Collective renovation decisions in multi-owned housing management: the case of public–private homeowners associations in Poland

The increasing popularity of multi-owned housing in various parts of the world and the need to constantly maintain these stocks in good condition raises issues of collective renovation decisions made by groups of unit owners in this housing sector. The main aim of the study was to identify the exogenous factors influencing collective decision-making actions of public–private homeowners associations regarding common property renovations in the management of housing condominiums in Poland. To achieve this aim, an original multi-stage top-down research methodology was applied. As the dependent variable representing the success of collective decision-making actions on common property renovations, the renovation fund rate passed by homeowners associations was adopted. Empirical research was conducted on representative samples of public–private homeowners associations with the participation of the municipality of Olsztyn in Poland. The results showed that building age, the number of unit owners and the share of the municipality in the common property significantly influenced renovation fund rates, ceteris paribus, with the influence of the former factor being positive and the others negative. This confirmed the hypothesis that the municipalities tend to negatively influence the collective renovation decisions of public–private homeowners associations by their shares in common properties. This study contributes to collective action theory when public and private actors co-participate in group decision-making and has far-reaching policy and practical implications. Hence, further research on this topic in different markets and institutional arrangements is recommended.

to the inability to exclude any HOA member from their costs and benefits. This prompts individuals to behave opportunistically and free-ride on the contribution of others (Hastings et al., 2006;Chu et al. 2012). The key condition for success in making such collective decisions is proper coordination and cooperation among unit owners (Chen & Webster, 2005;Yau, 2011) to obtain the consent of the qualified majority of HOA members for the renovation of the common property. But how can a group of interdependent unit owners get that majority when all face temptations to free-ride, shirk or otherwise act opportunistically (Ostrom, 1990). However, despite these difficulties, in practice, it is often possible to obtain the consent of the majority mentioned, and collective renovations decisions take force. The deviations between traditional group decision-making theory (Olson, 1965) and practical experiences of collective renovation decisions in Polish HOAs reveal the research gap in MOH management, which is analyzed more in-depth in this paper.
The subject of research in this study were HOAs that met two conditions: (1) they belonged to the category of large HOAs (OUA, 1994), i.e. HOAs owning buildings of more than seven premises, in which the said collective renovation decisions (classified as extraordinary) required the approval of the qualified majority of HOA members, i.e. members jointly holding a majority of shares in the common property, and (2) they were marked by their public-private nature, meaning that they consisted of both municipal (public) and private owners of dwellings who were also co-owners of the common property. Public-private HOAs have been created as a result of the privatization (sale) of social rented municipal dwellings to their occupiers (tenants). Municipalities, while owning a portion of the housing units in buildings belonging to such HOAs, in addition to the function of a coowner, i.e. an HOA member with certain rights and obligations over the common property, also perform the function of a social landlord. In this function, municipalities implement local housing policies aimed at meeting the housing needs of low-income households, as these units are part of the municipal housing stock. Moreover, municipalities often also act as property managers of common properties owned by such HOAs, appointing their own organizational units for this function, especially in cases where their shares in those properties are predominant. However, it should be emphasized that in the management of common property owned by public-private HOAs, decisions on how to manage this property (self-management or hiring a private or public third-party manager) are made autonomously by the qualified majority of HOA members (as for common property renovations). In contrast, operational decisions regarding the selection of a specific common property manager are made with the consent of the said majority of HOA members by the boards of these associations. This means that HOA members owning more than 50% of shares in the common property (such as municipalities) can impose on the HOA their preferred management entity and renovation policy for the common property.
The main aim of the study was to identify the exogenous factors influencing collective decision-making actions of public-private HOAs regarding common property renovations in the management of housing condominiums in Poland. Collective decision-making is perceived as the most important type of HOAs' collective actions in housing condominium management since it shapes the direction and style of such management and determines its effectiveness (Gao, 2015). The renovation fund rate (RFR) passed by public-private HOAs was used as the synthetic indicator (measure) of the success of collective decisionmaking actions taken by these HOAs in the analyzed domain. Given anecdotal evidence from observations of management practice, it is reasonable to suppose that municipalities, in the perspective of the privatization of all remaining dwellings in public-private HOAs, may behave opportunistically, seeking to reduce their spending on common property renovations resulting from their membership in those HOAs. Accordingly, two research hypotheses were posited that municipalities participating in public-private HOAs tend to negatively influence collective decision-making actions of these HOAs regarding common property renovations by: (1) Controlling property management services for such HOAs as owners of the companies performing those services; (2) Opposing resolutions of such HOAs seeking to increase common property renovation expenditures using their shares in common properties.
These hypotheses need to be verified with quantitative (statistical) methods, as such supposed behavior of municipalities has not yet been empirically studied in Poland.
