Political and financial factors motivating privatisation, its reverse and corporatisation of municipal waste collection

ABSTRACT

This paper analyses different modes of provision for household waste collection and changes in that by Dutch municipalities in the period between 1998 and 2018. Based on a transition matrix, the shifts in the mode of waste collection are made transparent. In more than half of the municipalities, shifts took place, 56% towards outside production and 44% towards inside production and in later periods, inside production has become important. Furthermore, in the most recent time, reverse privatisation is preferred. We find that ideological factors have an influence on choices of provision modes and that municipalities with a more political fragmented board are less likely to corporatize or privatise waste collection. Finally, there is evidence that municipalities with a relatively negative financial position are more likely to go outside and that in pre-election years there is some hesitation to corporatize.

KEYWORDS:

• Corporatisation
• privatisation and its reverse
• political fragmentation
• financial position municipality
• municipal waste collection

1. Introduction

In Europe, also due to fiscal pressure, provision of public services has been privatised in many municipalities since the 1980s. However, in the last decade, there is increasing evidence of a trend in the opposite direction. At the beginning of this century, Bel and Costas (2006) indicate that privatisation of solid waste collection has become rusty. Hall et al. (2013) claim that there was extensive re-municipalisation in the water sector in France and the energy sector in Germany, which can be analysed as a ‘pendulum’ reflecting political struggles. However, this evidence is anecdotal, and the authors point out that empirical studies of re-municipalisation across different sectors can be helpful to advance the practical knowledge of policymakers. In an overview of empirical literature, Clifton et al. (2021) show that re-municipalisation is part of a broader set of reverse privatisation and indicate that re-municipalisation is generally part of a pragmatic market management process. In this literature, there is a growing attention for corporations, which maintain majority public ownership but put the service on an enterprise basis (Andrews et al. 2020; Voorn et al. 2021). Also, for Dutch refuse collection, this form seems to be dominant (Gradus and Budding 2020).

Tavares and Camöes (2007) show for Portugal that services as refuse collection involving the collection of user fees are more likely to be transferred to municipal corporations. Also, in other countries as Germany and Italy, municipality-owned firms or corporations have been noted as an important delivery mode, where water, energy and waste disposal are the most frequent areas of corporative activities (Grossi and Reichard 2008). The Netherlands also faces a move towards municipality-owned firms or corporatisation for fee-financed municipal waste management. Voorn and Van Genugten (2022) have been examining a sample of six Dutch cases and find that municipal corporations are an appropriate delivery mode for these municipalities.

In this paper, we analyse different modes of provision for household waste collection and changes in that by Dutch municipalities. The basis of our analysis is a transition matrix, where the mode of production is ordered from inside towards outside production (see Hefetz and Warner 2004). This method enables us to compare stability and shifts in forms of service delivery. Furthermore, we use political, financial, and socio-economic characteristics to explain shifts. Thereby, we focus on political transaction costs in local service delivery, linking political and financial features to these shifts (Rodrigues, Tavares, and Araújo 2012; Tavares and Camöes 2007). In addition, as Andrews et al. (2020) suggest that local governments may use corporate entities to meet the demands of different age groups to raise additional revenue, we include several socio-economic variables such as ethnicity and the share of elderly people.

Our study aims to add to the literature in at least five ways. First, the dataset we use covers a longer time period (20 years; 1998–2018) than previous studies, and therefore, we are able to analyse changes in time periods with varying (political, socio-economic, etc.) characteristics. Furthermore, we do not only include corporatisation and privatisation as changes in production modes but also cooperatisation, reverse in-housesation, and reverse privatisation. Second, we aim to contribute to the debate about the influence of adverse financial situations on contracting out decisions. Whereas the dominant view in public administration literature is that poor financial conditions lead to the adoption of external arrangements rather than in-house solutions (López-Hernández et al. 2018; Warner 2023), and some studies indeed find results that confirm this hypothesis (e.g., Gradus and Budding 2020; Plata-Díaz et al. 2019), findings of other studies point into the opposite direction (Rodrigues, Tavares, and Araújo 2012; Warner and Hefetz 2012). Third, our study aims to provide further insight into the influence of ideology on production mode decisions. Based on a review of empirical studies explaining privatisation, Bel and Fageda (2007) find that most of them found no or only a weak influence of ideology. However, for the European context, there is mixed evidence (cf. Gradus and Budding 2020; de la Higuera-Molina et al. 2022). Therefore, the influence of ideology on production mode decisions is still an open issue. Fourth, as previous studies (e.g., Schoute, Gradus, and Budding 2021) find some indications that political fragmentation decreases the ability to take outsourcing decisions, but findings are generally statistically weak, we aim to deepen the analysis on this issue by using three variants of fragmentation indices. Fifth, and finally, we would like to contribute to the studies that take the influence of the political cycle into account. De la Higuera-Molina et al. (2022) hypothesise and find proof that in preelection years, there is more opposition again privatisation, and as a result, a lower probability to privatise or corporatize and in postelection years, this probability is higher. We further analyse this issue by investigating differences in production mode decisions in all (four) years in the election cycle.¹

The article is structured as follows. First, we describe the shifts over time. This is followed by a section that introduces our data and hypotheses. After that, logit analyses explaining the shifts are presented. Finally, we offer some concluding remarks and topics for future research.

