jurisdiction size and local government policy … · jurisdiction size and local government policy...

American Political Science Review of 20 November 2016

doi101017S0003055416000320 ccopy American Political Science Association 2016

Jurisdiction Size and Local Government Policy ExpenditureAssessing the Effect of Municipal AmalgamationJENS BLOM-HANSEN Aarhus UniversityKURT HOULBERG KORA Danish Institute for Local and Regional Government ResearchSOslashREN SERRITZLEW Aarhus UniversityDANIEL TREISMAN University of California

Across the developed world the last 50 years have seen a dramatic wave of municipal mergers oftenmotivated by a quest for economies of scale Re-examining the theoretical arguments invokedto justify these reforms we find that in fact there is no compelling reason to expect them to

yield net gains Potential savings in for example administrative costs are likely to be offset by oppositeeffects for other domains Past attempts at empirical assessment have been bedeviled by endogeneitymdashwhich municipalities amalgamate is typically nonrandommdashcreating a danger of bias We exploit theparticular characteristics of a recent Danish reform to provide more credible difference-in-differencesestimates of the effect of mergers The result turns out to be null cost savings in some areas were offsetby deterioration in others while for most public services jurisdiction size did not matter at all Givensignificant transition costs the finding raises questions about the rationale behind a global movementthat has already restructured local government on almost all continents

INTRODUCTION

O ver the last 50 years a wave of municipalmergers has swept the developed world FromScandinavia to New Zealand reforms have re-

drawn the map of local government combining smallunits to form larger ones Reformers have had sev-eral objectives including reinforcing democracy andbuilding local government capacity (Baldersheim andRose 2010b 242ndash5) But the main motivation has beeneconomicmdashto reduce costs by capturing economiesof scale Among the industrialized democracies thistrend has affected all types of regimesmdashfrom decen-tralized federations to unitary statesmdashand countries ofall sizesmdashfrom Luxembourg to the United States

For such a widespread phenomenon municipalamalgamation has undergone surprisingly little system-atic evaluation In part this reflects the difficulty ofdisentangling effects given endogeneity in the processIn most cases the choice of which local governmentsare merged is not random sometimes central politi-cians decide sometimes leaders of the municipalitiesthemselves Either way this may cause the mergedunits to differ from the unmerged ones complicatingthe evaluation

Jens Blom-Hansen is Professor Department of Political ScienceAarhus University Bartholins Alle 7 8000 Aarhus C Denmark(jbhpsaudk)

Kurt Houlberg is Director of Research KORA Danish Institutefor Local and Regional Government Research Koslashbmagergade 221150 Copenhagen K Denmark (kuhokoradk)

Soslashren Serritzlew is Professor Department of Political ScienceAarhus University Bartholins Alle 7 8000 Aarhus C Denmark(sorenpsaudk)

Daniel Treisman is Professor Department of Political ScienceUniversity of California 4289 Bunche Hall Los Angeles CA 90095-1472 USA (treismanpolisciuclaedu)

The authors would like to thank Peter Bjerre Mortensen NicolaiKristensen Poul Erik Mouritzen Larry Rose and five anonymousreferees for comments on various drafts The study was partly fundedby a grant from The Danish Council for Independent Research

The enthusiasm for enlarging local districts is sur-prising given the weakness and conditionality of thetheoretical rationale Economies of scale are only oneof the likely consequences of increased jurisdiction sizeSuch benefits may be offset by the loss of effects thatfavor small unitsmdashgreater ease of local monitoringmore effective accountability mechanisms or greaterTiebout-style competition for mobile voters and capi-tal At the same time the savings from economies ofscale will depend on the initial and postamalgamationsizes of the units and will also vary across the typesof public services supplied which have different costfunctions The net benefits are likely to be indeter-minate

In this article we examine the consequences for thecost of providing public services of an amalgamation re-form that occurred in Denmark in 2007 In this reform239 municipalitiesmdashessentially all those with popula-tions under 20000 peoplemdashwere combined to form 66new units An additional 32 municipalities were leftuntouched (Mouritzen 2010)

For several reasons the Danish reform is particularlywell-suited to test the effects of increasing jurisdictionsize First the universal nature of the change effectivelyruled out selection all municipalities below a certainsize were required to merge with others and 98 per-cent complied Second the 32 municipalities that wereleft untouched (and which had populations similar tothose of the 66 new units) constitute a control groupfor comparisons Third the governments in questionmatter Danish municipalities play important roles inmanaging schools child care infrastructure environ-mental regulation social spending and culture FinallyDenmarkrsquos official statistics are accurate and detailedwith broad coverage of local unit characteristics

A previous article examined the effect of thisreform on administrative costsmdashmostly wages ofmunicipal employees and maintenance of adminis-trative buildingsmdashand found that these fell after

1httpdxdoiorg101017S0003055416000320Downloaded from httpwwwcambridgeorgcore UCLA Library on 05 Dec 2016 at 183635 subject to the Cambridge Core terms of use available at httpwwwcambridgeorgcoreterms

Jurisdiction Size and Local Government Policy Expenditure November 2016

consolidation (Blom-Hansen Houlberg and Serrit-zlew 2014) That might seem at first to vindicate theenthusiasm for mergers However administrative costsamount to less than 10 percent of total municipalspending We focus here on the other 90 percent andask Do municipal mergers decrease the costs of pro-vision of public services such as schools roads andinfrastructure

We find no clear and systematic effects from amal-gamations We replicate the finding of Blom-HansenHoulberg and Serritzlew (2014) that administrativecosts declined We find also that spending on roadmaintenance per kilometer of road fell in the mergedunits although we cannot say whether this representsgreater efficiency or skimping on repairs Howeverthe economies of scale in administration and (possi-bly) road maintenance were offset by diseconomies ofscale for labor market programs In most policy areasmdashincluding elder care schools daycare and caring forchildren with special needsmdashjurisdiction size did notseem to matter at all Aggregating the effects the netimpact was null If the pattern in Denmark holds moregenerally the global amalgamation wave is unlikely toyield the savings its proponents anticipate We interpretour null finding as supporting the position of skepticswho contend on theoretical grounds that the quest foran optimal jurisdiction size is futile (Dahl and Tufte1973 Treisman 2007)

The article is organized as follows The next sectionprovides background on the global wave of municipalamalgamations of recent decades The third sectiondiscusses theoretical arguments about the effects ofjurisdiction size The fourth section outlines the Dan-ish reform The fifth section describes the data andmethods used in the analysis The sixth section presentsresults and the final section concludes

THE GLOBAL MERGER WAVE

Since the 1950s reforms to enlarge jurisdictions havetransformed the structure of local government acrossthe developed world As societies modernized and builtmore extensive welfare states the local governmentunits inherited from earlier periods were often thoughttoo small to capture economies of scale in service provi-sion (Baldersheim and Rose 2010a 2010b 242 Fox andGurley 2006 8 Keating 1995 118 Newton 1982 191Vetter and Kersting 2003 19)1 Almost everywhereprojects to merge municipalities were debatedmdashandin most cases adopted

These reforms spanned the globe Table 1 briefly re-views the main cases the dramatic scope of which mayhave escaped nonspecialists

From such a survey the extent of the phenomenonbecomes obvious municipal merger mania has sweptthe developed world Reforms have varied in their rad-

1 In Australia for instance reformers argued ldquothat lsquobigger is cheaperrsquodue inter alia to the existence of substantial economies of scalerdquo(Dollery Byrnes and Crase 2008) Similarly in Eastern Canada re-formers in the 1990s repeatedly emphasized anticipated cost savings(Sancton 1996)

icalism in some nations eg the UK the local govern-ment system has been comprehensively restructured inothers eg France the changes have been more lim-ited Countries startedmdashand endedmdashat quite differentpoints While in Mexico Ireland New Zealand Den-mark and Japan the average municipal populationis now more than 40000 residents in France TurkeySwitzerland Austria and Iceland it is still below 5000(OECD 2010 207) Even where mergers were notrapidly implemented demands for them dominatedthe intellectual agenda This is all the more intriguinggiven an opposite tendency among many developingand postcommunist countries where democratizationhas often prompted the division of administrative unitsinto ever smaller pieces (Swianiewicz 2010) In Sub-Saharan Africa for instance 29 countries saw the num-ber of administrative units grow by at least 20 percentbetween 1990 and 2012 Brazilrsquos roster of municipal-ities also increased by 50 percent after the transitionfrom military rule and there were major increases inIndonesia and Vietnam (Grossman and Lewis 2014196)

LOCAL JURISDICTION SIZE THEORY ANDEMPIRICAL SURVEYS

The optimal scale of local government jurisdictionsmdashor of government jurisdictions in generalmdashhas beendebated since the time of Plato Although the searchfor an ideal size that can be identified on theoreticalgrounds independent of context has consumed enor-mous intellectual energy over the years we believethat for several reasons it is a vain quest We brieflyreview the main arguments and explain why they failto yield general implications We suggest that withoutknowing the particular mix of tasks assigned to localgovernments and their technologies it is impossible topredict whether on balance enlarging municipalitieswill have positive or negative effects

Most scholars have conceptualized the optimal scaleof local government as a tradeoff between certain ef-fects that favor large size and others that favor smallerunits (Dahl and Tufte 1973 Hooghe and Marks 2009Treisman 2007) Oates (1972) in a famous analysissaw the main conflict as that between the more pre-cise matching of services to local tastes that is possiblewhen jurisdictions are small and the economies of scaleattainable when they are large

Since economies of scale are the most commonlycited advantage of large sizemdashand the dominant ar-gument for amalgamationsmdashwe discuss them in somedetail In both the private and the public sector returnsto scale are thought to increase for two main reasons(Boyne 1995 Hirsch 1959 Sawyer 1991 47ndash70) Firstthere are fixed costs associated with providing variouskinds of public service so the marginal cost will fallwith output at least up to a certain point Some publicgoods have elements of nonrivalry in consumption sothe marginal cost is zero (Bergstrom and Goodman1973 Borcherding and Deacon 1972) For instance dis-ease surveillance water quality control and restaurant


American Political Science Review

TABLE 1 Local Government Amalgamations in Developed Countries since 1950

Country Time Result References

Sweden 1952 1969 Massive amalgamation Lidstrom 2010Norway 1960s Massive amalgamation Baldersheim and Rose 2010cDenmark 1970 2007 Massive amalgamation Mouritzen 2010Finland 2006ndash2011 From 431 to 336 municipalities Sandberg 2010 OECD 2014a

271Iceland 2006 From 204 to 79 local units Sverrisson and Hannesson 2014UK 1960s and 1970s Massive amalgamation John 2010 Boyne 1998 15ndash61Ireland 2014 From 114 to 31 local authorities Forde 2005 Loughlin 2011

Cheney 2014West Germany 1960s and 1970s From 24000 to 8000

municipalitiesWalter-Rogg 2010

Former East Germany Since 1990 Elimination of 50 percent of localunits

Walter-Rogg 2010 Wollmann2003 OECD 2014a 272

Austria 1960s From 4000 to 2700 local units Pleschberger 2003 Fallend2011

Switzerland Since 1996 From 3000 to 2600 communes OECD 2014a 277 Kubler andLadner 2003 Ladner 2011

Belgium 1970s Elimination of 75 percent ofmunicipalities

OECD 2014a 271 Wayenberget al 2011

Netherlands Since 1950 Elimination of 50 percent of localunits

Boedeltje and Denters 2010Derksen 1988 OECD 2014a266

Luxembourg 2009ndash2017 Program to cut almost 40 percentof municipalities

OECD 2014a 271

France 1970s From 37000 to 36000communes

Kerrouche 2010

Spain 1977ndash2007 From 8800 to 8111 local units Dafflon 2013 191 Alba andNavarro 2003 Colino and DelPino 2011

Italy - No significant reduction Brunazzo 2010 Piattoni andBrunazzo 2011

Greece Since 1990s Massive amalgamation Hlepas 2003 Hlepas andGetimis 2011 OECD 2014a271ndash2

Turkey 2008 From 3225 to 2950municipalities plansannounced to reduce to 1395

OECD 2014a 271

Lithuania 1990s Elimination of 75 percent of localunits

OECD 2014a 271

Latvia 1990s Elimination of 75 percent of localunits

OECD 2014a 271

Estonia - Plans to reduce 226 units to lessthan 50 (not yet implemented)

OECD 2014a 272

Canada Since 1960s Amalgamations (scale variesacross provinces)