This study attempts to fill the gap in the literature on collective decision-making in public-private HOAs in particular. This is due to the fact that many previous studies in this area have focused on owners' groups composed exclusively of private persons, and the presence of public actors in HOAs was neglected. Moreover, the literature on commonpool resources (CPRs) does not explicitly address the specificity of public entities co-participating with private persons in collective decision-making. The importance of this issue also stems from the multifaceted relevance of keeping MOH buildings in good condition, not only for healthy, safe and satisfying lives of residents and the preservation of asset value for owners but also for creating sustainable urban environments (Eslinga and Hoekstra 2005;Wilhelmsson, 2008;Yau, 2012). The common parts of buildings as man-made CPRs (Ho & Gao, 2013) do not have self-regeneration abilities like natural CPRs, and their condition is constantly deteriorating due to wear and tear. Hence, it is imperative that HOAs make collective decisions to maintain and renovate common parts throughout their entire occupation period. It should be added that, due to the long-term nature of municipal housing privatization in Poland, the country provides a unique laboratory for studying this type of public-private HOAs, as they still create a significant share in the national housing market.
Following the Introduction, this paper contains four sections. The second section outlines a literature review on theoretical concepts and selected previous empirical research findings concerning collective decision-making in MOH management. The third section covers the material and methods used for achieving the aim of the study, while the fourth section presents and discusses the results of the empirical research. Finally, the fifth section concludes the research findings and outlines their limitations and implications with recommendations for further research.

Literature review
In the literature, it is widely recognized that decision-making in the management of housing condominiums, as special forms of MOH developments, is much more complicated (Shin et al., 2021;Treffers & Lippert, 2020) compared to single-owner developments because they consist of both private and public spheres (Yau, 2012). In the private sphere, unit owners have the exclusive right to own and use their dwellings, but in the public sphere, they are co-owners of the common property and jointly share the management rights and responsibilities associated with them. In other words, co-owners making voluntary decisions with respect to the common property within the MOH development are legally, economically and socially interdependent (Bailey & Robertson, 1997), and these decisions become collective rather than individual decisions. This means that a group of rational co-owners (individuals) need to act together with respect to jointly owned parts of the MOH development to achieve a common goal. However, rational choice theory implies that individuals will act rationally to achieve their own interests, but the assumption that groups will also act rationally to collectively realize a common interest is not valid as a group is composed of a set of individuals who will prioritize their own interest above that of the group. Therefore, despite it being in their best interest, individuals will not necessarily join together to solve a collective action problem and will only agree to collaborate when there is a sufficient incentive. This behavior, identified by Olson (1965), is referred to as "the collective action problem" of group decision-making (Hastings et al., 2006).
Collective action problems will occur "where a number of decision-makers have to unilaterally decide whether or not to contribute to the provision of some collective good" (Hovi & Foss, 1995), with a collective good being defined as one in which all group members will benefit, regardless of whether or not they contribute to the cost. The non-excludable nature of the common property as a particular type of collective good allows individual owners to behave opportunistically and free ride-ride on the contribution of others (Walters, 2002;Lai and Chen 2004;Hastings et al., 2006;Chu et al. 2012). Owners will only participate in the collective action when their transaction cost in doing so will be less than the benefit derived from this action. The problem of free-riding implies that without some authority to ensure participation by all group members to reduce transaction costs, collective action will be doomed to failure. This authority can rely on formal external institutions (e.g. laws and contracts) or informal institutions of social norms and customs internal to the group (Ostrom, 2005(Ostrom, , 2010Walters, 2002). The statutory condominium system in Poland provides a legal framework that acts as a formal coordination mechanism for collective decision-making by co-owners over the common property. However, research has shown that although a legal framework is important, it may not always guarantee successful collective management (Ostrom et al., 2006).
From the collective action theory perspective a housing condominium is seen as a common pool or common property resource (CPR), collectively managed by co-owners with the purpose of maintaining the quality of the built environment (Vergara et al., 2019). Hence, co-owners in condominiums have similar collective action problems to those in other communities involved in either public goods or CPR management (Agrawal, 2001;Orban, 2006;Ostrom, 2000). However, co-owners also tend to be apathetic towards the common property management due to their inadequate awareness, confusion about management responsibilities or lack of sufficient management expertise, time and money (Blandy et al., 2006;Easthope & Randolph, 2018;Goodman & Douglas, 2008;Ho & Gao, 2013;Hui, 2005;Yau, 2012). On the other hand, however, it is argued that Olson's (1965) pessimistic view of collective action theory is not sufficiently reflected in the condominium management practice, as some co-owners, often going beyond the logic of personal cost-benefit calculation, actively participate in collective decisions and actions with respect to common property; otherwise all such MOH developments would be dilapidated or suffer other consequences of mismanagement (Yau, 2011(Yau, , 2014. The literature review reveals various approaches to solving collective action problems of common property management in housing condominiums. The first includes selective incentives (e.g. government subsidies or rewards and punishments or individuals) that are necessary to motivate co-owners to cooperate (Olson, 1965). These incentives may be different in large and small groups of co-owners, although as Bengtsson (1998) stated, their introduction leads only to marginal or short-term solutions to the free-rider problem that are far from stable. The second approach is based on centralized solutions, which require the involvement of a central authority in collective action by co-owners that handles most of the common property management issues, including coordination of decision-making processes, enforcement of management rules, resolution of conflicts among co-owners, and