traditional measures of economic development such as employment, the number of business establishments, and sales tax revenue. This leads us to conclude that the Indiana TIF is not an economic development tool, but a county budget management tool. We offer the following policy recommendations. • The use of TIFs in Indiana should be reviewed by the legislature. The nature of these results implies that while the average Tiff has no meaningful impact, there are undoubtedly some with positive and some with negative effects on their counties. • County leaders considering TIFs should evalu- ate the potential tax shifting to non -TIF regions when considering a TIE This is especially critical in light of the effects of property tax caps, and the need to make local quality of place improvements in many places in Indiana. • TIF reporting could be improved to include a cri- terion for evaluating the potential impact—one that counts tax rates, employment, and capital 1 investment both before and after the TIF proj- ect. This should be done for both the TIF district and the non -TIP area of a county. These findings should be continuously made available (with at least annual updates) for all TIP districts. The legislature should limit the use of TIFs to those counties that exhibit at least minimally effec- tive fiscal management. Specifically, we recom- mend precluding the use ofTIFs in counties with unfunded pension liabilities (less than 80 percent actuarially funded), in school districts that have requested transportation waivers within the past five years and in counties or municipalities lacking an adequate Rainy Day Fund. Introduction Tax increment financing (TIF) is a form of economic development incentive—the property tax revenues associated with the growth of assessed value or 'increment' on a designated piece of property can be captured for redevelopment. In Indiana, a TIP may be formed by a county or municipal government through the creation of a redevel- opment commission. These redevelopment commissions may then identify a geography, which will act as a T'IF district and capture any growth in property taxes from new construction in that area for other uses.i1t The geography of a redevelopment commission is colloquially referred to as a 'TIF District.' The property tax revenues from new construction within aTIF district are designed to be for areas "need- ing redevelopment" under the language of the Indiana Code. The legislative language is broacl, and by the permitting of bonding repay- ment, effectively allows any project which might otherwise not be clearly within the scope of traditional redevelopment. Thus, the use of TIFs offers some of the most flexible financing tools available for local governments in Indiana. Debt within Indiana's TIP districts comprises some 20 percent of Indiana's $12 billion local debt out- standing as of 2013. There is not a comprehensive analysis of all TIP uses in Indiana, a product which niay not be possible given the data limitations. Still, the use of TIF monies by redevelopment commissions appears to tar- get activities designed to improve the economy of the region in which they arc spent. I-Iowever, there is legitimate concern that the intent ofTlF-related expenditures does not in fact generate better economic outcomes for counties. It is that issue upon which this study will focus. To do so, we construct and test a series of models that mea- sure the impact tax increment financing plays in affecting economic development measures, such as employment and assessed value. We also test the impact TMP plays on the effective tax rates in counties. We then offer routes to additional research and policy recommenda- tions germane to our findings. We begin with a review of the existing research on these issues. Other Research on Tax Increment Financing A number of authors have previously examined TIFs to evaluate their impact and efficacy, Several authors have examined economic activity within a TIP district. They have examined questions such as why and where TIF districts are located, the effect on residential industrial and commercial property value, and the effect on public services, wages, employment, and tax rates. The intent of most TIPs are to redevelop or refurbish a region, but the stated legislative intent, and the actual practice of loca- tion has seen some scrutiny. Diane Gibson (2003) examined Chi- cago TIFs in the 1990s, finding that distressed but not very poor regions were more likely to use TIFs, which were often connected to empowerment zones in the city. Felix and Hines (2013) found that areas oflower income, those proximal to state borders, and those with poor treasures of political performance tend to offer incentives in general. However, very poor and very troubled regions tend to offer TIFs less frequently than others. Man (1999) found no evidence that growing cities were more likely to adopt TIPS than those not grow- ing in her examination of Indiana cities over a decade. Mason and Thomas (2010) studied TIF use in Missouri, finding that there was geographic competition related to TIF adoption (cities were more likely to adopt a TIF district if their neighbors had done so), and that high levels of economic inequality between municipalities may be caused by TIF adoption. Byrne (2002) found that such strate- gic interaction with regard to TIF adoption also was evident in the Chicago metropolitan area. Together these studies suggest that TIPS locate in communities focusing on revitalization, but nor in terribly poor places. Also, there is a strong suggestion of inter -regional com- petition in TIP location. Several studies have identified property value growth in TIF dis- tricts and in counties overall using TIFs. These include a study of Michigan cities in the mid 1980s and in Wisconsin from 1990 to 2003 (see Anderson 1990; Merriman, Skidmore, and Kashian 2011). A follow-up study of Wisconsin by Kashlan and Skidmore (2012) found that TIF and non -TIF regions experienced differential effects within the same county. They report