1 Introduction

Public procurement is a crucial factor in modern economies: every year, public authorities in the EU spend around €2 trillion (14% of GDP) on the purchase of services, works, and supplies.Footnote 1 Public procurement requires a competitive tendering process, in which the winner of the contract has to deliver the tendered service. It is no subsidy or other direct financial support from governments responding to market failures (Catozzella and Vivarelli, 2016; Becker, 2015) but an alternative market-based instrument with which policymakers can support specific policy goals, such as fostering innovation activities. For example, the German 2021 government coalition agreement mentions procurement 35 times, underscoring specifically the government’s intention to use public procurement as an instrument to accelerate environmentally beneficial innovation in enterprises and the transformation towards an ecological (green) economy.

It is well known that a firm with innovative projects has project-related additional funding needs (Hottenrott and Peters, 2012 and Schäfer et al., 2017 for family firms). Such activities require initial expenditure, specifically in R &D. In addition, during the process of transforming innovative capabilities into output, further funding needs may arise. The relationship between innovation and financial constraints is an intensively discussed research topic (e.g., Li et al., 2019; Aghion et al., 2012; Acharya and Xu, 2017; García-Quevedo et al., 2018; Elshaarawy and Ezzat, 2022; Bonanno et al., 2022). Due to the specific nature of innovation, closing these funding gaps is often difficult, in particular for small and medium-sized enterprises (e.g., Lee et al., 2015; Schäfer et al., 2017).

Using public procurement to overcome innovation constraints is a key policy goal.Footnote 2 Primarily, research explores whether winning a public procurement contract fosters innovation (e.g., Czarnitzki et al., 2020; Krieger and Zipperer, 2022). However, little is known about how success in a competitive public tendering process is related to innovative firms’ financial constraints. Does a public procurement contract ease or increase the financial constraints of innovative firms? We focus on this research gap. Thereby, we pay special attention to firms, in particular SMEs, with environmentally beneficial innovative activities (“green” innovators), and we link the strand of literature on financial constraints and innovation with the strand on procurement and “green” innovation.

We use the Community Innovation Survey (CIS), which is conducted bi-annually in EU member states, EFTA countries, and EU candidate countries. The vast majority of observed firms in this survey are SMEs. The CIS provides micro firm data, including a wide range of indicators on environmental innovation activities, public funding, turnover from innovative products, and financial and other barriers to innovation. The 2014 wave also includes a range of indicators about the enterprises’ public procurement contracts.

The possible endogeneity of procurement success with respect to a firm’s innovative output is one of the challenges we face: are firms innovative because they have been awarded a public procurement contract, or have they been awarded a public procurement contract because they were innovative (Caravella and Crespi , 2021)? While previous research focuses exclusively on the effect of public procurement as a driver of innovation, we argue that public authorities have a clear incentive to select firms that are already more innovative than others in order to maximize the probability of success of their innovation policies. Moreover, innovative firms may be more likely to submit a bid for an innovative PP competition than their non-innovative peers (Guerzoni and Raiteri , 2015). To meet this challenge, we apply extended probit (eprobit) models considering endogenous selection into innovation and public procurement as endogenous treatment when assessing the importance of PP for an innovative firm’s risk of facing financial constraints.

Economic theory suggests that the certification effect of success in public procurement, particularly when mediated by the demand-pull effect, may lower a company’s funding constraints (Ghisetti , 2017; Li et al. , 2019; Dai et al. , 2021). Yet, we find contrasting results. Being successful in a PP significantly increases a firm’s probability of facing financial constraints. We attribute this result to a high pre-financing requirement as a result of winning the public tender. In particular, an increased need for working capital comes usually along with PP success, while the revenues from the project only flow later, after the goods or services have been delivered. The need of upfront funding appears to counteract the potentially constraint-reducing effects from a PP’s demand-pull or certification effect. The second main result of our study is that a green innovator firm with a PP contract is more likely than its peers without such a contract to express the need of public co-financing for innovation. This result also applies to firms with environmentally beneficial innovative activities related to renewable energies. We look at renewable energies separately because of their central importance in the transition to an economy low in CO2 emissions. Overall, the evidence implies that public subsidies or grants are important drivers of environmentally beneficial innovation.

Our research contributes to the vast literature on the importance of financial barriers for innovation, particularly for SMEs. We add to the rather new but increasingly more important strand of studies that focuses on the mediating role of public procurement in easing financial barriers to green innovation. Our results suggest that policies promoting green and renewable-resource-related innovation through PP can be successful. However, it is essential to recognize that winning a PP contract comes with additional requirements for funding and, consequently, a higher likelihood for SMEs to face financial constraints. Our evidence clearly supports regulatory reforms aiming at easing financial barriers and combining procurement with public grants to promote environmentally beneficial innovations (Hoekman and Taş , 2022; Caravella and Crespi , 2021).