Bish 2001 Sancton 2000 Slackand Bird 2013

USA Since 1930s Elimination of 123 multipurposemunicipalities in Kansas andNebraska since 2007 Between1930 and 1970 100000 schooldistricts eliminated Howeverother types of special districtsintroduced

OECD 2014b 78ndash9 Berry 200926ndash50 Foster 1997 1ndash28Berry and West 2010 Strang1987

Australia Since 1970s From 900 to 600 local councils Dollery et al 2008 Byrnes andDollery 2002

New Zealand 1980s From 200 to 74 city and districtcouncils

Boston et al 1996 183ndash202Dollery and Wallis 2001196ndash220

Japan 1953 1999 From 3232 to 1719 local units OECD 2014a 271South Korea 1990s Wave of amalgamations OECD 2005 141



inspections may not cost more to provide for multipleresidents than for just one (Santerre 2009) Secondincreasing the scale of service provision makes possi-ble a more fine-grained division of labor yielding theassociated benefits of specialization

However above a certain level such benefits oflarger size are offset by problems of communicationand control As output grows so does the need to trans-mit information through more layers of managementLarge production processes often suffer from bureau-cratic congestion (Williamson 1967) Consequentlyproduction processes normally exhibit first increasingthen constant and finally decreasing returns to scalethe typical cost curve is U-shaped It follows that thereis an optimal sizemdashat the bottom of the U-shapedcurvemdashat which unit costs are lowest Advocates ofmunicipal amalgamation usually suppose that this op-timum occurs at a relatively high local population

Influential as this approach has been it does not infact yield any clear implication about the optimal size ofmunicipalities There are two key problems First mostlocal authorities provide a range of services each withunique production characteristics Economies of scaleare specific to the particular technologies and goodsor services produced Thus there is not one optimalsize but many one for each of the services providedOf course if all municipal services had minimum costpoints at high population levels then amalgamatingsmall units might improve things on average But infact the technologies for different common local ser-vices differ a great deal (Bish 2001) To produce all atoptimal scale one would need to replace municipali-ties with multiple overlapping single-purpose unitsmdashwhich besides being highly complex would itself leadto redundancy of administrative personnel (Ostrom1972) For municipalities that provide multiple servicesthe efficiency consequences of amalgamation will de-pend on the initial and final size of their jurisdictionsand on the particular portfolio of tasks assigned tothem and their associated production technologies Ef-ficiency might either increase or decrease and a greatdeal of information is needed to predict which it willbe in a particular case

The second problem is even more fundamental Mostdebates relate the size of municipal districts to the coststructure for provision of particular servicesmdashfor ex-ample primary education But it is not municipal gov-ernments that educate children it is schools that do soThe most relevant cost effects relate to the size of theschool not that of the school district The same is trueof child care centers libraries and residential homesfor the elderlymdashin each case smaller organizations arethe direct providers of services and it is primarily thescale of these smaller organizations that determines ef-ficiency The distinction parallels that in the private sec-tor between plant-level and firm-level returns to scale(Boyne 1995 220 Sawyer 1991 50ndash1 Scherer and Ross1990) Any scale economies at the level of direct serviceproviders such as schools and child care centersmdashandthese seem to be meager at best according to a reviewof the empirical literature by Walker and Andrews(2015 111ndash2)mdashcan be harvested without altering lo-

cal government jurisdictions since one can resize theorganizations and their service areas withinmdashand evenacrossmdashexisting municipal boundaries For a subset oflocal government functions the costs of which occur atthe firm level (most notably administration) increasingjurisdiction size may confer economies of scale (seeBlom-Hansen Houlberg and Serritzlew 2014) Butsince enlarging municipal districts does not in itselfaffect the size of individual schools hospitals or otherplant-level organizations amalgamation will not affectplant-level efficiency at all

In short even setting aside Oatesrsquo (1972) argumentthat scale economies are offset by less precise matchingof services to local tastes the existence of economies ofscale does not imply any direct and universal prescrip-tions for the design of local government systems exceptperhaps in the case of certain single-purpose serviceproviders For municipalitiesmdashor other multipurposeentitiesmdashthere is simply no good reason to expect thatlarger size will generally lead to cost savings

A second argument in favor of amalgamations isthat larger jurisdictions may be able to capture notjust economies of scale but also economies of scopeIt may be more efficient to produce certain relatedservicesmdashsay sewerage and recycling of water cfDollery and Fleming (2006)mdashjointly than to producethem separately This does not in itself dictate largerjurisdictionsmdashit concerns the range of services pro-duced not the scale of productionmdashbut if some of theservices have a minimum efficient scale then achiev-ing the bundle of economies could require increas-ing government size In fact the relationship betweeneconomies of scale and scope is far from clear Theymay complement each other or conflict But they mayalso be unrelated (Dollery and Fleming 2006) Giventhis we should not expect increased size to lead to costreductions for this reason either

A third effect traditionally seen to favor larger sizeconcerns externalitiesmdashthe imposition by one indi-vidual of costs or benefits on others that are notcompensated via the market Allocative efficiency isincreased when government regulates taxes or sub-sidizes activities so that individuals internalize sucheffects However if the externalities affect mostly indi-viduals outside the given governmentrsquos jurisdictionmdashwhich is more likely to be the case when jurisdictionsare smallmdashthe governmentrsquos incentive to address themis weaker When units are larger local governmentswill be motivated and able to tackle more of the pre-vailing externalities A similar problem affects not actsof individuals but government policies If the positiveeffects of a local governmentrsquos policies spill over intothe neighboring jurisdictions rather than accruing tothe citizens that the given government represents thegovernment will undersupply this policy

The only way to eliminate all such cross-borderinfluences would be to expand jurisdictions withoutlimit not just enlarging local governments but merg-ing them into the central government Of coursesuch a ldquosolutionrdquo would forego all benefits of smallersize A more sensible approach is to assign serviceresponsibilities to tiers of government in a way that



balances the benefits of small size against the cost ofexternalities The optimal balance will be specific toparticular services As pointed out by Olson (1986) andTullock (1969) among others different public servicesproduce different externalities Consequently any at-tempt to address externalitiesmdashlike attempts to cap-ture scale economiesmdashwill involve tradeoffs

Thus on close examination the arguments that favorlarge municipal jurisdictions will only hold in particularcontexts At the same time other effects could rendersmaller jurisdictions more efficient (Boyne 2003 370ndash2) Various scholars argue that citizens will monitorgovernment more actively in smaller communities re-sulting in greater bureaucratic effort and less waste(Dahl and Tufte 1973 Denters Goldsmith LadnerMouritzen and Rose 2014) If yardstick competitionis part of the system for evaluating local governmentsthis may work best when there are more competingunits (Allers 2012) although some studies have failedto find empirical confirmation for this (Boyne 2003382) Meanwhile if the costs of moving to another ju-risdiction increase with distance Tiebout-style (1956)competition among local governments to attract resi-dents or mobile capital through government efficiencyand responsiveness will be stronger when units aresmaller Competition among a large number of smalljurisdictions may also serve to constrain them fiscallyforcing them to supply services efficiently (Brennanand Buchanan 1980 168ndash86) Finally Oatesrsquo argumentthat smaller jurisdictions enable governments to moreprecisely tailor public services to local tastes has foundechoes in subsequent analyses (Alesina and Spolaore2003 Oates 1972)

Just as with the arguments for large scale the logicbehind these various effects is not always as clear asit might seem (Treisman 2007) But even ignoring thisit is clear that the advantages of large and small sizewill aggregate and offset each other in context-specificways Rather than a presumption that amalgamationwill generally increase efficiency we hypothesize thatamalgamation should have no general effects it willincrease efficiency in some contexts and decrease it inothers (Fox and Gurley 2006 Treisman 2007 53ndash73)In short the most plausible hypothesis is a null one2

If the theoretical literature in public finance and po-litical science provides no compelling general reasonto expect efficiency gains from municipal mergers doesthe empirical literature detect such gains in practiceNumerous studies have sought to estimate the costfunctions for local services A number of articles havesurveyed their results (Bish 2001 Boyne 1995 Byrnesand Dollery 2002 Derksen 1988 Fox and Gurley 2006Holzer et al 2009 Martins 1995 Ostrom 1972) Themain conclusion from these reviews is that there is noconsistent evidence on economies of scale in local gov-ernment Some studies detect a tendency for very smallmunicipalities to be inefficient (eg Breunig and Ro-caboy 2008 Sole-Olle and Bosch 2005) and some havefound administrative efficiency gains from larger size

2 In addition to the question of optimal scale the costs of transitionfrom one size to another may be significant

(Blom-Hansen Houlberg and Serritzlew 2014) butthe general finding is that the evidence is inconclusiveMost studies report that optimal scale varies across dif-ferent servicesmdashwhile a few such as water and sewagehave considerable economies of scale others such asschools may exhaust such economies at populationsunder 10000 (eg Fox and Gurley 2006)

To explicate the findings of these review studies inmore detail we look more closely at those of two ofthe most recent and comprehensive ones The firstis Byrnes and Dollery (2002) who review 24 inter-national studies and eight Australian ones They findthat among the international studies 29 percent findevidence of U-shaped cost curves 39 per cent find nostatistical relationship between per capita expenditureand size 8 percent find evidence of economies of scaleand 24 percent find diseconomies of scale The eightAustralian studies they survey also reach mixed find-ings On this basis Byrnes and Dollery (2002 405)conclude that ldquoconsiderable uncertainty exists as towhether economies of scale do or do not existrdquo

The second review study is Holzer et al (2009) whoexamine 65 studies from a broad range of countriesThey find that there is little evidence for a relationshipbetween size and efficiency for municipalities with pop-ulations between 25000 and 250000 Among munici-palities with populations under 25000 they find somesuggestions that efficiency increases with size but onlyin certain contexts At the same time they note thatmuch of the literature argues that small municipalitiesare not less efficient except in specialized services Onthis basis they conclude that ldquo[t]he literature provideslittle support for the size and efficiency relationshipand therefore little support for the action of consol-idation except as warranted on a case-by-case basisrdquo(Holzer et al 2009 1)

In sum the empirical literature on the effects ofmunicipal mergers has failed to identify systematicpatterns that hold across time and space From ourvantage point this state of affairs is unsurprising Sincethe advantages of large and small size depend on con-text and since plant-level and firm-level scale effectsare at best weakly related the absence of systematicconsequences of jurisdiction size is what one shouldexpect Our re-examination of the theoretical argu-ments suggests why empirical researchers have comeup empty-handed

Another lesson from the existing studies is that it isdifficult to study scale effects Even a strong correlationbetween size and costs must be treated with cautionwhen studies are based on observational data (Boyne2003 388) A problem with observational studies isthat the size of jurisdictions is nonrandom Their scaleis determined by a variety of factors that also affectthe cost of public services Regional subcultures andlocal political histories will influence both jurisdictionsize and also levels of corruption and bureaucraticefficiency When large cities are poorly run districtssometimes secede to form smaller autonomous munic-ipalities (Anderson 2012) At the same time centralreformers eager to see a successful outcome to theirreform may choose to amalgamate municipalities that



are already for other reasons more efficient leadingto an association between size and performance

A solution to this endogeneity problem is the experi-mental approach (Walker and Andrews 2015 126) Weuse a recent Danish municipal reform which we intro-duce in greater detail in the next section to addressthis problem As will become clear we find evidenceconsistent with our hypothesis that no general relation-ship exists between jurisdiction size and public servicespending Even after accounting for endogeneity farmore precisely than is usually possible the finding ismdashas expectedmdashnull

THE DANISH MUNICIPAL REFORM

On January 1 2007 a major reform of Danish localgovernment changed the size of most of the countryrsquosmunicipalities3 Denmark a small unitary state with alarge welfare state (see Arter 2012) has three levelsof government Before the reform the lowest levelconsisted of 271 municipalities From 2007 large scalemergers left just 98 municipalities with an average pop-ulation of 57000 inhabitants4

Each municipality is governed by a city councilelected every four years with day-to-day administra-tion left to standing committees under the city counciland to the mayor who is elected by the city council Themunicipalities provide basic welfare services distributevarious social transfers and administer aspects of utili-ties culture and recreation In our analysis we focus oneight major policy areas schools daycare elder carechildren with special needs roads culture administra-tion and labor markets In Lowirsquos (1972) terms all ofthese involve distributive policies