that the presence of a TIF dis- trict did not affect the tax nates within the municipality in which it was located, but did in the surrounding taxing jurisdiction. A third study of these Wisconsin TM districts determined that about half of the annexations in Wisconsin at the time were of recent TIF districts. A study of Milwaukee TIF districts found that the value growth of property within the TIF district is attributable to the capitalization of higher quality public services offered in these districts. Brent (2009) reports the same effects in Chicago, with the quality of local public services influencing the magnitude of the impacts. Anderson (1990) found that property values grew faster in areas with TIP districts. This led him to suggest that TIF may be a budget manipulation, not an economic development tool. When studying Indiana, Man and Rosentraub (1998) found that residential property values were higher in TIF adopting cities than in those without TIF districts. These stud- ies focus primarily on residential property. Studies of commercial and industrial property are more mixed. Weber, Bhatta, and Merriman (2003) found that industrial TIF parcels in Chicago were of lower value than those outside the TIF district, but the opposite was true in mixed-use TIF districts. Byrne (2006) concludes otherwise, finding when controls for regional eco- nomic conditions are included, the industrial TIFs are the only types that experience property value growth within Chicago. Smith (2004) examined the same geography and found that property value growth among multifamily housing units occurred in TIF districts. I. Indiana Code 36-7-14 39(a) outlines this process, with definitions of the base and minimum taxes due to taxing units. 2 These two separate sets of studies focusing on residential and non- residential property growth tell one clear story, and an unclear one. It appears relatively clearly across several studies in several regions that residential property growth within TIF regions occurs, and that differences between TIF and non -TIF regions occur. In contrast, the differences in findings and the lack of diversity in geography provide no clear conclusion on property values for industrial and commercial property within or outside of TIP districts. In an examination of TIF district effects on Illinois rax rates and revenue, Weber, Hendrick, and Thompson (2008) found very little effect on revenues of surrounding school districts in suburban areas. In urban school districts, TIFs reduced available revenue, while in rural school districts, the presence of a T7F district boosted tax revenues. Studies of the impact of TIF on economic activity yield mixed results. Man (1999) reports that cities with TI is saw greater employ- ment growth than those without. This finding was echoed by Bynre (2010), who found industrial TIPS were accompanied by employment growth, while retail TIFs in Illinois saw negative growth of employ- ment. Dye and Merriman (2000) offer some die clearest exposition on TIF and economic development. These authors found that locations that adopted TIPS grew more slowly than those that did not. This appears counterintuitive, and they were able to reject empirically sam- ple selection bias in their data (slower growing places did not adopt TIF more heavily than faster growing places). They also offered a dear theoretical explanation that echoed the findings of Johnson (1990). While TIFs may boost investment and employment within a region, they also affect tax races in non -TIF areas, which in turn may reduce net economic growth within a municipality or county. TIFs have been in use for decades, and there is evidence that places with TIPS see some new invesunent, some higher residential property value, and some increased growth within the TIF borders (in income and employment). These are fairly intuitive results, but they do nor address the net effect within municipalities or counties who adoptTl Ps. The more sophisticated analyses reviewed here (and this is the bulk of research the matter), tell a more nuanced story about TTP adoptions and economic development. Since TIFs potentially shift tax burden, the net effect of a region's economic performance will not likely mimic that of the TIF district. Any proper analysis of TIF impacts will have to evaluate the net effects of TIF and non -TIF areas within a county or municipality. We next rum our attention to that issue. Tax Increment Financing in Indiana Table 1 shows basic statistics for TIF, in Indiana from 2003-2012 for 91 of Indiana's counties. The typical county had about $215 mil- lion of net assessed value within TIFs in 2012. The aggregate net assessed value in TIFs increased from just over $10 billion in 2003, to more than $19 billion in 2012. _Figure 1 shows snapshots of TIF usage from 2003 and 2012 These maps indicate that TIF usage has increased over time and Cuban coun- ries use TIFs more intensively. Table 1. Net Assessed Value in TIFs ($ M'Ilion, Inflation -Adjusted) i . Year Mean Median .Std Dev'• 'Min M µ'' SumO1s.-- 2003 111 23 296 0 2,430 10,056 91 h3 m0 21' ' 330 h 2 829; ,10 239 c:..11.121.. 2005 113 21 310 0 2,609 10,290 91 [2006.:'r. 113 24, 'X s 1267{ ny :.40. , .2:018, t10 274; =x'91 2007 155 33 364 0 2,447 14,082 91 2008 t µ 841 35` ". 437' 1 r'_0 3 06ff,te35 r16 489 2009 199 43 480 0 3,316 18,099 91 [ 010 1.0216 a 461 i'.F,ir,S 5-'t..:.sa0 3:4934 b1 69 54i aui.''9r.1't: 2011 228 45 554 0 3,573 20,718 91 912012 : w 215 -51, a,. 