The remainder of the paper is organized as follows. In Sect. 2, we briefly review the existing literature on public procurement and its impact on innovation, as well as derive demand-pull and certification hypotheses of relaxing financial barriers. Section 3 presents the data. We describe the empirical methodology and identification challenges as well as the foundations of the applied extended probit regression (eprobit) model in Sect. 4. In Sect. 5, we report and discuss the estimation results. Section 6 concludes.

2 Literature review and hypotheses

Financial constraints are particularly likely for innovative firms as information asymmetries are large and innovative success is highly uncertain. For small or medium-sized innovative firms, their financial opacity adds to the constraint problem. Another potentially exacerbating factor is environmental innovation. Furthermore, innovation is typically brought forward by highly qualified employees who are non-tangible assets that cannot be pledged as collateral (Zúãniga-Vicente et al. , 2014; Schäfer et al. , 2017). Although using public procurement to stimulate innovation is a key policy instrument, we know little about whether public procurement would reduce or exacerbate the financial constraints of an innovative company. This is even less known for innovators in the ecological (“green”) field.

This blind spot may have persisted so long because the research focuses almost entirely on the causality running from winning a public procurement contract to innovation. For instance, based on a large survey of German firms, Aschhoff and Sofka (2009) find that both public procurement and knowledge spillover from universities have similarly positive impacts on innovation success. Moreover, PP is found to be particularly useful for smaller firms in economically strained regions and in the field of distributive or technological services. Using the German part of the EU’s Community Innovation Survey in the period from 2010 to 2012, Czarnitzki et al. (2020) discover, across a wide set of specifications, a robust and significantly positive effect of innovation-directed public procurement (IPP) on firm turnover with innovative products and services. However, the effect seems to be restricted to incremental innovations rather than true market novelties. Guerzoni and Raiteri (2015) apply a matching approach to a data set of 5200 European firms. They observe that firms with IPP contracts are more likely to increase their private spending on innovation activities and point out possible reinforcement interactions between demand-side (PP) and supply-side (tax credits, subsidies) policy instruments.

Caravella and Crespi (2021) analyze the effects of regular public procurement (RPP) and IPP on private R &D investments for a sample of Italian firms. They conclude that IPP stimulates R &D investments only when combined with supply-push measures, such as soft loans, tax deductions, and grants. Shin and Lee (2022) observe that South Korean firms with IPP contracts experience a higher rate of change in total factor productivity as well as a higher growth rate of value-added than firms with RPP. Ghisetti (2017) applies a matching approach to a data set of 3000 European manufacturing firms and find that companies with a PP contract are more likely to innovate. The study proposes that PP, because it is an instrument to reduce the risks associated with innovation, particularly regarding uncertain demand, could be well suited to foster environmental innovation.

Despite the above-cited research, a causal relationship running from PP to innovation is not convincing if public authorities have an incentive to select firms as winners of the tender that are already more innovative than others. This incentive clearly exists because such awarding policy would increase a public authority’s chance of achieving its own innovation policy targets. Effectiveness in spurring innovation is more likely if tendering firms are able to demonstrate capability to innovate. In addition, firms with some experience in innovation may be more likely than their non-innovative competitors to submit a bid to a tender for innovation (Guerzoni and Raiteri , 2015), and those bids may be of a superior quality. Thus, the decision to participate in a public procurement tournament may occur simultaneously with the outcome (innovation) leaving open in which direction the causality runs.Footnote 3

The link between success in a public tender process and funding constraints is also poorly explored. Will an innovative firm with success in a procurement tournament face higher or lower financial barriers to innovative activities than its peers without a contract award? Will this relationship be different for firms with environmentally beneficial activities? While limited evidence is available on public procurement in general, the impact of success in a public procurement tournament on financial barriers of firms with environmentally beneficial innovative activities is a complete blind spot. Dai et al. (2021) examine the quantitative importance of the demand-pull effect of PP by applying a causal mediation analysis to a sample of high-tech firms in China. They find that the treatment effect of public procurement, particularly when mediated by a demand-pull effect, improves the firms’ access to external financing. They infer from their results that public agencies, if able to identify promising innovative projects or firms, could use public procurement to ease market failures in innovation financing.