Municipal spending accounts for more than half of allpublic expenditure in Denmark The local governmentsfund their activities from various income sources themost important of which is the local income tax Thistax finances about half of all municipal spending withthe remainder coming from user charges and centralgovernment grants The average local income tax ratewas 249 percent of citizensrsquo personal income in 2014In principle the municipalities are free to decide theirown income tax rate but in practice the central gov-ernment has imposed a number of controls over localtaxation Nevertheless compared to other countriesDanish municipalities still enjoy considerable auton-omy (Blom-Hansen and Heeager 2011)

The 2007 reform was quick and radical Before 2002municipal restructuring had not made it onto the Dan-ish political agenda When the idea of a centrally im-posed reform was floated in a parliamentary commit-tee discussion the government firmly rejected it Yetin 2004 a government-commissioned report recom-mended amalgamations One year later in the spring

3 The Danish reform is also described in Blom-Hansen Houlbergand Serritzlew (2014) This and the following section build upon thisdescription4 There is also a regional level in Denmark with five regions primarilyresponsible for health care In this article we only focus on the locallevel

of 2005 the national parliament approved a semivolun-tary merger program which had been forced throughwith the backing of a narrow majority (Bundgaardand Vrangbaeligk 2007 Christiansen and Klitgaard 2010Mouritzen 2010)

The reform had two main elements The first was areshuffle of functions across tiers involving income taxassessment services for handicapped rehabilitationhealth promotion primary education for children withspecial needs environmental protection and regionalroads Although this list may sound impressivespending on the new functions amounted to only about8 percent of the municipalitiesrsquo previous budgets Thereallocation of functions did not involve the traditionalmunicipal core tasks related to welfare and publicutilities

While the reshuffle of functions included allmunicipalities the second elementmdashthe municipalamalgamationsmdashdid not This part of the reform left 32municipalities that were already above the size thresh-old intact but required the other 239 to merge into66 new larger entities The reform stipulated that mu-nicipalities with fewer than 20000 citizens were to becombined with neighbors to form new units that shouldaim for the target size of about 30000 citizens The onlyway that municipalities with fewer than 20000 inhab-itants could avoid amalgamation was by concluding acooperative arrangement on service provision with alarge neighboring municipality This proved very dif-ficult in practice and only five of the 239 units tookthis path Three small municipalitiesmdashFarum Holms-land and Hvorslevmdashfailed to make arrangements forthemselves and were subjected to intervention by thecentral government which then organized their amal-gamations

METHODS AND DATA

We use the 2007 Danish municipal amalgamation re-form as a source of exogenous variation in jurisdictionsize to address the problem of endogeneity We treatthe case as a quasi-experiment A quasi-experimentshares many features with other types of experiment(Cook and Campbell 1979 56 Dunning 2012 15ndash21)It has at least in the ideal situation experimental andcontrol groups as well as pre- and post-treatment mea-sures of relevant variables In this case the ldquocontrolgrouprdquo consists of the 32 municipalities that were al-ready above the size threshold and so did not un-dergo amalgamation Their jurisdictions experiencedonly negligible demographic changes The ldquotreatmentgrouprdquo consists of the 66 municipalities formed by theexogenously decreed amalgamation of smaller units

In contrast to other experiments assignment to ex-perimental and control groups is not randomized inquasi-experiments This raises the possibility that dif-ferences in results might be caused by preexisting dif-ferences between the groups rather than by the ex-perimental intervention so such differences need tobe carefully controlled Still compared to traditionalobservational studies quasi-experiments have the



TABLE 2 Size of Municipalities in Control Group andTreatment Group before and after Reform (percent)

Control Group Treatment Group

Population Size Prereform Postreform Prereform Postreform

Under 5000 9 9 5 25001ndash10000 0 0 47 010001ndash20000 6 6 31 220001ndash30000 28 28 7 1430001ndash50000 31 31 5 4450001ndash100000 16 16 3 35More than 100000 9 9 0 5Total 100 100 100 100N 32 32 239 66

great advantage that the main independent variableis determined by some process that is exogenous to theone under study

Although the impetus for amalgamation in the Dan-ish program was clearly exogenous to the individualmunicipalitiesmdashall small ones were required to un-dergo reformmdashthe precise choice of partner and thusthe exact size of the new merged unit were left to localdecisions The reform gave the local governments sixmonths to settle the amalgamations The key issue forour research design is whether service provision costsplayed any significant role in shaping the individualmunicipalitiesrsquo choices

In fact the evidence clearly suggests that costs ofadministration and services were not very importantto amalgamation patterns Case studies reported inMouritzen (2006) of specific amalgamations demon-strate that other factors such as local identity and lo-cal politiciansrsquo ambitions for office in the future af-fected how municipalities were amalgamated Bhattiand Hansen (2011) show in a quantitative study ofall municipalities that social connections (measuredas commuting patterns) between municipalities had asignificant effect on the chance of amalgamation Allthis increases confidence that considerations of serviceprovision costs played little role in the outcomes Wetherefore proceed on the assumption that service pro-vision costs were exogenous to the amalgamations

In Table 2 we compare the growth in size foramalgamated (treated) and nonamalgamated (control)municipalities The size of the nonamalgamated mu-nicipalities in the control group changed little butin the amalgamated municipalities the changes weredramatic

The reform took effect in 2007 Our data span 2003ndash2014 ie four years before the reform and eight yearsafter To allow for pre- and postreform comparisonwe impose the postreform structure on the prereformstructure by aggregating prereform municipalities thatwould eventually be amalgamated to their postreformsize5 The municipalities of Koslashbenhavn Frederiksberg

5 A few municipalities were split among two or more new municipali-ties In these cases we divided the expenditure of the old municipality

and Bornholm had prereform status as both county andmunicipality and were therefore excluded This leavesus with 1140 observations (95 municipalities over 12years) Of these 95 municipalities 29 did not experiencea change in borders (the control group) and 66 resultedfrom mergers (the treatment group)6

Hence we have 116 prereform and 232 postreformobservations for the control group (29 units over fourand eight years respectively) and 264 prereform and528 postreform observations for the treatment group(66 municipalities over four and eight years respec-tively) Studying changes in service costs for the treat-ment group alone would confound the effect of changesin size with the general trend in service costs overtime Following Blom-Hansen Houlberg and Serrit-zlew (2014) we use the difference-in-difference (DiD)approach to isolate the causal effect of size comparingdata for the treatment group and the control group

The logic is this The difference in service costs forthe treatment group before and after the reform isan estimate of the combined effect of changes in sizeand time The difference in service costs for the controlgroup before and after the reform is an estimate ofthe effect of time but not of changes in size The dif-ference between these two differences constitutes theDiD estimator which estimates the average effect ofthe changes in size on service costs for the treated units(or the average treatment effect for the treated ATT)The DiD-estimator can be obtained from the followingregression analysis

Yi = α + β1TGi + β2Ti + β3TGi times Ti + εi (1)

where Yi is a measure of service costs for municipality iTGi is a dummy variable taking the value 1 if municipal-ity i belongs to the treatment group (0 otherwise) Ti isa dummy variable taking the value 1 if the observationis measured post reform (0 otherwise) and TGi times Ti

among the new ones in the same proportion as the division of theold municipalityrsquos population6 Including AEligroslashskoslashbing and Marstal which were amalgamated intoAEligroslash effective January 1 2006



is an interaction term It can easily be shown that β3 isthe DiD estimator (see Wooldridge 2009 or Lassen andSerritzlew 2011 or Blom-Hansen Houlberg and Ser-ritzlew 2014 for a similar application) Furthermore β1is an estimate of the differences between the treatmentand control groups before the reform If municipalitieswere assigned randomly (which of course they arenot) this should be close to zero β2 is an estimate ofthe general trend in service costs over time This maybe positive or negative depending on factors such asthe development in available technology changes inprices and wages or changes in service provision

Equation (1) operates with only two periods onepre- and one postreform However reforms have aninherent temporal component Reaction to shocks canbe slow (OrsquoToole and Meier 1999 514) and there maybe a delay between the time at which a change is im-plemented and that at which employees and organiza-tions perform differently (Oberfield 2014) To see howeffects develop over time we expand (1) with dummyvariables T2003i minus T2014i and corresponding interactionterms to estimate changes in service costs over timefor the span of data available We also include a set ofcontrol variables that capture changes in factors rele-vant to service costs (other than size) that may changedifferently for the control and the treatment group

Our dependent variable is a number of differentspecifications of spending per capita As noted byHolzer et al (2009 19) and Boyne (1995 219ndash20)this measure is used throughout the literature Andseen from the taxpayerrsquos perspective it is probably themost relevant concept to focus on But it should betreated with caution It does not measure effectivenessor efficiency (cf Boyne 2002 17ndash8) No valid generalindicators of service quality or effects on formal policyobjectives are available and accordingly our analysiscannot estimate size effects on quality or effectivenessFurthermore spending per capita does not measureefficiency since population is a poor proxy for ser-vice outputs (Boyne 1995 219) However to facilitatecomparison with previous literature we use spending-per-capita measures in our main analysis but we alsopresent a robustness analysis that breaks down spend-ing per capita into its two components quantity ofoutput and unit costs The latter is closer to measuringefficiency

To be more precise the dependent variable is netcurrent expenditure per user in eight policy areasmeasured in DKK in 2014 prices These eight policyareas include all major services that the municipalitiesprovided both before and after the 2007 reform Newfunctions transferred to the municipalities as part of thereform as well as some minor functions are excluded7

7 We exclude new functions (most notably care for disabled adultswhich accounts for 25 billion DKK out of a total of 425 billionDKK excluded) because we cannot study how these expenditureschange from before the reform We also exclude functions that areonly relevant to some municipalities (for example about 3 billionDKK spent on collective traffic and harbors) and minor functionsthat are very volatile (for example 1 billion DKK for snow clearingand 6 billion DKK for urban planning and environmental protectionwhich is sensitive to yearly fluctuations due to for instance storm

We include only current expenditure since capital ex-penditure in Denmark is fully accounted in the year ofinvestment (the cash flow principle) We use net expen-diture in order to focus on the expenditures financed bythe municipality itself Hence conditional grants fromthe central government user fees and cross-municipalpayments for services provided to other municipalitiesare subtracted Table 3 presents the eight policy areasin more detail For precise operationalizations pleaserefer to Appendix Table A1 in the online supplemen-tary material

As is evident from Table 3 total expenditures in-cluded in the analysis amounted to 2455 billion DKKin 2014 This constitutes 85 percent of all municipal ex-penditure that year8 Daycare schools elder care andlabor market activities (including income transfers) arethe major expenditure areas while roads culture andchildren with special needs constitute minor expendi-ture areas

Since assignment of municipalities to treatment andcontrol groups is not randomized we include a setof social economic environmental and political con-trol variables (Andrews et al 2005) used in previ-ous policy analyses of Danish municipalities (Blom-Hansen Houlberg and Serritzlew 2014 Serritzlew2005 Oslashkonomi- og Indenrigsministeriet 2012) Firstwe include two indicators for spending needs dis-persed settlements and socioeconomic expenditureneeds Dispersal of settlements is a potentially time-variant structural condition influencing costs Socioe-conomic expenditure needs is an index measure usedin the national equalization scheme for municipalitiesconstructed from a number of objective indicators suchas the number of unemployed the number of childrenof single parents etc We also control for location onan island this is a time-invariant but very importantdeterminant of spending needs Second an indicator offiscal pressure (an estimate of expenditure needs rela-tive to the tax base) controls for variations in economicpotential among the municipalities Finally we con-trol for two political factors that might influence localpolicy Greater political fragmentation as captured bythe effective number of political parties could increasegovernment spending if government resources are seenas common property subject to overuse by fragmenteddecision-makers (Velasco 2000) Meanwhile a higherproportion of socialist seats in the council might pre-dispose the municipality to spend more (Boyne 1996)The precise specifications of the control variables alsoappear in Appendix Table A1 in the online supplemen-tary material

RESULTS

Before turning to the DiD-based regression analyseswe present a first view of the data in Figure 1 which

damage and flooding) or very dependent on context (for instance 1billion DKK related to new refugees)8 Total municipal net current tax financed expenditures in 2014amount to 288 billion DKK (excluding cofinancing of regional healthservices and services for insured unemployed)



TABLE 3 Policy Areas

Policy Area Main Functions

Net CurrentExpenditures2014 in BillionDKK (percent) User Group

Daycare Daycare in private homesKindergartens

253 (103) Children aged0ndash5 years

Schools Public primary and lowersecondary schoolsCompulsory grants topupils in private schools

