DO BOSSES MATTER?The role of city managers and external variables in explaining efficiency differences of Finnish municipalities*
Heikki A. Loikkanen**, Ilkka Susiluoto*** and
Michael Funk****
KTTO ry, Tiedettä ja viiniä, Uusi ylioppilastalo, Mannerheim –Sali,
18.3.2009
** University of Helsinki, Finland*** City of Helsinki Urban Facts, Finland**** University of Fribourg, Switzerland
2
Two-stage study
First stage: application of Data Envelopment Analysis (DEA) gives cost efficiency scores for municipalities
Second stage: explain efficiency scores (regr.)
A. With characteristics of municipalities - These steps were done in our earlier study Loikkanen &
Susiluoto (2005), published in Urban Publ Econs Review
B. This study: Consider in addition the role of city managers, their work environment and attitudes
- here we use survey Finnish results from U.N.Di.T.E Leadership Study 1996 (15 countries were involved)
3
Two-stage presentation
First stage: Presentation of earlier results of both stages, where city managers were not included
- data 353 municipalities , 1994-2002
Second stage (some new results):
A. Discussion on do bosses matter?- CEOs and Public sector leaders - What are Finnish City Managers?
B. Presentation of results, when city managers are included in regression models
- data 192 municipalities, 1994-1996- efficiency scores for these 192 municipalities
come from our earlier study
4
First stage: application of Data Envelopment Analysis (DEA)
Municipalities are multi-service providers (producers or buyers)
With DEA you get: Municipal cost efficiency scores (max value 100) - How large are differences in scores between municipalities?
- Which municipalities are best/weakest?
-Size, location etc?
5
Data
353 Finnish municipalities included
Years 1994-2002
Excluded from data: municipalities with less than 2000 people Åland islands (small communities) municipal annexations in the study
period
6
OUTPUTS AND INPUT IN DEA APPLICATION
Altogether 10 outputs (services): most important basic health, social and educational services
Special health care, infrastructure and transportation excluded
One input: Sum of net costs of included activities (money)
Four different DEA models estimated. Their averages = final results for the municipalities
7
OUTPUT MEASURES IN DEA MODELS
.
1. Days in child care centers2. Days in family child care3. Open basic healthcare, visits4. Dental care, visits5. Bed wards, basic health care 6. Institutional care of elderly 7. Care of the handicapped, days8. Hours of teaching in comprehensive schools9. Hours of teaching in senior secondary schools10. Loans from municipal libraries
8
DEA results: Municipalities in order of efficiency 1994-2002
40
50
60
70
80
90
100
effi
cien
cy,
%
9
Efficiency of basic service provision during 1994-2002 (left) and 2000-2002 (right)
10
Municipalities with highest efficiencies
• DEA Popul- Province UEUrb. % ation raterate
• 1 Rusko 99,5 3 300 Vars.-Suomi 10,2 73• 2 Raisio 99,2 22 800 Vars.-Suomi 13,3 98• 3 Toijala 98,8 8 100 Pirkanmaa 17,8 96• 4 Kihniö 98,3 2 500 Pirkanmaa 19,1 34• 5 Lemi 96,6 3 100 Et.-Karjala 15,1 44
• 6 Karjaa 96,5 8 800 Uusimaa 13,3 81• 7 Masku 96,4 5 300 Vars.-Suomi 9,0 80• 8 Nakkila 96,4 6 100 Satakunta 16,2 75• 9 Karkkila 96,1 8 700 Uusimaa 14,0 86• 10 Rautjärvi 96,0 4 800 Et.-Karjala 17,0 58• Median 87,2 6 000 15,7 60
11
Municipalities with lowest efficiencies DEA Popul- Province UE
Urb. % ationrate rate
• 344 Vuolijoki 73,7 2 900 Kainuu 22,1 51• 345 Suomussalmi 73,2 11 400 Kainuu 28,4 56• 346 Puolanka 69,5 4 000 Kainuu 23,6 52• 347 Sodankylä 68,6 10 200 Lapland 26,9 59• 348 Kuusamo 67,7 18 200 P.-Pohjanm. 22,9 63• 349 Kolari 67,5 4 200 Lapland 26,8 45• 350 Inari 66,1 7 600 Lapland 25,2 63• 351 Enontekiö 65,6 2 300 Lapland 29,6 42• 352 Muonio 65,4 2 600 Lapland 21,4 55• 353 Kittilä 62,4 6 000 Lapland 24,5 49
• Median 87,2 6 000 15,7 60
12
DEA efficiency, 10 largest cities
• DEA Ra- Popul- Income/ Coll. Un- • score ting ation person, degree empl. • % /353 € eduction rate• % %