Dai et al. (2021) argue that winning a PP contract serves as a certificate or signal of the quality of recipient firms. This certification effect then stimulates innovation by attracting external investors and relaxing financial constraints. The effect is even more pronounced for financially constrained, small, and young firms. Baum et al. (2021) focus on European SMEs to explore the relationship between key financial ratios and success in public procurements. Specifically, they apply a difference-in-difference analysis to a sample in which firms from the TED databaseFootnote 4 are matched with comparable SMEs from the AMADEUS database. They find that a lower equity ratio and a higher short-term debt ratio increase the companies’ chance for success. Moreover, the PP-award firms continue to operate with a lower equity ratio than comparable firms without an award. These findings support the certification hypothesis. Success in a PP might certify a high potential to become financially strong, thus serving as a substitute for high equity levels.

A procurement contract award means stable demand and sales of goods, works, or services. In this sense, public procurement is a derisking tool to innovation. There is ample literature on a derisking effect of government subsidies when capital market imperfections make it difficult to raise finance for uncertain innovation projects due to information asymmetries. Lerner (1999) examines the long-run effects on firms’ financing of the United States Department of Energy’s Small Business Innovation Research (SBIR) program. The study shows that SBIR-award winners grow faster in terms of sales and employment than similar non-supported peers and are more likely to later receive venture capital funding. However, the effect is confined to areas with substantial venture capital activities and more pronounced in high-tech industries. Larger subsidies are found to not have a larger impact, supporting a certification effect of subsidies but also hinting at decreasing returns of an additional unit of subsidy.

Feldman and Kelley (2006) also argue that government subsidies may serve as a signal for project or company quality. They observe that award winners of the US Advanced Technology Program subsequently attract larger amounts of R &D funding from non-government actors like venture capitalists. Kleer (2010) shows in a simple signaling model that a government grant may signal a good investment opportunity to private investors and, therefore, ease funding constraints.

Using a data set of 1600 Belgian small and medium-sized firms, Meuleman and Maeseneire (2012) find further evidence that receiving R &D subsidies is a positive signal about SME qualityFootnote 5 and results in better access to long-term debt. They identify no effect on short-term debt and external equity finance. The certification effect of R &D grants is stronger in case of higher asymmetric information, while the relative size of the grant does not affect the likelihood of attracting external investors. Howell (2017) finds that early-stage grants of the SBIR program ease financing constraints, but expresses some doubt about the certification mechanism. Instead, it is emphasized that grants enable companies to manufacture prototypes, which they would otherwise not be able to finance.

Based on a sample of Chinese listed corporations, Wu et al. (2017) confirms evidence that receiving R &D subsidies increases the enterprises’ likelihood of raising external finance. This R &D subsidies’ certification effect is stronger in private enterprises than in state-owned ones. Chen et al. (2018) examine the impact of R &D and non-R &D government subsidies on initial public offering (IPO) performance using a sample of 269 Chinese enterprises. They argue that while R &D subsidies initially convey a positive signal, above a certain threshold, investors become increasingly concerned about the risks and uncertainties associated with R &D activities. Non-R &D subsidies, on the other hand, would serve as a positive signal of government confidence in companies’ capabilities. In contrast, Li et al. (2019) find that Chinese companies receiving public R &D subsidies enjoy a positive certification effect, which eases financial constraints of firms that borrow from banks. This effect is stronger for unlisted firms and in regions with weaker intellectual property rights protection. Overall, the majority of studies supports the hypothesis that derisking via subsidies eases funding constraints for innovative firms.

Public procurement and government subsidies are both derisking tools. We expect the signaling and certification effect of government subsidies and grants to apply analogously to public procurement. Yet, some research proposes contrasting effects of government subsidies. Instead of serving as a positive signal for potential external capital providers, receiving public funding may decrease the efficiency of innovation investments (Caravella and Crespi , 2021; Liu et al. , 2021). Subsidized firms may tend to over-invest, or subsidies may simply fund projects that had been conducted anyway. Even worse, public funding may make rent-seeking easier and actually increase the required funding per unit of innovative output (Catozzella and Vivarelli , 2016). Under such circumstances, financial constraints could even be beneficial to innovation since they force firms to be more innovative and efficient. On the other hand, Zhang and Guan (2018) report that financial slack positively moderates the relationships between government incentives and firms’ innovation performance in Chinese high-tech companies.

As mentioned earlier, procurement contracts are different from subsidies with respect to the in-flow of funds (Di Mauro et al. , 2020). In particular, success in a competitive tender means usually no immediate transfer of public funds to the firm. Payment occurs after the contracted deliveries have arrived with the procurer. On the contrary, award winners may need additional funds to pay wages and to purchase raw material and other inputs for the production of the deliveries. Based on these two concepts, we arrive at the central:

Research Question 1

Does success in public procurement mitigate or exacerbate innovative enterprises’ financial constraints?