541 (220) Children andyoung peopleaged 6ndash16years

Elder care Home helpNursing homes andsheltered housing

444 (181) People aged 65+

Children and youngpeople withspecial needs

Preventive activitiesResidential homes forchildren and youngpeople with special socialor functional needs


Roads Maintenance of publicroads

49 (20) All inhabitants

Culture Culture and leisureactivities (includingparks sport centers andgrants for cinemas andtheatres and local clubs)


Administration Administrative personnelcompensation forpoliticians maintenanceof buildings purchasingof administrative utensilsinsurance auditing etc


Labor market Labor market activities andsocial security includingincome transfers likesickness benefits earlyretirement benefits andcash benefits fornoninsured unemployed


Total expendituresincluded

Sum of the eight policyareas


shows the development over time in expenditure peruser in different functional areas for amalgamated andnonamalgamated municipalities The first eight panelsin the figure are the eight expenditure areas while thelast panel shows the sum of all expenditures (per in-habitant) These graphs present the raw data withoutany control for factors other than amalgamations Stillthey illustrate findings that we later confirm

First Figure 1 shows parallel trends for amalgamatedand nonamalgamated municipalities before the reformThis is crucial for the DiD-analyses presented belowThe different groups of units were evolving along simi-lar paths Second if the amalgamations affected spend-ing we should expect to see different trends for amal-gamated and nonamalgamated municipalities after thereform In fact we see no consistent differences For ex-ample in the school area amalgamated municipalitiesspent less per pupil than nonamalgamated ones bothbefore and after the reform But the trends over time

appear to be the same for the two groups Municipali-ties that were merged in 2007 neither converged withmdashnor diverged frommdashthe unmerged units Indeed the2007 reform seems to have left no mark

This makes sense given the distinction we noted be-tween firm level and plant level characteristicsmdashherethe size of the municipality and the size of schoolswithin it Even if larger schools were more efficientamalgamating municipalities would not in itself de-crease spending unless it somehow led to the amalga-mation of schools A similar pattern is found for spend-ing per user on daycare and elder care These policyareas are in many ways comparable to public schoolsin the Danish system Daycare is provided mainly inpublic kindergartens and elderly care in nursing homesand sheltered housing Each municipality has severalof these institutions to serve different geographical ar-eas Amalgamating a municipality does not in itselfincrease the size of the plant level institutions Culture



FIGURE 1 Group Means on Dependent Variables by Year

and total expenditure per inhabitant also follow thispattern

In some areas the time trends for the two groups ofmunicipalities do diverge after 2007 For instance in theroad area amalgamated and non-amalgamated mu-nicipalities had similar expenditure trends until 2007But then a gap appears and the amalgamated munic-ipalities start to spend less than the nonamalgamatedones until 2012 before converging in 2013 but thendiverging again in 2014 Danish municipalities are re-sponsible for the maintenance of local roads and make

decisions about quality levels Some of the work iscarried out by municipal maintenance divisions someis contracted out to private providers (Blom-Hansen2003) The same time pattern is also seen in the areaof administration where no subsequent convergenceoccurs

The opposite patternmdashin which amalgamated mu-nicipalities start to spend more than nonamalgamatedones after 2007mdashis found in two other areas care forchildren with special needs (municipalities are respon-sible for preventive activities such as counseling and



FIGURE 1 Continued

pedagogical support of families at risk as well as forthe forcible removal of children from their homes) andlabor market policy (municipalities distribute incometransfers such as sickness benefits run job centers andadminister eligibility for social benefits)

Based on the graphs it appears that in most func-tional areas the municipal amalgamations had no effecton spending per potential user In other areas mergersseem to have either reduced or increased spending rel-ative to the control group However these conclusionsare preliminary One needs to check that the same re-sults obtain holding constant other factors that mighthave influenced expenditure trends

We therefore now turn to the results of the DiDanalyses Table 4 first compares the average prereformexpenditure levels to the average postreform levels inrespectively the amalgamated and nonamalgamatedmunicipalities This table contains only one prereformand one postreform observation for each municipalityThe estimation method is OLS with clustered stan-dard errors The upper panel in Table 4 includes only adummy indicating units that underwent amalgamationin 2007 (the treatment variable) and a time dummy in-dicating whether observations are made pre- or postre-form According to the DiD logic the reform effect isidentified by the interaction of the treatment variableand the post-reform time measure The variable post-reformlowastamalgamated is therefore our DiD estimator

Since no controls are included in the upper panel inTable 4 it basically reproduces the graphs in Figure 1It confirms that in most areas the amalgamations leftno mark but in some areas they seem to have inducedeither increases or reductions in spending

The lower panel in Table 4 introduces our controlvariables None of them have effects in all analysesbut several are important for understanding expendi-ture developments in individual areasmdashnote the jumpin R-squared in all cases However the DiD estimatorstill indicates that in most areas the amalgamations leftno mark But again in some areas they seem to haveeither increased or reduced spending More preciselyin the areas of children with special needs daycareschools and elder care there is no evidence that theamalgamation reform mattered In the areas of roadsand administration the impression from the graphs inFigure 1 is confirmed Amalgamations seem to have ledto lower spending In the area of labor market services(and to a limited extent culture) the opposite is thecase Summing across all policy areas no amalgama-tion effect is found for total spending Our results thusparallel those of Allers and Geertsema (2014) whoalso failed to find any systematic effects on spending ofmunicipal amalgamations in the Netherlands

Table 5 presents a more detailed analysis WhileTable 4 compared average pre- and postreform ex-penditure levels Table 5 includes all our yearly


JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016

TABLE 4 Two-period Estimates for Eight Policy Areas With and Without Controls

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls

Daycare(per 0ndash5 year

old)

Schools(per 6ndash16 year

old)

Elder care(per 65+ year

old)

Children withspecial needs(per 0ndash22 year

old)Roads

(per capita)Culture

(per capita)Administration

(per capita)

Labor marketactivities

(per capita)

Totalexpenditures(per capita)

Amalgamated(dummy)

minus1293381lowastlowastlowast minus1025651lowastlowastlowast minus310914lowastlowast minus3152 4073 minus71663lowastlowastlowast minus45773lowastlowast 12856 minus346892lowastlowastlowast

(230265) (189567) (129465) (45486) (6218) (15892) (21917) (41575) (87980)DiD estimatorPostreform lowast

amalgamated197234lowast 169870 19437 48853 minus15350lowastlowastlowast 18511lowastlowastlowast minus33850lowast 49950lowastlowastlowast 58350(112587) (103434) (98566) (37319) (5457) (6056) (19300) (14486) (51422)

Time dummyPostreform 337246lowastlowastlowast 49495 minus654286lowastlowastlowast 175799lowastlowastlowast 17885lowastlowastlowast minus30383lowastlowastlowast 53358lowastlowastlowast 189467lowastlowastlowast 265324lowastlowastlowast

(105040) (89947) (86042) (32885) (5129) (5264) (18543) (11811) (47121)Constant 7134281lowastlowastlowast 7969805lowastlowastlowast 5391886lowastlowastlowast 675301lowastlowastlowast 86935lowastlowastlowast 271910lowastlowastlowast 575147lowastlowastlowast 714989lowastlowastlowast 4342236lowastlowastlowast

(213895) (176738) (119695) (39972) (5872) (15147) (20806) (38606) (83400)Observations 190 190 190 190 190 190 190 190 190Adj R2 0388 0275 0319 0174 0024 0250 0104 0293 0289

(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)

Schools (per6ndash16 year

old)

Elder care (per65+ year

old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)

minus177523 minus26326 minus145725 135770lowastlowast 8571 minus7377 14352 11306 47225(183190) (208147) (135438) (51911) (7796) (9946) (27200) (20900) (63433)

DiD estimatorPostreform lowast

amalgamatedminus19224 minus8270 minus14934 52844 minus16101lowastlowastlowast 8344 minus43450lowastlowast 76460lowastlowastlowast 13157

(102302) (115510) (97967) (34155) (5433) (5758) (18158) (18451) (43320)Time dummyPostreform 471743lowastlowastlowast 178281lowast minus574185lowastlowastlowast 158701lowastlowastlowast 21076lowastlowastlowast minus17465lowastlowastlowast 63550lowastlowastlowast 156434lowastlowastlowast 301708lowastlowastlowast

(92352) (105727) (89283) (30797) (5008) (5631) (18134) (15621) (40569)Control variablesSmall Island 937061lowastlowastlowast 1221581lowastlowastlowast minus277030 248156 31989lowastlowast minus6149 196077lowastlowastlowast minus3597 411861lowastlowastlowast

(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of

settlementminus174041lowastlowastlowast minus118968lowastlowastlowast 44900 minus8937 3718lowastlowastlowast minus13252lowastlowastlowast 13155lowastlowast minus5505 minus2154

(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam

bridgeorgcore UCLA Library on 05 D

ec 2016 at 183635 subject to the Cambridge Core term

s of use available at httpww

wcam

bridgeorgcoreterms

Am

ericanPoliticalScience

ReviewTABLE 4 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Fiscal pressure minus91601lowastlowastlowast minus75547lowastlowastlowast minus15854lowast minus5319 minus642 minus4897lowastlowastlowast minus5732lowastlowastlowast 8317lowastlowastlowast minus27484lowastlowastlowast

(11003) (12051) (8237) (3299) (464) (827) (1729) (1347) (3462)Socioec expenditure

needs020 052lowastlowastlowast 053lowastlowastlowast 035lowastlowastlowast 001 007lowastlowastlowast 006lowastlowast 031lowastlowastlowast 063lowastlowastlowast

(015) (016) (014) (005) (001) (002) (002) (003) (005)Party fragmentation 81470 23989 minus83303 55218lowastlowastlowast minus1435 minus837 6278 18643lowast 37819lowast

(63747) (87272) (81135) (20453) (4261) (5671) (12246) (10585) (22461)Share of socialist

seats13568lowastlowastlowast 11478lowastlowast minus4019 1439 minus535lowastlowastlowast minus549lowast minus551 2724lowastlowastlowast 2188(4064) (5007) (5401) (1394) (196) (314) (850) (682) (1819)

Constant 14732392lowastlowastlowast 13665763lowastlowastlowast 6349458lowastlowastlowast 305443 146202lowastlowastlowast 668468lowastlowastlowast 974297lowastlowastlowast minus777181lowastlowastlowast 5564145lowastlowastlowast


Notes Robust standard errors in parentheses (clustered at each municipality)lowastlowastlowast plt001 lowastlowast plt005 lowast plt010

13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016

TABLE 5 Single Year Estimates in Eight Policy Areas SUR Regressions (except model 9 which is an additive of the eight areas)

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)

minus203796lowast minus323686lowastlowast minus109456 114451lowastlowastlowast 7466 minus9759 8417 minus1564 minus10530(122018) (129471) (117335) (42096)dagger (5947) (8652) (16652) (19822) (64076)

DiD estimatorsAmalgamated lowast 2004 8245 141125 minus30229 11879 minus386 minus009 minus1204 minus2514 5469

(164983) (175060) (158651) (56918) (8041) (11698) (22516) (26802) (21578)Amalgamated lowast 2005 minus127783 475329lowastlowastlowast minus122672 35290 minus3652 minus3595 minus2248 15709 38647

(165440) (175546) (159091) (57076) (8063) (11731) (22579) (26877) (28301)Amalgamated lowast 2006 minus104294 382234lowastlowast minus102076 32799 9737 minus1439 minus3791 34320 57409lowast

(165510) (175620) (159158) (57100) (8067) (11736) (22588) (26888) (33543)Amalgamated lowast 2007 minus273088lowast 177656 minus92504 35414 minus3813 minus2433 minus4434 61174lowastlowast 23029

(165660) (175779) (159302) (57152) (8074) (11746) (22609) (26912) (40419)Amalgamated lowast 2008 minus186428 190169 minus163006 60240 minus15718lowast 3568 minus20501 84403lowastlowastlowast 20992

(165626) (175743) (159270) (57140) (8072) (11744) (22604) (26907)daggerdagger (42899)Amalgamated lowast 2009 minus71395 273537 minus203580 93567 minus18801lowastlowast 11625 minus41332lowast 82828lowastlowastlowast 22253

(165559) (175672) (159205) (57117) (8069) (11739) (22595) (26896)daggerdagger (47028)Amalgamated lowast 2010 minus49451 264224 minus62915 75730 minus18329lowastlowast 6624 minus54009lowastlowast 66957lowastlowast 15604