4 95 i -vi,,.0 r 3364 404.98 -^R 91'i Note: LaPorte County is not included due to missing data fo some years. Source: Author calculations from DGLF data. Figure 1. TIF Net Assessed Value by County, 2003 & 2012 (Inflation -Adjusted) Percent of Net Assessed Value 2003 Net Assessed Value ($ Million) 2003 Ott, D❑ ❑ $11sx<$50 M dQ�o ❑ $50 Msx<$100 M U off fl❑�O ® $100 Msx<$500 M ❑"ID �❑a-Q III $500 M -t ❑ 6�°gloQ°R k_11=1Lf-D-ZOLityA PeRe Net Assessed Value per Capita ($)($ 2003 ❑ $0 111 r9{9©❑ 0$1sx<$100 aQ�-ljjo�7 ❑ $1005x<$1,000 11LSO LIU ❑ $1,000sx<$5,000 ❑ �SJ❑J❑ El $5,000+ PRd� nkalat un a ao p 3 f2F4cA 2012 bQofa ❑ln17-41100 P❑00Q � Dpp �LJQI WaoN fj6SECIEJEdQoopa16L ar i�R 2012 2012 Dcgdact rknosro ©o goEc• 0 . opooring BaricaunD Aria°84eP Table 2. Descriptive Statistics Stmt sf ciw1,s +lam .4 Value of TIF (% net assessed value) tN a d value in TIF($$Jm Ilion) m Net assessed value in TIF per capita ($) n �s TM ,sec sea . smllion) +NomTIF�eba�aes�edvalue:($ million)e. Non TIF net assessed value per capita 1:Effectiveu o' taxrate(/) e" P.,...Pu6...,.,K.: Net assessed value ($ million) (-Net asserts yd,vlue,Per„zP,it�a ($) '^ Total employment 1:M 1r -sayers turingg employments Retail employment ,)Business establitnents� Sales tax due ($ million) Note: All dollar values were adjusted for inflation using the CPI. 2.00 it Std Dev• r 3.49 1,476 3377? 48,638 w1µ866 3,542 y50T114' 39,257 .9 4,285 36661 34.85 928 ),,115 80: 46,713 :h 15828 1,615 } 48016: 16,937 rfa 2813" 1,868 ;,54t: 11.55 1,856 I11cT5'794 A)aci13,730 Wit W.473 6,163 4113.894-1 78,743 asc.5 831 7,723 t§:(1•4'4.2764.6' 87.54 TIF Impact Modeling for Economic and Fiscal Measures We propose to treasure the impact of TIFs on activity ill Indiana counties from 2003 to 2012.121 This approach is motivated for two reasons. First, available data on TIPS and potential economic vari- ables are limited to these years. From a geographic standpoint, the county is the unit of interest for several reasons. Economic activity within a redevelopment region and surrounding areas will be primar- ily captured within the county. So, such variables as employment and assessed value within a county will capture much of the net impact of TIFs across the boundaries of the redevelopment commission. In this respect, we take seriously the analyses of Dye and Merriman (2000) and Johnson (1990), who report that there are impacts of TIP dis- tricts within and outside their geographic boundaries. To better understand the impact of TIPS, we compose a model where economic activity within a county is a function of economic activity in both TIP and non -TIF regions within the county. `This total economic activity is hi turn a function of total capital, total labor, and government services. The Appendix contains a brief theo- retical treatment of the issue. Not surprisingly, each of those variables is affected by taxes and other variables including regional spillovers, recessions, and a host of county -specific issues like the presence of an interstate, large manu- facturing firms, or a university. At the heart of the empirical issue is the combined effects of a TIF across both TIF and non -TIF regions. The reason for this is that with tax rates comprising an element of capital and labor formation, the effects of a TIF are not isolated to the TIF district if there are resulting changes to non -TIF areas. In particular, changes to tax rates in non - TIF areas effect capital accumulation and labor that in tum effect total production. So, our interest is the net observed effect of a TIF in these counties. •Yd Mna Mx Maximum 22.00 3 573 8,213 '"\52ha891 103,734 3:601 55,219 fi 104 156; 677,569 73 286. 66,127 xt h24 566' 827.82 gAmimum.a 0 } ' t d: 0 X266 24,740 667 266 art+637a 2,299 v 119 0.94 920 IK 3'915' 915 915 d .,915.. 915 r'91z5- 920 kitirb 897,4 920 920 DLGF 'Author's calculation using LGF data Authors calculation using DLGF data Author s�caleulahons ,e tA*NMe,, Author's calculations DLGF DLGF BEA, REIS tiBEA REIS) j�,yamr.xu.-a r.. -,.a BEA, REIS County B iea 11tte_rn Stats Indiana Unfortunately, we do not have data 011'1'IF and non -TIF areas, and very limited sub -county data. We do have a decade of information on tlhe size of the TIE and measures of capital accumulation, tax rates, and employment. To examine the overall effect of TIF, we perform nvo separate tests. First, we construct a simple statistical test of the effect of the TIP on taxes and capital accumulation in a county. We use a modelling approach similar to that of Greenstone and Moretti (2003) and Faulk and Hicks (2013), in which we offer a treatment model of the'1'1F in each county and year. In the first set of estimates, we seek to evaluate the impact of TIF on the effective property tax rate in each county. This is followed by an analysis of the relationship between TIFs and capital formation and employment. In each stage, we employ a model in which the affected variable (effective tax rate, capital, employment) is affected by TIFs within a county, the mean value of adjacent county TIFs, statistical controls for unrelated correlation across time and geography, a time trend, a random error term, and an error term controlling for those factors that do not change within each county over the observed period. The Appendix contains a brief technical explanation of these models. Sum- mary statistics for variables used in the models are listed in Table 2. In our test of rhe effect of'l'IF on effective tax rates, we estimate five different model variations to test the role TIE plays in county -level tax rates. In most of the models, we use the share of county -wide assessed value in TIFs as the TIF measure. In the first nvo models, we employ ordinary least squares and vary fixed or autoregressive terms. The third model removes the spatial aurocorrelation term. In the fourth model, we use an alternative TIP measure—the assessed value of property in the TIE The final model examines the growth of the effective tax rare as an alternative measure. As shown in Table 3, these models provide clear insight into the county -wide tax impact of TIP districts. In models that explain 2. Data on the net assessed value in TIFs are available from the DLGF for 2003-2013, and most economic variables of interest are not yet available for 2013. Data on the number of TIF districts in a county or the land area covered by TIF districts in a county are not available. 