Promoting environmental innovation using public procurement is a special case of Innovative Public Procurement (IPP). Zipperer (2019) analyzes a panel of 5400 German firms. She finds that public procurement contracts have a demand-pull effect on general innovation, but finds no conclusive evidence regarding environmental innovation. In addition, the study reveals heterogeneous effects for different sectors: while companies in the water supply and waste management industry are more likely to introduce product innovation, firms in the electricity and gas sector are more likely to introduce process innovations after winning a PP. Krieger and Zipperer (2022) apply a cross-sectional difference-in-differences approach. They reveal that winning a PP contract triggers a demand-pull effect for SMEs, which increases their probability to introduce environmental product innovations by 25 percentage points. In contrast, the authors find no significant effect for larger companies or for the introduction of environmentally friendly process innovations. Clearly, the specific immaturity of many markets for environmentally beneficial innovative products and technologies increases the innovation risk further. This has implications for financial constraints as Jensen et al. (2019) show. Studying data from the Mannheim Innovation Panel (MIP), the authors reveal that firms with environmentally beneficial innovative activities are more likely to have higher latent financial needs than other innovators. Early-stage markets combined with the increasing demand of societies and politicians to avoid or heal environmental damage also promote increased interest in risk-mitigating tools such as public procurement. Therefore, the relationship between a green innovator firm’s public procurement contract and its financial constraints is of particular interest and leads us to:

Research Question 2

Does success in public procurement mitigate or exacerbate financial constraints of enterprises with environmentally beneficial innovative activities?

3 Data source and sample description

The analysis is based on the European Community Innovation Survey (CIS), which aims to provide information about innovation activities in enterprises. The survey is conducted every 2 years in the European Union, EFTA countries, and EU candidate countries by the statistical office of the European Union, Eurostat. The CIS is implemented using a standardized core questionnaire to ensure cross-country comparability and is designed to provide information on a variety of innovation activities (product, process, organizational, and marketing innovation) at the firm level, broken down by country and economic sector.

The CIS 2014 wave comprises key information on public procurement contracts, barriers to innovative activities, and drivers of environmentally beneficial innovations in addition to a wide range of innovation variables and other firm-level indicators (size category, export sales, sales, employment, structure of workforce). The cross-sectional data set covers the activities of firms during the 3-year reference period 2012–2014.

To investigate the core question, whether innovative firms that own a public procurement contract are financially less constrained than firms without such a contract, we first need to identify the owners of a public procurement contract. For this, we exploit the question: During the three years 2012 to 2014, did your enterprise have any contracts to provide goods or services for Domestic public sector organisations [or/and] Foreign public sector organisations. The public-procurement dummy variable equals one if a firm responded yes to this question, and zero otherwise. Note that PP as well as the other independent variables (SME, share of export sales, turnover and employment growth, share of employees with tertiary education degree, main market, industry, and country) are observed for the entire sample of non-innovative and innovative firms. Table 7 in the Appendix presents the key variables and the related survey questions.

The dependent variable is the financial constraint (FC) indicator. In contrast to the PP variable, the financial constraints variables are only observed for a sub-group of firms. We distinguish between financial constraints in general (FC), internal financial constraints only (IFC), and external financial constraints only (EFC). FC, IFC, and EFC are coded as barrier to innovation and, therefore, only observed for potentially innovative but deterred firms. Following García-Quevedo et al. (2018), we define the financial constraint dummy variables in the following way. It takes the value of one if a firm valued Lack of internal finance for innovation and/or Lack of credit or private equity as highly important barriers, and zero otherwise. We infer from assessing the lack of funding as highly important barrier to innovation that the companies have innovation projects but cannot pursue them because of the financial gap (Hottenrott and Peters , 2012; Schäfer et al. , 2017). In contrast, companies that report being innovative have overcome the financial barriers to innovation; otherwise, they would not be innovators.

Fig. 1
figure 1

Distribution of environmental innovation across industries

Full size image
Table 1 Share of small and medium-sized enterprises (%) among innovator firms
Full size table

We then take advantage of the question only asked to firms with environmentally beneficial (green) innovative activities: During 2012 to 2014, how important were the following factors in driving your enterprise’s decisions to introduce innovations with environmental benefits? To analyse the impact of a public procurement contract on a green firm, we define such a firm as financially constrained if this firm expressed the need for public co-financing, or in other words, if the firm assessed Government grants, subsidies or other financial incentives for environmental innovations as either medium or highly important drivers of those activities. In this case, the green financial constraints dummy variable (GFC) takes the value of one, and zero otherwise.

