Financial Sustainability of Local Government Units and Their Investment Capacity: A Comparative Analysis of Municipalities and Cities with County Rights in Poland

Abstract

This study examines the relationship between the financial sustainability of local government units (LGUs) and their investment capacity, focusing on differences between municipalities and cities with county rights in Poland. It investigates how operating surplus, debt-servicing costs, and revenue autonomy are associated with investment capacity and whether these relationships differ across LGU types. The analysis uses data for Polish municipalities and cities with county rights for 2018–2024 obtained from the Local Data Bank of Statistics Poland. Various panel regression specifications were employed, supplemented by descriptive Pearson correlation analysis. The results indicate a consistent positive tendency in the relationship between operating surplus and investment capacity, whereas higher debt-servicing costs were significantly associated with lower investment capacity. Cities with county rights exhibited higher investment capacity than municipalities, although the strength of the evidence varied depending on the measure used. Revenue autonomy, in contrast, was found to be negatively associated with investment capacity. This study provides a comparative assessment of the relationship between financial sustainability and investment capacity, offering implications for local fiscal policy and future research.

1. Introduction

Local government units (LGUs) play a pivotal role in fostering socio-economic development at the local and regional levels [1,2,3,4,5]. The implementation of their statutory responsibilities, including the development of technical and social infrastructure, requires not only adequate institutional capacity but also stable and financially sustainable foundations [1,2,3,4]. In this context, the relationship between the financial sustainability of LGUs and their capacity to undertake and finance investment projects has become an important issue [5,6,7,8,9]. In Poland, research on local government finance has focused predominantly on municipalities, while considerably less attention has been paid to cities with county rights—urban municipalities that simultaneously perform municipal and county-level functions—and relatively few studies have examined counties or voivodeships. Existing studies on LGUs have primarily addressed their financial position, financial condition, financial stability, and financial autonomy, including revenue autonomy. Comparisons between different types of LGUs have also frequently been conducted from the perspective of financial autonomy. By contrast, relatively limited attention has been paid to the relationship between the financial capacity of LGUs and their investment capacity, particularly in studies comparing different types of local governments over a longer period.

This relationship is relevant because financial sustainability provides the fiscal capacity necessary to undertake and finance public investment, while investment capacity enables local governments to implement development projects that may contribute to long-term sustainable development objectives at the local level. Against this background, a comparative analysis of municipalities and cities with county rights provides an opportunity to examine whether the relationship between financial sustainability and investment capacity differs across these two types of LGUs.

The aim of this study is to identify the relationship between the financial sustainability of local government units (LGUs) and their investment capacity, with particular emphasis on differences between municipalities and cities with county rights. Specifically, this study examines how operating surplus, debt-servicing costs, and revenue autonomy are associated with investment capacity and considers whether the conclusions depend on the measure of investment capacity applied.

To address this objective, this study seeks to answer the following research questions:

RQ1. How are the financial sustainability dimensions of local government units—operating surplus, debt-servicing costs, and revenue autonomy—associated with their investment capacity?

RQ2. Does investment capacity differ between municipalities and cities with county rights, and do these differences depend on the measure of investment capacity applied?

RQ3. To what extent do the relationships between financial sustainability and investment capacity remain consistent across different measures of investment capacity?

To achieve the research objective and test the proposed hypotheses, an empirical analysis was conducted using data for municipalities and cities with county rights in Poland covering the period 2018–2024. The study employed statistical methods, including correlation analysis, group comparison tests, and panel regression models, enabling a comprehensive assessment of the relationships under investigation.

Existing research on public finance, and more specifically, on local government finance, including financial condition and financial sustainability, both internationally and in Poland, has predominantly focused on a single type of local government unit (e.g., municipalities or cantons) and relatively short observation periods. The contribution of this study lies in integrating the analysis of local governments’ financial sustainability with the assessment of their investment capacity over a multi-year period, while providing an empirical comparison of municipalities and cities with county rights. The use of several measures of investment capacity allows this concept to be assessed from different perspectives within the same analytical framework.

The article is both cognitive and applied in nature. Its findings may support local policymakers in designing financial policies that maintain fiscal stability while strengthening investment capacity. Furthermore, this study contributes to the literature on the sustainability of local public finances and the functional diversity of local government units.

2. Literature Review

2.1. Financial Sustainability of Local Government Units: Concept and Determinants

2.1.1. Concept of Financial Sustainability of Local Government Units

Financial sustainability is one of the key concepts in the analysis of local government units. In Poland, it is shaped by the legal framework, particularly the requirement to maintain a current budget balance (current revenues ≥ current expenditures) and compliance with the individual debt limit, both of which enforce financial policy consistent with the principle of fiscal responsibility. Financial sustainability is a multidimensional concept [10] that can be examined from both a static (annual) and a dynamic (long-term) perspective [6,11]. In empirical research, it is commonly operationalized using indicators such as the operating surplus, the level of debt and debt-servicing costs, own-source revenues, and the structure of budget expenditures.

Financial sustainability refers to the ability of LGUs to pursue a stable and responsible fiscal policy. It also determines their capacity to perform both current and development tasks [12]. In practice, this requires that local governments are able to finance their statutory responsibilities with the resources at their disposal [13]. In the short term, this ability is expected to ensure a balance between revenues and expenditures, enabling local government units to finance current expenditures from current revenues, generate an operating surplus (defined as the positive difference between current revenues and current expenditures, should not be confused with the budget surplus, which is the positive difference between total budget revenues and total budget expenditures in a given fiscal year), and meet their debt obligations.

From a broader, dynamic (long-term) perspective, financial sustainability refers to the fiscal sustainability of local government units [5,7,14], understood as their ability to perform public tasks over the long term, including financing future investments [15,16], without excessive borrowing [17] or the risk of losing financial liquidity. In this context, the level and structure of debt, as well as debt-servicing costs, are of particular importance, as they may constrain the development potential of local government units [18]. Therefore, financial sustainability does not mean an absence of debt but rather maintaining it at a safe and manageable level, keeping in mind that its fluctuations are correlated with the electoral cycle [19,20].

2.1.2. Determinants of Financial Sustainability of Local Government Units

Long-term financial sustainability should also be associated with budget stability [1,5,16], particularly with regard to own-source revenues, which enhance the financial (revenue) autonomy of local government units [21] and strengthen their financial position and development potential [22]. Greater revenue autonomy also increases the capacity of local governments to make investment decisions [23]. However, their impact on the investment activity of local government units is not unequivocal and may depend on the specific characteristics of a given unit [24]. However, financial sustainability should be assessed not only in terms of revenue-generating capacity [25] but also with regard to the ability to pursue development policies and maintain debt within prudent and sustainable limits [26].