2. Different production modes and shifts between them

Dutch municipalities have a legal obligation to provide a waste-collection infrastructure for household waste. Municipalities are free to choose whether to carry out this task in-house, to do it in cooperation with other municipalities or to contract it out to outside contractors (municipal or private enterprise). This decision is made by the municipal council, which is elected every 4 years by the inhabitants. Of all observations in our 1998–2018 data set, 35% represent waste collection contracted out to a private firm and 26% to a municipal enterprise. Municipal enterprises operate under commercial law, while their shares are owned by municipalities. A third group of observations (15%) represent collection in cooperation with other municipalities – labelled ‘“IMC”’ in the figures and tables that follow; they are organised as public WGR (Law on Common Arrangements) entities, where the (executive) board is directed by the mayors and aldermen of participating municipalities (see Bel et al. 2010). A fourth group of observations (3%) has the task executed by the public entity of a neighbouring municipality – labelled ‘neighbour’ in the figures and tables. This is a kind of cooperation but different from a WGR, which is a separate legal entity. The remaining observations (20%) represent in-house collection by municipalities themselves. In the last three cases, the services are run publicly and thus not under commercial law. The shares of these five modes of Dutch waste collection between 1998 and 2018 are illustrated in Figure 1.

Figure 1. Shares, for different collection modes, 1998–2018.

Remarkably, the share of corporations increased substantially from 5% in 1998 to 35% in 2011 and stayed relatively stable afterwards. The share of municipalities collecting in-house decreased from 31% in 1999 to 17% in 2018. In 1998, 43% of municipalities used private firms, while this percentage was 26% in 2018. This decrease was at least partly the result of the merging of small municipalities, as private collectors were especially active in these environments. IMC decreased slightly from 18% in 1998 to 12% in 2010 but increased after to 19% in 2018. Finally, the percentage of municipalities that hire a neighbouring municipality is rather stable, about 5% in all years. So, the percentage of municipalities operating under commercial law (private and corporation) increased in the Netherlands from 48% to 60% over the period.

Shifts

Like Hefetz and Warner (2004), Gradus and Budding (2020) and Warner (2023), a transition matrix is constructed, where the mode of production is ordered from left to right (and from top to bottom) towards outside production. This matrix includes all shifts, even if a municipality shifts several times in the period 1998–2018. Importantly, the number of municipalities decreased from 548 in 1998 to 380 in 2018. In the Netherlands, mergers take place gradually every year.² This method enables us to compare stability in forms of service delivery and to assess shifts towards outside production or towards inside production. We define shifts to inside as shifts towards more inside production (i.e., from neighbour, IMC, corporation and private firm to in-house, from IMC, corporation and private firm to neighbour, from corporation and private firm to IMC, and from private firm to corporation) and shifts to outside as shifts towards more outside production (i.e., from in-house to neighbour, IMC, corporation and private firm, from neighbour to IMC, corporation and private firm, from IMC to corporation and private firm, and from corporation to private firm).

Shifts from 1999 to 2018 are available for 564 Dutch municipalities. Three hundred and two Dutch municipalities (i.e., 54%) shifted to a different mode of production one or more times. In total, there were 637 switches, of which 309 (italics in Table 1) were towards outside production or to the market, and 241 (bold in Table 1) were towards inside production or from the market. Thus, 56% of shifts were towards outside production, and 44% towards inside production. Moreover, corporatisation was important as 206 such shifts took place (37% of all shifts). In addition, in 262 municipalities, the mode of production did not change during the period 1999–2018 (this is the sum of the numbers on the diagonal). As a result, the mode of production was stable over this period in less than half of the Dutch municipalities.

Table 1. Number of municipalities switching mode and staying the same, 1999–2018.

		To					Number shifts	Observations
		In-house	Neighbour	IMC	Corporation	Private
From	In-house	63	11	28	62	38	139	202
	Neighbour	6	4	11	4	14	35	39
	IMC	17	10	41	51	21	99	140
	Corporation	8	0	4	36	69	81	117
	Private	33	14	60	89	118	196	314
Number shifts		64	35	103	206	142	550	-
Observations		127	39	144	242	260	-	812

Numbers in italics indicate shifts towards outside production and in bold shifts towards inside production.

Let us focus in more detail on both privatisation and reverse privatisation (see the first four items in the fifth column and the first four items in the fifth row of Table 1). Remarkably, there were 142 shifts to privatisation, but there were also 196 shifts away from privatisation (i.e., reverse privatisation). Interestingly, almost half of these reverse shifts (89) are towards corporations. So reverse privatisation is far more important than privatisation in 1998–2018. Finally, it should be noted that remunipalisation (from the private to the public domain) is small with 119 shifts (i.e., 22%).

Pattern over time

The shifts in four equal periods (1999–2003, 2004–2008, 2009–2013, and 2014–2018) are given in Table 2. It should be noticed that the number of shifts in 2014–2018 is only 93, which is small, compared with 135, 191 and 131 in 2009–2013, 2004–2008 and 1999–2003, respectively. Interestingly, the relative number of shifts towards outside production (bold in Tables) is dropping from 62% in 1999–2003 to 35% in 2014–2018. Interestingly, the dominant trend towards outside production in the first three periods has changed into towards inside production in the last period.

Table 2. Number of municipalities switching mode and staying the same, 1999–2018.