The GFC dummy variable provides an indirect measure of a financial gap and may, therefore, be affected by some limitations. However, medium or high importance of financial support from public sources signals that the firm would not have undertaken the environmental innovation if this support had not been accessible. In other words, the derisking of innovative projects via government grants, subsidies, or other financial incentives ensured that funding restrictions did not stand in the way of realizing the innovative project with environmental benefits.

Table 2 Innovator firms with procurement contracts
Full size table
Table 3 Firms with environmentally beneficial innovative activities facing financial constraints
Full size table

Figure 1 illustrates how environmentally beneficial innovative activities are distributed across industries. Environmental innovation is found in every industry; however, the number of active companies differs widely. Environmental innovator firms (EIF) are especially frequent in electrical equipment/machinery, plastic/glass/ceramic, and metals, but are rare in hotels/ restaurants, mining/quarrying, and construction. In addition, we observe a higher number of other innovator firms (OIF) relative to EIF in less energy-intensive industries like finance/insurance/real estate, professional, scientific & technical activities, and wholesale/retail trade.

Apart from a procurement contract, our main interest is on the impact of a firm’s size on financial constraints. Table 1 illustrates the paramount importance of SMEs in the overall sample as well as among firms with activities in product or process innovation, be they environmentally beneficial (EIF) or not (OIF). A total of 89.7% of the firms in the sample are SMEs. The share of SMEs among OIF is about 87% compared to 78% among EIF. The SME-share in the category of firms with environmentally beneficial innovative activities related to renewables (EIFR) is 74 % compared to about 86% in the group of environmental innovator firms without relation to renewables (EIFO).

Table 2 shows the percentage of innovative enterprises with a PP contract from 2012 to 2014 broken down by firm and innovator type. Among the SMEs, a lower percentage possesses a procurement contract than among the non-SMEs. The share of enterprises with a procurement contract is not significantly different among EIF and OIF. Among EIFRs, the percentage of companies with a procurement contract is higher (29%) than among EIFO (27%). The difference is statistically significant at the 10% level. To put these numbers into perspective, about 18% of all enterprises in the sample reported having a procurement contract, so the proportion is higher among firms with innovative activities than among non-innovators.

Finally, Table 3 focuses on the share of firms with environmentally beneficial innovative activities facing financial constraints. The share of EIFR facing financial constraints is higher than the share of EIFO. The break down by size reveals that the difference can be mainly attributed to a higher share of SMEs facing financial constraints than non-SMEs.

Table 4 presents the summary statistics of the main variables used in the multivariate empirical analysis. We control for firm characteristics, such as export sales, turnover and workforce growth, share of employees with tertiary education degree in the workforce, the main market, and macro indicators like industry and home country. We pay special attention to small and medium-sized enterprises (SME).

Table 4 Summary statistics for entire CIS 2014 sample
Full size table

The number of observations per variable differs substantially. About 43% of the 96,703 respondents have introduced product, process, organizational, or marketing innovations, while about 29% have introduced product or process innovations (referred to as innovator). A total of 28,243 firms, about one-third, are EIF, while two-thirds are OIF. Of the EIF, 70% are EIFR and 30% EIFO. Almost 6% of respondents declare that the lack of internal and/or external finance is a highly important barrier to innovation. A total of 25% out of the firms with environmentally beneficial innovative activities are financially constrained. Almost 18% of 61,262 respondents possess a PP contract. The overwhelming majority of firms are SMEs.

4 Econometric approach and identification strategy

The centerpiece of our research is the question of whether possessing a PP contract affects the financial constraints of innovative companies. Possessing a PP contract is considered to be a treatment to the firm, and the aim of the empirical analysis is to identify the impact of this treatment on firms’ financial constraints.Footnote 6 If PP contracts would be randomly assigned to firms, one could apply common probit models including binary treatment to identify the effect of winning a contract on the likelihood of being financially constraint. To put it differently, if winning a PP contract is an exogenous “treatment,” both constrained and unconstrained firms were equally likely to be successful in the tender. In our case, however, the conditional independence assumption is unlikely to hold, because the treatment status is likely to be related to potential outcomes. Unobserved factors may affect the public procurement decision because they affect the likelihood of being subject to financial constraints. For example, the unobserved propensity (or resistance) of a house bank to accept a successfully acquired procurement contract as a sign for a high creditworthiness of the company would affect the likelihood of having a procurement contract but also the likelihood of financial constraints for a firm.