The literature on fiscal sustainability and public debt is extensive and continues to expand. A recent bibliometric review identified more than 320 publications in this field published between 1999 and 2023 [27]. Most of these studies originated from the United States, Japan, China, Germany, and Poland. Among the most influential contributors to this field is the Portuguese economist Antonio Afonso, who recently examined the causal links between fiscal and external sustainability in the EU [28].

Early empirical studies on fiscal sustainability primarily focused on individual countries or small groups of countries and examined causality and the relationship between government revenues and expenditures [29]. Subsequent research increasingly employed standard panel data techniques and panel cointegration analysis [30]. Annual data are most commonly used in empirical analyses [31], although studies based on quarterly data are also available [32].

More recent studies have adopted wavelet analysis as an innovative approach to examining fiscal sustainability, providing insights into long-term sustainability trends, fiscal policy responses to budget deficits, and the evolving dynamics of fiscal sustainability across countries [33]. Recent evidence also indicates that digital government enhances fiscal sustainability primarily through two mechanisms: improving administrative efficiency and strengthening government integrity. These mechanisms effectively curb the growth of local government debt while expanding fiscal space [34].

Demographic factors have also become an increasingly important theme in the fiscal sustainability literature [35,36,37,38]. A common finding of these studies is that demographic changes, including population ageing, fertility rates, labor market developments, and other demographic factors, substantially affect fiscal policy and fiscal sustainability, highlighting the need for well-informed policy decisions and timely public interventions [39].

Compared with the extensive literature on fiscal sustainability at the national level, research focusing on fiscal sustainability in the local government subsector remains considerably less developed, both internationally and in Poland [40]. Nevertheless, some studies have examined the impact of national fiscal policy on local fiscal sustainability (e.g., Xia & Zhang [41]).

Poland provides an important example of a decentralized fiscal system that remains highly dependent on intergovernmental transfers while exhibiting relatively low fiscal autonomy [42]. Similar institutional arrangements can also be observed in other countries [43]. Moreover, intergovernmental transfers have been shown to stimulate local investment, highlighting their developmental role [42], as demonstrated also in the Philippines [44]. In Ghana, such transfers have additionally been found to improve the quality of public services, increase local government revenues, and facilitate the reinvestment of resources in infrastructure development [45].

The operating surplus is a key component of financial sustainability [16] because it serves as an indicator of the current financial condition of local government units, provides the basic fiscal space for investment, secures debt servicing [3,10,46,47,48], and reflects the ability to finance development from own resources [49]. However, the limited revenue autonomy of local government units may hinder the generation of operating surpluses, thereby constraining their capacity to undertake investments financed through repayable sources. Moreover, the low degree of fiscal decentralization and the growing use of off-budget financing mechanisms may further undermine the fiscal sustainability of local government units [50]. In contrast, an excessive debt-servicing burden may reduce the investment capacity of local government units [6,9,17,46].

In recent years, the issue of local government fiscal sustainability has increasingly been framed within the broader concept of Sustainable Public Finance, which integrates fiscal stability with long-term social, environmental, and economic objectives. From this perspective, public finances should not be evaluated solely in terms of budgetary balance or debt levels, but also with regard to their capacity to support the sustainable and inclusive development of local communities [5,14]. In this context, particular importance is attached to Sustainable Development Goal 11 (SDG 11—Sustainable Cities and Communities), which aims to make cities and human settlements inclusive, safe, resilient, and environmentally sustainable.

As demonstrated by Kumar, Shandilya, and Varghese [51], the growing body of research on SDG 11 reflects increasing scholarly interest in issues related to the financing of urban infrastructure, the fiscal resilience of local governments, the quality of public service delivery, and the use of financial instruments supporting sustainable development. Recent studies further indicate that the fiscal sustainability of local governments has become a prerequisite for achieving the Sustainable Development Goals, as it enhances their capacity to respond effectively to challenges arising from climate change, demographic shifts, and the energy transition [5,6,52].

Consequently, the fiscal sustainability and investment capacity of local governments should be viewed not only as economic concepts but also as fundamental components of implementing sustainable development at the local level and strengthening the long-term resilience of cities and communities in line with the objectives of SDG 11 [51].

In summary, financial sustainability should not be regarded as an end in itself but rather as the foundation of the investment capacity of local government units. This relationship, however, is bilateral. A high operating surplus supports investment activity, whereas excessive debt-financed investment may undermine financial sustainability.

2.2. Investment Capacity of Local Government Units: Concept and Determinants

Investment capacity is one of the key dimensions of the functioning of local government units. Its importance has increased with the growing role of local governments in supporting sustainable territorial development through infrastructure investment and spatial planning [53]. In general terms, it refers to the ability (capacity) of a local government unit to plan (prepare projects), finance, and implement investment projects while maintaining financial sustainability. This capacity is determined by a range of factors, including financial, organizational, institutional, and legal conditions [2,54,55].

Financial factors are primarily associated with financial autonomy and the ability of a local government unit to obtain funds for investment projects. These include the use of own-source revenues, non-repayable external funds (grants, including EU funds), as well as repayable sources of financing, which often require a financial contribution from the beneficiary (e.g., loans, bonds, and EU financial instruments). The use of repayable financing is also linked to the ability of local government units to incur liabilities and service their debt [4,17,56]. However, financial resources alone are not sufficient to ensure successful implementation of development projects. Effective planning frameworks, strategic and managerial capacities, cross-sectoral cooperation, and adequate administrative capability are also essential for translating financial resources into completed investments [57,58,59].

Organizational factors are related to differences in the scope and scale of responsibilities assigned to particular types of local government units [24,55]. The type of local government unit and its level of socio-economic development [60] are, in turn, associated with its financial potential, reflected in the sources and structure of its revenues [3,56,61]. This potential is generated by the residents and economic entities operating within a given area, that is, by its demographic potential and tax base [4,62]. In this context, particular attention should be paid to the legal regulations governing the revenue system of individual local government units, as well as to the design of the public finance system as a whole. Special consideration should also be given to the unique position of cities with county rights within this system.

