Multinomial Logistic Analysis of SMEs Offering Green Products and Services in the Alps–Adriatic Macroregion

Abstract

We investigate what drives small and medium-sized enterprises (SMEs) in the Alps–Adriatic macro-region to offer green products and services. A sample of 2305 SMEs from Flash Eurobarometer 498 is analyzed with a multinomial logit model that distinguishes firms that already offer green products/services, those planning to do so within two years, and those with no such intention. SMEs implementing ≥ 5 resource-efficiency actions are 75% more likely to offer green products/services (RRR = 1.75, p < 0.05). An increase in the share of green jobs to >30% of staff triples that likelihood (RRR = 3.65, p < 0.01). Selling only services reduces the probability by 17%. Country dummies show Austria and Slovenia as early movers, while Croatia lags. This is the first study to apply a three-outcome model to green market activity in this macroregion, thus revealing nonlinear and country-specific patterns that a binary approach would mask. This study has multiple implications for entrepreneurial practice: (i) entrepreneurs should focus on high-leverage resource-efficiency bundles (energy, waste, circular design) rather than single actions; (ii) policymakers should combine financial incentives with green-skills vouchers to accelerate adoption in service-oriented SMEs.

1. Introduction

Global environmental trends urge prompt economic reaction and business shift towards sustainable climate paths [1]. Therefore, governments across the world have set a green transition as a sustainable development strategy. Previous experiences suggested that the important cause of environmental degradation lies in production and consumption systems [2,3]. SMEs play an important role in many economics and are recognized as important actors in the green transition. Namely, in the EU in 2023, 65,2% of employment was provided by SMEs, while the SMEs participated with 53,1% in enterprise value added [4]. Also, SMEs are important for economic growth across EU member states, as shown by [5].

This paper provides a new contribution to the extant literature on the SMEs’ path toward ensuring environmental sustainability and developing a circular economy in three distinct ways:

• We isolate this (post)transitional group of SMEs in a specific region and show its distinct drivers;
• We introduce the green jobs intensity variable to test its incremental effect;
• We provide the cross-country comparison within the Alps–Adriatic macro-region.

Our objective is to identify the factors that relate to the likelihood of SMEs offering green products and services in the context of the Alps–Adriatic macroregion [6] (p. 627) by addressing three research questions (RQs):

• RQ1. How does engagement in resource-efficiency actions affect an SME’s likelihood to offer (or plan to offer) green products/services?
• RQ2. How does the share of green jobs within an SME influence that likelihood?
• RQ3. Are there unspecified heterogeneities in green offer across the five Alps–Adriatic countries?

SMEs operating in the following countries are included in the analysis: Croatia, Italy, Slovenia, Austria, and Hungary, which is confirmed by Grulja et al. [7]. Specifically, the listed countries participate in the Alps–Adriatic–Alliance [8]. While Lippot [9] points out that “there is no clear or commonly accepted definition of the region geographically and politically described as the Alps–Adriatic”, he defines the Alps–Adriatic as a broader region that involves parts of Croatia, Italy, Slovenia, Austria, and Hungary. Our empirical analysis identifies the characteristics of SMEs in the Alps–Adriatic macroregion that offer or plan to offer green products or services. The research relied on the secondary data from Flash Eurobarometer 498 survey on “SMEs, resource efficiency and green markets” conducted in 2021.

Similarities and differences in meanings of goods, products, and services are discussed in [10] (p. 25), and the authors conclude that “there is no clear distinction for separating goods from services”. Thus, in order to keep consistency with the formulation from the questionnaire [11] and with the related work of Hoogendoorn et al. [12] (p. 764), the term “products and services” is used in the present study, which is also in line with the approach used by Maxwell et al. [13].

The categorical dependent variable on green products and services offerings had three modalities (answers: yes; no, but planning to do so and in the next two years; no and not planning to do so) is constructed and related to the resource efficiency actions, types of support needed, types of business output (products or services) and business relationship (B2B and/or other). The multinomial logit model was applied initially in an extended form that included firm size in terms of number of employees and scale of total revenues, as well as its age. However, further testing suggested a choice of the reduced model.

This paper continues with the description of the theoretical background (in Section 2), presentation of our methodological approach and the research materials (in Section 3), detailed explanation of the empirical results (in Section 4), their discussion (in Section 5) and the overall conclusion (in Section 6).