This research design also entails a further complication regarding the observation of financial constraints for firms in the CIS. FC are only observed for firms that are prevented from innovation because of high innovation barriers. Similarly, the green financial constraints (GFC) indicator is only observed for firms with environmentally beneficial innovative activities but not for firms with other innovative activities. However, whether a firm is an innovator or not, or whether the innovation is environmentally beneficial or not, is not a randomly assigned outcome across firms, but is likely to be affected by unobserved characteristics of the firm. This creates the so-called endogenous selection problem, which is also known as “selection on unobservables.”Footnote 7 To tackle these identification challenges, we formulate a structural parametric model that includes both endogenous treatment and endogenous selection aspects by specifying additional equations. This recursive modelFootnote 8 identifies the impact of PP on financial constraints by employing instrumental variables.Footnote 9

We start the model formulation with describing the endogenous binary treatment effect assignment. Let \({PP}_i\) be a binary indicator, which describes whether firm i wins a tender or not. We use a probit model to describe the likelihood of \({PP}_i=1\) conditioned on control variables \(x_i\) and instrument variables \(z_i\). The equation becomes then

$$\begin{aligned} \textit{PP}_i=1 \qquad \left( \beta _1 x_i+\gamma _1 z_i + \varepsilon _{1i}>0\right) , \end{aligned}$$
(1)

where the assumption is that the error term \(\varepsilon _1\) is normally distributed.

According to our research questions, we want to estimate the impact of the endogenous treatment status \({PP}_i=1\) on the main variable of interest, indicating whether firm i is financially constraint, denoted as \({FC}_i=1\). This equation can be written as

$$\begin{aligned} \textit{FC}_i=1 \qquad \left( \beta _2 x_i+ \theta _2 \textit{PP}_i + \varepsilon _{2i}>0\right) , \end{aligned}$$
(2)

where \(\theta _2\) indicates the treatment effect of PP on FC and \(\varepsilon _2\) is normally distributed.Footnote 10

The following assumptions are required to identify the endogenous treatment effect \(\theta _2\):

  1. 1.

    The instrument variables z only appear in the treatment assignment Eq. (1) and not in the outcome Eq. (2), which is the so-called exclusion restriction.

  2. 2.

    Treatment assignment is related to instrumental variables, that is, the instrumental variable(s) in Eq. (1) are relevant.

  3. 3.

    \(cov(z,\varepsilon _2)=cov(z,\varepsilon _1)=cov(x,\varepsilon _1)=0\), which means that instruments and control variables are exogenous.

Equations (1) and (2) form a canonical system of equations. Assuming that \(\varepsilon _1\) and \(\varepsilon _2\) are jointly multivariate normally distributed, we have

$$\begin{aligned} \left( \begin{matrix}1&{}\rho _{\varepsilon _1 \varepsilon _2}\\ \rho _{\varepsilon _1 \varepsilon _2}&{}1\\ \end{matrix}\right) . \end{aligned}$$

If \(\rho _{\varepsilon _1 \varepsilon _2}=0\), the treatment is considered to be exogenous.

As an additional extension to the endogenous treatment model formulated in the previous equations, we also consider the endogenous selection \(s_i\) mentioned above, which can be handled within the endogenous treatment model as well. Companies may have innovation projects they cannot pursue due to lack of finance (Hottenrott and Peters , 2012; Schäfer et al. , 2017; García-Quevedo et al. , 2018). On the other hand, companies that have innovations are assumed to have overcome the barriers to innovation, including financial constraints. Thus, financial constraints are assumed to be a barrier to innovation, but this is only observed among potentially innovative but deterred (those which label the lack of funding highly important) and other non-innovative companies.

We introduce formal notation regarding selection. Denote firm i is selected if binary indicator \(s_i=1\). Since we can only consider firms that are not innovators for the main outcome equation, in our case, \(s_i=1\) indicates that a firm is a non-innovator firm. Specifying \(w_i\) as exogenous covariates which determine selection, the likelihood of selection can be modeled using a probit model

$$\begin{aligned} s_i=1 \left( \beta _{3i} w_i+\varepsilon _{3i}>0\right) . \end{aligned}$$
(3)