		To
		In-house	Neighbour	IMC	Corporation	Private	Number shifts	Observations
Panel A: 2014–2018
From	In-house	53	0	0	8	5	13	66
	Neighbour	2	13	3	0	0	5	18
	IMC	2	3	54	2	1	8	62
	Corporation	2	0	1	108	12	15	123
	Private	9	4	19	20	94	52	146
Number shifts		15	7	23	30	18	93	–
Observations		68	20	77	138	112	–	415
Panel B: 2009–2013
From	In-house	67	3	8	8	6	25	92
	Neighbour	1	9	3	2	1	7	16
	IMC	6	0	42	5	12	23	65
	Corporation	3	0	1	119	27	31	150
	Private	4	5	16	24	119	49	168
Number shifts		14	8	28	39	46	135	–
Observations		81	17	70	158	165	–	491
Panel C: 2004–2008
From	In-house	78	4	9	19	15	47	125
	Neighbour	2	10	1	0	7	10	20
	IMC	6	3	52	35	7	51	103
	Corporation	3	0	2	78	25	30	108
	Private	11	4	15	23	139	53	192
Number shifts		22	11	27	77	54	191	–
Observations		100	21	79	155	193	–	548
Panel D: 1999–2003
From	In-house	106	4	11	27	12	54	160
	Neighbour	1	7	4	2	6	13	20
	IMC	3	4	73	9	1	17	90
	Corporation	0	0	0	37	5	5	42
	Private	9	1	10	22	173	42	215
Number shifts		13	9	25	60	24	131	–
Observations		119	16	98	97	197	–	527

Numbers in italics indicate shifts towards outside production and in bold shifts towards inside production.

Trends

We distinguish different (reverse) shifting trends. Hereby, we combine IMC entities, where the executive board is directed by the mayors and aldermen of participating municipalities and execution by a neighbouring municipality as they have similarities with respect to governance. We label these four shifting (to) trends ‘in-housesation’, ‘cooperatisation’, ‘corporatisation’ and ‘privatisation’, respectively. Figure 2 shows the number of shifts to the four service delivery modes (from any other mode) as a percentage of the total number of shifts in different periods.

Figure 2. Shifting trends 1999–2018 (left) and sub-periods.

Overall, corporatisation (with 36% of the shifts) is the most important trend, and in-housesation (with 12%) is the less important trend. For the different sub-periods, in-housesation, corporatisation and cooperatisation are of almost equal size. For privatisation, we see a drop from 35% in 2009–2013 to 18% in 2014–2018. Overall, re-municipalisation (in-housesation and cooperatisation) has become more important in later periods (in 2014–2018 50%).

Figure 3 shows the number of reverse shifts from the four service delivery modes (to any other mode) as a percentage of the total number of shifts in different periods. We label these four reverse shifting (from) trends as ‘reverse in-housesation’, ‘reverse cooperatisation’, ‘reverse corporatisation’ and ‘reverse privatisation’, respectively.

Figure 3. Reverse shifting trends 1999–2018 (left) and sub-periods.

Overall, reverse in-housesation (with 33%) and reverse privatisation (32%) are the most important reverse trends. Moreover, the number of shifts from corporatisation (13%) is very small, so corporatisation seems a stable trend. For the different sub-periods, we see a large increase for reverse privatisation to 54% in 2014–2018. As a result, more than half of all shifts in the last period was re-privatisation.

3. Data and hypotheses

Like Warner and Hefetz (2012) and Gradus and Budding (2020), we apply a logit model to test theories about shifting trends. The dependent variable takes the value 1 in years (and 2 years before and after) in which the municipality shifts the mode of production (e.g., to corporatisation) and zero otherwise. Five models are estimated, differing with respect to the trend (or direction of the shifts). We choose the three dominant trends in Figure 2 (corporatisation, privatisation and cooperatisation) and the two dominant trends in Figure 3 (reverse privatisation and reverse in-housesation).

Hypotheses

The dominant view in public administration literature is that adverse financial situations lead to the adoption of external arrangements rather than in-house solutions (López-Hernández et al. 2018; Warner 2023). Gradus and Budding (2020) indeed find some evidence that municipalities with high debt have a higher probability to contract out. Plata-Díaz et al. (2019) demonstrate the existence of a wide-ranging and negative relationship between the elements of financial condition (poor liquidity, absence of flexibility, budget deficit) and the contracting out. However, some empirical findings suggest that contracting out at the local level may actually be less prevalent if there is financial stress (Warner and Hefetz 2012). Rodrigues et al. (2012) point out that local public sector unions may be especially vigorously opposed to contracting out during economic decline in fiscal crises. Therefore, it is interesting to extent the period from 2015–2018, as in aftermath of the European crisis (2011–2012) there were budget cuts in the Netherlands and municipalities with a weak financial position are more likely to choose external solutions.

Based on a review of empirical studies explaining privatisation, Bel and Fageda (2007) find that most of them found no or only a weak influence of ideology. This is confirmed by Warner and Aldag (2021) who show that only 21% of respondents in their survey reported ‘there was strong political support to bring back the service delivery’ and this is the fifth ranked factor. However, for the European context, there is mixed evidence. For Spain, Zafra-Gómez et al. (2016) find that progressive parties are reluctant to initiate privatisation processes, but de la Higuera-Molina et al. (2022) observe that progressive parties, in fact, are more likely to favour contract renewal. For the Dutch setting, Gradus and Budding (2020) find some evidence for political preferences as an explanation for changing the mode of production. Councils with many conservative liberals are in favour of change, particularly towards the market and privatisation, whereas councils with many social democrats are against change (see also Schoute, Gradus, and Budding 2021). Gradus and Dijkgraaf (2019) find strong evidence that political fragmentation is negatively related to the implementation of tough environmental decisions as unit-based pricing systems. Especially for the physical domain (with services such as the maintenance of roads/green spaces and garbage collection), Schoute et al. (2021) give indications that political fragmentation decreases the ability to outsource. In this article, we analyse whether privatisation and corporatisation decisions of waste management are related to fragmentation of the council or board. Finally, following de la Higuera-Molina et al. (2022), we hypothesise that in preelection years, there is a lower probability to privatise or corporatise, and in postelection years, the probability is higher.