The investment capacity of a local government unit facilitates the implementation of development projects, particularly infrastructure and social investments that improve the quality of life of residents [4,5,63,64] and contribute to the achievement of sustainable development objectives [65]. By supporting sustainable development, these investments also strengthen the long-term resilience of local communities and their ability to respond to environmental, social, and economic challenges [66]. In empirical research, investment capacity is not directly observable and is therefore operationalized using proxy indicators [10,64]. The use of several indicators simultaneously makes it possible to capture different dimensions of investment capacity and reduces the risk of a one-sided interpretation of the results [5,6]. The most commonly used measures include the share of capital expenditure in total expenditure (reflecting the scale of a local government unit’s investment commitment and its ability to allocate resources to development purposes), capital expenditure per capita (assessing investment activity on a per capita basis and enabling comparisons across local government units of different sizes), and the level of the operating surplus, including operating surplus per capita. The diversity of measurement approaches highlights the multidimensional nature of investment capacity [11] and the need to analyze it within the broader context of public finance [5,7].

A review of the literature on financial stability, including the financial sustainability of local government units, as well as on the role and responsibility of local governments in promoting sustainable local development, not only confirmed the relevance of these issues but also provided the basis for formulating the following research hypotheses:

H1. 

A higher operating surplus is associated with greater investment capacity in local government units (LGUs).

H2. 

Higher debt-servicing costs are associated with lower investment capacity in LGUs.

H3. 

Greater revenue autonomy is positively associated with the investment capacity of LGUs.

H4. 

Cities with county rights exhibit higher investment capacity than municipalities without county status.

Despite the growing literature on local government financial sustainability, relatively limited attention has been paid to how its different dimensions are associated with investment capacity, particularly in longitudinal studies comparing different types of local government units. This study addresses this gap by examining the associations between selected indicators of financial sustainability and different measures of investment capacity over the 2018–2024 period and by comparing municipalities and cities with county rights operating within the same national institutional and fiscal framework.

3. Materials and Methods

3.1. Data and Variables

This study was conducted using data obtained from the Local Data Bank (Bank Danych Lokalnych, BDL) maintained by Statistics Poland for the period 2018–2024. The source data cover all municipalities and cities with county rights in Poland. At the end of 2024, there were 2477 municipalities in Poland, including 66 urban municipalities with the status of cities with county rights. Data for both categories of LGUs are available in the BDL at the voivodeship level, covering all 16 voivodeships. The dataset therefore contains 32 observations for each year (16 voivodeships × 2 LGU types), resulting in a total of 224 observations over the seven-year study period. Each observation refers to a specific LGU type within a given voivodeship and year rather than to an individual local government unit. Financial sustainability and investment capacity were operationalized using three variables for each construct (

Table 1. Variables used in the analysis.

Area of Analysis	Function of Variables	Description
sustainable finance	independent variables	Share of the operating surplus * in total revenues (operating_surplus_total_revenues)
		Share of debt-servicing costs in total expenditure (debt_servicing_total_expend)
		Share of own-source revenues in total revenues (own_to_total_revenues)
investment capacity	dependent variables	Share of capital expenditure in total expenditure (invest_capacity_1)
		Capital expenditure per capita (invest_capacity_2)
		Operating surplus per capita (invest_capacity_3)
LGU type	time-invariant explanatory variable	Type of local government unit: municipality or city with county rights (lgu_type)

* Operating surplus = current revenues − current expenditure.

). Statistical analyses were performed using R (version 4.4.1) in RStudio (version 2026.06.0).

Descriptive statistics by LGU type and year are presented in

Table 2. Descriptive statistics by LGU type.

	LGU Type	n	Mean	Median	SD	Min	Max
operating_surplus_total_revenues	municipality	112	0.0748	0.0771	0.0198	0.0266	0.116
	city with county rights	112	0.0485	0.0483	0.0355	−0.0795	0.116
debt_servicing_total_expend	municipality	112	0.00801	0.00675	0.00414	0.00236	0.0183
	city with county rights	112	0.0145	0.0113	0.0101	0.00212	0.0445
own_to_total_revenues	municipality	112	43.9	44.6	5.64	31.5	54.0
	city with county rights	112	55.1	54.9	6.92	42.3	71.2
invest_capacity_1	municipality	112	20.3	18.5	8.37	6.37	59.5
	city with county rights	112	16.8	15.7	6.71	4.63	46.9
invest_capacity_2	municipality	112	1243	1016	481	592	2940
	city with county rights	112	1402	1291	440	542	2845
invest_capacity_3	municipality	112	459	453	127	161	771
	city with county rights	112	399	386	308	−728	1193

and

Table 3. Descriptive statistics by year.

	Year	n	Mean	Median	SD	Min	Max
operating_surplus_total_revenues	2018	32	0.0801	0.0816	0.0208	0.0227	0.116
	2019	32	0.0725	0.0752	0.0168	0.0381	0.100
	2020	32	0.0557	0.0643	0.0245	−0.00102	0.0919
	2021	32	0.0947	0.0955	0.0150	0.0466	0.116
	2022	32	0.0573	0.0661	0.0282	−0.00599	0.107
	2023	32	0.0249	0.0316	0.0287	−0.0795	0.0609
	2024	32	0.0465	0.0535	0.0268	−0.0223	0.0881
debt_servicing_total_expend	2018	32	0.00754	0.00656	0.00306	0.00398	0.0145
	2019	32	0.00778	0.00671	0.00267	0.00466	0.0145
	2020	32	0.00597	0.00492	0.00215	0.00346	0.0114
	2021	32	0.00408	0.00327	0.00178	0.00212	0.00886
	2022	32	0.0149	0.0122	0.00680	0.00767	0.0361
	2023	32	0.0200	0.0164	0.00928	0.00980	0.0445
	2024	32	0.0186	0.0152	0.00942	0.00830	0.0435
own_to_total_revenues	2018	32	49.9	49.0	10.1	32.2	71.2
	2019	32	48.7	47.4	8.60	32.9	68.4
	2020	32	48.0	46.4	7.20	34.2	64.5
	2021	32	47.5	46.7	8.17	31.5	65.3
	2022	32	51.0	49.7	7.96	37.2	70.1
	2023	32	50.2	49.3	8.12	34.1	70.4
	2024	32	51.0	50.3	8.78	34.4	70.1
invest_capacity_1	2018	32	20.1	19.1	6.34	8.58	44.7
	2019	32	18.0	17.1	5.74	6.81	39.0
	2020	32	16.0	15.0	6.41	4.63	38.0
	2021	32	16.1	15.1	6.48	5.82	34.9
	2022	32	17.7	17.3	6.54	5.66	38.9
	2023	32	23.6	23.6	10.3	6.58	59.5
	2024	32	18.5	18.7	9.33	4.84	54.
invest_capacity_2	2018	32	1135	1077	268	785	1864
	2019	32	1118	979	347	745	2215
	2020	32	1053	998	396	542	2448
	2021	32	1129	1006	463	671	2758
	2022	32	1332	1278	261	838	2102
	2023	32	1869	1797	401	1224	2845
	2024	32	1623	1596	424	708	2940
invest_capacity_3	2018	32	448	425	147	148	923
	2019	32	444	436	95.5	292	754
	2020	32	362	391	144	−9.78	579
	2021	32	712	703	166	390	1122
	2022	32	429	463	210	−51.5	878
	2023	32	179	226	235	−728	544
	2024	32	427	456	270	−246	1193

.