Combining the three probit models of Eqs. (1)–(3), one gets an extended canonical model with variance-covariance matrix \(\sum \) of the three error terms denoted as

$$\begin{aligned} \left( \begin{matrix}1&{}\rho _{\varepsilon _1 \varepsilon _2}&{}\rho _{\varepsilon _1 \varepsilon _3}\\ \rho _{\varepsilon _1 \varepsilon _2}&{}1&{}\rho _{\varepsilon _2 \varepsilon _3}\\ \rho _{\varepsilon _1 \varepsilon _3}&{}\rho _{\varepsilon _2 \varepsilon _3}&{}1\\ \end{matrix}\right) . \end{aligned}$$
(4)

If \(\rho _{\varepsilon _1 \varepsilon _3}=0\), selection is exogenous.Footnote 11

5 Estimation results

We have now arrived at the identification of the impact of a public procurement contract on an innovative firm’s financial constraints. Table 5 presents the estimation results using the methodology described in Sect. 4. The columns show the marginal effects. For reasons of comparison, we report in Columns (1), (3), and (5) the coefficients obtained from the simple probit model. Column (2) with dependent variable FC in general (either internal or external) and Columns (4) and (6) with dependent variable internal financial constraints (IFC) and external financial constraints (EFC), respectively, present the core specifications with endogenous treatment and endogenous selection.

Table 5 The impact of a public procurement contract on a firm’s financial constraints
Full size table

The specification in Columns (2), (4), and (6) combine the main equation on the likelihood of financial constraint, the treatment effects model of public procurement, and the endogenous selection into innovation. The tests for error correlation between main equation and treatment equation, \(\rho _{12}\), and for the error correlation between main equation and selection equation, \(\rho _{13}\), are both significant, indicating that we can reject the null hypotheses of no endogenous selection and no endogenous treatment.Footnote 12

Table 6 The impact of a public procurement contract on a green innovator’s financial constraints
Full size table

The main findings regarding Research Question 1 are clear. Possessing a PP contract significantly increases the probability of facing general financial constraints and external financial constraints. The extended probit regression models confirm the positive coefficients of the simple probit model and reveal that the true marginal effects of PP would be underestimated in the simple probit model. The estimated coefficients for the average treatment effect of the treated (ATET) are obtained by using the sub-sample of potentially innovative but deterred firms with no procurement contract and by applying it to the potentially innovative firms with a PP contract, in order to achieve an estimate of the potential financial constraints which the PP-contract firms would have faced had they not gained the PP contract. Average treatment effect (ATE) and ATET are highly significant and positive in all three specifications.

In interpreting the results, we focus only on Columns (2) and (6) because the treatment effect is non-significant for internal financial constraints IFC (Column (4)). The ATET is 0.286 and 0.187 in Columns (2) and (6), respectively. For those who possess a PP contract, the probability of facing financial constraints is almost 29 or 19% higher than if the firm did not have a PP contract. Since we corrected for endogenous selection into innovation, this result applies to both potential but deterred innovators as well as to innovator firms.

The results show that PP contracts exacerbate funding gaps for innovative firms. A likely explanation for this finding is that processing a PP contract triggers additional needs to fund the purchase of preliminary products and working capital for the production of the contract deliveries. Either existing funding gaps are deepened or new financing gaps are opened up. Both lead to additional loan or equity demand and even greater challenges in realizing potential innovation projects.

At first sight, this result seems to contradict Baum et al. (2021) who find that a lower equity ratio and a higher short-term debt ratio is no bottleneck for a company’s chance to win a public procurement contract and to conduct the project. However, it is important to note that Baum et al. (2021) consider SMEs in general while we evaluate innovative firms. Financial institutions are particularly restrictive if innovative firms apply for funding. Moreover, the authors focus on information that outsiders, such as banks, have access to for assessing a company’s financial strength, while we use the CIS data which provides an individual assessment of a company’s financial means required for pursuing an innovative project. The comparison with Baum et al. (2021) suggests that—given that a procurement contract exists—a potentially innovative company’s perception of the importance of financial gaps as barrier to innovation might be different from the ex-post assessment by banks using information from balance sheet data. The coefficient of the SME variable in Table 5 shows the sign that is expected from the literature. Smaller enterprises are more likely to face financial barriers to innovation than large firms (19 and 13 percentage points, respectively). A high turnover share from exports increases the probability of facing general financial constraints. A share of employees with tertiary education degree above 50% significantly lowers the probability of facing financial constraints. The European Union as the main market eases financial constraints as well.

Table 6 provide results which address Research Question 2. Due to data constraints, we use a distinct financial constraint indicator for firms with environmentally beneficial innovative activities (green activities). A green firm is considered as financially constrained (GFC = 1) if it expresses that government grants, subsidies or other financial incentives for environmental innovations were important drivers for conducting such innovative activities.