For the political variables, it should be noticed that for subsequent years without an election, the results of the last election are taken. In the year of the election (1998, 2002, 2006, 2010, 2014, and 2018), the previous election is taken. Therefore, the percentage of seats of the following national parties, based on the local council elections in 1998, 2002, 2006, 2010, and 2014 is included: Social Democrats (PvdA), Conservative Liberals (VVD), Christian Democrats (CDA), Progressive Liberals (D66), Green Left (GL), Socialist Party (SP), Christian Union (CU), and Reformed Political Party (SGP). For each political party, its number of seats divided by the total number of seats on the municipal council is included. We combine both small Protestant parties (CU and SGP) in one variable. Furthermore, in our tables, these political parties are ordered from the right to the left.³ The so-called independent local parties are used as a benchmark as in general they are difficult to pinpoint on a left–right scale. Moreover, as on a daily basis, aldermen form the executive board of a municipality, the share of aldermen from each political party is also included as an explanatory variable.

In this paper, we measure fragmentation in three ways. First, it is operationalised as the number of political parties in the council as well as in the board of aldermen divided by the number of seats in the council and the total number of aldermen (Goeminne, Geys, and Smolders 2008; Gradus and Budding 2020). Second, it is calculated as the effective number of political parties by taking the quadratic sum of the relative size of parties in the municipal council and the board of mayor and aldermen. In political science, this is often called the Laakso – Taagepera index. Third, and finally, it is (simply) operationalised by looking at the number of absolute parties in these entities. In our calculations, independent local parties were included, the latter being defined as a party not represented in the National Parliament. Note that in the Netherlands, the local political landscape became more fragmented over time: the number of absolute parties rose from 6.5 in 1999 to 8.0 in 2018. For the effective number of parties, a similar effect can be observed ((increased from 4.8 in 1998 to 6.0 in 2018).

Therefore, we evaluate the following hypotheses:

1. Municipalities with a negative financial position are more likely to shift to corporatize or privatise.

2. Municipalities of which the political bodies (i.e., the municipal council or the board of aldermen) have a more right-wing orientation are more likely to corporatize or privatise.

3. Municipalities of which the political bodies (i.e., the municipal council or the board of aldermen) are less fragmented are more likely to corporatize or privatise.

4. In preelection years, there is a lower probability to privatise or corporatize and in postelection years, there is a higher probability to privatise or corporatize.

Control variables

We use several control variables. Andrews et al. (2020) indicates that socioeconomic factors may be more important predictors of corporatisation because they can retain control over the companies that they create. In addition, Gradus and Budding (2020) shows that a municipality is less willing to privatise if unemployment in a municipality is high. Therefore, we include four socio-economic variables: household size, unemployment, share of elderly people (above 65 years) and ethnicity (defined as the share of non-western inhabitants). We include a year trend variable to test whether there is a persistent trend. The reason for this is that there might be societal trends pro or con a specific mode of production. For instance, the general opinion about privatisation or corporatisation might influence the mode of production.

Data

Data for the number of inhabitants, population density, municipal debt, income, waste collection modes and other municipal characteristics come from Statistics Netherlands. Data about municipal political characteristics (seats and aldermen) in the election years 1998, 2002, 2006, 2010 and 2014 were collected from the Gids Gemeentebesturen (Guide to Municipal Boards) (Gradus, Dijkgraaf, and Budding 2023). See Table 3 for descriptive statistics. The number of observations is 6,300.

Table 3. Descriptive statistics.

	Mean	Maximum	Minimum	Std. Dev.
Corporatisation	0.12	1.00	0.00	0.32
Cooperatisation	0.06	1.00	0.00	0.23
Reverse in-housesation	0.09	1.00	0.00	0.28
Privatisation	0.09	1.00	0.00	0.29
Reverse privatisation	0.10	1.00	0.00	0.30
Inhabitants (/10,000)	3.87	83.36	0.09	6.28
Household size	2.42	3.67	1.66	0.21
Population density	0.71	5.35	0.00	0.86
Elderly people (share)	0.16	0.31	0.06	0.03
Ethnicity (share)	0.05	0.38	0.00	0.05
Income per inhabitant (€,/10,000)	13.94	24.90	7.80	2.00
Unemployment	0.02	0.11	0.00	0.01
Municipal debt per inhabitant (€,/1,000)	1.90	11.68	0.00	1.40
Fragmentation I: Parties aldermen/parties	0.92	2.00	0.33	0.19
Fragmentation II: Effective number of aldermen	3.00	8.33	1.00	0.82
Fragmentation III: Absolute number of parties in board	3.11	7.00	1.00	0.86
Aldermen_VVD (share)	0.16	0.75	0.00	0.17
Aldermen_CDA (share)	0.25	1.00	0.00	0.19
Aldermen_CU/SGP (share)	0.07	1.00	0.00	0.15
Aldermen_D66 (share)	0.04	0.50	0.00	0.10
Aldermen_PvdA (share)	0.18	1.00	0.00	0.19
Aldermen_GL (share)	0.04	1.00	0.00	0.09
Aldermen_SP (share)	0.01	0.60	0.00	0.06
Fragmentation I: Parties council/Council seats	0.33	0.71	0.13	0.08
Fragmentation II: Effective parties council	5.28	11.31	1.69	1.35
Fragmentation III: Absolute number of parties council	6.92	16.00	3.00	2.03
Council_VVD (share)	0.15	0.48	0.00	0.08
Council_CDA (share)	0.21	0.74	0.00	0.09
Council_CU/SGP (share)	0.07	0.59	0.00	0.11
Council_D66 (share)	0.04	0.31	0.00	0.06
Council_PvdA (share)	0.16	0.54	0.00	0.10
Council_GL (share)	0.04	0.38	0.00	0.05
Council_SP (share)	0.02	0.43	0.00	0.05

4. Estimation results

Table 4 shows the estimation results, hereby presenting fragmentation index I based on the number of parties divided by the number of seats. In the appendix, the estimation results for the other fragmentation indexes (II: effective and III: absolute number of parties) are shown.