the provision of building management services (Chu et al., 2012;Hastings et al., 2006;Yiu et al., 2006). A central authority could be an owners' organization, a property manager, or a combination of the two. One advantage of this approach is that it reduces transaction costs in MOH management (Christudason, 2008;Walters, 2002;Yiu et al., 2006). Another advantage is that the central authority has the professional expertise to mobilize owners to cooperate and can represent the owners as a whole when dealing with a third party. In contrast, the major drawbacks of this approach include rent-seeking and principal-agent problems (Chen & Webster, 2005;Yau et al., 2021;Yip et al., 2007). Rent-seeking refers to special-interest group behavior that generates significant personal benefits for a limited, identifiable number of group members while imposing a small individual cost on all members of the group. However, the essence of the principal-agent problem is how to structure the relationship to ensure the agent acts in the best interests of the principals. Furthermore, institutional solutions that regulate owners' behavior with formal rules and/or informal constraints have been proposed to reduce uncertainty in personal interactions (North, 1990). However, they are more likely to guarantee only a minimum level of collective action (Bengtsson, 1998). Other researchers have proposed a communitarian approach (e.g. social norms, enhanced mutual trust, sense of community, etc.) as a promising solution to the collective action dilemma in MOH management (Borisova et al., 2014;Wu, 2012;Yau, 2011Yau, , 2012. They derive their thesis from the long-term nature of neighborhood relationships, the geographic proximity and daily face-to-face contacts of the co-owners, their shared interests and belief in success as well as the limited group size, which can help initiate cooperation among them. Other authors have advised a coercive approach, which resorts to management institutions mandated by the authorities (Bailey & Robertson, 1997;Chan & Choi, 2015;Paulsson, 2010). They assume that the fear of punishment for noncooperation will stimulate co-owners to cooperate. Some authors further emphasize the supporting role of the integrated use of multiple information and communication technologies (ICTs) for the success of collective action (Cardoso et al., 2019). The listed solutions should not be regarded as mutually exclusive but rather as potentially complimentary, as confirmed by the practice of condominium management.
Based on the literature, several empirical studies can be identified on the exogenous factors influencing the collective decision-making of co-owners in MOH developments. It has been shown that the number of dwellings in a co-ownership regime influences maintenance and organizational outcomes, with smaller groups likely to be more effective (Olson, 1965). However, as Ostrom (2005) has demonstrated, group size is only one factor that should be taken into account. Different incentives and perceptions of the benefits and costs might influence the sustained organization that is required to collectively maintain the common property in a condominium (Chen & Webster, 2005). Orban (2006), studying the main difficulties of condominium management in privatized social housing in Hungary, found that condominium size (measured by the unit number) is negatively related to the cooperative potential of the building's residents. Additionally, that author investigated the role of the law and of some of the residents, revealing how more positive outcomes are achieved when a leader in the HOA is present. Some studies indicated that management regimes are important for solving collective action problems (Ho et al., 2006;Walters & Kent, 2000); others focused on the impact of property rights' delineation on building management (Blandy et al., 2006;Yiu et al., 2006). There are also studies that empirically showed that individual participation is of paramount importance to the success of collective action. Yau (2014) examined factors that influence how actively an owner participates in MOH management in Hong Kong. It was found that, in addition to the value of the common good and the selective benefits and costs of participation, individual residents' perceptions of self, group and proxy efficacies are important determinants of their participation behavior. In a recent study, it was found that residents' level of participation in collective action correlates positively with self-and group-efficacy beliefs but negatively with perceived proxy efficacy (Yau, 2018). Borisova et al. (2014) assessed the performance of HOAs in Russia using the stochastic frontier technique. They found that cultural traits enabling tenants to make proper use of the HOA decision-making procedures are essential for resolving the collective action problem and ensuring the accountability of management bodies and outside contractors. It was also stated that building size (number of tenants), building age affecting physical conditions of common property and tenants' social capital and inequality in the tenant community could be relevant for HOAs performance, but the magnitude and even direction of their impact cannot be established a priori and, as it is often the case with CPRs, could be context-specific. Ho and Gao (2013), examined the impact of factors such as building condition, building location, number of owners, and presence of owners' organizations on collective decision actions by owners in apartment buildings in Hong Kong on the appointment of a PM company. The results showed that the impact of these factors was statistically significant but different in strength and direction. Gao (2015) developed a method to measure the cooperative abilities of an owners' group using survey data from apartment buildings in Hong Kong. It was found that building location and building type posed an insignificant impact on an owner group's cooperative abilities. Moreover, these abilities would be affected by group size, group income level, age level, management style, and building age. The author concluded that if an owner group whose members are old and financially incapable is required to manage an aged and large-scale MOH building, there is a good chance the group will fail in such a management task. In turn, Gao and Chen (2016) examined different aspects of owner heterogeneity in terms of how they affect the owners' collective actions in MOH management. They found that the overall owner heterogeneity-in terms of knowledge level, age level, tenure mode, and length of time living therehad a significant negative impact on owners' collective actions. The authors observed that if owners of an MOH development differed in one aspect, they could still manage to find common ground to advance their common interests; but if the owners differed in many or all aspects, they would have great difficulty in working together. In other words, the impact posed by an individual aspect of owner heterogeneity would be small and unimportant, but the mixture of these negligible effects could be significant and detrimental to MOH management. Thus, owner heterogeneity does matter in this management as long as the extent to which owners are different from each other is considerable.