2. Theoretical Background

Constant economic growth globally has already led to concerns about natural resources and broader climate changes [14]. Sustainable development from the economic (and not personal usage of resources) point of view is closely connected to SMEs, especially their green practices. Our research framework concerns sustainability, sustainable working environment, and green products and services in small and medium-sized companies. SMEs are among the most important stakeholders in the European economy and have contributed highly to employment and GDP in the European Union [15]. SMEs can be agile and play a vital role in economic development. They can respond quickly to environmental demands and must adopt sustainable and green practices to save costs, enhance market positions, and achieve economic success and stability [16,17]. This is also why SMEs are so important in the research field of resource efficiency, green services and products, and green jobs.

Several aspects can be analyzed (from governmental to technological) [18]. However, our research will mainly focus on market aspects (or pressure to act resource efficiently from a green environmental point of view).

The following

Table 1. Key research constructs.

Term	Operational Definition (Source)
Green product	A tangible good whose primary purpose is to reduce environmental risks and minimize resource use across its life cycle [11].
Green service	An intangible offering designed and delivered with measurably lower environmental impact relative to the conventional alternative [11].
Green job	A position in which ≥50% of working time is devoted to activities that improve resource efficiency or lower emissions [19].
Resource-efficiency action (REA)	Any of the nine practices in Eurobarometer Q6–Q7 (energy saving, water saving, … circular design) [11].

provides operational definitions of our research constructs.

2.1. Green Products and Services

Future employment growth in the European Union is expected to be mainly driven by several public services, such as health and social work and the education system [20]. The education system is related to finding and developing talent and skills and, therefore, in direct and indirect connection with technological development, the correct way of using resources and data (in a green way) [21]. Also, business, transport and communication services, distribution, and retail are connected to fast technological development and other intensive research and development activities. The increase in service-sector employment was forecasted to be strongest in newer Member States and part of the Alps–Adriatic macroregion. This is also connected to a shift from polluting to cleaner sectors and increasing automation and robotization in this macroregion [20].

Compared with physical goods, services exhibit a markedly different environmental profile, with the bulk of environmental impact stemming from operational energy use and logistics rather than material inputs [14]. The intangibility of services complicates the communication and certification of environmental performance—service-oriented SMEs are less likely to adopt product eco-labels and instead gravitate toward process-based or organizational certifications [22]. We, therefore, expect service-only firms to display lower observed green-offering rates.

Other studies [14,22] have already shown that SMEs have implemented green practices within their business. They also notice two main drivers: internal companies’ motivation and achieving (or keeping) a good public image (as part of marketing strategy). Green practices also concern promoting resources, and knowledge that can help support environmentally friendly businesses [22] by developing, producing, and consuming green products and services.

2.2. Green Jobs

A green economy (inclusive of SMEs) is usually defined as one that minimizes negative environmental impacts (and maximizes environmental sustainability) while improving the well-being of local and global society through jobs and growth [14]. According to sustainable development awareness, companies must use their internal capacities to include environmentally sustainable practices in their working environment [21].

Green policies and business practices have positive and negative effects on employment. They can also create new jobs or preserve the existing ones. However, at the same time, environmental regulations can have several negative job consequences (by raising costs, reducing demand, and therefore pushing companies into an uncompetitive stage). Green jobs will be created by developing new technologies and their implementation in the business environment [23,24].

Green jobs can be short- or long-term-oriented. Construction and installation jobs are temporary. On the other hand, manufacturing and maintenance jobs are, in principle, longer-lasting [23,24]. It can be argued that human capital and innovation activities further drive green development [25]. Previous research [26] has also shown a positive correlation between green HRM practices and employees’ green behavior. This is also indirectly connected to better understanding and consuming green products and services.

2.3. Resource Efficiency

Innovation in a business environment has become an essential strategy for economic development and achieving competitive advantage [10]. Innovations are closely connected to the efficient use of resources, goods, and services. In addition, investment in green human-resource practices—such as eco-training, green performance metrics, and recruiting sustainability-focused staff—builds organizational capabilities for spotting inefficiencies and innovating resource-saving workflows [21,25]. De Andrade et al. [19] show that SMEs with higher shares of employees in green jobs adopt a broader set of resource-efficiency actions, underscoring the role of human capital in driving sustainability [19]. Accordingly, a greater green-job intensity within an SME can be expected to increase its likelihood of commercializing green products/services.

The primary purpose of resource efficiency is to ensure the most effective and sustainable way to minimize and reduce negative environmental impact while maximizing value for all stakeholders. In a business context, it means optimizing the inputs (for example, raw materials, energy sources, working and labor market) in the most productive manner and way while ensuring that the companies’ outputs (products or services) are produced in ethically and sustainably appropriate ways with the least possible quantity of resources. Resource efficiency can be achieved in three broad domains: technological (production processes and equipment), social (organizational behavior concerning resource use), and institutional (governance frameworks) [18]. Coordination of those three domains improves resource efficiency [27,28]; consequently, this generally improves companies’ environmental performance.