(165360) (175460) (159013) (57049) (8059) (11725) (22568) (26863) (56782)Amalgamated lowast 2011 8716 239655 minus16987 78684 minus18149lowastlowast 4324 minus57082lowastlowast 96701lowastlowastlowast 46487

(165621) (175737) (159264) (57138) (8072) (11743) (22603) (26906)daggerdaggerdagger (63961)Amalgamated lowast 2012 minus130426 192446 27324 82648 minus24229lowastlowastlowast 6313 minus60686lowastlowastlowast 110737lowastlowastlowast 42104

(165909) (176043) (159541) (57238) (8086) (11764) (22642)dagger (26953daggerdaggerdagger (54916)Amalgamated lowast 2013 72228 329923lowast minus11565 78142 minus7665 16314 minus54226lowastlowast 104628lowastlowastlowast 96197

(165488) (175597) (159137) (57093) (8065) (11734) (22585) (26884)daggerdaggerdagger (59957)Amalgamated lowast 2014 167078 371238lowastlowast minus44418 73532 minus13006 14685 minus59689lowastlowastlowast 99320lowastlowastlowast 87396

(165462) (175568) (159112) (57084) (8064) (11732) (22581)dagger (26880)daggerdaggerdagger (58970)Control variablesSmall Island 867066lowastlowastlowast 1104194lowastlowastlowast minus285506lowastlowastlowast 300412lowastlowastlowast 35248lowastlowastlowast minus7639 198169lowastlowastlowast minus4862 399776lowastlowastlowast

(99300)daggerdaggerdagger (105365)daggerdaggerdagger (95489)daggerdagger (34258)daggerdaggerdagger (4840) (7041) (13552)daggerdaggerdagger (16132) (95794)daggerdaggerdaggerDispersal of

settlementminus170282lowastlowastlowast minus102486lowastlowastlowast 47756lowastlowastlowast minus8375lowast 4405lowastlowastlowast minus12830lowastlowastlowast 15518lowastlowastlowast minus3410 2562(13254)daggerdaggerdagger (14064)daggerdaggerdagger (12745)daggerdaggerdagger (4573) (646) (940)daggerdaggerdagger (1809)daggerdaggerdagger (2153) (9631)

Fiscal pressure minus83154lowastlowastlowast minus71255lowastlowastlowast minus12542lowastlowastlowast minus4331lowastlowastlowast minus723lowastlowastlowast minus4532lowastlowastlowast minus5111lowastlowastlowast 8422lowastlowastlowast minus23980lowastlowastlowast

(3517)daggerdaggerdagger (3731)daggerdaggerdagger (3382)daggerdaggerdagger (1213)daggerdaggerdagger (171) (249)daggerdaggerdagger (480)daggerdaggerdagger (571)daggerdaggerdagger (3023)daggerdaggerdaggerSocioec expenditure

needs021lowastlowastlowast 058lowastlowastlowast 055lowastlowastlowast 037lowastlowastlowast 001lowastlowastlowast 006lowastlowastlowast 005lowastlowastlowast 032lowastlowastlowast 064lowastlowastlowast

(005)daggerdaggerdagger (005)daggerdaggerdagger (005)daggerdaggerdagger (002)daggerdaggerdagger (000) (000)daggerdaggerdagger (001)daggerdaggerdagger (001)daggerdaggerdagger (004)daggerdaggerdagger

14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Party fragmentation 64797lowastlowastlowast 32604 minus82247lowastlowastlowast 35568lowastlowastlowast minus1973lowast minus1122 5883lowast 13660lowastlowastlowast 23167(24061)dagger (25531) (23137)daggerdaggerdagger (8301)daggerdaggerdagger (1173) (1706) (3284) (3909)daggerdaggerdagger (16708)

Share of socialistseats

13043lowastlowastlowast 11933lowastlowastlowast minus3448lowastlowast 1090lowastlowast minus519lowastlowastlowast minus378lowastlowastlowast minus438lowastlowast 2458lowastlowastlowast 2272(1602)daggerdaggerdagger (1700)daggerdaggerdagger (1541) (553) (078) (114)daggerdagger (219) (260)daggerdaggerdagger (1540)

Year dummies2004 29762 minus93642 69864 minus15252 1728 869 13029 51001lowastlowast 84816lowastlowastlowast

(137513) (145913) (132236) (47442) (6702) (9750) (18767) (22340) (20281)daggerdaggerdagger2005 82944 minus471790lowastlowastlowast 171315 minus32813 2295 3996 18990 74535lowastlowastlowast 95974lowastlowastlowast

(137755) (146169)daggerdagger (132468) (47525) (6714) (9768) (18800) (22379)daggerdagger (25826)daggerdaggerdagger2006 341932lowastlowast minus463534lowastlowastlowast 131720 minus30769 minus23285lowastlowastlowast minus1231 minus18990 70775lowastlowastlowast 55050lowast

(137784) (146200)daggerdagger (132496) (47535) (6715)daggerdagger (9770) (18804) (22384)daggerdagger (30435)2007 695972lowastlowastlowast minus44349 60357 87431lowast 11202lowast minus525 28993 73488lowastlowastlowast 262598lowastlowastlowast

(137965)daggerdaggerdagger (146392) (132670) (47597) (6724) (9783) (18829) (22413)daggerdagger (36074)daggerdaggerdagger2008 756711lowastlowastlowast 57147 minus61612 136541lowastlowastlowast 17032lowastlowast minus1337 45393lowastlowast 93656lowastlowastlowast 328926lowastlowastlowast

(137955)daggerdaggerdagger (146381) (132660) (47594)daggerdagger (6724) (9782) (18827) (22411)daggerdaggerdagger (38551)2009 863071lowastlowastlowast 187968 minus107124 166146lowastlowastlowast 16219lowastlowast minus13681 61418lowastlowastlowast 132039lowastlowastlowast 412635lowastlowastlowast

(137836)daggerdaggerdagger (146255) (132546) (47553)daggerdaggerdagger (6718) (9773) (18811)daggerdagger (22392)daggerdaggerdagger (41587)daggerdaggerdagger2010 712887lowastlowastlowast 89405 minus430745lowastlowastlowast 177495lowastlowastlowast 10733 minus16172 77441lowastlowastlowast 180111lowastlowastlowast 394354lowastlowastlowast

(139230)daggerdaggerdagger (147735) (133887)daggerdagger (48034)daggerdaggerdagger (6786) (9872) (19002)daggerdaggerdagger (22619)daggerdaggerdagger (54651)daggerdaggerdagger2011 382949lowastlowastlowast minus153133 minus776496lowastlowastlowast 139314lowastlowastlowast 17947lowastlowastlowast minus21668lowastlowast 63542lowastlowastlowast 264150lowastlowastlowast 348080lowastlowastlowast

(139440)dagger (147958) (134089)daggerdaggerdagger (48106)daggerdagger (6796)dagger (9887) (19030)daggerdagger (22653)daggerdaggerdagger (60979)daggerdaggerdagger2012 499831lowastlowastlowast minus209719 minus758687lowastlowastlowast 131457lowastlowastlowast 24526lowastlowastlowast minus23794lowastlowast 74468lowastlowastlowast 280005lowastlowastlowast 388838lowastlowastlowast

(139648)daggerdaggerdagger (148178) (134288)daggerdaggerdagger (48178)dagger (6806)daggerdaggerdagger (9902) (19058)daggerdaggerdagger (22686)daggerdaggerdagger (50994)daggerdaggerdagger2013 366694lowastlowastlowast minus448297lowastlowastlowast minus899975lowastlowastlowast 160982lowastlowastlowast 16154lowastlowast minus32369lowastlowastlowast 79390lowastlowastlowast 322778lowastlowastlowast 357318lowastlowastlowast

(139376)daggerdaggerdagger (147889)daggerdagger (134026)daggerdaggerdagger (48084)daggerdagger (6793) (9883)daggerdagger (19021)daggerdaggerdagger (22642)daggerdaggerdagger (56287)daggerdaggerdagger2014 329738lowastlowast minus231745 minus946800lowastlowastlowast 174369lowastlowastlowast 19055lowastlowastlowast minus31713lowastlowastlowast 91422lowastlowastlowast 318802lowastlowastlowast 382505lowastlowastlowast

(139413) (147928) (134062)daggerdaggerdagger (48097)daggerdaggerdagger (6795)dagger (9885)daggerdagger (19026) (22648)daggerdaggerdagger (55046)daggerdaggerdaggerConstant 13893344lowastlowastlowast 13337278lowastlowastlowast 5889011lowastlowastlowast 268823lowastlowast 159152lowastlowastlowast 632684lowastlowastlowast 912390lowastlowastlowast minus836848lowastlowastlowast 5194830lowastlowastlowast

(347760)daggerdaggerdagger (369002)daggerdaggerdagger (334414)daggerdaggerdagger (119976) (16949)daggerdaggerdagger (24658)daggerdaggerdagger (47461) (56495)daggerdaggerdagger (296603)daggerdaggerdaggerObservations 1140 1140 1140 1140 1140 1140 1140 1140 1140R2 0697 0589 0498 0547 0355 0611 0552 0862 0804

Notes Standard errors in parentheses For model 9 robust standard errors (clustered at each municipality) and R-squared is adjusted R2Level of significance is marked by asterisks after the parameter estimate lowastlowastlowast plt001 lowastlowast plt005 lowast plt01Level of significance Bonferroni-corrected for ten simultaneous tests daggerdaggerdagger plt001 daggerdagger plt005 dagger plt01

15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms


observationsmdashthat is four prereform years and eightpostreform years for all municipalities This analysisthus makes it possible to identify the exact timing ofa reform effect Since a reform effect is not likely tomaterialize immediately after the reform Table 5 canshow whether it occurs with a time lag In addition weintroduce one more methodological adjustment Sinceour data are expenditure allocations from the sameoverall budget to different policy areas they are notlikely to be completely independent across policy areasWe therefore run the analyses as seemingly unrelatedregressions (SUR) (Zellner 1962) Table 5 is thereforealso a robustness check of the results in Table 4

Again according to the DiD logic reform effectsare identified by interaction terms of the treatmentvariable (amalgamation) and post-treatment timemeasures In Table 5 the DiD estimators are conse-quently Amalgamatedlowast2007 Amalgamatedlowast2008 Am-algamatedlowast2009 Amalgamatedlowast2010 Amalgamatedlowast-2011 Amalgamatedlowast2012 Amalgamatedlowast2013 andAmalgamatedlowast2014

Table 5 confirms the results from Table 4 In the ar-eas of daycare schools elder care and children withspecial needs there is no evidence that the amalgama-tion reform made a difference to spending In the areasof roads and administration mergers seem to have ledto lower spending while the opposite is the case in thearea of labor market services The suggestion in Table 4of higher spending on culture is not reproduced Incontrast to Table 4 Table 5 allows the timing of thesereform effects to be identified In the road area reformeffects start in 2008 and grow over the following yearsuntil the effect ceases to be statistically significant in2013 In the administrative area they do not materi-alize until 2009 but then also grow over the followingyears9 In the labor market area permanent negativereform effects appear already in 2007

To briefly comment on the remaining findings inTable 5 the year dummies estimate the general timetrend including changes in how functional respon-sibilities are assigned for each year relative to theinitial year 2003 As is evident these dummies arestatistically significant in most analyses indicating thatthe municipalities experience common influences overtime This confirms the impression from the graphs inFigure 1 which showed parallel expenditure trends forthe amalgamated and nonamalgamated municipalitiesTurning to the control variables municipalities on smallislands face extraordinary diseconomies of scale in theprovision of services for daycare schools roads chil-dren with special needs and administration The vari-able dispersal of settlement shows that thinly populatedmunicipalities spend more on elder care roads andadministration but less on all other areas Fiscal pres-sure leads to lower spending in all policy areasmdashexceptthe labor market probably because fiscal pressure ispartly caused by unemployment Next socioeconomicexpenditure needs are cost drivers in all policy areasFinally expenditure in Danish municipalities may also

9 This particular result corresponds to Blom-Hansen Houlberg andSerritzlew (2014)

reflect political factors Both party fragmentation andparty ideology measured as the share of socialist seatshave nontrivial but unsystematic effects across policyareas