Two more studies using Ostrom's (Ostrom, 2005) Framework for Institutional Analysis and Development (IAD) to explain exogenous factors influencing the collective MOH management are worth noting. In the first, Gao and Ho (2016) synthesized the key exogenous factors of MOH management from a collective action perspective and then examined the relationship between these factors and management outcomes using survey data. The authors found that apart from the physical features of an MOH development (e.g. building age and scale), the attributes of the owner group (e.g. group size and agent) and the institutions governing the MOH management played an important role in shaping the management outcomes. More specifically, it was found that these outcomes were negatively related to building age but positively correlated with building scale, group factor, and institutions. In addition, it has been argued that although building location and group heterogeneity play a role in shaping management outcomes, no conclusive evidence on these two factors was found. In a second study by Donoso and Elsinga (2018), the IAD framework was applied to explain exogenous factors that might affect the decision-making process of maintenance activities. That study examined the links between three different institutional characteristics related to common property management: the communities, the governance structure, and the physical environment with the perceived level of condominium maintenance. Based on results obtained in low-income condominium settings in two cities from Colombia and Ecuador, the authors stated that if there is (1) trust in the leaders of the community, (2) agreement about who is responsible for maintenance, (3) participation in assembly meeting and (4) adequate physical conditions of the building that can be maintained, owners are likely to have higher benefits than those who do not trust others and do not go meetings. In addition, it was emphasized that owners collectively need information, as well as knowledge of the rules, to manage condominium complexity.
The theoretical concepts and empirical findings presented above specifically address one of the most challenging collective action domains (arenas) in MOH management, which constitute collective decisions made by owners' groups (HOAs) regarding common property renovations (Easthope & Randolph, 2009;Yip & Forrest, 2002;Yip et al., 2007). This is due to the high costs of their implementation, as well as their economic, sociocultural, organizational and technological implications for co-owners, co-residents and the entire urban environment. One of the detailed issues in this domain is the choice of how to finance expenditures for common property renovations. As suggested by Lujanen (2010) and Arkcoll et al. (2013), the preferred way of financing such expenditures is through the use of a renovation fund that is collectively passed and successively contributed by all members of an owners' association over an extended period of time. The rationale is that since the benefits of major repairs are of a long-term nature, their costs should also be spread over a longer period of time. Polish HOAs accumulate money in these funds according to specific (unit) renovation fund rates (RFRs), which are passed by the qualified majority of HOA members and apply to all unit owners in the housing condominium. The RFR is calculated by dividing the expected cost of renovation work by the number of months until the start of the planned renovation work and the sum of the usable floor area of all dwellings in the HOA. Since renovation fund rates are found in the widespread practice of financing renovations in housing condominiums in Poland, they were used to develop the research methodology of this study.

Methodology and data
For empirical verification of the posited research hypotheses and achievement of the main goal of the study, an original research methodology was applied. The fundamental problem concerned the selection of an appropriate dependent (response) variable that would represent (measure) synthetically the success of collective decision-making actions of public-private HOAs regarding common property renovations. In considering possible approaches to measuring collective action success in CPR management (Ruttan, 2008), it was assumed that the success of the collective actions in question would be measured on the basis of the quantity of collective goods produced by these actions. Accordingly, the aforementioned renovation fund rate (RFR) actually passed by HOA members was employed as the dependent variable representing (measuring) the success of collective decision-making actions taken by public-private HOAs in the analyzed domain. The arguments for this selection are at least threefold. First, renovation funds are commonly used in the management practice of housing condominiums owned by public-private HOAs under consideration in Poland. Second, the RFR is passed voluntarily at a general HOA meeting and is the result of a fully collective decision-making action of a qualified majority of HOA members (unit owners) jointly holding more than 50% of all shares in the common property. Once the RFR is established by such a collective decision, all HOA members are obliged to pay the appropriate renovation fund fees, which means that none of them can be excluded from either the costs or benefits of renovating the common property. Third, the housing stock owned by the public-private HOAs in Poland consists predominantly of old and deteriorated multi-family buildings with serious technical and functional deficiencies, in which renovation works of varying scope are necessary. The relatively low-quality standard and poor technical condition of this housing stock results from its defective age and many years of renovation neglect resulting in a successively increasing renovation gap in this stock (Muczyński et al., 2021).
In order to identify the exogenous factors influencing collective decision-making actions on common property renovations by public-private HOAs, a multi-stage topdown research approach was developed. The research procedure used to accomplish the main aim of the study involved the following stages: 1. Specifying an initial set of factors as candidates for the role of independent variables explaining renovation fund rates (RFRs) in housing condominiums (MOH developments) owned public-private HOAs, followed by defining measurement scales for these variables and determining ways of their quantification. 2. Collecting empirical data on RFRs (as the dependent variable) and pre-selected factors potentially influencing them (as potential explanatory variables) in the sample MOH developments owned by public-private HOAs during the time period under study in a given housing market.