This can lead to better productivity, long-term economic, environmental, and social benefits, and overall successful business. Vučković and Čučković [29] applied a multinomial analysis on the World Bank Enterprise Survey 2019 subset of Western Balkan countries. They provide a closer insight into the characteristics of SMEs using less capital-intensive and more capital-intensive resource efficiency measures.

For SMEs, starting resource efficiency with strategies that include less financial input at the beginning (such as recycling materials, reducing waste, etc.) is more manageable. Such actions positively affect emissions reduction since they directly support local jobs (by contracting other organizations) [14,30].

From a business point of view, resource efficiency can be influenced by internal and external factors. Internal are mostly connected to financial and technological capacities and external to public support and knowledge networks [31]. This suggests that we can expect differences in testing B2B or B2C relationships between green jobs, green services, green products, and resource efficiency.

2.4. Managerial Motives and Policy Incentives

SMEs pursue environmental innovations for multiple overlapping reasons:

• Cost savings via reduced energy and material expenditures [14];
• Stakeholder and market pull, with green offerings strengthening the brand image and meeting customer demand [22];
• Access to external support, including financial incentives, technical consultancy, and market-identification assistance [18];
• Regulatory compliance under EU and national directives on resource efficiency [3].

These drivers frame the expectations on the impact of expressed support needs and help explain the empirically identified cross-country differences.

3. Materials and Methods

Flash Eurobarometer 498 survey was conducted to investigate factors determining the likelihood of a firm offering green products or services by applying multinomial logit analysis in Stata 18.5.

3.1. Data

Flash Eurobarometer 498 is the fifth wave of a survey on SMEs, Resource Efficiency, and Green Markets conducted from 8 November to 10 December 2021, by Ipsos European Public Affairs, applying quota sampling, cross-sectional design, and computer-assisted telephone interviews [11]. The sample selected for this study comprised the firm samples in five selected countries representing the Alps–Adriatic macroregion: Croatia, Italy, Hungary, Austria, and Slovenia. Furthermore, in line with Majid et al. [32] and de Andrade et al. [19], the sample was reduced only to firms reporting 1 to 250 employees (full-time employees) in order to obtain results relevant to SMEs. Of the 9159 establishments contacted in Flash Eurobarometer 498, 2784 belonged to the five target countries. We excluded enterprises with >250 employees (n = 423) and those lacking a response to variable Q9 (n = 56). The final analytical sample is 2305 SMEs (see

Table 2. Summary statistics.

Country	N	%
Croatia	461	20.0
Italy	502	21.8
Hungary	471	20.4
Austria	445	19.3
Slovenia	528	22.9
Sector	N	%
Retail (NACE G)	718	31.1
Manufacturing and Services	1587	68.9
Sell only services	N	%
Yes	911	39.5
No	1394	60.5
B2B focus	N	%
Yes	1808	78.5
No	497	21.5

below and

Table A1. Variable definitions, coding, and descriptive statistics.

N	2305
Does your company offer green products or service? (Dependent variable)
Yes—1 No, but you are planning to do so in the next 2 years—2 No, and you are not planning to do so—3 (base outcome)	818 (34.0%) 308 (12.8%) 1281 (53.2%)
Independent variables
What actions is your company undertaking to be more resource-efficient?
0—no action	138 (5.8%)
1—few actions (1–2 selected activities)	508 (21.3%)
2—some actions (3–4 selected activities)	796 (33.4%)
3—many actions (5 or more than 5 selected activities)	942 (39.5%)
What additional resource efficiency actions will your company plan to implement over the next two years?
0—no action	467 (20.1%)
1—few actions (1–2 selected activities)	498 (21.4%)
2—some or many actions (3 or more selected activities)	1362 (58.5%)
Are financial incentives for developing products, services, or new production processes needed?	1069 (44.4%)
Assistance with identifying potential markets or customers needed?	432 (17.9%)
Is technical support and consultancy needed for the development of products, services, and production processes?	553 (23.0%)
Consultancy services needed for marketing or distribution?	386 (16.0%)
The ratio between employees’ full time equivalent and full-time employees working in green jobs some or all of the time
≥0 and ≤5% (base value)	1395 (58.0%)
>5% and ≤10%	120 (5.0%)
>10% and ≤30%	228 (9.5%)
>30% and ≤100%	664 (27.6%)
Industry
0—Manufacturing (C), Services (H/I/J/K/L/M) or Industry (B/D/E/F)	1689 (70.2%)
1—Retail (G)	718 (29.8%)
Country
Croatia (base category)	461 (19.2%)
Italy	502 (20.9%)
Hungary	471 (19.6%)
Austria	445 (18.5%)
Slovenia	528 (21.9%)
Is your company selling services?
No	1496 (62.2%)
Yes	911 (37.8%)
Is your company selling its products or services to other companies?
Yes	599 (24.9%)
No	1808 (75.1%)

in Appendix A for complete definitions and summary statistics).