Table 4. Estimation results.

	Corporatisation		Privatisation		Cooperatisation		Reverse in-housesation		Reverse privatisation
	Council	Aldermen	Council	Aldermen	Council	Aldermen	Council	Aldermen	Council	Aldermen
Inhabitants	−0.05*	−0.10***	0.06**	0.04	−0.03	−0.04	0.08*	0.04	0.03	0.00
Inhabitants^2	0.001**	0.001***	0.000	−0.000	−0.001	−0.000	−0.004*	−0.003	0.000	0.000
Household size	−0.60	−0.94***	2.43***	2.48***	−0.31	−0.23	−0.73*	−0.55	2.24***	2.02***
Population density	0.16**	0.13**	−0.07	−0.09	0.16*	0.12	0.08	0.09	−0.12	−0.15*
Elderly people	3.73*	3.23*	3.94	3.24	−9.23***	−9.40***	2.82	3.27	6.85***	6.64***
Ethnicity	−3.35*	−2.52	−1.81	−0.34	3.679	2.39	−2.70	−2.44	−1.62	−1.37
Income per inhabitant	−0.11***	−0.07**	0.18***	0.18***	0.02	0.01	−0.05	−0.03	0.12***	0.12***
Unemployment	−52.54***	−58.80***	−12.75	−19.75*	−42.36***	−44.25***	−58.82***	−56.58***	−10.90	−18.13*
Municipal debt per inhabitant	0.02	0.02	0.03	0.03	−0.04	−0.04	0.17***	0.17***	0.03	0.01
Fragmentation council I	0.75	-	0.53	-	0.85	-	−0.82	-	1.63**	-
Council_VVD	2.22***	-	2.00***	-	0.58	-	1.50**	-	0.12	-
Council_CDA	−1.39***	-	−1.08*	-	2.43***	-	1.00	-	−0.28	-
Council_CU/SGP	−0.26	-	0.15	-	−1.12*	-	1.58***	-	−0.86**	-
Council_D66	0.02	-	−2.37**	-	−3.54***	-	0.17	-	−4.32***	-
Council_PvdA	0.10	-	−0.20	-	−0.96	-	0.84	-	−1.32**	-
Council_GL	−1.73**	-	−2.01**	-	−2.32*	-	−0.88	-	−0.08	-
Council_SP	−1.10		3.62***		−0.29		0.16		1.99**
Fragmentation board I	-	−0.46*	-	−0.51*	-	0.27	-	−1.09***	-	−0.15
Aldermen_VVD	-	0.07	-	0.05	-	−0.53	-	0.97***	-	−1.44***
Aldermen_CDA	-	−0.14	-	−0.59**	-	0.32	-	0.61**	-	−0.57**
Aldermen_CU/SGP	-	0.17	-	−0.50	-	−0.80*	-	1.73***	-	−0.88***
Aldermen_D66	-	1.25***	-	0.15	-	−1.22*	-	0.21	-	−0.01
Aldermen_PvdA	-	−0.15	-	0.00	-	−0.82**	-	0.99***	-	−0.49*
Aldermen_GL	-	−0.48	-	−0.45	-	−1.64**	-	−0.81	-	−0.91
Aldermen_SP	-	0.89		1.30*		−0.69		0.87		1.12*
Year trend	−0.05***	−0.06***	−0.05***	−0.04**	0.10***	0.09***	−0.09***	−0.10***	−0.01	−0.01
Constant	1.50	2.90**	−11.03***	−10.15***	−1.74	−1.06	0.55	0.61	−10.29***	−8.50***

***, **, * indicate significance at the 0.01, 0.05 and 0.10 levels, respectively.

When corporatisation is estimated, inhabitants squared, household size, population density, income, unemployment, and several political variables are significant (at more than 5% level). The population variables are significant at the 1% level in the board estimation. Therefore, we found a U-relationship of managerial capabilities (measured with population) with corporatisation (Andrews et al. 2020),⁴. If the number of inhabitants increases up to this minimum, the probability of corporatisation decreases; with a population higher than this maximum, it increases. The household size variable is significant and negative in the aldermen estimation: if household size increases, the probability of corporatisation decreases. For population density, the estimation is significant and positive in both estimations: if the population density increases, the probability of corporatisation increases. The income variable is significant and negative in both estimations: if income increases, the probability of corporatisation decreases. For unemployment, it is significant and negative at 1% level in both estimations, so the higher unemployment, the lower the chance that corporatisation occurs. For the political variables, conservative liberal councillors are in favour of corporatisation and Christian Democrats’ (CDA), and Green-Left (GL) councillors are against it. In addition, Progressive Liberal (D66) aldermen are in favour of corporitzation, but no effect is found for Conservative Liberals (VVD) aldemen. For fragmentation index I, we found a negative and significant effect in the board estimation (at 10%) but not in the council estimation. Fragmented boards are thus less in favour of corporatisation. These results are robust, and even statistically stronger, if we take the other fragmentation indexes (refer to Tables A1 and A2). Notice that there is a significant and negative effect for a year trend.