3.2. Analysis of Relationship Between the Financial Sustainability Indicators and Investment Capacity

For this analysis, panel regression models were employed. The choice of this specification was motivated by the panel structure of the data and the results of the Breusch–Pagan Lagrange Multiplier (LM) test and the F test (Chow test). These tests indicated that panel regression models were more appropriate than pooled ordinary least squares (pooled OLS) models. The Hausman test was used to choose between fixed-effects (FE) and random-effects (RE) panel models. The FE model accounts for unobserved heterogeneity by allowing unit-specific effects to be correlated with the explanatory variables. In contrast, the RE model assumes that unit-specific effects are random and uncorrelated with the explanatory variables.

In the study, the fixed-effects (FE) panel model took the following form:

$$\begin{gathered} {invest\_capacity}_{it}= {\propto }_i+{\beta }_1{operating\_surplus\_total\_revenues}_{it} \\ +{\beta }_2{debt\_servicing\_total\_expend}_{it} \\ +{\beta }_3{own\_to\_total\_revenues}_{it}+\sum _{r=1}^{R-1}{\gamma }_r{year}_{r,it}+{\epsilon }_{it} \end{gathered}$$

(1)

where

• ${invest\_capacity}_{it}$—investment capacity of LGU i in year t;
• ${operating\_surplus\_total\_revenues}_{it}$—operating surplus to total revenues ratio of LGU i in year t;
• ${debt\_servicing\_total\_expend}_{it}$—debt-servicing costs to total expenditure ratio of LGU i in year t;
• ${own\_to\_total\_revenues}_{it}$—own-source revenues to total revenues ratio of LGU i in year t,
• ${year}_{r,it}$—year dummy variables controlling for time-specific effects;
• ${\propto }_i$—unit-specific fixed effect, representing all unobserved, time-invariant characteristics of local government unit i;
• ${\beta }_1,{\beta }_2,{\beta }_3$—coefficients associated with the explanatory variables;
• ${\gamma }_r$—coefficients associated with the year dummy variables (reference year omitted);
• ${\epsilon }_{it}$—idiosyncratic error term.

The random-effects (RE) panel model took the following form:

$$\begin{gathered} {invest\_capacity}_{it }= {\beta }_0+\sum _{k=1}^{K-1}{\propto }_k{lgu\_type}_{k,it} \\ +{\beta }_1{operating\_surplus\_total\_revenues}_{it} \\ +{\beta }_2{debt\_servicing\_total\_expend}_{it} \\ +{\beta }_3{own\_to\_total\_revenues}_{it}+\sum _{r=1}^{R-1}{\gamma }_r{year}_{r,it}+u_i+{\epsilon }_{it} \end{gathered}$$

(2)

where

• ${\beta }_0$— the overall intercept (constant term);
• ${lgu\_type}_{k,it}$—dummy variables representing the type of local government unit;
• ${\propto }_k$—coefficients associated with the categories of LGU type (reference category omitted);
• $u_i$—unit-specific random effect.

3.3. Analysis of Differences Between Municipalities and Cities with County Rights

To examine differences in investment capacity between municipalities and cities with county rights while accounting for the panel structure of the data, correlated random-effects (CRE) models based on the Mundlak [67] approach were estimated. This approach allows time-invariant differences associated with LGU type to be estimated while accounting for potential correlation between the time-varying covariates and unobserved unit-specific heterogeneity.

4. Results

4.1. Panel Model for Invest_Capacity_1

For investment capacity measured as the share of capital expenditure in total expenditure, the Hausman test indicated that the random-effects (RE) model was more appropriate. The p-value of the test was 0.7875, providing no grounds for rejecting the null hypothesis that the unit-specific effects are uncorrelated with the explanatory variables. The Breusch–Pagan LM test confirmed the superiority of this specification over the pooled ordinary least squares (pooled OLS) model, yielding a p-value of 2.2 × 10⁻¹⁶. This allowed the null hypothesis of no significant individual effects to be rejected, indicating that the use of a panel model (RE or FE) was justified.

The estimated RE-1 model was statistically significant overall (p < 0.001), and it explained approximately 43% of the variance in the dependent variable (R² = 0.429), indicating a moderate goodness of fit. Given that invest_capacity_1 (capital expenditure as a share of total expenditure) and the debt-servicing-cost indicator (debt-servicing costs as a share of total expenditure) share total expenditure as a common denominator, the possibility of a mechanically induced association between these variables was additionally examined. As a robustness check, an alternative specification including only the debt-servicing-cost indicator was estimated. This specification yielded an R² of 0.020, indicating that the debt-servicing-cost ratio alone accounted for only a small proportion of the variation in invest_capacity_1. Although this check does not rule out an association arising partly from the common denominator, it reduces the concern that the explanatory power of the full RE-1 model is primarily driven by the shared-component structure of the two indicators.

The following diagnostic tests were performed for the model:

• The Durbin–Watson and Breusch–Godfrey tests for residual autocorrelation yielded a chi-square statistic of 41.028 (df = 7, p = 7.996 × 10⁻⁷), indicating the presence of within-unit serial correlation;
• The Breusch–Pagan test for heteroskedasticity yielded a BP statistic of 18.392 (df = 10, p = 0.0487), indicating heteroskedasticity of the error terms;
• Pesaran’s test for cross-sectional dependence in panel data yielded z = −0.89151 (p = 0.3727), providing no significant evidence of cross-sectional dependence in the model residuals across local government units.

Additionally, multicollinearity among the explanatory variables was assessed using the Variance Inflation Factor (VIF). For this purpose, an auxiliary OLS model was estimated. The highest VIF value was 3.645354, while the GVIF^1/(2Df) value for the factor variable (year) was 1.125718. As all values were below 5, no significant multicollinearity was detected.