The dependent variable is categorical and based on answers to the survey: Does your company offer green products or services? Possible answers were: Yes—1 (base outcome); No, but planning to do so in the next 2 years—2; No and not planning to do so—3. Since it can be supposed that the two subgroups of firms that do not offer green products and services (firms that plan and firms that do not plan) significantly differ, it was decided to keep the multinomial structure of the variable leading to the choice of multinomial logit as a modeling strategy.

Instead, Vasilescu et al. [33] constructed a similar dependent variable using data from a previous wave Eurobarometer 456, keeping the binomial variable structure and applying the binomial logit model. Hoogendoorn et al. [12] also used one of the more distant waves of the survey Flash Eurobarometer 342 to investigate “greening products or services”. The relevant dependent variable was constructed based on questions: “How much did these green products or services represent in your turnover…?” [12].

Survey items are related to the present and planned actions of firms that increase resource efficiency: “What actions is your company undertaking to be more resource efficient?” and “Over the next two years, what are additional resource efficiency actions that your company plans to implement?” Nine types of activities were suggested as possible answers to both questions: saving water, saving energy, using predominantly renewable energy, saving materials, switching to greener suppliers of materials, minimizing waste, selling residues and waste to another company, recycling and designing products that are easier to maintain, repair or reuse [11]. One variable measures current resource efficiency and refers to the number of activities firms are currently undertaking. In contrast, another variable measures the number of resource efficiency actions firms intend to implement over the next two years.

In order to improve the understanding of firms that offer or intend to offer green products and services, their need for external support was investigated. Dummy variables indicated business needs for financial incentives, assistance identifying potential markets or customers, technical support, and consultancy services for marketing or distribution.

Furthermore, the importance of green jobs is measured by using a ratio between employees’ full-time equivalent and full-time employees working in green jobs some or all of the time, falling into ordinal categories no green job, up to 10%, between 10% and 30%, and above 30% (similarly constructed as in [19]).

The country and sector of activities were controlled by introducing relevant dummy variables. The initial model suggested no significant differences except for the retail sector. This was taken into account in the final model. Also, dummy variables for selling only services and for selling to other companies were added to the model.

3.2. Hypotheses

Our empirical analysis uncovers the characteristics of SMEs in the Alps–Adriatic macroregion that offer green products or services or of those that are planning to do so in the next 2 years compared to SMEs that do not offer green products and services and are not planning to do so. The following hypotheses were proposed and tested by estimating the multinomial logit model:

Hypothesis 1 (H1).

Engagement in resource efficiency actions increases the likelihood of offering green products and services in the Alps–Adriatic macroregion.

Vasilescu et al. [33] suggest that engagement in resource efficiency measures is positively associated with green products and services. Namely, firms that save water, energy, and materials, pursue circular economy practices, redesign their products, and select greener suppliers are more likely to offer green products and services [11].

Hypothesis 2 (H2).

SMEs declaring the need to use support mechanisms are more likely to feature green products and services in their offerings.

The extant literature recognizes the relationship between greening SME business and policy support. Vasilescu et al. [33] found that access to resources, including financial aid, expert consultancy, market identification assistance, and new product development support increases the likelihood that a company will green their business. They particularly emphasized financial incentives, identifying them as the strongest predictor of such transitions. Ayoungman et al. [34] show that financial considerations are pivotal in shaping a firm’s propensity to invest in environmentally sound technologies. Similarly, Akhtar et al. [35] documented a positive relationship between governmental support and outcomes in green innovation. Wang et al. [36] established that existing market conditions and regulatory frameworks play an important role in encouraging green production.

Hypothesis 3 (H3).

An increase in the proportion of green jobs within an SME influences the offer of green products and services by that firm.

The development of human capital is considered a conduit for the adoption of ecologically sound technologies and practices by the existing workforce [21]. In addition, increase in green jobs in SMEs can be linked to higher levels of resource efficiency [19], which suggests higher levels of sustainability awareness and orientation.