If the shifts to privatisation are estimated, inhabitants (for the council estimation), household size, income per inhabitant, unemployment (for the aldermen estimation at 10%) and several political variables are significant. Interestingly, the larger the household size, the higher the chance of privatisation. For unemployment, the effect is now negative at 10% for the aldermen estimation, so the higher unemployment, the lower the chance that privatisation occurs. The income variable is significant and positive in both estimations (at 1%): if income increases, the probability of privatisation increases. For political variables, Conservative Liberal’s and Socialist Party’s councillors and aldermen (at 10%) are in favour of privatisation. For the Socialist Party, this is surprising, but it should be noted that it is a small party with a seat share of 0.01. Christian Democrats’ councillors and aldermen and Progressive Liberals’ councillors are against it. Also here, for fragmentation index I, we found a negative and significant effect at 10% for the board estimation, so fragmented boards are less in favour of privatisation. However, for the other fragmentation indices, we found no significant results for the board estimation, but the fragmentation index is positive and significant for the council estimation at 5% for fragmentation index II and 10% for fragmentation index III. Notice that there is a significant and negative effect for a year trend.

If shifts to cooperatisation are estimated, the variables inhabitants, household size, income and population density are not significant anymore. Unemployment is significant at 1% level in both estimations, so the higher unemployment, the lower the chance that cooperatization occurs. The elderly variable is now significant and negative (at 1%) in both estimations: the higher the share of elderly the lower the probability for cooperatisation. For the political variables, Christian Democrats’ councillors are in favour of cooperatisation, Progressive Liberals’ councillors are against it, while the effect for Christian Union/Reformed Political Party and Green-Left is negative at 10%. For aldermen, we find negative effects for Social Democrats and Green-Left and at 10% for Christian Union/Reformed Political Party’s and Progressive Liberal aldermen. For fragmentation I, we do not find significant effects. However, for fragmentation index II (effective number of parties) we find a positive and significant result for the board estimation (at 10%) and the same is true for fragmentation index III (absolute number of parties) for the council estimation (at 1%). The other variables are also positive, but not statistically significant. We get a significant and positive year trend effect for cooperatisation (at 1% level).

If the shifts to reverse in-housesation are estimated, unemployment, municipal debt and a number of political variables are significant. Unemployment is significant at 1% level in both estimations, so the higher unemployment, the lower the chance that reverse in-housesation occurs. Debt has a positive effect on reverse in-housesation (significant at the 1% level in both estimations), so the higher the debt, the higher the chance that reverse in-housesation occurs. The number of inhabitants and household size are only significant at 10% in the council estimation. For the political variables, Conservative Liberals and Christian Union/Reformed Political Party councillors are in favour of reverse in-housesation. In addition, Christian Democrats, Social Democratic, Christian Union/Reformed Political Party and Conservative Liberal aldermen are in favour of reverse in-housesation. Also here, for fragmentation index I, we found a negative and significant effect for the board estimation (at 1%), so fragmented boards are less in favour of reverse in-housesation. For the other fragmentation indices of the board negative but insignificant effects are found. For the fragmentation of the council, we do not find statistically significant results, except fragmentation index III (refer to Table A2), where we find a positive and significant effect (at 1%). Notice that there is a significant and negative effect for a year trend.

Finally, the shifts from privatisation or reverse privatisation are given in the last column of Table 5. Now, household size, elderly people and income per inhabitant are significant. Also here, the larger the household size, the higher the chance of reverse privatisation. The elderly variable is significant (at 1% level) in both estimations: the higher the share of elderly, the lower the probability for reverse privatisation. The income variable is significant and negative in both estimations (at the 1% level): if income increases, the probability of reverse privatisation increases. For the political variables, Progressive Liberals, Social Democrats and Christian Union/Reformed Political Party councillors are against reverse privatisation. Interestingly, Socialist Party councillors and aldermen are in favour of reverse privatisation. In addition, Christian Union/Reformed Political Party, Christian Democratic, Social Democrats, and Conservative Liberals’ aldermen are against reverse privatisation. A positive effect is found for the council for fragmentation index I (at 5%), but not for the board. For the other fragmentation indices, no significant effects are found. Interestingly, only for reverse privatisation no trend effect is encountered.

Table 5. Effect dummies years before elections.

Years before election	Three	Two	One
Model with council variables
Corporatisation	−0.07	−0.05	−0.28**
Privatisation	0.08	0.20	0.12
Cooperatisation	−0.01	0.15	0.06
Reverse in-housesation	−0.03	−0.04	0.12
Reverse privatisation	0.17	0.21*	0.09
Model with aldermen variables
Corporatisation	−0.09	−0.07	−0.22**
Privatisation	0.08	0.20	0.12
Cooperatisation	0.00	0.17	0.08
Reverse in-housesation	−0.04	−0.05	0.11
Reverse privatisation	0.17	0.21*	0.11

Estimation model is equal to those in Table 4 except that trend variable is replaced by dummies years before election.

*/**/*** means significance of 10%/5%/1%.

For testing hypothesis IV, we add three dummies (i.e., three, 2 years and 1 year before the election). For clarity reasons, we only present the estimation of these dummies.

In both aldermen and council equations, there are only two coefficients significant (see Table 5). Our results show that in preelection years the likelihood of corporatisation is lower (at 1% level) and the chance of reverse privatisation is higher (at 10% level) in the second year after the election. For the other estimations, no significant effects are found.