Determining the strength and direction of statistical relationships between observed
RFRs and pre-selected factors (as potential explanatory variables) based on the values of calculated correlation coefficients, and performing the final selection of explanatory variables by reducing the initial set of factors by those that are not correlated with the dependent variable at the assumed level of statistical significance. 4. Choosing the analytical form of the multiple regression model describing the hypothetical functional relationships between the RFRs rates and the finally established set of explanatory variables, followed by estimating the structural parameters of the model. 5. Statistical verification of the estimated model, which involved testing whether it meets standard statistical postulates regarding the significance of the structural parameters and the desired characteristics of the residual components and assessing the degree of its adaptation to empirical data. 6. Substantive verification of the estimated model involving the assessment of its consistency with theoretical and practical knowledge about the phenomenon under study, including the assessment of the strength and direction of the influence of explanatory variables on the dependent variable. 7. Identifying the exogenous factors influencing collective decision-making actions of public-private HOAs on common property renovations in the studied housing market on the basis of the estimated regression coefficients between the RFRs and their statistically significant explanatory variables.

3
In specifying an initial set of factors potentially influencing renovation fund rates (RFRs) collectively passed by public-private HOAs in housing condominiums, the relevant literature on the determinants of common property management was considered (e.g. Easthope et al., 2020;Hastings et al., 2006;Lujanen, 2010;North, 1990;Olson, 1965;Ostrom, 2000Ostrom, , 2005Walters, 2002;Yau, 2014Yau, , 2018. Selected previous empirical research on collective actions of owners in MOH developments were also analyzed (e.g. Borisova et al., 2014;Gao, 2015;Gao & Chen, 2016;Gao & Ho, 2016;Ho & Gao, 2013). In addition to theoretical considerations and empirical findings, practical knowledge of the housing market under study and the availability of relevant data on examined MOH developments and public-private HOAs as owners' groups were important premises in specifying the initial set of factors discussed. The preliminary (pioneering) nature of the conducted research in the Polish context should also be taken into account. Given these circumstances and the aim and subject of the study, as well as the research hypotheses posited, six factors were selected for the initial set of independent variables explaining the renovation fund rates collectively passed by public-private HOAs in housing condominiums, which were divided into two categories. The first contained two factors describing the type of property manager appointed by the HOA, including municipal property management (MPM) and private property management (PPM), while the second consisted of factors that describe both the common property and the member group of the HOA, such as property location (LOC), building age (AGE), number of unit owners (NUO) and share of the municipality (SMU). The list of the factors (variables) used in the research and their descriptions are presented in Table 1.
One of the problems was the choice of the analytical form of the regression function, especially since many various forms of modelled functional relationships are used in housing research, and hardly any consensus was reached in this regard. Searching for the form of the multiple regression model describing the hypothesized relationships concerning the collective decision-making actions under study, previous research in this area and the available statistical material were analyzed. The common practice of various forms of regression function among housing researchers and their adequacy to the specifics of the studied phenomena were also taken into consideration. Pretesting of several frequently used forms of regression function in housing studies was also conducted. As a result of the above analyses and tests, the linear form of the multiple regression function was chosen for the model describing the hypothesized functional relationships between the renovation fund rates passed by public-private HOAs and their potential explanatory variables. The theoretical regression model was thus described by the formula: where RFR-renovation fund rate representing the degree to which a public-private HOA succeed in collective decision-making actions regarding common property renovations as the dependent variable (in Polish zlotys monthly per m 2 of the useable floor area of dwellings in the HOA); X i -independent variables (i = 1,…,k) explaining the renovation fund rate (RFR) passed by a public-private HOA; β 0 , β i -structural parameters (regression coefficients) of the model, including the coefficient β 0 as the intercept and the coefficients β i (i = 1,…,k), with each interpreted as the absolute change (increase or decrease) in the dependent variable (RFR) resulting from a unit change of the corresponding independent variable, ceteris paribus. ε-the random error.
Data on renovation fund rates (RFRs) actually passed by public-private HOAs surveyed and the exogenous factors potentially influencing them were collected in the housing market of (1) RFR = 0 + 1 X 1 + 2 X 2 + ⋯ + k X k + Table 1 The factors (variables) used in the research (Source: Own study) the city of Olsztyn in Poland. This is a province (voivodeship) capital located in the northeast of Poland, which at the end of 2020 had 171,249 inhabitants and a housing stock of 80,967 dwellings in total. The ownership structure of housing in Olsztyn was dominated by private ownership of individuals comprising a total of 58,190 (71.9%) dwellings, 32,246 (55.7%) of which belonged to various types of HOAs. The cooperative housing sector included a total of 17,581 (21.7%) dwellings, while the municipal housing stock consists of 4017 (5.0%) housing units, 2608 (64.9%) of which (located in 638 buildings) belonged to the public-private HOAs that are the subject of this study. The housing stock owned by the considered public-private HOAs in Olsztyn consists mainly of old multi-family buildings in poor technical condition, which in the vast majority require renovation actions. In 2020, about 75% of buildings owned by public-private HOAs in Olsztyn underwent minor renovations, while 24.8% of these buildings underwent major renovations. Residents' primary motivations for renovations are related to improving the functional, economic, environmental and aesthetic qualities of the common properties. The main research sample (S1) included the representative set of 135 housing condominiums (MOH developments) owned by HOAs with the participation of the Olsztyn municipality. This sample was drawn up using a stratified random sampling technique, hence it replicates the structure of the general population. It consisted of two component subsamples, where the property management (PM) services contracted by these HOAs were provided by different types of property managers. The first subsample (sample S2) included 80 objects in which PM services were provided by municipal property management companies (MPMCs), and the second subsample (sample S3) included 55 objects where these services were provided by private property management firms (PPMFs) (Fig. 1).