Hypothesis 4 (H4).

The service-only nature of an SME’s market offer influences the likelihood of green products or service provision.

Significant distinctions exist between goods and service categories, particularly concerning their production and delivery processes. These inherent differences may lead to varying propensities for green offerings among service-only companies. Hoogendoorn et al. [12] observed that firms concentrating exclusively on services tend to present fewer green offerings, a proposition that will be further scrutinized in our empirical analysis.

Hypothesis 5 (H5).

Engagement primarily in business-to-business (B2B) markets is hypothesized to be associated with a diminished likelihood of SMEs offering green products and services.

According to Hoogendoorn et al. [12], a primary focus on the B2B market tends to reduce a company’s inclination to engage in green business practices. This view is further supported by Hoejmose et al. [37], whose research indicates that business-to-consumer (B2C) companies tend to have more developed green supply chains than their B2B counterparts. It is also important to note that the type of market could unevenly influence different sustainability efforts. Bassi and Dias [38] observed that this effect can differ based on the specific sustainability-oriented action.

Our empirical analysis employs the multinomial logit model (MNLM). As an extension of the binomial logit model, the MNLM is particularly suitable for situations where the dependent variable is categorical and presents multiple, non-ordered choices (formally presented in Equation (1)) [39]:

$$ln{\Omega }_{\left(m|b\right)}\left(X\right)=\mathit{ln}{\displaystyle \frac{\mathit{Pr}(y=m|X)}{\mathit{Pr}\left(y=b|X\right)}}=X{\beta }_{m|b} for m=1 to J,$$

(1)

where b denotes a reference group (base category). In this study, the reference group (base category) has been set at the value of y = 3. This group consists of entities that currently ‘do not offer green products or services’ and, importantly, ‘have no current plans to undertake such initiatives’. The predicted probabilities have been computed using Equation (2):

$$\mathit{Pr}\left(y=m|X\right)={\displaystyle \frac{\mathit{exp}\left(X{\beta }_{m|b}\right)}{{\sum }_{j=1}^J\mathit{exp}\left(X{\beta }_{j|b}\right)}}.$$

(2)

To capture our three-category outcome (offerers vs. planners vs. non-offerers) without collapsing “planners” into a binary split, we employ a multinomial logit model (MNLM). This approach preserves the distinct transitional behavior of firms planning to go green (Long and Freese [39]; Greene et al. [40]). We confirmed the assumption of independence of irrelevant alternatives via Hausman and Small–Hsiao tests. Multicollinearity checks yielded a mean variance-inflation factor of 1.89—well below the standard threshold of 10—indicating no serious collinearity issues.

It should be noted that using secondary, cross-sectional data precludes causal inference and restricts us to variables included in the Eurobarometer. We control for size, sector and country dummies to mitigate omitted-variable bias.

4. Results

Apart from the presented variables, the initial model also included the independent variables of firms’ age, size, and revenue. It also involved several disjointed categories merged in the final model.

Our definition of the reduced model (as described by

Table 3. Fitness statistics.

Tests	Extended (Initial) Model	Reduced (Final) Model
Akaike’s Information Criterion (AIC)	4080.684	4034.298
Bayesian Information Criterion (BIC)	4563.083	4264.011

,

Table A2. Variance inflation factor.

Variable	VIF	Variable	VIF	Variable	VIF
re_present		sup_fin = 1	1.06	nace = 1	1.11
1	3.96	sup_mark = 1	1.06	ipscntry = 12	1.89
2	4.97	sup_tech = 1	1.07	17	1.71
3	4.39	sup_cons = 1	1.07	20	1.73
re_plan		green_job = 10	1.07	24	1.69
1	1.91	30	1.09	serv = 1	1.11
2	1.89	100	1.13	b2b = 1	1.02
Mean VIF = 1.89

, and

Table A3. Independent irrelevant alternatives (IIA) tests.

N = 2305; df = 20	Hausman Test		Suest–Based Hausman Tests		Small–Hsiao Tests Set Seed 1673029581
Alternatives	chi2	P > chi2	chi2	P > chi2
1	6.208	0.999	27.879	0.112	−328.641	−317.581	22.120	0.334
2	7.732	0.993	28.063	0.108	−572.194	−561.634	21.119	0.390
3	18.195	0.575	41.540	0.003	−274.385	−263.151	22.469	0.316
Ho: Odds (Outcome-J Vs. Outcome-K) are independent of other alternatives

, and visualized by Figure 1) is based on analyzing the goodness of fit. Akaike’s Information Criterion (AIC) of the final reduced model is 4034.298, while for the initial extended model, it amounts to 4080.684. The final model’s Bayesian Information Criterion (BIC) is 4264.011, while it amounts to 4563.083 in the initial model. All VIF values remain under 7, and the mean VIF is 1.89, suggesting no severe presence of multicollinearity. Housman and Small–Hsiao’s independent irrelevant alternatives (IIA) tests are satisfied in the final model.