Let us summarise the estimation results based on four hypotheses given at the end of paragraph of section 3. For reverse in-housesation, we find that municipalities with a negative financial position have a higher chance that this mode is chosen. Local Dutch governments are exploring alternative service delivery modes if debt is high and there is less in-house production. This seems to be in accordance with hypothesis I. For political parties, we found that Conservative Liberals’ councillors are in favour of privatisation and corporatisation and also reverse in-housesation and Conservative Liberal’s aldermen are in favour of reverse in-housesation. Interestingly, Socialist Party’s councillors are the only party in favour of reverse privatisation. However, we did find that Social Democrats are not in favour of reverse privatisation and as they are the bigger party (eight times as big as the Socialist Party in terms of the council and eighteen times in terms of the board), there is only very weak evidence for hypothesis II. We found that more political fragmented boards are less likely to corporatise, privatise or do reverse in-houseation. Interestingly, this result is robust for corporatisation if the other fragmentation indices are used. However, if instead of the board of aldermen the council is used, this result disappears in most cases. This can make sense if the board is more in charge of the choice of mode of waste collection. Only for reverse privatisation an effect is found of a more fragmented board, and this effect is positive. Overall, there is evidence for hypothesis III. There is only weak evidence for hypothesis IV. In preelection years, there is a lower probability to corporatize and in two years after the election, and there is a higher probability to reprivatise.

5. Conclusions and discussion

As other Western countries, the Netherlands faces a trend towards municipality-owned firms or corporatisation and cooperatisation. Full administrative data for waste collection between 1998 and 2018 shows that almost 40% of the shifts was corporatisation and there was not much reverse corporatisation. Therefore, the number of municipalities with a corporation increased substantially from 5% in 1998 to 35% in 2011 and stayed more or less constant afterwards. Remarkably, reverse privatisation is far more important than privatisation in 1998–2018. Moreover, reverse privatisation has increased substantially especially during 2015–2018.

In this study, we obtain empirical evidence in support of some influence of ideological factors, political fragmentation and opportunistic behaviour. First, we found that Conservative Liberal councillors are in favour of corporatisation and privatisation, which is in line with earlier studies (Gradus and Budding 2020). Second, we found that municipalities with a more political fragmented board are less likely to corporatise or privatise waste collection. This confirms Goeminne et al. (2008) who show that political fragmentation decreases the ability to take ‘tough’ decisions on budget. For corporatisation, this conclusion is robust if other fragmentation indices are used. Third and in line with de la Higuera-Molina et al. (2022), in pre-election years there is some hesitation to corporatise. In addition, local Dutch governments are exploring alternative service delivery modes if debt is high as there is less in-house production. This can be an indication that efficiency arguments still play a role in the Netherlands. However, we are not able to determine which alternative mode should be chosen as for privatisation or corporatisation such effect for debt is not found. Overall, the negative sentiment towards privatisation and corporatisation seems more important as also estimations of year trend are showing.

There are many avenues to explore for future research. First, whereas we included several elements that can have an influence on choices of production modes, other elements can also play a role. Perhaps, personal characteristics of key actors in the decision-making process can have an influence, such as their age and years of experience (cf. Andrews et al. 2020). The choice of these personal characteristics can be based upon insights from the Upper Echelon Theory, which basically argues that personal characteristics can have an impact on how organisations operate, and which decisions are made (Hambrick and Mason 1984). Second, corporatisation and cooperation take different forms (Aldag and Warner, 2023). For some, the task of the corporation is collecting waste, while for others also, energy production based on the incineration of unsorted waste is done. Therefore, it is important to analyse what kinds of activities corporations and also cooperations do. Third, we advise to further explore the issue of political transaction costs. Following Sørensen (2007) we expect that these will be larger if more municipalities are involved with corporatisation and cooperation. Fourth and finally, we recommend to investigate the costs of different modes of production.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. As shifting from one mode to another takes time, we assume that decision makers need on average two years preparation time, and we assume that they look forward two years after the legal shift. Therefore, we specifically analyse this period of two years before until two years after the shift.

2. When municipalities merge to form a new municipality, shifts are only included if they take place after the new municipality is installed.

3. This order is based on the right-left dimension of the Chapel Hill Expert Survey Trend questionnaire for these National Parties.

4. The minimum of the polynomial is at approximately 500,000 inhabitants (i.e., 0.10/(2 × 0.001) × 10,000). In the council estimation (at 10% level) this is 250,000.

 

Appendix

Table A1. Estimation results for effective number of parties as fragmentation index.