The empirical data as of the end of 2020 were obtained from primary information sources, mainly including the detailed analytical register on HOAs with the participation of the Olsztyn municipality in MS Excel spreadsheet format maintained by the Municipal Housing Management Unit (MHMU) of the City of Olsztyn. Data contained in this register enabled establishing for each of the surveyed HOAs, in addition to the renovation fund rate, the type of

Results and discussion
Descriptive statistics of the collected empirical data on the pre-selected factors (as their potential explanatory variables) and renovation fund rates are presented in Table 2. The first two factors, such as municipal property management (MPM) and private property management (PPM), were dummy variables (as X 1 and X 2 , respectively), taking the value 1 if the given type of property management was present in the HOA and 0 if it was not. The other factors, such as property location (LOC), building age (AGE), number of unit owners (NUO) and share of the municipality (SMU), were quantitative variables (as X 3 , X 4 , X 5 , and X 6 , respectively) that were measured on ratio scales according to the key shown in Table 1. Finally, the renovation fund rate (RFR) as the dependent variable was also a quantitative variable measured on a ratio scale. The statistics presented in Table 2 refer to data on all objects in the main research sample (S1) of public-private HOAs in the Olsztyn housing market at the end of 2020. Based on such processed data, the strength and direction of statistical relationships between RFRs and their pre-selected explanatory variables were determined using correlation analysis.  Due to measuring independent variables on ratio scales and the adopted form of the regression function, Pearson's linear correlation coefficient (r) (Cohen, 1988) was employed. This coefficient expresses the statistical interdependence between studied variables numerically. It is sensitive to extreme and outlier observations, and values close to zero do not imply the lack of correlation at all, but only the lack of linear correlation. The key issue in correlation analysis in a research sample is the statistical significance of calculated coefficients (Aczel, 2006). The results of the correlation analysis (r coefficients) are presented in Tables 3 and 4.
The correlation results showed in Table 3 indicate that in the main sample (S1), the strength of statistical relationships between RFR and both dummy variables (X 1 , X 2 ) was noticeable but statistically not significant (at p < 0.05). In contrast, the direction of correlational dependencies between these variables was, as expected, the opposite. This means that despite some positive symptoms, in the light of the correlation results, it cannot be stated that the type of property manager (MPMCs or PPMFs) appointed by the sampled HOAs influenced their collective decisions on renovation fund rates. Such a result resonates well with contemporary findings, as renovation decisions (including RFR rates) in public-private housing condominiums are taken directly by HOA members (unit owners), who act independently of suggestions from any type of common property managers, who play only an auxiliary role in this regard. Therefore, the first hypothesis that municipalities participating in public-private HOAs tend to negatively influence collective decision-making actions of these HOAs on common property renovations by controlling PM services provided by MPMCs has not been empirically confirmed. The variable X 3 (property location) was also found to be an insignificantly correlated variable with the RFR, which is consistent with the findings of Gao (2015) for apartment buildings in Hong Kong. This also resonates well with contemporary findings on the subject in Poland. The reason for this is the commonly low quality standard and poor technical condition of this housing stock in the surveyed market regardless of the location qualities of the buildings in this stock. In contrast, the linear correlation of variables such as X 4 (AGE), X 5 (NUO) and X 6 (SMU) with the RFR was found to be statistically significant (at p < 0.05). This was found not only in the main research sample (S1) but also in the component subsamples of MOH developments (HOAs) in which common properties were managed either by MPMCs (sample S2) or by PPMFs (sample S3) ( Table 4).
The results presented in Tables 3 and 4 indicate that the strength of the statistical relationship between RFR and the individual variables (X 4 , X 5 and X 6 ) was at a moderate level (Rafer et al., 2003). Comparison of correlation coefficients for these variables showed that building age (X 4 ) and number of unit owners (X 5 ) had a noticeably greater influence on RFR in MOH developments managed by PPMFs (sample S3) than by MPMCs (sample S2). In contrast, the influence of the share of the municipality (X 6 ) on RFR was found to be larger in sample S2 than in sample S3. Directions of the revealed relationships between RFR and each of the statistically significant variables are expressed by the signs of corresponding correlation coefficients. The positive correlation of building age (X 4 ) is as expected because as a building ages, its physical condition deteriorates due to wear and tear, resulting in increasing renovation needs that require growing expenditures to cover. In contrast, the negative correlation of the number of unit owners (X 5 ) finds justification explicitly in collective action theory (Olson, 1965;Ostrom, 2005) and is consistent with the results of previous studies in this area (e.g. Borisova et al., 2014;Orban, 2006). This also resonates well with contemporary findings in this regard in Poland. In turn, the negative correlation of municipality's share (X 6 ) is an original finding from the correlation analysis that corresponds to observed practice but has not yet been empirically proven in Poland. This finding positively verified the second hypothesis that municipalities participating in public-private HOAs tend to negatively influence collective renovation decisions of these HOAs using their shares in common properties. Based on the correlation results, the initial set of explanatory variables was reduced by those that were not correlated with the dependent variable at the accepted level of statistical significance. Positively verified explanatory variables were highly independent of each other, which was important when building regression models.