Figure 2 provides a visual summary of the direction and statistical significance of the relationships between our main predictors and the likelihood of offering green products/services. Resource-efficiency actions, expressed support needs and green-jobs intensity all have significant positive associations with green offering status (solid gray arrows), whereas “sell only services” and B2B sales do not reach significance (dashed arrows).

Reported results in

Table 4. Multinomial logistic regression results.

Base Outcome: 3 = Not Offering and Not Planning to Offer Green Products/Services		(1 = Yes)	(2 = No, but Planning)	(1)	(2)
		Coefficient	Coefficient	RRR	RRR
Present resource	Few actions (1–2)	−0.0891	−0.176	0.915	0.839
efficiency	(re_present = 1)	(0.261)	(0.331)	(0.238)	(0.278)
base: no actions	Some actions (3–4)	0.344	−0.0841	1.410	0.919
	(re_plan = 2)	(0.252)	(0.329)	(0.355)	(0.303)
	Many actions (>4)	0.557 **	−0.0130	1.746 **	0.987
	(re_plan = 3)	(0.251)	(0.330)	(0.439)	(0.326)
Planned resource	Few actions (1–2)	0.294 *	0.777 ***	1.342 *	2.176 ***
Efficiency	(re_plan = 1)	(0.174)	(0.243)	(0.233)	(0.529)
	Some or many actions (>2)	0.0528	0.577 ***	1.054	1.781 ***
	(re_plan = 2)	(0.140)	(0.217)	(0.147)	(0.386)
Support measures	Financial incentives	0.815 ***	0.733 ***	2.260 ***	2.082 ***
	(sup_fin = 1)	(0.104)	(0.138)	(0.235)	(0.288)
	Potential markets and custom.	0.850 ***	0.697 ***	2.339 ***	2.007 ***
	(sup_mark = 1)	(0.132)	(0.179)	(0.308)	(0.360)
	Technical support	0.223 *	0.594 ***	1.250 *	1.812 ***
	(sup_tech = 1)	(0.123)	(0.156)	(0.153)	(0.282)
	Marketing and distribution	0.576 ***	0.407 **	1.778 ***	1.502 **
	(sup_cons = 1)	(0.136)	(0.194)	(0.243)	(0.291)
Green jobs (%)	>5% and ≤10%	0.722 ***	0.383	2.059 ***	1.466
	(green_job = 10)	(0.226)	(0.296)	(0.466)	(0.434)
	>10% and ≤30%	1.026 ***	0.507 **	2.789 ***	1.660 **
	(green_job = 30)	(0.174)	(0.230)	(0.485)	(0.381)
	>30% and ≤100%	1.295 ***	0.329 **	3.650 ***	1.389 **
	(green_job = 100)	(0.117)	(0.164)	(0.427)	(0.228)
Sector of activity	Retail (NACE-G)	0.467 ***	0.0392	1.595 ***	1.040
	(nace = 1)	(0.114)	(0.157)	(0.182)	(0.163)
Country	Italy	0.206	−0.312	1.229	0.732
(base: ipscntry = Croatia)		(0.175)	(0.210)	(0.215)	(0.154)
	Hungary	−0.496 ***	−1.100 ***	0.609 ***	0.333 ***
		(0.174)	(0.212)	(0.106)	(0.0704)
	Austria	0.703 ***	−0.909 ***	2.020 ***	0.403 ***
		(0.172)	(0.247)	(0.348)	(0.0997)
	Slovenia	0.369 **	−0.531 ***	1.446 **	0.588 ***
		(0.158)	(0.193)	(0.228)	(0.114)
Does your company	sell only services	−0.190 *	−0.268 *	0.827 *	0.765 *
	(serv = 1)	(0.109)	(0.148)	(0.0904)	(0.113)
Selling products or	Yes	−0.165	−0.187	0.847	0.830
services to companies	(b2b = 1)	(0.117)	(0.155)	(0.0994)	(0.129)
Constant		−2.166 ***	−1.917 ***	0.115 ***	0.147 ***
		(0.293)	(0.373)	(0.0337)	(0.0548)
Observations		2530	2530	2530	2530
LR chi2 (38) = 512.02 Pseudo R2 = 0.1146

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

consist of β coefficients of multinomial logit model that are interpreted as changes in log odds ratio of being in the considered category vs. being in the base category, and of relative risk ratios (RRRs) which in STATA report coincide with odd ratios (OR).