	Corporatisation		Privatisation		Cooperatisation		Reverse in-housesation		Reverse privatisation
	Council	Board	Council	Board	Council	Board	Council	Board	Council	Board
Inhabitants	−0.05**	−0.06***	0.04	0.05**	−0.05	−0.07*	0.13***	0.10**	−0.01	0.01
Inhabitants^2	0.00**	0.00**	0.00	0.00	0.00	0.00	−0.01**	0.00*	0.00*	0.00
Household size	−0.68*	−0.86**	2.53***	2.50***	−0.38	−0.27	−0.72	−0.53	2.12***	2.03***
Population density	0.15**	0.13**	−0.08	−0.08	0.15	0.12	0.08	0.09	−0.14	−0.15*
Elderly people	3.43*	3.97**	4.29*	3.34	−9.49***	−9.89***	2.83	3.25	6.48***	5.65***
Ethnicity	−3.20*	−1.87	−1.86	−0.39	3.58	2.01	−2.64	−2.74	−1.70	−1.42
Income per inhabitant	−0.11***	−0.08**	0.19***	0.17***	0.02	0.02	−0.05	−0.04	0.12***	0.11***
Unemployment	−53.82***	−57.32***	−11.66	−19.00*	−42.73***	−44.94***	−58.65***	−55.84***	−11.05	−17.92*
Municipal debt per inhabitant	0.01	0.02	0.03	0.03	−0.04	−0.04	0.17***	0.17***	0.02	0.01
Fragmentation council II	−0.04	-	0.10**	-	0.03	-	−0.07	-	0.07	-
Council_VVD	2.02***	-	2.12***	-	0.51	-	1.50**	-	0.01	-
Council_CDA	−1.74***	-	−0.78	-	2.35***	-	0.88	-	−0.41	-
Council_CU/SGP	−0.14	-	0.05	-	−1.09*	-	1.64***	-	−0.82*	-
Council_D66	0.33	-	−2.76***	-	−3.46***	-	0.29	-	−4.28***	-
Council_PvdA	−0.09	-	−0.03	-	−1.00	-	0.76	-	−1.39***	-
Council_GL	−1.50*	-	−2.32**	-	−2.26*	-	−0.81	-	−0.06	-
Council_SP	−0.99	-	3.51***	-	−0.28	-	0.18	-	1.91**	-
Fragmentation board II	-	−0.23***	-	−0.06	-	0.13*	-	−0.06	-	−0.01
Aldermen_VVD	-	0.07	-	0.04	-	−0.53	-	0.92***	-	−1.45***
Aldermen_CDA	-	−0.21	-	−0.58**	-	0.40	-	0.66**	-	−0.56**
Aldermen_CU/SGP	-	0.27	-	−0.58*	-	−0.85**	-	1.49***	-	−0.92***
Aldermen_D66	-	1.42***	-	0.10	-	−1.31*	-	0.00	-	−0.04
Aldermen_PvdA	-	−0.21	-	−0.02	-	−0.79**	-	1.00***	-	−0.49*
Aldermen_GL	-	−0.36	-	−0.52	-	−1.73**	-	−1.05	-	−0.95*
Aldermen_SP	-	0.83	-	1.30*	-	−0.66	-	0.86	-	1.12*
Year trend	−0.05***	−0.06***	−0.05***	−0.04**	0.10***	0.09***	−0.09***	−0.10***	−0.01	−0.01
Constant	2.41*	2.74**	−11.76***	−10.52***	−1.26	−0.99	0.49	−0.29	−9.53***	−8.65***

Table A2. Estimation results for absolute number of parties as fragmentation index.

	Corporatisation		Privatisation		Cooperatisation		Reverse in-housesation		Reverse privatisation
	Council	Board	Council	Board	Council	Board	Council	Board	Council	Board
Inhabitants	−0.06**	−0.05**	0.03	0.05**	−0.10**	−0.07*	0.02	0.11**	−0.02	0.01
Inhabitants^2	0.00**	0.00**	0.00	0.00	0.00	0.00	0.00	0.00**	0.00*	0.00
Household size	−0.67*	−0.85**	2.48***	2.50***	−0.31	−0.26	−0.57	−0.53	2.11***	2.02***
Population density	0.15**	0.14**	−0.08	−0.08	0.16*	0.12	0.11	0.09	−0.14	−0.15*
Elderly people	3.49*	3.99**	4.00*	3.28	−9.52***	−9.75***	3.10	3.29	6.35***	5.70***
Ethnicity	−3.29*	−1.75	−2.08	−0.44	2.83	2.06	−3.93*	−2.63	−1.80	−1.35
Income per inhabitant	−0.11***	−0.07**	0.18***	0.17***	0.03	0.01	−0.04	−0.04	0.12***	0.11***
Unemployment	−53.45***	−56.72***	−12.50	−18.96*	−42.11***	−45.52***	−56.45***	−55.57***	−11.65	−17.78*
Municipal debt per inhabitant	0.02	0.02	0.04	0.03	−0.03	−0.04	0.18***	0.17***	0.02	0.01
Fragmentation council III	−0.01	-	0.06*	-	0.13***	-	0.13***	-	0.04	-
Council_VVD	2.08***	-	2.11***	-	0.86	-	2.13***	-	0.04	-
Council_CDA	−1.60***	-	−0.96*	-	2.87***	-	1.85***	-	−0.51	-
Council_CU/SGP	−0.20	-	0.12	-	−1.18*	-	1.47***	-	−0.79*	-
Council_D66	0.19	-	−2.58**	-	−3.91***	-	−0.54	-	−4.21***	-
Council_PvdA	−0.01	-	−0.13	-	−0.81	-	1.27**	-	−1.42***	-
Council_GL	−1.61*	-	−2.21**	-	−2.71**	-	−1.33	-	−0.01	-
Council_SP	−1.05	-	3.56***	-	−0.42	-	0.05	-	1.95**	-
Fragmentation board III	-	−0.25***	-	−0.04	-	0.12	-	−0.09	-	−0.04
Aldermen_VVD	-	0.07	-	0.05	-	−0.53	-	0.91***	-	−1.44***
Aldermen_CDA	-	−0.19	-	−0.56**	-	0.37	-	0.65**	-	−0.56**
Aldermen_CU/SGP	-	0.31	-	−0.60*	-	−0.87**	-	1.53***	-	−0.88***
Aldermen_D66	-	1.46***	-	0.08	-	−1.30*	-	0.05	-	0.00
Aldermen_PvdA	-	−0.20	-	−0.01	-	−0.80**	-	0.98***	-	−0.49*
Aldermen_GL	-	−0.28	-	−0.54	-	−1.75**	-	−0.99	-	−0.89
Aldermen_SP	-	0.84	-	1.31*	-	−0.67	-	0.84	-	1.12*
Year trend	−0.05***	−0.06***	−0.05***	−0.04**	0.10***	0.09***	−0.09***	−0.10***	−0.01	−0.01
Constant	2.16*	2.73**	−11.38***	−10.58***	−2.27	−0.96	−1.38	−0.23	−9.39***	−8.57***