4 Table 3. Effect of Tax Increment Financing on ruble ' uPanel Least Squares r 0.484516*** 0 .TIF percent of county assessedvalue,+ t' TIF assessed value ($ million) a ialats* ?cau xanw x guag Average TIF assessed value inat tltacent county ($ milOmn)q na Spatial autocorrelation Time trend Temporal autocorrelation Cw : Sra: st ounty, fixed effects Period fixed effects Ad ustechR-s uared F -statistic [0.0000] 64-.61Y-7"65,: [0.07741) 1 o00191f , r0!172311b;„ County Effective Property Tax Rates 1°d n vn�lx.j� .+es r _RG ..°,4'a Panel Least Squar.sPanel Least Squares 1.037055*** 1.788927*** [0.0087] [0.0000] 0009914 c .9.012763* [ .[0:08451 W. 4;] *' + [0:07571 0( 0.754467*** [0.0000] 0-06031 ';[(5 -1.89 -33M -33b 0.257077*** [0.0000] Yes ,'ft No 63.77902 [0.0000] fDurbinWatson'stat Note: Parentheses indicate standard deviation. 0 000198..44; 0.379711* [0.00001 t + 0 0227171y v� [07861] t Yes 65.42415 [0.0000] kl¢4335493,�; 9191593 LoSsa 0.155984*** [0.0014] No 0.833459rV' 1'_1 44.30247 [0.00001 2:121852 <i ...� Panel GLS Lagged 2.098207*** [0.0000] tt Growth of ETR' Panel , n Least Squares' -0.827356*"* [0.00001 013172 s[0 0301f{4 0.000156** [0.0234] 17:82E:05( [0 39811p'y. 0.138957*** [0.0064] 0.010331 1 '*,36E05 0.498111*** [0.00001 0 016766 *. [000001+- , -0.359907*** [0.00001 1,909-92.611MllLik : 0.170534*** [0.00011 No 0[869968 51.87487 [0.00001 2095394, r No '�0�145105,; 2.290690 [0.00001 7417171... Significance: * at 0.10 level, ** at 0.05 level, *** at 0.01 level. This model uses White's (1980) corrected t -statistic, asymptotically efficient p -values or F -statistic, Table 4. Effect of TIF on Capital Accumulation (Assessed NetAssessed Value, ,.. ;,per Sapitao-, 0.057517*** [0.00001 yt?5i258914 3.11E-06 [0.1543] tT 3177E 06 OAOOOI * -1.50E-07 [0.0001 1 ;1•0 ;005334* -0.002369*** [0.00001 0'.000396*. 0A000] 0.469771*** [0.00001 ` Yes No c abs s TIF assessed val($ m ue illion)., ^���, a,.}or�ts:3aA....i.SWfl]L�.,w3P Average TIF assessed value in adjacent county ($ million) 1 tt! o 5patiaial auocorrelatwn`i>>.t µ,y Population tura �E�TF2 TIF percent of county assessed value =Lore it gu Temporal autocorrelation f CountyAfaed;et..Vs Period fixed effects :Adjusted;R-squared,t F -statistic t.Durbm..Watson stat '. Value) etAssessetlyalueue! er,capita:b ,a 0.064203*** [0.0000] 44261 a,Y [00000] - E+jj 1.01 E-06 [0.53581 ri's1144E 06)4, l' ?�C000017r,.,. at -1.20E-07 [0.0002) ,0 006332*4 -0.001845*** [0.0000] No r Ass ss : cae NontTl FJNebAssesseud�Value '4A46 t1PeFCapital''4.,e-. .f.„ 0.057504*** [0.00001 7[2692/A ,_ NonAIFiNet Assessed Value per Capital df4a \,..', 0.064171*** [0.00001 Iroxr,34443044 [6:0000]L ' 2.44E-06 [0.1754] 1 44E 06'rt.:' [00001] - -1.20E-07 [0.0002] y 0:006331 -0.001853*** [0.00001 6.88E-06 [0.0038] ;77E-,06 +[0.0002] 6. -1.50E-07 [0.00011 279.7803 [0.00001 Note: Parentheses indicate standard deviation. 0 005331* [Of00'00] ii -0.002374* [0.0000] :0!000397,* [0'0000]1 ig 0.469755*** [0.00001 Yes 248.0735 [0.00001 Significance: * at 0.10 level, ** at 0.05 level, *** at 0.01 level. This model uses White's (1980) corrected t -statistic, asymptotically efficient p -values or F -statistic. 5 the bulk of tax rate variation, the presence of a TIF demonstrates an across-the-board, statistically meaningful impact under different measures of TIF and different model specifications. Using the value of estimated coefficients, an increase of TIF share of assessed value of 1.0 percentage point is associated with an effective tax rate increase of around 0.01 percentage point in the county. The fourth model suggests that each $10 million increase in assessed value within a TIF increases effective tax rates by 0.01 percentage points. This result has two possible explanations. First, it may be that the use of TIFs increases the cost of public services for adjacent taxpayers. Second, it may be that communities with higher effective tax rates use TIPS more intensively than communities with lower effective tax rates. Existing research allows for both possibilities, and in this set- ting we cannot clearly isolate the extent to which either of these two effects dominates. The strongest conclusion that we can draw is that higher effective rax rates are associated with TIFs. These results lead to the second analytical step, which is the county- wide effect ofaTIF on capital accumulation. We know that increases in assessed value within TIE districts occur as new facilities are con- structed in these districts. Here we examine a broader impact—tie impact of assessed value in TIF districts on the overall assessed value in a county, and on assessed value outside ofTIF districts in a county. We use a more fully parameterized model and estimate nvo specifica- tions that provide similar results. The results show that a $1 million increase in TIF assessed value is associated with an increase of $4.43 to $5.26 per capita in assessed value in the county. This is a small but positive relationship. In the typical county, the increase in assessed value associated with TIFs would be in the range of $306,000 to $365,000. Again, we are unable to draw a causal link between TIFs and overall assessed value growth. It may be that counties that have higher growth in assessed value use more 17Fs or that TIFs cause overall assessed value growth. The results of the models showing the relationship between TIF assessed value and non -TIE assessed value suggest that most of the assessed value growth is occurring in non - TIF areas. A $1 million increase in TIF assessed value is associated with non -TIF assessed value growth of $3.44 to $4-27 per capita. Again, this is a small but positive effect. See Table 4. Next, we examine the impact of TIFs on employment, the number of business establishments and sales tax revenue in a county. Por brev- ity, we report only the influence of TIF on each of these variables in Table 5. While TIFs are associated with a small but positive increase in assessed value, the same is not true for their impact on employ- ment and other economic indicators. A 1.0 percent increase in TIF assessed value in a county is associated with lower employment levels and a lower number of business establishments. Due to statistical issues, we are not able to conclude that TIFs cause lower employment, only that they are associated with lower employment. It may be that counties experiencing employment/ establishment decreases are more likely to use Ties. Alternately, it could be that TIFs do lead to employment/establishment declines as competing businesses in non -TIF districts reduce employees or close 6 Table 5. Effect of Tax Increment Financing on County -wide Economy (Dependent Variable) b e , 470WAIT,j't Effect of a 1 0/ Increase of Total employment Manufactuhn employment- .tan. Retail employment BusmessAsv_ Festablishmen -78.99643* [0.0504] X32:70521* [0 0169] mwegga -15.66258** [0.0137] Sales tax due ;;;assessed value'm TIF„'8 A 1.0% increase in TIF assessed value is associated with 78 fewer jobs in the county. esxnn:**xxr.))