Implementing many resource efficiency measures (compared to implementing no measures) increases the log odds ratio of offering green products and services vs. not planning to do so by 0.577 (at 5% probability). Planning only a few resource efficiency measures increases the log odds ratio by 0.294 of offering green products and services vs. no and not planning to do so (at 10% probability).

This threshold effect echoes the “critical mass” logic in which firms internalize sustainability routines only after building sufficient capabilities (Bassi and Dias [38]). In practice, bundling multiple measures (e.g., energy, water, waste, circular design) appears necessary to trigger visible green-offering behavior. Therefore, Hypothesis 1 is partly accepted for SMEs in the Alps–Adriatic macroregion.

In line with Vasilescu et al. [33], SMEs that expressed the need for any support are more likely to offer green products and services or to plan to do so than not and not plan to do so. Among the four support types, financial incentives have the largest effect (RRR = 2.26), followed by assistance in market identification (RRR = 2.34 for planners). For example, expressing the need for financial incentives or assistance with the identification of potential markets or customers to launch the range of services increases the risk of planning to offer green products and services in the next two years by more than 100% compared to the SMEs that do not express such needs. Therefore, Hypothesis 2 is accepted.

The green job ratio positively and increasingly affects the likelihood of offering green products and services. A higher green job ratio increases the green jobs ratio by more than 30%. It increases the risk of offering green products and services by 365% compared to a green job ratio of 5% or less (at the 1% significance level). A green job ratio between 10% and 30% is associated with the most significant increase in the risk of planning to offer green products and services in the next two years (66%) compared to a green job ratio of 5% or less (significant at 5% level). These results do confirm the findings of de Andrade et al. [19] (p. 10), who find that a positive % of green jobs are associated with a reduced likelihood of belonging to lower resource efficiency (sustainability practice) clusters. Therefore, Hypothesis 3 is accepted.

Reverse causality cannot be ruled out: SMEs already actively offering green products may hire more green-job staff to support expanded offerings. A longitudinal design would be required to untangle whether green jobs drive green offerings or vice versa.

The risk of offering green products and services and planning to offer green products or services in the next two years decreases by 17% and 23% if an SME sells only services compared to other firms (10% significance level). Therefore, Hypothesis 4 is accepted.

Finally, selling products or services to other companies does not affect the likelihood of offering green products in line with the results of Vasilescu et al. [33]. Therefore, Hypothesis 5 is rejected. Contrary to past findings [12], we find no adverse B2B effect. A likely explanation is the broader diffusion of EU green procurement guidelines since 2015 [3], which has extended environmental requirements into purely B2B transactions.

The likelihood of offering green products and services is also controlled by the country and its business sectors. Notably, Austria and Slovenia’s stronger propensities align with their early uptake of European Green Deal measures and complementary national support programs (European Commission [3]; Flash Eurobarometer 498 [11]), whereas Croatia’s main SME green loan scheme was only introduced after the 2021 survey wave [11]. The results show that both factors are statistically significant, which will be addressed in the Discussion section.

The overall results of our empirical research are presented in

Table 5. Results of hypotheses testing.

Hypothesis	Acceptance
H1: Engagement in resource efficiency actions increases the likelihood of offering green products and services.	Partially accepted
H2: SMEs that express a support need are more likely to offer green products and services.	Accepted
H3: Increasing the green jobs ratio increases the likelihood of offering green products and services.	Accepted
H4: Offering only services affects the green product offering.	Accepted
H5: Selling products or services to businesses (B2B) is associated with a lower likelihood of offering green products and services.	Rejected

.

5. Discussion

This study aims to increase the understanding of drivers related to the likelihood of SMEs offering green products and services in the context of the Alps–Adriatic macroregion. The research relied on the Flash Eurobarometer 498 survey on “SMEs, resource efficiency and green markets” conducted in 2021.

A categorical dependent variable on green products and services offerings with three modalities (answers: yes; no, but planning to do so and in the next two years; no and not planning to do so) is constructed and put into relation with resource efficiency actions, types of support needed, types of business output (products or services) and business relationship (B2B and/or other). The multinomial logit model was applied initially in an extended form that took into account firm size in terms of number of employees and scale of total revenues, as well as its age. However, further testing suggested a choice of the reduced model.