The results reported in Figure 1 and Tables 4 and 5constitute our core findings However before draw-ing final conclusions we conduct three robustnesschecks First in Appendix Table A2 in the online sup-plementary material we break down our dependentvariablemdashspending per potential usermdashinto its twocomponentsmdashthe quantity of outputs supplied (per po-tential user) and the cost of each unit of output Lowerspending per user might indicate either a reduction insupply (fewer units) or an increase in efficiency (lowercost per unit) rendering the previous results a littleambiguous In the six functional areas for which suchbreakdowns are possible10 we find no evidence of anychangemdasheither positive or negativemdashin the efficiencyof provision after amalgamation11 As for the amountsupplied this is significantly higher for labor marketactivities and roads but it is significantly lower for eldercare In the case of roads this reflects a greater transferof regional roads to the newly merged municipalitiesthan to the control group municipalities and not somemunicipal decision It is hard to think of any generallogic that would explain this pattern For children withspecial needs we observe an interesting change Thereis some tendency for amalgamated municipalities tosupply more units (that is to forcibly remove morechildren) after the reform Since we control for socioe-conomic expenditure needs this is unlikely to reflectdisproportionate changes in the composition of citizensin amalgamated and nonamalgamated municipalitiesThis could be produced by a tendency for smaller units(ie later-amalgamated municipalities before the re-form) to hesitate to forcibly remove children becausethe major long-term expense of this intervention canhave serious budgetary consequences for a small mu-nicipality12 This is offset by a statistically insignificanttendency for unit costs to be smaller resulting in thenet result that expenditure does not change In sumincreased jurisdiction size seems to have had mixedeffects if any on spending levels and no discernibleeffect on efficiency

Second in Appendix Table A3 in the online sup-plementary material we rerun the analysis for sub-groups of municipalities of different (prereform) sizesAlthough most studies find that the evidence oneconomies of scale in local government is inconclusivesome find a tendency for very small municipalities to

10 The measurement of the number of units supplied varies acrosspolicy areas depending on the type of task and the most appro-priate available data For daycare for instance the supplied unitsare measured by the number of children aged under six enrolled inmunicipal daycare whereas for roads the number of units refers tothe length of municipal roads maintained by the municipality andfor elder care it is a weighted average of the number of housing unitsoperated and the number of hours of home help for the elderly SeeAppendix Table A1 in the online supplementary material for thespecific measurement for each policy area11 Spending per unit of output is significantly lower for roads in oneyear but insignificant in all others and the sign flips back and forth12 We thank one of the referees for suggesting this interpretation



be inefficient (eg Bodkin and Conklin 1971 Breunigand Rocaboy 2008 Sole-Olle and Bosch 2005) Wetherefore investigate whether small municipalities gainmore from amalgamation than somewhat larger onesAppendix Table A3 reports results rerunning Model9 of Table 5 for just those amalgamated municipalitieswhose prereform size averaged respectively less than10000 citizens less than 12000 citizens and less than15000 citizens In each case the results were not sys-tematically different from those of our main analysis(for amalgamated municipalities with prereform aver-age size of up to 20000 citizens)

Third in Appendix Table A4 in the online supple-mentary material we report results for two groups ofmunicipalities based on the similarity of their prere-form spending levels The first group consists of pairs ofamalgamating municipalities that had relatively similarspending levels while the second contains pairs withmore different prereform spending levels The aim isto see if the results could be driven by a tendency formunicipalities with similar spending to merge For pairsof municipalities with very different spending levelsone might imagine that spending in the low-spendingmunicipality would converge upward to that of its high-spending counterpart However we find that results arevery similar in the two groups

DISCUSSION AND CONCLUSION

Since the 1950s a wave of municipal amalgamationsmotivated largely by a belief in readily attainableeconomies of scale has expanded the jurisdictions oflocal governments across the developed world Ex-ploiting the exogenous imposition of a reform toamalgamate all Danish municipalities with populationsunder 20000 inhabitants and using a difference-in-differences design to compare these merged munici-palities with other relatively large ones untouched bythe reform we provide stronger evidence than previ-ously available about the effects of jurisdiction size onspending

We show that increasing local governmentsrsquo jurisdic-tion size had no systematic consequences on spendingIn one or two functional areas amalgamation led tolower spending in one it led to higher spending andin most areas spending was unaffected From the lo-cal taxpayersrsquo perspective total spending per capitais probably the most salient variable But spendingper capita can also be usefully decomposed into twocomponent partsmdashthe number of units supplied (percapita) and the cost per unit Although like the rest ofthe literature on this topic we lack compelling across-the-board indicators of service quality cost per unitcan serve as a reasonable proxy of efficiency In noneof the service categories for which we could estimatecost per unit did larger jurisdiction size result in eithersignificantly higher or lower efficiency measured in thisway

Our design does not allow us to see exactly why thisis so The lack of an effect certainly does not mean thatfixed costs are irrelevant to production in the eight

policy areas studied or that no economies of scale ex-ist On the contrary previous literature suggests thatfixed costs can be considerable (Boyne 1995 Hirsch1959 Sawyer 1991) A more plausible interpretationis that the relevant kind of fixed costs are difficult toreduce by municipal amalgamation Some of the mostexpensive public services are produced at units withinlocal government jurisdictions such as schools kinder-gartens and nursing homes Increasing the scale of localgovernments does not automatically increase the scaleof such service providers (Boyne 1995 Sawyer 1991)As in private production firm size does not equateto plant size Besides multipurpose governments canalmost never be optimally sized for all the services theyprovide since different services have different produc-tion functions and externalities (Olson 1986 Tullock1969) Any systematic effect in one area may be offsetby countervailing effects in another (Treisman 2007)These empirical findings are consistent with the weak-ness of the theoretical rationale for consistent scaleeffects

We have abstracted here from the direct costsof amalgamation reforms Various evidence suggeststhese can be large not just because of the transi-tion costs but alsomdashand probably more importantlymdashbecause municipalities about to merge often indulge ina last-minute flurry of spending (Blom-Hansen 2010Hansen 2014 Hinnerich 2009 Jonsson 1983 Jordahland Liang 2010) If mergers have no general positiveeffects the costs of implementing them should givepause to reformers We conclude that if Denmarkrsquosexperience is typical the global amalgamation wavewill probably not result in real savings This has policyimplications Prospective reformers of the architectureof government should not build plans to consolidatelocal government upon an expectation that larger sizewill lead to cost reductions

This result may also have implications for how thequestion of optimal size should be investigated empir-ically If jurisdiction size has no unequivocal effect oncosts for multipurpose units it makes little sense tolook for a unique context-free answer The optimalscale for a political entity depends on what servicesit provides Consider for example Australia wherelocal government is only ldquoengaged in the most mini-mal property-oriented services (primarily ldquoroads andrubbishrdquo)rdquo (Boadway and Shah 2009 276) It maywell be that the economically optimal size in such acase is small perhaps 5000 inhabitants (the Australianmunicipalities are in fact larger than that) Or imag-ine another country in which local governments areresponsible for elementary schools elderly care andchild care How large municipalities are is not very rel-evant to the costs of providing these goods since whatmatters most is the size of schools retirement homesand daycare centers Of course this does not mean thatone should ignore scale effects Rather it suggests theneed to direct attention to questions that are likely tohave answers such as the optimal size of a particularservice at the plant level The accumulation of knowl-edge on such questions promises both academic andpolicy payoffs



Drawing lessons from one countryrsquos experience re-quires care The quasi-experimental nature of the Dan-ish reform offers unusual opportunities to identifycausal relationships but the results cannot be general-ized without caution First the world of municipalitiesis diverse Some countries (for example France Aus-tria and Switzerland) have very small municipalitieswell below the smallest included in the data analyzedhere Although we expect that a similar logic appliesto them too we cannot rule out that some munici-palities are so small that amalgamation would in factproduce economies of scale across the board Since thevariance in the pre- and postreform size of Danish mu-nicipalities is limitedmdashwith only a few below 5000 orabove 100000 citizensmdashit will require further researchto see whether the results extend to systems with muchsmaller or larger units Second Danish municipali-ties aremdashas in most countriesmdashmultipurpose serviceproviders However in some countriesmdashespecially theUSAmdashsingle-purpose entities are also important Insuch cases the difficulty of aggregating optimal scalesfor multiple services disappears although one is stillleft with the disconnect between firm and plant levelcosts (eg those of the school and those of the schoolboard)

Further research will also be needed to pin downwhy economies of scale failed to materialize in this caseand in others If one key factor ismdashas we conjecturedmdashthe disconnect between firm size and plant size effectsthen we might expect to see consistent divergencesin the effect of amalgamations on plant level costs(for instance of schools and hospitals) and firm levelcosts (for instance of administration in city hall) Thesewill not necessarily correlate and of course enlargingmunicipal jurisdictions will not make the schools andhospitals within them either bigger or smaller At thesame time analyses of this question must take seri-ously the endogenous way in which local governmentjurisdictions evolve If future well-designed studies ofadditional countries also fail to find clear evidence forscale effects this will deepen doubts about the wisdomof the global movement for municipal amalgamation

SUPPLEMENTARY MATERIAL

To view supplementary material for this article pleasevisit httpdxdoiorg101017S0003055416000320

REFERENCES

Alba Carlos and Carmen Navarro 2003 ldquoTwenty-five Years ofDemocratic Local Government in Spainrdquo In Reforming LocalGovernment in Europe eds Norbert Kersting and Angelika Vet-ter Opladen Leske amp Budrich 197ndash221

Alesina Alberto and Enrico Spolaore 2003 The Size of NationsCambridge MA MIT Press

Allers Maarten A 2012 ldquoYardstick Competition Fiscal Disparitiesand Equalizationrdquo Economics Letters 117 4ndash6

Allers Maarten A and J Bieuwe Geertsema 2014 ldquoThe Effects ofLocal Government Amalgamation on Public Spending and ServiceLevels Evidence from 15 Years of Municipal Boundary ReformrdquoUniversity of Groningen unpublished paper (httpirsubrugnldbi53ad249381b25)

Anderson Michelle Wilde 2012 ldquoDissolving Citiesrdquo Yale Law Jour-nal 121 1364ndash446

Andrews Rhys George A Boyne Jennifer Law and Richard MWalker 2005 ldquoExternal Constraints on Local Service StandardsThe Case of Comprehensive Performance Assessment in EnglishLocal Governmentrdquo Public Administration 83 639ndash56

Arter David 2012 Scandinavian Politics Today ManchesterManchester University Press

Baldersheim Harald and Lawrence E Rose 2010a ldquoTerritorialChoice Rescaling Governance in European Statesrdquo In Ter-ritorial Choice The Politics of Boundaries and Borders edsHarald Baldersheim and Lawrence E Rose Houndsmills Pal-grave 1ndash20

Baldersheim Harald and Lawrence E Rose 2010b ldquoA Compara-tive Analysis of Territorial Choice in Europe ndash Conclusionsrdquo InTerritorial Choice The Politics of Boundaries and Borders edsHarald Baldersheim and Lawrence E Rose Houndsmills Pal-grave 234ndash60

Baldersheim Harald and Lawrence E Rose 2010c ldquoThe StayingPower of the Norwegian Peripheryrdquo In Territorial Choice ThePolitics of Boundaries and Borders eds Harald Baldersheim andLawrence E Rose Houndsmills Palgrave 80ndash101

Bergstrom Theodore C and Robert P Goodman 1973 ldquoPrivateDemands for Public Goodsrdquo The American Economic Review 63(3) 280ndash96

Berry Christopher R 2009 Imperfect Union Representation andTaxation in Multilevel Governments Cambridge UK CambridgeUniversity Press

Berry Christopher R and Martin R West 2010 ldquoGrowing PainsThe School Consolidation Movement and Student OutcomesrdquoJournal of Law Economics amp Organization 26 1ndash29

Bhatti Yosef and Kasper Moslashller Hansen 2011 rdquoWho MarriesWhom The Influence of Societal Connectedness Economic andPolitical Homogeneity and Population Size on Jurisdictional Con-solidationsrdquo European Journal of Political Research 50 (2) 212ndash38

Bish Robert L 2001 Local Government Amalgamations Discred-ited Nineteenth-Century Ideals Alive in the Twenty-First C DHowe Institute Commentary No 150 Toronto C D Howe In-stitute

Blom-Hansen Jens 2003 ldquoIs Private Delivery of Public ServicesReally Cheaper Evidence from Public Road Maintenance inDenmarkrdquo Public Choice 115 419ndash38

Blom-Hansen Jens 2010 ldquoMunicipal Amalgamations and CommonPool Problems The Danish Local Government Reform in 2007rdquoScandinavian Political Studies 33 51ndash73