-. :A807 increase in TIF assessed; :value rstassociated wdhs321fewer}' ?manufacttuuunng.Iobs m thet unty �$ A 1.0% increase in TIF assessed value is associated with 15 fewer retail jobs in the county. A 1 07 me ease in TIF assessed ::r value ishassociated wdh 2Sewer ibusme'sstestabhsh ments in y± 1 l T �T ()nnt .. ,t.,P.u:fitib9iY,1J+1W..( -2.266806 Negative, but not statistically [0.1277] significant Note: Parentheses indicate standard deviation. Significance:* at 0.10 level, ** at 0.05 level, ***at 0.01 level. This model uses White's (1980) corrected t -statistic, asymptotically efficient p -values or F -statistic. Source: Author calculations. While TIFs are associated with a small but positive increase in assessed value, there is a negative association with employment and other economic indicators. in response to new businesses opening within TIF districts. Or, some combination of these factors could be at work. TIF5 have no discernible statistical impact on sales tax revenues in a county. This later result would occur if retail activity shifts from non - TIF to TIF areas as new retail establishments are developed in TIF districts so that retail sales in TIF locations are substituting for retail sales in non -TIF locations. Our set of models assessed the effect ofTIF on total capital growth, total employment, manufacturing employment, retail employment, the number of business establishments and sales tax revenue. In all cases, the effects of TIFs were negative, but in the case of sales rax revenue the impacts were not statistically meaningful due to a high variability of effects across counties. Summary and Recommendations The use of TIFs in Indiana has increased substantially over the period examined in this analysis. To better understand this, we exam- ined TIF districts in Indiana 2003-2012 in an effort to evaluate the impact of IF on capital growth, employment, and tax rates in coun- ties. This approach is designed to test the efficacy ofTlFs as an eco- nomic development tool designed to boost employment and capital investment in communities. Our findings are very clear and echo those of many other researchers. First, we find that the size of a TIF within a county is associated with higher effective tax rates within the county. This is not surpris- ing given that Indiana's local property tax system would necessarily shift burdens of taxation from'FIP to non -TIF taxpayers to maintain constant levels of public services. We find that TIFs increase tax rates by about 0.01 percent on average, which is large enough to influence the tax variation across Indiana counties over the past decade. While we cannot conclusively report: that TIFs are the cause of higher tax rates on existing taxpayers, that is a very likely effect. Second, this finding establishes the need to examine TIF impacts in both TIF and non -TIF regions, since higher tax rates levied on non - TIF businesses and households will necessarily alter the net effect of any new capital accumulation or employment within a TIF district. We conducted this analysis asking whether the size ofTIFs within a county influenced net county capital accumulation or employment. There was evidence that, on average, there was a small, positive cor- relation between the size of a TIP district and capital accumulation (measured as assessed value). TIFs were negatively correlated with other measures or economic development such as employment, business establishments and sales tax revenue. I-Iowever, in no case was the statistical certainty or size of the lost employment sufficient to conclude that TIFs caused these negative impacts in Indiana counties. The strongest conclusion that we can draw is that TIFs are associated with these negative outcomes, which is a finding that is consistent with our first result—that TIFs are associated with higher effective tax rates in the counties in which they are used. This study offers a very robust and critical Ending regarding the efficacy of TIFs in Indiana: The presence and size of a TIP district within a county is associated with higher overall tax burdens, which likely is due to a shift of the costs of public services to other taxpayers. This finding holds across multiple variations in our statistical models 7 County leaders considering TIFs should evaluate the potential tax shifting to non -TIF regions. This is especially critical in light of the effects of property tax caps and the need to make local quality of place improvements in many towns. and measures of TIF presence. The result of this tax shifting is that an examination of the economic effects within TIF and non -TIF areas is necessary to evaluate the net impact ofTIFs on economic develop- ment. In so doing, we found that the net impact of TIFs on a county economy is modest, but on average negative in measures of economic development other than assessed value. This suggests that the Indiana TIF is not an effective economic development tool, but is instead a budget management tool for local governments. Policy Considerations Our findings point to some specific policy considerations for Indiana: • TIF usage should be reviewed by the state legislature. The nature of these restdts implies that, while the average TIF has no meaningful impact, there are undoubtedly some with positive effects and some with negative