To account for heteroskedasticity and within-unit serial correlation, inference was based on Arellano-type cluster-robust standard errors, clustered at the panel-unit level. Given the relatively limited number of panel units, the HC3 (Heteroskedasticity-Consistent type 3) correction was used to provide more conservative finite-sample inference (

Table 4. RE-1 model estimates with HC3 cluster-robust standard errors.

Coefficients	Estimate	Std. Error	z-Value	p-Value	Significance
Intercept	50.45178	7.87630	6.4055	9.446 × 10−10	***
factor (lgu_type)	6.26877	2.86696	2.1866	0.0298643	*
operating_surplus_total_revenues	11.75853	16.53175	0.7113	0.4776955
debt_servicing_total_expend	−358.61529	158.60958	−2.2610	0.0247701	*
own_to_total_revenues	−0.63612	0.18583	−3.4231	0.0007423	***
factor (year) 2019	−2.66899	0.41635	−6.4104	9.195 × 10−10	***
factor (year) 2020	−5.52803	0.71582	−7.7226	4.392 × 10−13	***
factor (year) 2021	−6.93261	1.10118	−6.2956	1.722 × 10−9	***
factor (year) 2022	1.21436	1.00696	1.2060	0.2291668
factor (year) 2023	8.81663	1.89934	4.6419	6.031 × 10−6	***
factor (year) 2024	3.45298	1.99771	1.7285	0.0853527
Total Sum of Squares:	4704.3
Residual Sum of Squares:	2685.1
R-Squared:	0.42923
Adjusted R-Squared:	0.40243
Chisq: 160.179 on 10 DF, p-value: <2.22 × 10−16

Note: Significance levels: *** p < 0.001; * p < 0.05.

).

4.2. Panel Model for Invest_Capacity_2

The Hausman test for investment capacity measured as capital expenditure per capita strongly rejected the null hypothesis that individual effects are uncorrelated with the explanatory variables (p = 2.982 × 10⁻¹¹), confirming that the fixed-effects (FE) model is appropriate. Additionally, the F test (Chow test) confirmed the superiority of this specification over the pooled OLS model (p = 1.353 × 10⁻¹⁰), allowing the null hypothesis of no significant unit-specific effects to be rejected.

The estimated FE-2 model was statistically significant overall (p < 0.001). According to the within R² value of 0.633, it explained approximately 63% of the within-unit variation in the dependent variable, indicating a good fit to the data.

The following diagnostic tests were performed:

• The Durbin–Watson and Breusch–Godfrey tests for residual autocorrelation yielded a chi-square statistic of 51.97 (df = 7, p = 5.919 × 10⁻⁹), indicating the presence of within-unit serial correlation;
• The Breusch–Pagan test for heteroskedasticity yielded a BP statistic of 6.1814 (df = 9, p = 0.7216), indicating no heteroskedasticity of the error terms;
• Pesaran’s test for cross-sectional dependence in panel data yielded z = −0.55326 (p = 0.5801), providing no significant evidence of cross-sectional dependence.

Multicollinearity among the explanatory variables had already been assessed for the first model (RE-1). As the FE-2 model included the same set of explanatory variables, there was no need to repeat the analysis.

Although the diagnostic tests did not indicate heteroskedasticity in the FE-2 model, within-unit serial correlation was detected. Accordingly, Arellano-type standard errors clustered at the panel-unit level were used. The HC3 correction was applied consistently across all model specifications, providing a uniform and conservative approach to inference given the relatively limited number of panel units (

Table 5. FE-2 model estimates with HC3 cluster-robust standard errors.

Coefficients	Estimate	Std. Error	t-Value	p-Value	Significance
operating_surplus_total_revenues	1636.648	1486.753	1.1008	0.2724206
debt_servicing_total_expend	−29,177.906	10,586.544	−2.7561	0.0064410	**
own_to_total_revenues	−78.508	22.787	−3.4452	0.0007075	***
factor (year) 2019	−95.18	34.173	−2.7853	0.0059111	**
factor (year) 2020	−237.755	63.316	−3.7550	0.0002327	***
factor (year) 2021	−318.963	95.377	−3.4447	0.0007089	***
factor (year) 2022	536.274	91.432	5.6227	6.932 × 10−8	***
factor (year) 2023	1210.358	143.187	8.4530	8.735 × 10−15	***
factor (year) 2024	950.427	144.344	6.5844	4.684 × 10−10	***
Total Sum of Squares:	36,776,000
Residual Sum of Squares:	13,479,000
R-Squared:	0.63348
Adjusted R-Squared:	0.55337
F-statistic: 35.1433 on 9 and 183 DF, p-value: <2.22 × 10−16

Note: Significance levels: *** p < 0.001; ** p < 0.01.

).

4.3. Panel Model for Invest_Capacity_3

For investment capacity measured as operating surplus per capita, the fixed-effects (FE) model was also found to be more appropriate (p = 8.242 × 10⁻⁵ in the Hausman test). The F test (Chow test) for individual effects confirmed that this specification was preferred over the pooled OLS model (p = 1.799 × 10⁻¹⁵).

The model was statistically significant overall (p < 0.001). With an R² value of 0.9643, it explained approximately 96% of the within-unit variation in the dependent variable. However, the high explanatory power of this specification should be interpreted with caution because the dependent variable, invest_capacity_3 (operating surplus per capita), and one of the explanatory variables, operating_surplus_total_revenues (operating surplus as a share of total revenues), share operating surplus as a common component. This construction may mechanically strengthen the observed association between the two variables. Taking this into account, additional specifications were estimated to assess the extent to which this shared component might affect model fit. In a fixed-effects specification containing only the operating-surplus-to-total-revenues ratio as an explanatory variable, the within R² was 0.804. However, the model excluding this variable still yielded a relatively high within R² of 0.694, while a specification containing year effects alone produced a within R² of 0.570. These results indicate that the shared operating-surplus component may contribute to the strong explanatory power of the full model but does not account for it entirely. A substantial part of the within-unit variation in operating surplus per capita is also associated with common time-specific variation captured by the year effects. Accordingly, the coefficient on the operating-surplus ratio can be interpreted cautiously as an association rather than as evidence of an independent causal effect.

The following diagnostic tests for the model were performed:

• The Durbin–Watson and Breusch–Godfrey tests for residual autocorrelation yielded a chi-square statistic of 28.429 (df = 7, p = 0.0001838), indicating the presence of within-unit serial correlation;
• The Breusch–Pagan test for heteroskedasticity yielded a BP statistic of 31.283 (df = 9, p = 0.0002647), indicating heteroskedasticity of the error term;
• Pesaran’s test for cross-sectional dependence in panel data yielded z = −0.9307 (p = 0.352), providing no significant evidence of cross-sectional dependence.