Blom-Hansen Jens and Anne Heeager 2011 ldquoDenmark Be-tween Local Democracy and Implementing Agency of the Wel-fare Staterdquo In The Oxford Handbook of Local and RegionalDemocracy in Europe eds John Loughlin Frank Hendriks andAnders Lidstrom Oxford Oxford University Press 221ndash41

Blom-Hansen Jens Kurt Houlberg and Soslashren Serritzlew 2014ldquoSize Democracy and the Economic Costs of Running the Politi-cal Systemrdquo American Journal of Political Science 58 (4) 790ndash803

Boadway Robin and Anwar Shah 2009 Fiscal Federalism Cam-bridge UK Cambridge University Press

Bodkin Ronald J and David W Conklin 1971 ldquoScale and OtherDeterminants of Municipal Expenditures in Ontario A Quantita-tive Analysisrdquo International Economic Review 12 465ndash81

Boedeltje Mijke and Bas Denters 2010 ldquoStep-by-Step Territo-rial Choice in the Netherlandsrdquo In Territorial Choice The Pol-itics of Boundaries and Borders eds Harald Baldersheim andLawrence E Rose Houndsmills Palgrave 118ndash38

Borcherding Thomas E and Robert T Deacon 1972 ldquoThe De-mand for the Services of Non-Federal Governmentsrdquo The Amer-ican Economic Review 62 (5) 891ndash901

Boston Jonathan John Martin June Pallot and Pat Walsh 1996Public Management The New Zealand Model Auckland OxfordUniversity Press

Boyne George A 1995 ldquoPopulation Size and Economies of Scale inLocal Governmentrdquo Policy and Politics 23 (3) 213ndash22

Boyne George A 1996 Constraints Choices and Public PoliciesLondon JAI Press

Boyne George A 1998 Public Choice Theory and Local Gov-ernment A Comparative Analysis of the UK and the USAHoundsmills MacMillan



Boyne George A 2002 ldquoConcepts and Indicators of Local Author-ity Performance An Evaluation of the Statutory Frameworks inEngland and Walesrdquo Public Money amp Management 22 2

Boyne George A 2003 ldquoSources of Public Service Improvement ACritical Review and Research Agendardquo Journal of Public Admin-istration Research and Theory 13 367ndash94

Brennan Geoffrey and James B Buchanan 1980 The Power to TaxAnalytical Foundations of a Fiscal Constitution Cambridge UKCambridge University Press

Breunig Robert and Yvon Rocaboy 2008 ldquoPer-capita Public Ex-penditures and Population Size A Non-parametric Analysis usingFrench Datardquo Public Choice 136 (3-4) 429ndash45

Brunazzo Marco 2010 ldquoItalian Regionalism A Semi-Federationis Taking Shape ndash Or is itrdquo In Territorial Choice The Poli-tics of Boundaries and Borders eds Harald Baldersheim andLawrence E Rose Houndsmills Palgrave 180ndash98

Bundgaard Ulrik and Karsten Vrangbaeligk 2007 ldquoReform by Co-incidence Explaining the Policy Process of Structural Reform inDenmarkrdquo Scandinavian Political Studies 30 491ndash520

Byrnes Joel and Brian Dollery 2002 ldquoDo Economies of ScaleExist in Australian Local Government A Review of ResearchEvidencerdquo Urban Policy and Research 20 391ndash414

Cheney Peter 2014 ldquoReforming Local Governmentrdquo Eolas Maga-zine (httpwwweolasmagazineiereforming-local-government)

Christiansen Peter Munk and Michael Baggesen Klitgaard 2010ldquoBehind the Veil of Vagueness Success and Failure in InstitutionalReformsrdquo Journal of Public Policy 30 183ndash200

Colino Cesar and Eloisa Del Pino 2011 ldquoSpain The Consolidationof Strong Regional Governments and the Limits of Local De-centralizationrdquo In The Oxford Handbook of Local and RegionalDemocracy in Europe eds John Loughlin Frank Hendriks andAnders Lidstrom Oxford Oxford University Press 356ndash84

Cook Thomas D and Donald T Campbell 1979 Quasi-Experimentation Design amp Analysis Issues for Field SettingsBoston Houghton Mifflin

Dafflon Bernard 2013 ldquoVoluntary Amalgamation of Local Gov-ernments The Swiss Debate in the European Contextrdquo In TheChallenge of Local Government Size Theoretical Perspectives In-ternational Experience and Policy Reform eds S Lago-Penas andJ Martinez-Vazquez Northampton MA Edward Elgar Publish-ing 189ndash220

Dahl Robert A and Edward R Tufte 1973 Size and DemocracyStanford Standford University Press

Denters Bas Michael Goldsmith Andreas LadnerPoul Erik Mouritzen and Lawrence E Rose 2014 Size andLocal Democracy Cheltenham Edward Elgar

Derksen Wim 1988 ldquoMunicipal Amalgamation and the Doubt-ful Relation between Size and Performancerdquo Local GovernmentStudies 14 31minus47

Dollery Brian and Joe L Wallis 2001 The Political Economy ofLocal Government Cheltenham Edward Elgar

Dollery Brian and Euan Fleming 2006 ldquoA Conceptual Note onScale Economies Size Economies and Scope Economies in Aus-tralian Local Governmentrdquo Urban Policy and Research 24 (2)271ndash82

Dollery Brian Joel Byrnes and Lin Crase 2008 ldquoStructural Reformin Australian Local Governmentrdquo Australian Journal of PoliticalScience 43 333ndash9

Dunning Thad 2012 Natural Experiments in the Social Sciences ADesign-Based Approach Cambridge UK Cambridge UniversityPress

Fallend Franz 2011 ldquoAustria From Consensus to Competition andParticipationrdquo In The Oxford Handbook of Local and RegionalDemocracy in Europe eds John Loughlin Frank Hendriks andAnders Lidstrom Oxford Oxford University Press 173ndash96

Forde Catherine 2005 ldquoParticipatory Democracy or Pseudo-Participation Local Government Reform in Irelandrdquo Local Gov-ernment Studies 31 137ndash48

Foster Kathryn A 1997 The Political Economy of Special-PurposeGovernment Washington DC Georgetown University Press

Fox William F and Tami Gurley 2006 Will Consolidation ImproveSub-national Governments World Bank Policy Research WorkingPaper 3913

Grossman Guy and Janet I Lewis 2014 ldquoAdministrative Unit Pro-liferationrdquo American Political Science Review 108 (1) 196ndash217

Hansen Sune Welling 2014 ldquoCommon Pool Size and Project Sizean Empirical Test on Expenditures Using Danish Municipal Merg-ersrdquo Public Choice 159 3ndash21

Hinnerich Bjorn Tyrefors 2009 ldquoDo Merging Local GovernmentsFree Ride on their Counterparts when Facing Boundary ReformrdquoJournal of Public Economics 93 721ndash8

Hirsch Werner Z 1959 ldquoExpenditure Implications of MetropolitanGrowth and Consolidationrdquo Review of Economics and Statistics41 (3) 232ndash41

Hlepas Nikolaos-Komnenos 2003 ldquoLocal Government Reformin Greecerdquo In Reforming Local Government in Europe edsNorbert Kersting and Angelika Vetter Opladen Leske amp Budrich221ndash41

Hlepas Nikos and Panagiotis Getimis 2011 ldquoGreece A Case ofFragmented Centralism and lsquoBehind the Scenesrsquo Localismrdquo InThe Oxford Handbook of Local and Regional Democracy in Eu-rope eds John Loughlin Frank Hendriks and Anders LidstromOxford Oxford University Press 410ndash34

Holzer Marc John Fry Etienne Charbonneau Gregg Van RyzinTiankai Wang and Eileen Burnash 2009 Literature Review andAnalysis Related to Optimal Municipal Size and Efficiency Re-port prepared for the Local Unit Alignment Reorganizationand Consolidation Commission httpwwwnjgovdcaaffiliatesluarccpdffinal optimal municipal size amp efficiencypdf

Hooghe Liesbet and Gary Marks 2009 ldquoDoes Efficiency Shape theTerritorial Structure of Governmentrdquo Annual Review of PoliticalScience 12 225ndash41

John Peter 2010 ldquoLarger and Larger The Endless Search for Effi-ciency in the UKrdquo In Territorial Choice The Politics of Boundariesand Borders eds Harald Baldersheim and Lawrence E RoseHoundsmills Palgrave 101ndash18

Jonsson Ernst 1983 ldquoMeasures Taken by Municipalities Undergo-ing Amalgamationrdquo Scandinavian Political Studies 6 231ndash4

Jordahl Henrik and Che-Yuan Liang 2010 ldquoMerged MunicipalitiesHigher Debt on Free-Riding and the Common Pool Problem inPoliticsrdquo Public Choice 143 157ndash72

Keating Michael 1995 ldquoSize Efficiency and Democracy Consoli-dation Fragmentation and Public Choicerdquo In Theories of UrbanPolitics eds David Judge Gerry Stoker and Harold WolmanLondon Sage 117ndash35

Kerrouche Eric 2010 ldquoFrance and Its 36000 Communes An Impos-sible Reformrdquo In Territorial Choice The Politics of Boundariesand Borders eds Harald Baldersheim and Lawrence E RoseHoundsmills Palgrave 160ndash80

Kubler Daniel and Andreas Ladner 2003 ldquoLocal Government Re-form in Switzerland More For than By ndash But What about OfrdquoIn Reforming Local Government in Europe eds Norbert Kerstingand Angelika Vetter Opladen Leske amp Budrich 137ndash57

Ladner Andreas 2011 ldquoSwitzerland Subsidiarity Power-sharingand Direct Democracyrdquo In The Oxford Handbook of Local andRegional Democracy in Europe eds John Loughlin Frank Hen-driks and Anders Lidstrom Oxford Oxford University Press196ndash221

Lassen David Dreyer and Soslashren Serritzlew 2011 ldquoJurisdiction Sizeand Local Democracy Evidence on Internal Political Efficacyfrom Large-scale Municipal Reformrdquo American Political ScienceReview 105 (2) 238ndash58

Lidstrom Anders 2010 ldquoThe Swedish Model under Stress The Wan-ing of the Egalitarian Unitary Staterdquo In Territorial Choice ThePolitics of Boundaries and Borders eds Harald Baldersheim andLawrence E Rose Houndsmills Palgrave 61ndash80

Loughlin John 2011 ldquoIreland Halting Steps Towards Local Democ-racyrdquo In The Oxford Handbook of Local and Regional Democracyin Europe eds John Loughlin Frank Hendriks and Anders Lid-strom Oxford Oxford University Press 48ndash71

Lowi Thodore J 1972 ldquoFour Systems of Policy Politics and ChoicerdquoPublic Administration Review 32 (4) 298ndash310

Martins M R 1995 ldquoSize of Municipalities Efficiency and CitizenParticipation A Cross-European Perspectiverdquo Environment andPlanning C Government and Policy 13 (4) 441ndash58

Mouritzen Poul Erik ed 2006 Stort er Godt Otte Fortaeligllinger omTilblivelsen af de nye Kommuner Odense Syddansk Universitets-forlag

Mouritzen Poul Erik 2010 ldquoThe Danish Revolution in Local Gov-ernment How and Whyrdquo In Territorial Choice The Politics



of Boundaries and Borders eds Harald Baldersheim andLawrence E Rose Houndsmills Palgrave 21ndash41

Newton Kenneth 1982 ldquoIs Small Really so Beautiful Is Big Reallyso Ugly Size Effectiveness and Democracy in Local Govern-mentrdquo Political Studies 30 190ndash206

Oates Wallace E 1972 Fiscal Federalism New York HarcourtBrace Jovanovich

Oberfield Zachary W 2014 ldquoAccounting for Time Comparing Tem-poral and Atemporal Analyses of the Business Case for DiversityManagementrdquo Public Administration Review 74 777ndash89

OECD 2005 OECD Territorial Reviews Busan Korea 2005 ParisOECD

OECD 2010 OECD Territorial Reviews Sweden 2010 ParisOECD

OECD 2014a OECD Territorial Reviews Netherlands 2014 ParisOECD

OECD 2014b OECD Regional Outlook 2014 Regions and CitiesWhere Policies and People Meet Paris OECD

Olson Mancur 1986 ldquoTowards a More General Theory of Govern-mental Structurerdquo American Economic Review 76 (2) 120ndash5

Ostrom Elinor 1972 ldquoMetropolitan Reform Propositions Derivedfrom Two Traditionsrdquo Social Science Quarterly 53 (3) 474ndash93