effects on their counties. • County leaders considering TIPS should evaluate the potential tax shifting to non -TIF regions. This is especially critical in light of the effects of property tax caps, and the need to make local quality of place improvements in many places in Indiana. • TIF reporting could be improved to include a criterion for evaluat- ing the potential impact—one that counts tax rates, employment, and capital investment before and after the TIF project. This should be done for both the TIF district and the non- fIF area of a county. These findings should be continuously made available (with at least annual updates) for all TIP districts. • The legislature should limit the use ofTlFs to those counties exhib- iting, at the stew least, minimally effective fiscal management. Spe- cifically, we recommend precluding the use ofTIFs... • in counties with unfunded pension liabilities (less than SO per- cent actuarially funded), • in school districts that have requested transportation waivers within the past five years, and • in counties or municipalities lacking an adequate `rainy day fund.' • Tax increment financing itself is a budget management tool, which when used prudently may boost the economic prospects of a region. On average, its use does not improve economic conditions, but instead is associated with higher effective tax rates, and less employment and income. Therefore, its use should be limited to communities with effective local fiscal policies. References Anderson, J.E. 1990. Tax increment financ- ing: Municipal adoption and growth. National TaicJournal 43 (2): 155-163. Brueckner, J.K. 2001. Tax increment financ- ing: A theoretical inquiry. Journal of Public Economics 81(2): 321-343. Byrne, PE 2006. Determinants of property value growth for tax increment financing districts. Economic Development Quarterly 20(4): 317-329. Byrne, P.F. 2010. Does tax increment financ- ing deliver on its promise ofjobs? The impact of tax increment financing on municipal employment growth. Economic Development Quarterly 24(1): 13-22. Cattail, D.A. 2008.' Fax increment financ- ing and property value an examination of business property using panel dam. Urban Affairs Review 43(4): 520-552. Dye, ILE, and J.O. Sundberg. 1998. A model of tax increment financing adoption incen- tives. Growth and Change 29(1): 90-110. Dye, RE, and D.F. Merriman. 2000. The effects of rax increment financing on economic development. Journal of Urban Economia 47(2): 306-328. Farris, S., and J. Horbas. 2010. Creation vs. capture: Evaluating the true costs of tax increment financing. Journal ofl3-operty Tax Assessment &Administration 6(4): 5-28. Felix, RA., and J.R. Hines Jr. 2013. Who offers tax -based business development incentives? Jou nal of Urban Economia 75: 80-91. ,,BALL STATE UNIVERSITY CENTER Fernandez, G.B. 2004. Tax increment financ- ing: Interaction between two overlapping jurisdictions. Journal of Urban Economia 55(1): 151-164. Gibson, D. 2003. Neighborhood character- istics and the targeting of tax increment financing in Chicago. Journal r f Urban Economics 54(2): 309-327. }Cashion, RD. and M.L. Skidmore. 2012. Interim report summarizing significant changes in tax increment finance policy in Wisconsin. University of Wisconsin - Whitewater Fiscal and Economic Research Center. 1-16. Man, J.Y., and M.S. Rosentraub. 1998. Tax increment financing: Municipal adoption and effects on property value growth. Public Finance Review 26(6): 523-547. Man, J.Y. 1999. Fiscal pressure, tax competi- tion and the adoption of tax increment financing. Urban Studies 36(7): 1151-1167. Man, J.Y. 1999. Effects of tax increment financing on economic development. Journal of Budgeting, Accounting er FinancialManagement 11(3): 417-430. Mason, S, and K.P. Thomas. 2010. Tax incre- ment financing in Missouri: An analysis ofdetemninants, competitive dynamics, equity, and path dependency. Economic Development Quarterly 24(2): 169-179. Merriman, D.F., M.L. Skidmore, and R.D. Kashian. 2011. Do tax increment finance districts stimulate growth in real estate val- ues?. Real _Lune Economics 39(2): 221-250. Skidmore, M., D. Merriman, and R. Kashian. 2009. The relationship between tax increment finance and municipal land annexation. Land Economics 85(4): 598-613. Skidmore, M., and R. Kashian. 2010. On the relationship between tax increment finance and property taxation. Regional Science and Urban Economics 40(6: 407-414. Snaith, B.C. 2006. The impact of tax incre- ment finance districts on localized real estate: Evidence from Chicago's multifam- ily markets. Journal ofHousingEconomics 15(1): 21-37. Smith, B.C. 2009. Ifyou promise to huild it, will they come? The interaction between local economic development policy and the real estate market: Evidence from tax incre- ment finance districts. Real Estate Economics 37(2): 209-234. Weber, R., S.D. Bhatti, and D. Merriman. 2003 Does tax increment financing raise urban industrial property values} Urban Studies 40(10): 2001-2021. Weber, R, R Hendrick, and J. Thompson. 2008. The effect of tax increment financing on school district revenues: Regional varia- tion and interjurisdictional competition. State &Local Government Review 40(1): 27-41. Photo Credits Flickr Creative Commons. Northeast Indiana Regional Partnership (pg 1). Michigan Municipal League (pg 7). Publication Credits Victoria Meldrum, manager of publications and web services, Center for Business and Economic Research, Ball State University. the State University's research;: analysis, regional. economics, CBER Data Center, newsletter with ^^• Business for Business and forecastingafor a Research Researchiconducts timelyheconomicpolicy CBER research i dlude3`Ip b1c5'"' o' a oergy�sector studies.. prrooduce,thhee r •'. Business Bulletin,.a weekly indicators.. - tos, and the Indiana iodated economic CENdiERyFOR,BUSINESS-AND ECON0MTV RIE SEE,AAIR`CH Centerppoi.Brfsmess State,Bniversity 2000 1AT liiniversity Ave., Muncie, +U3063600 765-285-5926 www.bsu :dutch& Appendices Appendix A. Methodology Appendix B. Literature Review Appendix