Can We Have Both? The Challenges of Efficiency Outcomes in a Public Procurement System and Reaching the Strategic Goals of Procurement

Abstract

This article explores the cost efficiency of public procurement procedures, techniques and strategic policies. Member States in the EU spend large sums of money on the procurement of various items, yet public procurement in the EU struggles with issues of competitiveness, efficiency and overlapping goals. The competitiveness and efficiency of various factors, procedures, techniques and sustainable measures employed in public procurement procedures that influence cost efficiency were examined in this study using data from Slovenian public procurement over the past eight years. The analysis employed a binary logit model that facilitates the efficient utilization of public funds and predicts substantial price reductions in public procurement procedures. The findings underscore the significance of transparency and open markets. The optimal cost-efficiency outcomes are achieved through combinations of transparent procedures and an undisclosed market environment during the tendering phase, complemented by less binding secondary procurement objectives. Conversely, procurement techniques that tend to disclose more information about competitors in the procurement process are found to be less cost-effective. Non-mandatory sustainable policies in public procurement tend to lead to better results in terms of cost efficiency than those imposed by an obligatory legal framework.

1. Introduction

Public procurement is an important part of public expenditure, accounting for between 12 and 15% of GDP [1]. In Slovenia, the share of total public procurement in 2022 amounted to 14.11 percent of GDP [2]. With over EUR 8 billion of public procurement contracts awarded in that year, public procurement is an important part of the public finance system.

The efficiency of public procurement systems has recently become the focus of European legislators after a report by the European Court of Auditors. The European Court of Auditors (hereinafter ECA) [3] carried out an audit of the European public procurement system at the end of 2023 and produced a special report entitled Public Procurement in the EU, which found that the level of competition in public procurement in the EU single market has declined over the last ten years. The ECA made recommendations in its report to improve the public procurement system, and the first (to the European Commission as caretakers of the procurement system of the European Union) is to clarify and prioritize public procurement objectives. The challenges of the competitiveness of the European Union’s internal market and public procurement have also attracted the attention of the European Council, which led to former Italian Prime Minister Enrico Letta preparing a special report on the necessary strategy for the modernization of the internal market. Due to the increased complexity and number of rules at the EU level, among other factors, he points to the need to efficiently and effectively implement and, inter alia, simplify these rules [4]. On this basis, the European Commission has launched an evaluation of whether public procurement rules in the EU are working as intended [5]. Our research on cost-effectiveness has theoretical assumptions and important practical notions and can help legislators answer the question of the efficiency of procurement rules and prioritize the objectives of procurement.

The public sector tends to meet the needs of the public interest and in this respect is significantly different from the private sector. This characteristic forms the basis for the functioning of public markets, where the public interest is replaced by profit maximization [6]. Value for money is itself an element of public expenditure. As an important element of the public interest and one of the fundamental principles of public procurement, it is used to realize the interest of the contracting authority in maximizing the benefits of the purchases. Value for money is the key to the optimal allocation of public goods and one of the main goals of public procurement [7]. In the past, several authors have made suggestions for a sound system of public procurement, for example, [8].

By proposing a new procurement directive in 2011, the European Commission sought to achieve two main goals: enable procurers to better use public procurement to support common societal goals and increase the efficiency of public spending to ensure the best possible procurement outcomes in terms of value for money [9]. At the same time, the European Parliament [10] adopted a resolution on the modernization of public procurement in the EU which advocates six tasks, half of which focus on increasing the efficiency of public procurement markets. Additionally, the European Commission [11] has issued a communication on the future challenges of public procurement in Europe, where the need to ensure a wider uptake of strategic procurement is recognized as one of the six strategic priorities that could lead to significant benefits in procurement outcomes.

In the study of the exchange mechanism of public procurement, it is important that public procurement procedures and techniques can be classified among the known different types of auctions, as well as the exchange mechanism of negotiation. Therefore, auction theory is an appropriate starting point to address aspects of the efficiency of the public procurement system as a function of the value for money of public money. Auctions in public procurement, which are the predominant exchange mechanisms in modern procurement [12,13,14,15,16], are usually conducted in a manner such as the English auction type, where the price (bid) is increased until one bidder remains. In public procurement, the process is reversed, as it is a buying process and not a selling process; hence, we speak of reverse auctions. However, formally, the process is the same.

An important aspect of auction theory is the economic analysis of different auctions and the commonly addressed question of whether those auctions are equivalent in terms of expected prices. Auction theory explains that auctions are to be revenue-equivalent as they result in the same expected prices. As Klemperer [13] (similarly [17]), as well as [18] and [19], have explained, Vickrey’s results on the equivalence in expected revenue of different auctions apply very generally, and buyers are indifferent to the choice of different types [20] except in negotiations where, as we have seen, research is undecided on the efficiency outcomes (e.g., [21,22]).

Efficient public procurement can lead to significant monetary savings and can therefore be understood as having better value for money, and public procurement has, among other factors, a built-in system of redress to prevent inappropriate decisions by contracting authorities [23]. Public procurement reforms that aim to increase economic growth will need to thoroughly account for the complex nature of the public procurement system as well as simultaneously address a range of challenges in the areas of accountability, transparency, fairness and economic efficiency [24].