The table reports average marginal effects. For completeness, we show the coefficients of the simple probit model in Column (1). The coefficients in Columns (2) and (4) are computed from an extended probit regression with endogenous treatment only. In addition, Columns (3) and (5) address the possibly endogenous selection into environmentally beneficial innovation and environmentally beneficial innovative activities related to renewables, respectively. We are looking at renewable energies separately because they are of central importance in the transition to an economy with low CO2 emissions. The non-significance of the correlation between the errors of main and selection equation indicates that there is no need to include the selection equation in the extended probit model.

Again, the main findings are clear. The average marginal effect of having a PP contract significantly increases the likelihood that a firm with environmentally beneficial innovative activity faces financial constraints. ATE and ATET are highly significant and positive in all eprobit regressions. In interpreting the results, we focus on Columns (2) and (4) because only the treatment is endogenous. The ATET in Column (2) is 0.3, which implies that, for green innovators with a PP contract, the average probability of facing financial constraints is 30% higher than if the firm did not have a PP contract.

The ATET in Column (4) is 0.32, which means that for a firm with renewable-resources-related environmentally beneficial innovative activities, the average probability of being exposed to financial constraints is 32% higher than if such a firm did not have a PP contract. The coefficients of the SME variable again show the expected result. Smaller green innovators are more likely to face financial constraints to innovation than larger firms. This is even more pronounced if a firm’s environmental innovation activity is related to renewables (around 6 vs. 7 percentage points).

6 Conclusions

This study aims to clarify whether public procurement mitigates or exacerbates innovative firms’ financial constraints. Thereby, we distinguish between two main types of innovative firms, those with general and those with environmentally beneficial innovative activities. Theory suggests that possessing a public procurement contract may lower an enterprise’s funding constraints for innovation. However, increased up-front funding required to finance the production of contract deliveries may deepen already existing funding gaps or create new ones.

We test these opposing conjectures separately for innovating firms in general and for firms with environmentally beneficial innovative activities in particular. We use firm-level data from the European Community Innovation Survey and apply extended probit models with endogenous treatment and selection to control for omitted variable bias. Our analysis reveals a significantly positive treatment effect of public procurement on the probability of facing financial constraints. In both areas, general and environmentally beneficial innovative activities, enterprises with public procurement contracts are more likely to face financial constraints than innovative companies without such a contract. We explain this evidence with a higher impact of increased pre-funding requirements, which often arise with a public procurement contract, in comparison to the potentially constraint-reducing demand-pull or certification effect of PP.

Our findings also confirm that being an innovating SME worsens financial constraints in both cases, if the SME is a general or an environmental innovator. This result in combination with the exacerbating impact of a public procurement contract on financial constraints may explain why SMEs’ participation in public tenders is still relatively low. Already financially constrained SMEs may fear additional funding gaps from successfully bidding in a public procurement process. The results highlight again the importance of removing financial barriers for SME innovators. However, it objects the expectation that strengthening SMEs’ participation in European public tenders would contribute to that. On the contrary, complementary grants or other financial incentives might be necessary to substantially increase the SMEs’ bidding rates in public tenders. Against the background of our results and the key role that public procurement plays in the implementation of ambitious climate protection goals and in the transition to a green economy, the pre-financing of innovating SMEs that are successful in the bidding process by the public sector is of particular importance.

Our research adds a missing piece to the vast literature on financial barriers to innovation. In particular, we contribute to the rather new, but increasingly more important, strand that focuses on the mediating role of public procurement in alleviating financial barriers to green innovation. With respect to policy implications, it is essential to recognize that companies winning a PP contract may face additional financing requirements that may exacerbate financial constraints. Even though, on the one hand, financial constraints could be beneficial to innovation when they force firms to be more innovative and efficient, on the other hand, they may prevent SMEs—which are already suffering disproportionately from financing restrictions—from taking part in the tendering process in the first place. Under these circumstances, financial constraints would work against reforms that explicitly aim at increasing the bidding rates of SMEs in public procurement. The good news is that public derisking of innovative projects via government grants, subsidies, or other financial incentives can be a complement to a public procurement contract. These instruments could help companies to meet the increased funding requirements arising from a public procurement contract.

Up-to-date and more detailed data on the involvement of innovative firms in public procurement tenders and their funding are necessary to shed more light on the relationship between the success of innovative European firms—in particular SMEs—in public tenders and their subsequent funding needs. This research is urgent in both areas, general and environmentally beneficial innovative activities. After all, public tenders are about to become a key policy tool in fighting both climate change and economic uncertainty.