A. Methodology Specimen Al. Brief Theoretical Outline Yi = YT1F YNnF = f [ K(N,T) N(K,w), GS(N,K) J where output, Y in county i, is a function of output in both TIF and non -TIF regions within the county. Output, or total economic activity, is a function of capital, K, labor, N and government services, GS. Each of those variables are affected by taxes, T, and wages w. We focus on the net effects of the TIF on capital formation, labor and taxes between the two locations (TIF and non -TIF). Here we assume that the interest rate is independent of local conditions. We also assume that labor and capital are gross complements, so that an increase in capital will lead to increased employment. Both of these assumptions have a great deal of empirical backing we simply do not test them in this model and that makes them assumptions. aN dY�=f'[K'(aNK+aTT)� N\aKK+dwaN w+2TT�, GS'[aKK+ BNSN�1 Assumptions explicitly involve only the elements of: DK 2N 2N aGS 2K aN acs 8N, 8K, 8w, 2K > 0 and 3T, 8T, 2N <0 BGS ON for which only aK and 8w > 0 offers a plausible challenge, but in assuming higher levels of private capital lead to higher levels of government services we make a conservative assumption regarding output. In assuming that higher wages leads to more employment we are assuming the supply effect dominates in a region. Specimen A2. Brief Empirical Specification Effective Tax Rate Estimate ETR,r=c.,+c+(3,TIF,,+(3�WTIF„+y1WY+T +eETR,1+e�� Economic Impact Estimates Estimate Y„=c�+c+(3, TIF„ +RWTI Fig+y1 WY+T+eYn-,+end where impacts on effective tax rates (ETR) or economic impacts (Y) in county i, in year t, are a function of fixed c. and common c, intercepts, TIF, adjacent county TIF, weighted by W, the first order contiguity matrix for each county, a spatial and time auto-correllation measure and a white noise error term, e. 9 Appendix B. Literature Review Table 61. Articles 1990-2006 IrYea`r,✓ Article.Author,; tGeggraphic±Area Y^,'..Research Quesiier i " ' TIFsMeasure' .;, StatisticaliTechniquer I'. , , 1Malor,Fmdmg, Cities using TIF had greater property value growth, all other things being equal. Owner occupiied housing values were - 11 / greater n TIF adopting cities' thar in non TIF, cities t Cities with TIF created an average of 4% more jobs than those without. sv raAteM an -1 :444,i :.. ctlesare mt,+r Pooeltoad,.. hkelytoadoptTIF.t "T ...; a � . Municipalities that adopt TIF grow more slowly than those that do not. A. muenicipR ahttthat tiadopts T IFwhen\ s d ats neighboro"not makes itself4 vf; ,a4preferred'location for businesses " ' aiiu'xcluingN". arewhichutrn puts, pressure on,nearby municipalities;to i, also ado t TIF ArO i ' l' rr` p. .... xj �*:�+��i;�^`xh'k?,Snu tin; TIFs in Chicago tend to be located in economically distressed tracts but not the most severely disadvantaged. Chicago has used TIF to complement empowerment zones. k L"%[R„ .r ro ck +8d4centip4opemes appregated at at :;. higher rate than both other groups''',€ Findingsdid,notfullyrsupport orfullya:, conbadicfthe'Ihypotheeis 'vyt+ %'' In industrial TIF districts, parcels sell for less than identical parcels outside of district. Industrial parcels in mixed - use TIF districts sell for no less, sometimes significantly more than those outside the district. - Appreciiatio prat es within •TlF221istractss"." exceeded those offproperties outside ' TIF Soundanes' and the designation _ofj F dsslncts stimulates market value mcrersesin areas that are, s t. ultimately, designated as TlFa distracts Industrial TIF designation is the only classification having an impact on success. TIF area/location, population density, race/ethnicity, and recency of creation all influence growth. Results also suggest a positive relationship between blight and property value growth in TIF districts. 1990 1998 1999 1999 2000 %2002a 2003 2003 t ir -21 Mt k s 2003 12004 2006 Anderson, John E. V a 1 j. MarJoyce Y &'%Mark S rl i . Rosenub tra Man, Joyce Y Man,;Joyce Y Dye, Richard F. & David F. Merriman Byrne 1Paul r a} Gibson, Diane Smith Breni C * s resentati? ),fr Weber, Rachel, Saurav Dev Bhatta, & David Merriman Smith BrentC =Chicago Byrne, Paul F. Michigan cities, 1985-1986 Indiana cities „ t y `� Indiana cities, 1985-1992 lana cities- Chicago metropolitan area Chicago tt 'metropolitan � area (256 municipahbes edst Chicago) Chicago, IL census tracts, 1990-2000 'Chicago/4U �, yAq++ 4 r ritt, 3 Chicago, IL IL • Chicago metropolitan area Are property value growth and TIF adoption related? t.. Fl How does TIF affect . r property: value growth '+ rttt t 't,_ : + Do TIF districts create an increase in local employment/jobs? • Ayrtre F ; How does TIF affect the property value growth rates of municipalities? [51;\municipal' tiess engage in strategic interaction ,,t � when engaging miTIF' �t adoption] lt decisions Why are TIFs located where they are within the city? ( Does proximty topl I44 ddistricts:h avec positive i impactoncommercial nM property,value?erii.t I` Does TIF raise urban industrial property values? Do properties wilhma TIF 'Pi strict ggh i ri higher t rates of appreciation after the TIF,is designated ".'I ; compared tio,pLopertieskirff outside.the TIF and ±.•j^ compared to properties sold inithedistractbefore designation' t t WhichdistrictDummy characteristics are important in influencing the success of TIF as measured by the growth rate of property values in the TIF district? Dummy variable for TIF Sones of dummy " Vvanables corresponding to the various years of TIF.: implementation TIF as a dummy variable mmyvaable for Dummy variable for TIF adoption, TIF district share of total equalized assessed property value of 1992 TFa s•ata tenS t variable r it'",i '�,.riP p > i, t,fi �` < - a t fi tt"3 tattLy Was tract included in TIF district designation between 1990-2000? t Q.11WTl boun45i, withint250 yarrdsfof i id IF,(controlgroup)orl"t'' notproxlmatetoTIF; TIF as a dummy variable Dichotomous vvariablese s representing whether or not a.property sale r occurred within at - designated,ThErdistriati ogina TIF project designation ,( variables for TIF classification Structural probit model First difference model fi fi ; ii , a Cross-section time series regression r�Cfra':prob moel{ sugstailicalgrow' Regression Linilear7i prtoba bihty model , r rzkT - �Tk tegeneral tllookmgtoepndorrelocaemth Weibull duration model W do He*thea y ilt y� ir} a}y ^ ,r ' ,ra'�'ta't '' iha Regression* using a two-stage procedure to correct selection bias. 1.) Multinomial logit model 2.) Adds selectivity correction factors derived for the logit model Hedonic models odels e 1t ix a ,, 1 ;: " r t t y t5 OLS regression model 10 Table B2. Articles 2008-2013 Year ArtrcleAuthor (G oe graphic Area M aukey e Wl,' ,1980'1999 e -x Illinois school districts (782 of 896 total) Wisconsin x ') municipalities 1961:520651.