As the set of explanatory variables remained unchanged, there was no need to reassess multicollinearity.

Diagnostic tests indicated both heteroskedasticity and within-unit serial correlation in this model. Accordingly, inference was based on Arellano-type cluster-robust standard errors, clustered at the panel-unit level. The HC3 correction was applied consistently with the other model specifications, providing a uniform and conservative approach to inference given the relatively limited number of panel units (

Table 6. FE-3 model estimates with HC3 cluster-robust standard errors.

Coefficients	Estimate	Std. Error	t-Value	p-Value	Significance
operating_surplus_total_revenues	9594.0525	860.4181	12.0933	<2.2 × 10−16	***
debt_servicing_total_expend	−4397.9090	2004.3132	−2.1942	0.02948	*
own_to_total_revenues	−8.1717	3.6615	−2.2318	0.02684	*
factor (year) 2019	60.1103	6.3961	9.3979	<2.2 × 10−16	***
factor (year) 2020	126.4552	17.5105	7.2217	1.334 × 10−11	***
factor (year) 2021	88.9335	13.2706	6.7016	2.469 × 10−10	***
factor (year) 2022	241.7924	14.8647	16.2663	<2.2 × 10−16	***
factor (year) 2023	317.8076	26.3531	12.0596	<2.2 × 10−16	***
factor (year) 2024	358.8797	23.8464	15.0496	<2.2 × 10−16	***
Total Sum of Squares:	8,296,300
Residual Sum of Squares:	296,190
R-Squared:	0.9643
Adjusted R-Squared:	0.95649
F-statistic: 549.198 on 9 and 183 DF, p-value: <2.22 × 10−16

Note: Significance levels: *** p < 0.001; * p < 0.05.

).

4.4. Correlated Random-Effects (CRE) Models Based on the Mundlak Approach

The CRE models were estimated separately for each of the three measures of investment capacity. Municipalities were used as the reference category; therefore, positive coefficients for cities with county rights indicate higher conditional values of the respective investment-capacity measures relative to municipalities. The estimated coefficients for cities with county rights, HC3 cluster-robust standard errors, corresponding p-values, and the results of the joint Mundlak tests (χ²(3) and p-values) are reported in

Table 7. CRE/Mundlak model results of differences in investment capacity by LGU type.

Investment Capacity Measure	City Coefficient (β)	HC3 Robust Std. Error	p-Value	Joint Mundlak χ2(3)	Mundlak p-Value
invest_capacity_1	4.20	3.50	0.231	3.679	0.298
invest_capacity_2	338.17	166.95	0.044	25.385	0.00001283
invest_capacity_3	141.53	24.41	2.440 × 10−8	20.047	0.000166

Note: The joint Mundlak test evaluates the null hypothesis that the coefficients on the unit-specific means of the time-varying covariates are jointly equal to zero.

.

As shown in Table 7, the coefficients for cities with county rights were positive across all three measures of investment capacity, although their statistical significance differed across specifications. The difference was not statistically significant for the share of capital expenditure in total expenditures (invest_capacity_1), whereas cities with county rights exhibited significantly higher capital expenditure per capita (invest_capacity_2) and operating surplus per capita (invest_capacity_3). The joint Mundlak tests were statistically significant for invest_capacity_2 and invest_capacity_3, indicating that the unit-specific means of the time-varying covariates were jointly significant and favoring the CRE specification over conventional RE.

The result for invest_capacity_3 should additionally be interpreted in light of the shared operating-surplus component between the dependent variable and one of the explanatory variables.

4.5. Empirical Assessment of Research Hypotheses

H1. 

A higher operating surplus is associated with greater investment capacity in local government units (LGUs)—Hypothesis partially confirmed.

The descriptive correlation analysis indicated a positive association between the operating surplus and investment capacity measured as operating surplus per capita, with Pearson correlation coefficients of 0.61 for municipalities and 0.69 for cities with county rights. This observation is consistent with the literature, which identifies the operating surplus as the primary source of investment financing and an important determinant of the development potential of local government units [3,10]. The operating surplus was also positively correlated with capital expenditure in both municipalities and cities with county rights:

• Municipalities: r = 0.58;
• Cities with county rights: r = 0.62.

The results from the panel models showed that the coefficients for the variable operating_surplus_total_revenues were positive across the estimated models; however, they were not statistically significant in the specifications using the first two measures of investment capacity. A positive and statistically significant coefficient was obtained only for invest_capacity_3; however, as discussed above, this result should be interpreted with caution because the dependent variable and explanatory variable, operating_surplus_total_revenues, share operating surplus as a common component, which may mechanically strengthen the observed association. Nevertheless, additional model specifications showed that the high explanatory power of the FE-3 model was not attributable solely to this shared component, as substantial within-unit variation in invest_capacity_3 was also explained when operating_surplus_total_revenues was excluded from the model (R² = 0.694).

H2. 

Higher debt-servicing costs are associated with lower investment capacity of LGUs—Hypothesis confirmed.

Higher debt-servicing costs were significantly associated with lower investment capacity of local government units. This finding is consistent with the literature, which indicates that rising debt-servicing burdens may reduce the fiscal space available to local governments and constrain their capacity to finance development projects [6,9,18]. In all panel regression models, the coefficient for the debt-servicing costs indicator was negative and statistically significant based on HC3 cluster-robust standard errors:

• Model RE-1: coefficient = −358.62; p = 0.0247701;
• Model FE-2: coefficient = −29,177.91; p = 0.006441;
• Model FE-3: coefficient = −4397.91; p = 0.02948.

H3. 

Greater revenue autonomy is positively associated with the investment activity of LGU—Hypothesis not confirmed.

The analyses did not provide sufficient evidence that there was a positive association between greater revenue autonomy and the investment activity of LGUs. The descriptive Pearson correlation coefficients indicated generally weak or very weak bivariate associations between the share of own-source revenues in total revenues and the measures of investment capacity. For municipalities, the strongest correlation was observed for investment capacity measured as the share of capital expenditure in total expenditure, but the association was negative (r = −0.37). In cities with county rights, the highest correlation was observed between the share of own-source revenues in total revenues and operating surplus per capita, although its magnitude remained weak (r = 0.24).