Regression models were estimated using the selected form of the regression function and the reduced sets of explanatory variables. These models were estimated separately for the main sample S1 (model M1) and the subsamples S2 and S3 (models M2 and M3, respectively) of MOH developments owned by public-private HOAs in the studied housing market. Estimation of the structural parameters of the models was performed using the ordinary least squares (OLS) technique. Since outlier observations can significantly bias the regression equations, the estimation procedures for each of the models were conducted in two steps. First, the originally estimated models were subjected to residuals analysis, which allowed to eliminate the most outlier observations from the primary observation sets S1-S3 based on the Cook's distance criterion (Cook & Weisberg, 1982). The final regression models M1-M3 were then estimated using the outlier-cleaned samples. These models showed a better fit to empirical data and lower estimation errors compared to the originally estimated models. Standard statistical tests (Aczel, 2006), such as Fisher-Snedecor (F) statistics, Student's tests, and Durbin-Watson (DW) statistics, were used to statistically verify the finally estimated models. In addition, coefficients of determination (R-squared and adjusted R-squared) and standard errors of estimation (S e ) were calculated. The regression results in the estimated models are presented in Table 5. Based on these results, it was stated that the models met standard statistical postulates as the regression coefficients were statistically significant (at p < 0.05), the residuals were symmetric, random in nature and revealed no autocorrelation. Moreover, given the complex nature of collective decision-making actions in public-private HOAs, it can be considered that the estimated models are characterized by the required (satisfying) adaptation to empirical data. This is particularly evidenced by the relatively high values of R-squared coefficients (54.3% to 58.8%). The obtained results, therefore, confirmed the statistical validity of the estimated models, including the correctness of the choice of the analytical form of the regression function and the sets of explanatory variables that significantly influence the dependent variable. Thus, the models provide a reasonable tool to assess the influence of the individual explanatory variables ceteris paribus on the RFR.
Substantive verification of the estimated models involved an assessment of their consistency with theoretical and practical knowledge about renovation fund determination as a result of collective decisions of HOAs as owner groups. It included, in particular, the assessment of the strength and direction of the influence of each explanatory variable on the RFR based on the values and signs of the regression coefficients at these variables. It should be noted that the regression analysis in models M1, M2, and M3 did not reduce the sets of explanatory variables (X 4 , X 5 , X 6 ) previously verified by correlation analysis. Furthermore, the obtained values and signs of regression coefficients showed structural similarity to the pattern of correlation coefficients. Therefore, given the arguments raised in the correlation analysis and the interpretation of regression coefficients in linear models, the values and signs of the calculated regression coefficients were considered reasonable, as they were consistent with a priori expectations in light of theoretical and practical knowledge in the area under study. The estimated values of these coefficients indicate that the average strength of the influence of a unit change in the identified explanatory variables on the RFR ranged, ceteris paribus, respectively: 0.0126 PLN (0.4%) for X 4 (building age), 0.0301 PLN (1.0%) for X 5 (number of unit owners), and 0.0354 PLN (1.2%) for X 6 (share of the municipality), with the direction of the influence of the first variable mentioned being positive, while the other two were negative. The negative influence of the municipality's share (X 6 ) on the observed renovation fund rates confirmed the second hypothesis that municipalities participating in public-private HOAs tend to negatively influence collective decision-making actions on common property renovations by opposing resolutions of such HOAs seeking to increase common property renovation expenditures using their shares in common properties.

Conclusions
Due to the increasing popularity of housing condominiums as the most prevalent form of MOH developments in many housing markets and the need to maintain these housing resources in good condition throughout their entire life cycle, collective decision-making actions regarding common property renovations taken by unit owners is a topical research problem with a global scope and is complex in nature. This is evidenced both by the formation of new theoretical concepts for solving collective action problems in the management of common properties as man-made CPRs, and by the growing number of empirical studies on the topic within different institutional arrangements. These studies have focused mainly on owners' groups composed of private persons, while the specificity of the behavior of public actors co-participating in different collective decision-making processes of those groups has been neglected. In order to contribute to filling this gap, in this preliminary study, the exogenous factors influencing collective decision-making actions of public-private HOAs on common property renovations in the management of housing condominiums under Polish conditions using a selected housing market were identified. Therefore, two research hypotheses on the supposed influence of municipalities as owners of MPMCs and participants of HOAs on collective renovation decisions of these organizations (owners' groups) were posited. For testing these hypotheses and achieving the main aim of the study, an original multi-stage top-down research methodology based on the multiple regression method was developed. The methodological achievement of this study is the conceptualization and operationalization of the research approach applied, including, in particular, employing the renovation fund rate (RFR) as the dependent variable representing the success of collective decision-making actions of public-private HOAs in the domain of common property renovations.