- Chicago, IL Illinois municipalities, -1981219994..- Chicago, IL, & Cook County Missbit riL„-. ;t5 ,. 3 v - 1 . �; ,. Wisconsin municipalities, 1990-2003 Wisconsin - municipalities 2990-2003 c' ' 'tr U.S. municipalities and counties 4' ResearchiQuestion ; . TIF Measure J c StatisticalrTechnique+ Semllog econometric `model wdh fixed + "' effects regression f} Heckman selection x, model to correct for,7 'selection 610,..c5:1,'" 3found i+ �' " Three -stage least squares and ordinary least squares regression Regression . k +,�' y "x� St * A 'ra Hedonic model with two-stage regression approach to address selection bias. 1.) Probit regression 2.) Linear regression [FIxedTeffect r +Y regression fixed yvtxs -effect estimatrony k++s 'a6 ty b, No real statistical analysis performed Four binomial' , ° logistic regression„'=ci ten ast ordinary -IPIR'.. -„square regression s + analyses .` ` b� 'Y i'a< `x° i is _ .. � Regression ;Regression } ri Regression 'Major/Finding?` Theterowsion of public services - loerewithin Tl Fkdistnctxs is capitlizediintOjbusinesS'ipropelity valuover timePdsdrve and es statrshcally signdicant relationships y;� between placement of &parcel.! ix within a TlF districtsand its assessed.' uF. i value , x. l..�k. .._._�.:��.... Urban school districts outside Chicago were most negatively affected by TIE Rural school districts were positively affected. Little effect on Chicago metropolitan area and other suburban areas upstate. T1Fuse is doselVairiked with n' - '~•i annexation a new TIFa leads to rincrease on average Over the studied -j timeperiod ,+TIF is'res onsible.for y i= morethan half of the annexed Land 't,z 1.) Properties inside TIF appreciate at a higher average rate than those outside. 2.) influence of TIF is dependent on the economic state of the neighborhood as compared to others without TIE fIndustnalkTIFs have a,posilve effecte on ;employ ment retail TIFs haveia negative effect onemployment ba 4 rr �ws 5 {(• ,+ Cannot be sufficiently addressed by simply looking at property values and money spent, more sophisticated statistical research techniques must be applied. 1) Adoption of TIF made neighboring ;;, city 2.5 hmesm aore likely tto adopt one 2) Some evidence found that+TlF -.� adoption patterns contribute to inter municipal inequality ° ” , i3 YEarIy adopters of TIF adopted 3'.62tik time more TIFs and 3.81 times more iik retail iIIFs ei f . ,Essuv, ., ..... .. The addition of a TIF district will not affect the tax rates of those within the municipality, but will increase the rates of those just outside the municipality, within the jurisdiction. Propertyvvalue grew more rapidlyin cihes/villages with TIF than those'. without Nan TIF areas of cihes/towns•.' wdh TIF grew jslightfyi more slowly, than the area in the TIFs but more rapidly than cihes/towns without TI Fs'' - Areas near state borders, of low income, or with troubled political cultures are most likely to provide incentive in general. Those with very poor residents or with troubled political cultures are less likely to offer TIFs specifically. 12.008 t 2008 :2009j 2009 20101 2010 2010 4 ` 1 •' i . _... 2010 2011 2013 Carroll,:kbeboreh Weber, Rachel, Rebecca, Hendrick, & Jeremy Thompson Skidmore Mark i David Merriman & Russ, Kashiani'i Smith, Brent C. Bryne PaulF Farris, Sherri & John Horbas Masori .Susan & Kenneth P !:)Ma1' t . S oa+5 - r : tci ..� .,.. Skidmore, Mark & Russ Kashian Merriman David Mark Skimore & Russ Kashian y Felix, R. Alison & James R. Hines What effect doesiTIF have on business property value over hme?i ,-` How does TIF affect the property tax rates and revenues for school districts? Does iTIF..encourage ` annexation"+k §hr �"n� rt 'r 1.) Do properties located inside a TIF district have a higher rate of appreciation than those outside? 2.) Is the rate of change in prices is higher once area is designated as TIF? Do TIFF districts i&creases employment in., ;i municipalities „-i„ Does TIF cause growth? +1') Do cities use TIF to , compete with other cities :- for inv'estment7 5 Does the pattern of TIF use in MOvameliorateeia 0 exaccerbateunequaldy between imffnicipalities k 3) IsTiHuse path,1/2 dependenTiin MOz..... What is the long term relationship between the use of TIF and property taxation in the municipality and overlying jurisdiction? Has TIF increased the - total property value in Wiscon m muff leipalities?', i �. ". - ' Why do cities and counties offer the tax- based incentives they do? Dichotomous vanablest for within'TIF young k TIFs (created aft jerk.,3 h 1989) and stataffilyijk. changed, Tl?s (created `tv after 1995) TIF age R+ e' TIF intensity: The proportion of the school district's tax base tied up in 1 or more TIF districts TIF NC/enable-ad' running} ally on districts created J since J990 t TIF as a dummy variable T!Fas a dummy rs xs va moble in general and i per classrfication Mall mdn5lrlal 'honsing 3 4 B11: --iii ixed ase`other� Looks at TIF implementation and value 1IF amount; value " and type used as w x ; dependent variables ,Proximity to other mi.' cities TIFused as an iii.J. independent Variable ) ffi sG r i'° .. Number of districts created TIF as an.mdependent. variable value of all real estate within 4 TIF districts id rnunicipality # of TIF cl istricts:per 10 000 u TIF and other tax incentives as dummy variables 11