• Helsinki 79,4 317 543 000 20 300 32,3 13,2• Espoo 82,5 268 205 000 24 100 39,5 9,5• Tampere 91,4 65 190 000 16 900 27,5 18,4• Vantaa 83,8 242 173 000 19 900 26,1 11,5• Turku 84,4 233 170 000 16 500 25,7 17,7• Oulu 90,6 80 116 000 17 400 30,3 17,9• Lahti 91,2 68 96 000 15 600 21,3 20,8• Kuopio 91,1 70 86 000 16 000 27,5 17,9• Jyväskylä 80,6 303 77 000 16 200 29,5 20,4• Pori 89,8 100 76 000 15 400 20,6 21,9
• Median 87,2 177 6 000 13 600 15,5 15,7
13
Second stage (regression) results; without city managers
Variable explained: efficiency scores (range 63-100)
Summary of factors which were
theoretically worth and possible to test with data available and
statistically significant in explaining efficiency differences
14
Results: Factors improving cost efficiency of municipal services
High education level (maximum effect on efficiency almost 10 %)
Dense urban structure ( max. effect 4 %)
Employees 35-49 years of age (2 %)
Using private sector as producer? (about 2 %)
15
Results: factors weakening cost efficiency
High income (labour cost) level (-10 %)
Big population (-12 % for Helsinki)
Peripheral location (-20 % for Lapland)
Wide variety of provided services (-5 %)
High unemployment (-5 %)
Purchases from joint munic. organisations, (-3%)
Large state (matching) grants in beginning of study, (-10%)
during lump sum grant era no effect
16
Factors which were statistically insignificant
Size of central (core) municipality relative to surrounding ones in the functional area Change in population size (5-year relative change)
Tax revenue per inhabitant
Political variables - party composition or its dispersion in municipal councils - turnover in municipal elections
17
Satisfaction with services (scores 1-5) in 2005 and DEA-efficinecy score (2000-2002) in 30
municipalities
75
80
85
90
95
100
3,4 3,5 3,6 3,7 3,8 3,9
satisfaction score
eff
icie
ncy s
co
re
R=+0,370
18
Do CEOs or City Managers matter?
Previous studies on
CEOs
Public sector leaders
City managers (loc. gov. CEOs)
19
Do CEOs in private firms matter?
Some examples of previous studies
Family and related shocks (Bennedsen et al.)
Gender and age (Kotiranta et al.)
Leadership type (Waldman et al. )
20
Does City Manager matter ?
City Manager vs. Elected City Mayors (many studies)- Hayes and Chang (1990), Deno and Mehay (1987), O’Brien (1995)
- In Kreft (2007) and Eliakonopov (2007) City Managers more efficient
Gender: impact on leadership and performance - No differences: Donnell and Hall (1980), Duerst-Lahti and Johnson
(1992), Karsten (1994) and Duerst-Lahti and Kelly (1995)
- Women better performance: Johansen (2007) on schools sup.intendents
Theory of career concerned public sector managersTirole (1994), Dixit (1994, 2002), Dewatripont (1999), Besley (2004) Implisiittiset sopimukset ja kannustimet (career concerns) tärkeitä
21
What is the Finnish City Manager?
Governance structure of municipalities electorate (voters) council government (executive board)
- all parties in council are proportionately represented in government
city (municipal) managerNote:
chairs of council and government are lay politicians (not the same person)
city managers are civil servants elected by councils, not political mayors elected by voters
22
Does Finnish City Manager Matter? City managers have quit a lot power
If they are W. Niskanen type bureaucrats, then we expect them to create slack (inefficiency) and all positive factors are used to enhance slack
If they are career concerned bureaucrats, then we expect them to serve the public by providing
services efficiently especially early in career and all positive factors like education, networks etc
are used to enhance efficiency Problem: we don’t know which type they are
If all or most of them are of the same type, we should get some results
23
U.Di.T.E Leadership study data
52 questions to all mainland city managers in mainland Finland,
Response rate 74 %
Information on following types of variables education, gender, party, tenure,… working environment, network, conflicts … attitudes towards reforms etc
Next, some of these were tested preliminarily.
24
Explaining DEA efficiency of Finnish municipalities with city
manager characteristics, OLS 1994-1996 (page 1).