The panel regression results provided stronger evidence against the direction proposed in the hypothesis. Contrary to the hypothesized positive association, the coefficient for the revenue autonomy indicator was negative and statistically significant in all three models:

• Model RE-1: coefficient = −0.63612; p = 0.0007423;
• Model FE-2: coefficient = −78.508; p = 0.0007075;
• Model FE-3: coefficient = −8.1717; p = 0.02684.

Thus, after accounting for the other variables included in the respective panel specifications, greater revenue autonomy was consistently associated with lower, rather than higher, values of the investment capacity measures. Accordingly, H3 was not supported; instead, the panel estimates indicated a statistically significant negative association between revenue autonomy and investment capacity. This finding may point to the complex nature of the relationship between revenue autonomy and investment activity [68] and suggest that additional financial and institutional factors may influence investment capacity [5,24].

H4. 

Cities with county rights exhibit higher investment capacity than municipalities without county status—Hypothesis confirmed.

The positive and statistically significant coefficient for lgu_type = “city with county rights” in the RE-1 model (p = 0.0299) supports H4 for invest_capacity_1. For invest_capacity_2 and invest_capacity_3, the assessment of H4 was based on the CRE models because the corresponding FE-2 and FE-3 models cannot estimate the coefficient for LGU type, which is time-invariant. As shown in Table 7, the coefficients for cities with county rights in both CRE models were positive and statistically significant. Moreover, the joint Mundlak tests were statistically significant for invest_capacity_2 and invest_capacity_3 (p < 0.001), indicating that the unit-specific means of the time-varying covariates were jointly significant and supporting the use of the CRE specification rather than the conventional RE specification for these models. Accordingly, H4 was supported for invest_capacity_2 and invest_capacity_3. As discussed above, the result for invest_capacity_3 should be interpreted in light of the previously identified shared-component issue.

For invest_capacity_1, the coefficient for cities with county rights remained positive in the CRE model but was not statistically significant. However, the joint Mundlak test was not statistically significant (p = 0.298), providing no evidence that the conventional RE specification needed to be extended to the CRE specification. This result is consistent with the earlier model-selection procedure supporting RE-1. Therefore, inference for invest_capacity_1 was based primarily on the preferred RE-1 specification, while the non-significant CRE estimate indicates some sensitivity to model specification.

Overall, the results supported H4 for all three measures of investment capacity, although the evidence for invest_capacity_1 was sensitive to model specification.

It is worth noting that the unadjusted mean values of invest_capacity_1 show a different pattern: municipalities recorded a higher average share of capital expenditure in total expenditure (20.3%) than cities with county rights (16.8%). These descriptive means, however, represent an unconditional comparison between the two LGU types and therefore differ from the conditional comparison provided by the RE-1 model. To examine this difference more closely, a post-estimation marginal comparison was performed based on the RE-1 specification. The model-adjusted difference between cities with county rights and municipalities was 6.27 percentage points in favor of cities (95% CI: 0.84–11.70; p = 0.024). Thus, after adjustment within the RE-1 specification, the direction of the difference was reversed relative to the unadjusted group means.

To further investigate this reversal, sequential RE specifications were estimated. The initial coefficient for cities with county rights (−4.684) remained negative after the inclusion of the financial covariates (−2.637) and became positive (+6.269) only after year effects were introduced. This indicates that the reversal was associated primarily with accounting for common year-specific variations rather than with adjustments for the financial covariates alone. Accordingly, the higher unadjusted mean for municipalities and the positive conditional coefficient for cities with county rights in RE-1 are not contradictory; they reflect unconditional and model-adjusted comparisons, respectively.

4.6. Results Summary

The results indicate a consistent positive tendency in the relationship between operating surplus and investment capacity, while higher debt-servicing costs were significantly associated with lower investment capacity. The evidence was insufficient to support the hypothesized positive associations of revenue autonomy with investment capacity. In fact, revenue autonomy was found to be negatively associated with investment capacity across the estimated panel models. The application of panel regression models with time effects made it possible to examine these relationships while accounting for both unit-specific heterogeneity and changes over time. The findings provide a basis for further research on the links between the fiscal sustainability of local governments and their investment capacity, particularly with regard to the mechanisms underlying the negative association between revenue autonomy and investment capacity. They may also provide useful evidence for the development of policy solutions supporting the financial management and investment capacity of local government units.

5. Discussion

The results suggest a positive tendency in the relationship between operating surplus and the investment capacity of local government units, although the evidence is not consistent across the different measures of investment capacity. Descriptive correlation analysis indicated positive bivariate associations between operating surplus and selected measures of investment activity; however, these correlations alone cannot provide robust evidence of such a relationship. The panel regression models showed that this relationship was not equally evident across all measures of investment capacity used in the study. This suggests that the relationship between the operating surplus and the investment capacity of local government units should be interpreted cautiously and in light of the measure of investment capacity adopted, as well as other financial conditions.

These findings are consistent with previous studies showing that the level of the operating surplus is most strongly associated with the investment and development capacity of municipalities and influences the scope of local government tasks performed [69]. Similar conclusions were reached by Cyburt and Gałecka [70], who identified the operating surplus as one of the key determinants of the financial sustainability and investment potential of local government units. At the same time, they emphasized that the relationship between the operating surplus and investment activity was not uniform across different types of local government units. The findings of Wójtowicz and Hodžić [71], highlighting the importance of the operating surplus for the financial resilience of large cities, further complement this interpretation by emphasizing the role of internal financial capacity in shaping local governments’ development opportunities.

The findings of the present study provide limited support for Hypothesis H1. Operating surplus was not significantly associated with the two direct investment-expenditure measures in the panel models using cluster-robust standard errors. A significant positive association was observed only when operating surplus per capita was used as the measure of investment capacity. This result should, however, be interpreted with particular caution because the operating-surplus-to-total-revenues ratio and operating surplus per capita share operating surplus as a common component. The strength of this association may therefore partly reflect this shared component rather than an independent relationship between financial sustainability and actual investment activity. The results therefore indicate that the relationship between operating surplus and investment capacity depends on the measure adopted and should not be interpreted as general evidence of a positive association with investment activity. The analysis showed that a higher share of debt-servicing costs in total expenditure was significantly associated with lower investment capacity in local government units. This finding suggests that, in addition to the level of indebtedness, the way debt is financed may also be important, particularly the repayment period and the cost of capital. This is consistent with the observations of Kotlińska [72], who argues that financing investments with debt characterized by excessively short repayment periods may place an excessive burden on the current budgets of local government units and reduce their development potential. Similar conclusions can also be drawn from the study by Chen et al. [9], which showed that the ability to finance investments is closely related to the revenue-generating capacity of local governments and to the legal framework governing local public finances. Consequently, the recommendation made by Kotlińska [10], that the repayment period of debt should be aligned with the economic useful life of the financed assets, appears well justified. Such an approach may reduce the risk of placing an excessive burden on current budgets and help maintain the investment capacity of local government units. The results consistently support H2, as higher debt-servicing costs were significantly associated with lower investment capacity across all estimated models.