OrsquoToole Larry J and Kenneth J Meier 1999 ldquoModeling the Im-pact of Public Management Implications of Structural ContextrdquoJournal of Public Administration Research and Theory 9 505ndash26

Piattoni Simona and Marco Brunazzo 2011 ldquoItaly The SubnationalDimension to Strengthening Democracy since the 1990srdquo In TheOxford Handbook of Local and Regional Democracy in Europeeds John Loughlin Frank Hendriks and Anders Lidstrom Ox-ford Oxford University Press 331ndash56

Pleschberger Werner 2003 ldquoCities and Municipalities in the Aus-trian Political System since the 1990s New Developments betweenlsquoEfficiencyrsquo and lsquoDemocracyrsquordquo In Reforming Local Governmentin Europe eds Norbert Kersting and Angelika Vetter OpladenLeske amp Budrich 113ndash57

Sancton A 1996 ldquoReducing Costs by Consolidating MunicipalitiesNew Brunswick Nova Scotia and Ontariordquo Canadian Public Ad-ministration 39 (3) 267ndash89

Sancton Andrew 2000 Merger Mania The Assault on Local Gov-ernment Montreal McGill-Queenrsquos University Press

Sandberg Siv 2010 ldquoFinnish Power-Shift The Defeat of the Periph-eryrdquo In Territorial Choice The Politics of Boundaries and Borderseds Harald Baldersheim and Lawrence E Rose HoundsmillsPalgrave 42ndash61

Santerre Rexford E 2009 ldquoJurisdiction Size and Local PublicHealth Spendingrdquo Health Services Research 44 (6) 2148ndash66

Sawyer Malcolm C 1991 The Economics of Industries and FirmsTheories Evidence and Policy London Routledge

Scherer F M and David Ross 1990 Industrial Market Structure andEconomic Performance Boston Houghton Mifflin

Serritzlew Soslashren 2005 ldquoBreaking Budgets An Empirical Examina-tion of Danish Municipalitiesrdquo Financial Accountability amp Man-agement 21 (4) 413ndash35

Slack Enid and Richard Bird 2013 ldquoMerging Municipalities Is Big-ger Betterrdquo IMFG Papers on Municipal Finance and GovernanceToronto University of Toronto

Sole-Olle Albert and Nuria Bosch 2005 ldquoOn the Relationship be-tween Authority Size and the Costs of Providing Local ServicesLessons for the Design of Intergovernmental Transfers in SpainrdquoPublic Finance Review 33 (3) 343ndash84

Strang David 1987 ldquoThe Administrative Transformation of Amer-ican Education School District Consolidation 1938-1980rdquo Ad-ministrative Science Quarterly 32 352ndash66

Sverrisson Sigurdur and Magnus Karel Hannesson 2014 LocalGovernments in Iceland Reykyavik Association of Local Author-ities in Iceland

Swianiewicz Pawel 2010 ldquoIf Territorial Fragmentation is a Problemis Amalgamation a Solution An East European PerspectiverdquoLocal Government Studies 36 183ndash203

Tiebout Charles M 1956 ldquoA Pure Theory of Local ExpenditurerdquoJournal of Political Economy 64 416ndash24

Treisman Daniel 2007 The Architecture of Government RethinkingPolitical Decentralization Cambridge UK Cambridge UniversityPress

Tullock Gordon 1969 ldquoFederalism Problems of Scalerdquo PublicChoice 6 (1) 19ndash29

Velasco A 2000 ldquoDebts and Deficits with Fragmented Fiscal Poli-cymakingrdquo Journal of Public Economics 76 105ndash25

Vetter Angelika and Norbert Kersting 2003 ldquoDemocracy ver-sus Efficiency Comparing Local Government Reforms acrossEuroperdquo In Reforming Local Government in Europe edsNorbert Kersting and Angelika Vetter Opladen Leske amp Budrich11ndash29

Walker Richard M and Ryes Andrews 2015 ldquoLocal GovernmentManagement and Performance A Review of Evidencerdquo Journalof Public Administration Research and Theory 25 101ndash33

Walter-Rogg Melanie 2010 ldquoMultiple Choice The Persistenceof Territorial Pluralism in the German Federationrdquo In Ter-ritorial Choice The Politics of Boundaries and Borders edsHarald Baldersheim and Lawrence E Rose Houndsmills Pal-grave 138ndash60

Wayenberg Ellen Filip De Rynck Kristof Steyvers andJean-Benoit Pilet 2011 ldquoBelgium A Tale of Regional Di-vergencerdquo In The Oxford Handbook of Local and RegionalDemocracy in Europe eds John Loughlin Frank Hendriks andAnders Lidstrom Oxford Oxford University Press 71ndash96

Williamson Oliver E 1967 ldquoHierarchical Control and OptimumFirm Sizerdquo Journal of Political Economy 75 123ndash38

Wollmann Hellmut 2003 ldquoGerman Local Government under theDouble Impact of Democratic and Administrative ReformsrdquoIn Reforming Local Government in Europe eds Norbert Ker-sting and Angelika Vetter Opladen Leske amp Budrich 85ndash113

Wooldridge Jeffrey M 2009 Introductory Econometrics A ModernApproach Canada South-Western Cengage Learning

Zellner Arnold 1962 ldquoAn Efficient Method of Estimating Seem-ingly Unrelated Regressions and Tests for Aggregation BiasrdquoJournal of the American Statistical Association 57 (298) 348ndash68

Oslashkonomi- og Indenrigsministeriet 2012 Kommunale Udgiftsbehovog andre Udligningssposlashrgsmal Betaelignkning nr 1533 Oslashkonomi-og Indenrigsministeriet marts


INTRODUCTION


LOCAL JURISDICTION SIZE THEORY AND EMPIRICAL SURVEYS


METHODS AND DATA

RESULTS



REFERENCES


consolidation (Blom-Hansen Houlberg and Serrit-zlew 2014) That might seem at first to vindicate theenthusiasm for mergers However administrative costsamount to less than 10 percent of total municipalspending We focus here on the other 90 percent andask Do municipal mergers decrease the costs of pro-vision of public services such as schools roads andinfrastructure

We find no clear and systematic effects from amal-gamations We replicate the finding of Blom-HansenHoulberg and Serritzlew (2014) that administrativecosts declined We find also that spending on roadmaintenance per kilometer of road fell in the mergedunits although we cannot say whether this representsgreater efficiency or skimping on repairs Howeverthe economies of scale in administration and (possi-bly) road maintenance were offset by diseconomies ofscale for labor market programs In most policy areasmdashincluding elder care schools daycare and caring forchildren with special needsmdashjurisdiction size did notseem to matter at all Aggregating the effects the netimpact was null If the pattern in Denmark holds moregenerally the global amalgamation wave is unlikely toyield the savings its proponents anticipate We interpretour null finding as supporting the position of skepticswho contend on theoretical grounds that the quest foran optimal jurisdiction size is futile (Dahl and Tufte1973 Treisman 2007)

The article is organized as follows The next sectionprovides background on the global wave of municipalamalgamations of recent decades The third sectiondiscusses theoretical arguments about the effects ofjurisdiction size The fourth section outlines the Dan-ish reform The fifth section describes the data andmethods used in the analysis The sixth section presentsresults and the final section concludes


Since the 1950s reforms to enlarge jurisdictions havetransformed the structure of local government acrossthe developed world As societies modernized and builtmore extensive welfare states the local governmentunits inherited from earlier periods were often thoughttoo small to capture economies of scale in service provi-sion (Baldersheim and Rose 2010a 2010b 242 Fox andGurley 2006 8 Keating 1995 118 Newton 1982 191Vetter and Kersting 2003 19)1 Almost everywhereprojects to merge municipalities were debatedmdashandin most cases adopted

These reforms spanned the globe Table 1 briefly re-views the main cases the dramatic scope of which mayhave escaped nonspecialists

From such a survey the extent of the phenomenonbecomes obvious municipal merger mania has sweptthe developed world Reforms have varied in their rad-

1 In Australia for instance reformers argued ldquothat lsquobigger is cheaperrsquodue inter alia to the existence of substantial economies of scalerdquo(Dollery Byrnes and Crase 2008) Similarly in Eastern Canada re-formers in the 1990s repeatedly emphasized anticipated cost savings(Sancton 1996)

icalism in some nations eg the UK the local govern-ment system has been comprehensively restructured inothers eg France the changes have been more lim-ited Countries startedmdashand endedmdashat quite differentpoints While in Mexico Ireland New Zealand Den-mark and Japan the average municipal populationis now more than 40000 residents in France TurkeySwitzerland Austria and Iceland it is still below 5000(OECD 2010 207) Even where mergers were notrapidly implemented demands for them dominatedthe intellectual agenda This is all the more intriguinggiven an opposite tendency among many developingand postcommunist countries where democratizationhas often prompted the division of administrative unitsinto ever smaller pieces (Swianiewicz 2010) In Sub-Saharan Africa for instance 29 countries saw the num-ber of administrative units grow by at least 20 percentbetween 1990 and 2012 Brazilrsquos roster of municipal-ities also increased by 50 percent after the transitionfrom military rule and there were major increases inIndonesia and Vietnam (Grossman and Lewis 2014196)

LOCAL JURISDICTION SIZE THEORY ANDEMPIRICAL SURVEYS

The optimal scale of local government jurisdictionsmdashor of government jurisdictions in generalmdashhas beendebated since the time of Plato Although the searchfor an ideal size that can be identified on theoreticalgrounds independent of context has consumed enor-mous intellectual energy over the years we believethat for several reasons it is a vain quest We brieflyreview the main arguments and explain why they failto yield general implications We suggest that withoutknowing the particular mix of tasks assigned to localgovernments and their technologies it is impossible topredict whether on balance enlarging municipalitieswill have positive or negative effects

Most scholars have conceptualized the optimal scaleof local government as a tradeoff between certain ef-fects that favor large size and others that favor smallerunits (Dahl and Tufte 1973 Hooghe and Marks 2009Treisman 2007) Oates (1972) in a famous analysissaw the main conflict as that between the more pre-cise matching of services to local tastes that is possiblewhen jurisdictions are small and the economies of scaleattainable when they are large

Since economies of scale are the most commonlycited advantage of large sizemdashand the dominant ar-gument for amalgamationsmdashwe discuss them in somedetail In both the private and the public sector returnsto scale are thought to increase for two main reasons(Boyne 1995 Hirsch 1959 Sawyer 1991 47ndash70) Firstthere are fixed costs associated with providing variouskinds of public service so the marginal cost will fallwith output at least up to a certain point Some publicgoods have elements of nonrivalry in consumption sothe marginal cost is zero (Bergstrom and Goodman1973 Borcherding and Deacon 1972) For instance dis-ease surveillance water quality control and restaurant


















OECD 2014a 271


Kerrouche 2010





OECD 2014a 271


OECD 2014a 271


OECD 2014a 271


OECD 2014a 272













































METHODS AND DATA































RESULTS









































FIGURE 1 Continued








JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

















OECD 2014a 271


Kerrouche 2010





OECD 2014a 271


OECD 2014a 271


OECD 2014a 271


OECD 2014a 272













































METHODS AND DATA































RESULTS









































FIGURE 1 Continued








JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES




































METHODS AND DATA































RESULTS









































FIGURE 1 Continued








JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES




























RESULTS









































FIGURE 1 Continued








JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES






































FIGURE 1 Continued








JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)

Without controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)







(1) (2) (3) (4) (5) (6) (7) (8) (9)

With controls


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)






(331925) (375100) (317625) (167725) (12324) (20833) (57374) (52414) (92226)Dispersal of


(54308) (33161) (33980) (23751) (1289) (4617) (6267) (8247) (10669)

12httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

Am



(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)











13httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

JurisdictionSize

andL

ocalGovernm

entPolicyE

xpenditureN

ovember

2016


(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)


Amalgamated(dummy)




















14httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

Am


Review

TABLE 5 Continued

(1) (2) (3) (4) (5) (6) (7) (8) (9)


old)


old)


old)


old)Roads

(per capita)Culture


(per capita)


(per capita)



















15httpdxdoiorg101017S0003055416000320D

ownloaded from

httpww

wcam




wcam

bridgeorgcoreterms




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES




























REFERENCES




















































































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES
























































































INTRODUCTION




METHODS AND DATA

RESULTS



REFERENCES

jurisdiction size and local government policy … · jurisdiction size and local government policy...

Documents