Model A Model B
1994 1995 1996 1994 1995 1996
Constant 75.9 79.4 82.8 79.5 75.5 78.0
(9.4***) (10.2***)(10.7***) (14.5***) (13.8***) (12.6***)
Structural factors:
population, 1000 -0.030 -0.033 -0.027
(-3.92***) (-7.14***) (-4.38***)
unemployment rate, % -0.278 -0.294 -0.314
(-1.84*) (-2.11**) (-2.59**)
education level of 0.070 0.046 0.049
population, index (4.28***) (3.15***) (3.35***)
City manager’s assessment of -0.425 -0.476 -0.539
structural factors (-2.07**) (-2.23**) (-2.99***)
25
Explaining DEA efficiency of Finnish municipalities (page 2)
Model A Model B
1994 1995 1996 1994 1995 1996
Location, physical structure:
economic distance*2 -0.024 -0.020 -0.017 -0.028 -0.027 -0.025
(-3.21***) (-3.22***) (-3.33***) (-4.52***) (-4.91***)(-6.36***)
urbanization rate, % 0.0013 0.0011 0.0013
(4.50***) (4.43***) (5.17***)
Local political structure:
left parties in -0.133 -0.162 -0.114 -0.155 -0.181 -0.149
municipal council, % (-3.01***) (4.14***) (-2.89***) (-3.69***) (-4.96***) (-4.00***)
concentration of -0.046 -0.137 -0.129
party structure in (-0.60) (-2.03**) (-1.85*)
municipal council, index
26
Explaining DEA efficiency of Finnish municipalities ( page 3)
Model A Model B
1994 1995 1996 1994 1995 1996
Characteristics of CM:
gender (m=0,f=1) -3.43 0.489 -0.65 -3.67 1.09 0.095
(-1.55) (0.17) (-0.28) (-1.56) (0.37) (0.04)
education level, years 0.166 0.407 0.297 0.143 0.375 0.272
(1.16) (3.07***) (2.16**) (0.97) (2.91***) (1.99**)
Attitudes of CM towards:
worker participation in 1.28 1.34 1.28 1.60 1.74 1.63
organizational change (1.41) (1.83*) (1.67*) (1.95*) (2.49**) (2.28**)
relative efficiency -0.293 -0.649 -0.780 -0.439 -0.675 -0.795
of public sector (-0.75) (-1.81*) (-2.23**) (-1.14) (-1.95*) (-2.27**)
co-operation 0.119 0.179 0.149 0.086 0.160 0.133
with other actors (1.97*) (3.18***) (2.66***) (1.46) (2.77***) (2.25**)
27
Explaining DEA efficiency of Finnish municipalities (4)
Model A Model B
1994 1995 1996 1994 1995 1996
CM’s assessment of negative
factors’ effect on his work
Various factors, -0.268 -0.110 -0.210
index (a) * (-2.93***) (-1.28) (-2.48**)
Group contradictions, -0.895 -0.072 -0.228
index (b) ** (-3.01***) (-0.26) (-0.90)
R2 (adj) 0.274 0.306 0.320 0.256 0.284 0.309
Ramsey 0.102 0.469 0.154 0.543 0.147 0.190
Jarque-Bera 0.078 0.133 0.028 0.082 0.277 0.039
Max VIF 1.79 1.94 2.11 1.51 1.51 1.51
N 192 192 192 192 192 192
* (a) Includes organizational, political and bureaucratic problems, contradictions between actor groups, problems in
organising work and lack of clearly defined political goals.
** (b) Contradictions between political parties, contradictions between sector managers and city offices.
28
On the basis of CM variables
CM’s ’s education level had a positive impact on efficiency
NO effect was found for Gender, age, political party membership
CM’s positive attitude towards workers’
participation in decision making (especially under organizational change) had a postive relation to efficiency.
29
CM’s positive attitude and experience concerning
cooperation with other stakeholders was positively related
to efficiency Partners: local politicians, other city management and
employees, state and regional administration
representatives, business firms, trade unions and media.
Municipalities where the CM regarded the private sector
generally more efficient than the public sector were more
efficient.
30
Efficiency was lower in municipalities where the CM saw a lot of
Contradictions between parties and municipalities and their government
Byrocracy and work organization problems
31
Factors having no effect on DEA efficiency
age of city manager
number of years as CM
length of work week
membership in local organizations, trade unions or political parties
living in the municipality now or as a child
recent privatization or centralization of service activities
years of the chair of municipal board in his task or his membership in municipal council
planning to look for work elsewhere
experiencing personal contradictions with other actors
large number of other actors influencing the budget
32
THESE WERE OUR VERY FIRST RESULTS, reported also in Kunnallistieteellinen aikakauskirja 4/2008 in Finnish, and there is also a conference paper in English, available upon request
More later
Thank you
Where is my wine??