No positive relationship was found between the level of revenue autonomy and the investment activity of local government units. A high share of own-source revenue does not therefore automatically translate into a greater scale of investment. In fact, the panel regression results consistently indicated a negative and statistically significant association between revenue autonomy and investment capacity, regardless of the measure of investment capacity used. This finding is consistent with the results reported by Surówka and Rechul [73], who emphasize the importance of revenue autonomy for the development activity of municipalities, while also pointing out that the developmental potential of own-source revenues should be assessed in conjunction with the operating surplus and the level of indebtedness. Similar conclusions were reached by Kluza and Wójtowicz [74], who showed that the investment activity of municipalities is shaped not only by own-source revenues but also by transfers from the state budget. The findings therefore suggest that the level of revenue autonomy alone does not fully explain differences in the investment activity of local government units. Units with similar levels of revenue autonomy may exhibit different levels of investment activity, indicating the importance of other sources of financing as well as additional financial conditions. Consequently, H3 was not supported; however, the panel models consistently indicated a statistically significant negative association between revenue autonomy and investment capacity.

Cities with county rights recorded higher capital expenditure per capita than municipalities, while the results for the share of capital expenditure in total expenditure were more nuanced, as discussed above. Overall, the comparison between LGU types depends on the measure of investment capacity adopted and, in the case of the capital expenditure share, on whether unconditional or model-adjusted differences are considered. The findings therefore do not indicate a universally higher investment capacity of either LGU type. Similar conclusions were reported by Przybyła et al. [75], who found that administrative status may contribute to higher investment activity in cities, although its importance should be considered together with revenue potential and local development policy. A comparable pattern was identified by Haraldsvik et al. [76], who demonstrated that the level of local government investment is primarily associated with the fiscal potential of local government units and with infrastructure needs resulting from demographic changes. The findings of the present study are consistent with these observations. The higher capital expenditure per capita recorded in cities with county rights may be explained by the broader scope of public tasks performed by these units and their greater investment needs. The results therefore provide support for Hypothesis H4.

The comparison between municipalities and cities with county rights further indicates that differences in investment capacity depend on the measure applied. Overall, the panel model results supported H4 for all three measures, indicating higher conditional values of investment capacity for cities with county rights than for municipalities. However, the magnitude and robustness of this advantage varied with the measure employed and the model specification, as discussed above. This is consistent with Oprea and Dascălu [77], who emphasize that local government investment activity is associated with a combination of financial, organizational, and institutional conditions. The results therefore support a multidimensional assessment of investment capacity rather than conclusions based on a single indicator. The correlation analysis also showed that the positive association between operating surplus and investment capacity was slightly stronger in cities with county rights than in municipalities. However, the differences in the correlation coefficients were small and should not be interpreted as evidence of systematically different relationships between the two types of LGUs. Rather, they provide additional descriptive evidence that the strength of the observed associations may vary across LGU types.

The findings also have practical relevance for the financial management of local governments. They indicate that investment capacity should be considered in the context of the current budgetary position of LGUs, particularly the burden of debt-servicing costs. The consistently negative association between debt-servicing costs and investment capacity across the analyzed specifications highlights the importance of careful debt management when planning investment expenditure.

6. Conclusions

The findings indicate a consistent positive tendency in the relationship between operating surplus and investment capacity, whereas a higher share of debt-servicing costs was significantly associated with lower investment capacity. By contrast, no evidence was found of the hypothesized positive association between revenue autonomy and investment activity (instead, a statistically significant negative association was observed across all three panel models), suggesting that this relationship is more complex. The results also confirm the existence of differences between municipalities and cities with county rights with respect to selected measures of investment capacity. In practice, this implies the need to conduct a financial policy that allows for maintaining a balance between financial stability and the implementation of investment projects.

This study contributes to the literature on local government finance by integrating the analysis of local governments’ financial sustainability with the assessment of their investment capacity over a multi-year period, while providing an empirical comparison of two types of local government units operating in Poland—municipalities and cities with county rights. This approach made it possible to examine how selected dimensions of financial sustainability are associated with investment capacity. The findings also contribute to the broader discussion on the long-term fiscal resilience of local government units, understood as their ability to maintain financial sustainability, adapt to changing socio-economic conditions, and continue delivering public services and implementing investment projects despite financial, demographic, or economic shocks [6]. The results indicate that maintaining a sustainable operating surplus while limiting the budgetary burden associated with debt servicing constitutes an important prerequisite for strengthening such fiscal resilience.

Nevertheless, this study has several limitations. The analysis was conducted for municipalities and cities with county rights in Poland over the 2018–2024 period and was based on financial indicators derived from data available in official public statistics. Consequently, this study did not account for qualitative factors, such as the quality of financial management, organizational capacity, the quality of investment project preparation, the ability to absorb external funding, or political and institutional conditions, all of which may also influence the investment activity of local government units.

The results should be interpreted as conditional associations rather than causal effects. The panel models account for common time effects, while the different panel specifications address unobserved unit-specific heterogeneity under different assumptions. In particular, the FE models control for time-invariant unobserved unit-specific characteristics, whereas the CRE specifications allow for correlation between unit-specific effects and the observed time-varying covariates through the inclusion of their unit-specific means. However, other time-varying factors not included in the models may be associated with both financial sustainability and investment capacity. Potential reverse relationships between investment activity and the analyzed financial indicators also cannot be excluded.

The findings may be useful for local government authorities in designing financial policies aimed at maintaining investment capacity while ensuring long-term financial sustainability. Furthermore, they may provide a rationale for further discussion on the directions of fiscal architecture reforms for local government units (LGUs) in Poland, particularly regarding the enhancement of their fiscal resilience, developmental capacity, and potential to finance investments that support Sustainable Development Goals (SDGs). In future research, it appears justified to incorporate a broader range of organizational and institutional factors, extend the analysis to remaining categories of LGUs, and cover a longer time horizon, which will allow for a more comprehensive explanation of the interdependencies between the financial stability, fiscal resilience, and investment activity of local governments.