The empirical achievement of this study involves empirically verifying the set research hypotheses and evidencing by means of correlation and regression analysis the influence of such variables as building age, number of unit owners in the HOA and the share of the municipality in the HOA, on the renovation fund rates, and thus on the collective decisionmaking actions of public-private HOAs on common property renovations in MOH developments. By synthesizing the findings obtained in the studied housing market, the following empirical conclusions can be drawn: 1. There was no significant impact of the type of property manager (private or municipal) on the renovation fund rates passed by the public-private HOAs. This may indicate, on the one hand, that PPMFs appointed by these HOAs did not have the opportunity to implement more extensive common property renovations than MPMCs, and on the other hand, that PPMFs did not motivate unit owners to pursue an active renovation policy to a greater extent than MPMCs. 2. The research hypothesis that municipalities tend to negatively influence collective renovation decisions of public-private HOAs by controlling the PM services provided to these HOAs by municipal-dependent MPMCs has not been empirically confirmed. This may be due to the fact that the research sample was predominantly HOAs with minority municipal shares in the common property. 3. The location factor of MOH developments owned by the public-private HOAs had an insignificant influence on their collective decision-making actions on renovation fund rates. This means that unit owners in these HOAs were not guided in making such decisions by the potential benefits of common property renovations dependent on location advantages, like the substantial increase in the market value of dwellings in attractive locations. 4. The identified exogenous factors (building age, number of unit owners, and share of municipality) significantly influencing collective renovation decisions in the surveyed HOAs explained 52.4%-56.3% of the variation in the dependent variable in the estimated regression models and the strength of their total impact in extreme cases reached 2.5%-3.0% of the average RFR values with cumulative unit changes of all those factors. 5. The factor with the relatively largest influence on the RFR in MOH developments managed by MPMCs was the share of the municipality, while in those managed by PPMFs, it was the number of unit owners. The negative correlation and regression coefficients of the former factor with the RFR confirmed the hypothesis that municipalities as members of public-private HOAs tend to negatively influence collective renovation decisions of such HOAs using their shares in common properties as decision-forcing instruments. 6. The revealed behavior of municipalities in public-private HOAs cannot be categorized as either material or political free-riding (Borisova et al., 2014). It can be called transformational free-riding because the MOH developments owned by such HOAs are in a long-term transition phase (Johnstone and Reid 2013) in which ownership of individual (rental) dwellings is gradually being transferred from the municipality as the original owner to their tenants.
Generalizing the above research results, it can be concluded that in the process of selling municipal (public) dwellings to tenants, it is recommended to use rapid privatization by a selective method which involves the sale of all municipal dwellings in thoroughly selected buildings. Such a solution is particularly preferred in MOH buildings with high renovation needs and a large number of housing units.
This research has its limitations as the built regression models are location-specific, so it is difficult to generalize them to different locations, all the more so since this study is pioneering in the Polish context and the scope of empirical research was limited to the selected housing market. Therefore, these models are primarily used for an understanding of the particular housing market. However, they also provide insights into exogenous factors that can be consistently considered by members of public-private HOAs when making collective renovation decisions in MOH developments across different locations. Other limitations of this research may be due to the relatively small sample sizes and the problem of omitted variables. Given the set modeling aim, the adopted sample sizes were considered sufficient. In turn, limiting the set of factors potentially influencing collective renovation decisions of public-private HOAs to those of key importance can be justified by the preliminary (exploratory) nature of this study, as well as the restricted access to personal data of unit owners.
Good insights into which factors influence collective decision-making processes on common property renovations in MOH developments owned by public-private HOAs, and by how much, have far-reaching policy and practical implications. The policy implications relate to the use of such knowledge to support sustainable development of urban housing stock and create more integrated residential communities by improving local housing policies, particularly concerning privatization and public management of municipal housing stock included in those MOH developments. In turn, the practical implications of knowing the influence of the factors mentioned relate primarily to the considered HOAs as owners' organizations, municipalities and private individuals as dwelling owners and common property co-owners in MOH developments belonging to these HOAs, and PM entities appointed by them. Such knowledge gained by HOAs can drive changes in decisionmaking processes and property management regimes leading to improved efficiency and effectiveness of common property renovations. It can become an attractor for municipalities to make more rational housing privatization decisions and for private dwelling owners to adopt more proactive renovation attitudes, as well as increasing social awareness in caring for common property. These effects may therefore create the need to improve PM services for public-private HOAs in terms of streamlining renovation procedures and ways to motivate unit owners to make collective renovation decisions. Finally, it should be added that this study contributes to the collective action theory when public and private actors co-participate in group decision-making. Further research on this topic is recommended to explore a broader spectrum of factors and conditions determining collective renovation decisions taken by HOAs and other similar owners' groups, including factors more in-depth characterizing those groups as such (e.g. group heterogeneity), in MOH developments located in different markets and institutional arrangements.
Author contributions The paper has a single author, which means conceptualization, methodology, formal analysis and investigation, writing-original draft preparation, writing-review and editing, Funding acquisition, resources, supervision, etc. were done by AM.
Funding Partial financial support was received from the University of Warmia and Mazury in Olsztyn.

Conflict of interest
The author has no financial or non-financial interests to disclose that are directly or indirectly related to the work submitted for publication.

Ethics approval
The author ensures objectivity and transparency in research and that accepted principles of ethical and professional conduct have been followed. Hence, the Author includes information regarding sources of funding, and stated that there are no potential conflicts of interest (financial or non-financial) related to the work submitted for publication.
Informed consent All authors agreed with the content and that all gave explicit consent to submit and that they obtained consent from the responsible authorities at the institute/organization where the work has been carried out.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.