Hoogendoorn et al. [12] and Vasilescu et al. [33] compared results of previous surveys. They designed similar dependent variables and applied ordered and binomial logit, respectively. However, our research exploits the richness of the data collected. It assumes that there is a statistically significant difference between SMEs that plan to offer green products or services in the next two years and those that do not plan to do so. Both Wald and LR tests of combined alternatives suggest that all three should remain. Thus, unlike previous papers, this research applies a multinomial logit structure and the more recent wave of “SMEs, resource efficiency and green markets” survey. De Andrade et al. [19] also applied a multinomial logit on Flash Eurobarometer 498 to investigate the relationship between green jobs and resource efficiency actions.

Findings of Hoodgendoorn et al. [12] and Vasilescu et al. [33] that firms in the wholesale and retail sector (NACE-G) are more likely to offer green products or services are also confirmed for the sample of Alps–Adriatic countries and the more recent dataset. This is also consistent with the positive effect of tangible services in Hoodgendoorn et al. [12]. Our research also confirms the findings of previous papers about a positive relationship between resource efficiency actions and green product offerings. However, selling green products or services to businesses (B2B) does turn out insignificantly in our research, although Hoodgendoorn et al. [12] found that it was a statistically significant factor affecting the likelihood of offering green products. The research also confirms Vasilescu et al. [33] finding that support needs are positively associated with offerings of green products and services.

Additionally, previous research is extended by new insights that show that companies selling only services are less likely to offer or plan to offer green products or services (10% level of statistical significance). In contrast, green job share has a positive impact (in line with [19]). Finally, the research demonstrates that (despite various controlled factors) country-specific characteristics still significantly affect the likelihood of selling or planning to sell green products or services.

After controlling for firm characteristics, Italian small firms behave much like Croatian ones. Slovenian and Austrian small firms are already more likely than Croatian firms to sell green products today, but they are less likely to say they will add new green products in the next two years, since many have already completed that transition. Such results may indicate that green markets in Slovenia and Austria have reached a mature phase where the SMEs have already passed through the basic process of green transition.

There are several managerial and policy implications of our study:

• SME managers should coordinate multiple measures (energy, waste, water, design) to surpass the capability threshold that unlocks green product/service offerings.
• As market-identification aid yield the largest RRR (2.34) and financial incentives the second (2.26), firms should actively seek grants and consultancy programs.
• Building green job intensity fosters innovation routines and sustains green offerings.
• Policymakers should integrate grant funding with hands-on advisory services and tailor outreach to service-only SMEs, which lag behind product-oriented peers.

This study is not free of limitations:

• Causality cannot be inferred due to the study’s cross-sectional design; panels or experiments are needed to validate directional effects.
• We use a secondary data source and, thus, rely on the existing survey data from the Eurobarometer. In future studies, research might wish to develop original survey instruments.
• Detailed subsector or thematic case studies (e.g., tourism vs. manufacturing services) should be performed to provide an in-depth analysis of the “service-only” firms.
• Follow-up studies should explore mixes of national policy and cultural factors underlying the heterogeneity among the five countries.

Therefore, future research could additionally investigate the role of the sector of activity and market relations for the green offer. Finally, the hypotheses can be tested further on different datasets. Finally, since there are significant differences among countries, the sources of those variations may be explored.

6. Conclusions

SMEs in the Alps–Adriatic region should improve their orientation toward sustainability, which can be achieved by shifting focus from single “green” projects to integrated approaches, ensuring quick adaptation to the regulatory challenges and market feedback. Policymakers must move ahead of these developments by designing adaptive support ecosystems, which put in place bundles of support mechanisms, including grant programs, technical assistance, and regional innovation hubs.

Over the next decade, we expect the emergence of service-sector eco-labels and digital certification platforms, which could empower the service-oriented SMEs in the region to become leaders of the circular economy. At the same time, the growing importance of green skills will drive demand for training and micro-credentials, enabling workers to rapidly enter into new sustainability jobs.

There are several implications of our research results. Firstly, SME owners and managers in the Alps–Adriatic region should prioritize formalization of Human Resources Management activities and routines, which prove as relevant, even for the smallest (micro) companies [41]. Green innovation and their diffusion, using supply chains, should be supported by fostering Corporate Social Responsibility (CSR) mindsets, especially at the top management level(s) [42]. Secondly, policymakers should develop the bundled support instruments for SMEs, consisting of multiple incentives, both financial and nonfinancial, including support to HRM and leadership programs, CSR development initiatives, etc.

Concerning further research, longitudinal and experimental studies should explore how firms transition from initial resource-efficiency measures to circular economy and how multi-stakeholder networks support this transformation.