High efficiency is one of the important objectives of public procurement [25], although the means and ends of public procurement (and its legal–economic regulation) may be interchangeable, as these means and ends and their hierarchy are not always clearly distinguishable [26]. Different countries within the European Union are trying to find the best balance between the normative regulation of individual procurement procedures and the discretion of contracting authorities. The right balance also means a more efficient public procurement system. As [23] points out, there are three key elements of an efficient public procurement system, namely, transparency, objectivity and non-discrimination. Additionally, the ECA [3] in its report has made recommendations to improve the public procurement system, and one of importance is to clarify and prioritize public procurement objectives. A reform of legislation and procedures that reduce prices in the procurement market can substantially reduce costs for the public sector and have a sizable macroeconomic benefit [27].

The public procurement system has been criticized over the years (e.g., [28]) for overemphasizing compliance and cost minimization to the detriment of public benefit and social welfare objectives. Some 15 years later, it is clear that the primary objective of value for money has somehow given way to secondary strategic policies. Today, the theory explains the components of public benefit from the perspective of the public procurer as the generation of innovation, the good functioning of the supplier market, the efficient (effective) functioning of the public procurement system and sustainable public procurement [29].

To test the cost efficiency of public procurement variables that can result in reduction, we employed the binary response model, which we believe to be not only theoretically important but also a practical and useful tool for possible future legislation. In this study, we address two questions. First, does the use of transparent public procurement procedures and different techniques increase the probability of the significant efficiency of government spending? Understanding this relationship is an important contribution to governments in designing their regulatory frameworks. Second, does the use of strategic procurement policies, such as green and social procurement, affect the efficiency of government spending? If so, to what extent?

2. Materials and Methods

Due to the size of public expenditure in public procurement, legislators have designed legislative frameworks to improve efficiency and to foster strategic goals of public spending, such as green and socially responsible procurement [11]. In the literature, the efficiency of public procurement is a matter of frequent scrutiny, for example, [30,31]. National procurement strategies often carry policy notes of delivering value for money in addressing efficiency, for example, [32]. Public procurement has long been used as a policy tool. It has been used to emphasize strategic goals in a wide range of fields. These include reducing unemployment, improving employment conditions, supporting small businesses, employing disabled workers and ensuring equal pay for men and women [33] and enhancing green and increasingly prevalent circular procurement [34]. Governments use their purchasing power to procure goods and services that foster innovation, respect the environment and combat climate change while improving employment, public health and social conditions. With its increasing popularity, corporate social responsibility has become linked to the use of public procurement [35,36,37]. Policies that focus on the operation or regulation of relationships between procuring entities and other entities outside the direct contract tend to be more burdensome for suppliers and also lead to higher costs [38]. The authors of [39] explain that the more complex the requirements of the secondary procurement policy, the higher the markup in the contract price paid by the procuring entity. Additional requirements (in a specific procurement) that are only indirectly related to the subject matter of the procurement and are defined and required by legislation increase both direct and transaction costs and lead to higher prices [40,41].

Previous studies have examined aspects of procurement efficiency mainly from the perspective of price or cost-effectiveness, and studies have used the difference between the estimated value (usually referred to as the reserve price, estimated price or expected price) and the final value of the best bid as a starting point and a model for calculating the efficiency of the different variables, for example, in [30,31,42,43,44,45]. Our study builds on previous research and will help policy makers address the challenges of reaching different goals of public procurement, efficiency and competitiveness on the one hand and pursuing strategic societal goals on the other.

The dependent variable employed in the present research was the same as that of many studies mentioned above, the difference between the estimated value, which is the average of the market prices (the contracting authority must determine the estimated value of each contract and base it on market prices), and the final (winning) price. This can be seen as an indicator of better auction performance, i.e., a more efficient use of public money.

2.1. Data

The primary sources are based on data from the Public Procurement Portal, owned by the Ministry of Public Administration and managed by the Slovene national gazette, Uradni list d.o.o. The data cover 120,455 public procurement procedures, which represent individual lots awarded by contracting authorities, i.e., state authorities, self-governing local authorities, public institutions and public enterprises between 2016 and 2023, and were processed by the authors. The computations were conducted using the EViews 12 University Edition for Windows software.

2.2. Variables

For each procurement, we examined at least 15 different factors, which are our explanatory variables. We chose existing types of procedures and the use of different public procurement techniques, such as centralized and joint procurement, electronic auctions and framework agreements. We also included the usage (or presence) of strategic procurement, such as green and socially responsible procurement. These different factors of procurement procedures can be measured on the scale of the whole population.

The dependent variable represents the effect of the difference between the estimated value of the public contract and the difference between the estimated market prices, based on a survey carried out by the contracting authority, and the price paid by the most advantageous tenderer. The estimation of the price is conducted by internal or external procurement experts, especially for large investment projects, and must be included in the project’s investment documents, so it should be considered representative of market prices. The difference between the estimated price and the value of a bid represents the cost-effectiveness of a specific procurement procedure, which explains whether the price in a procurement is lower than the contracting authority’s estimate [30]. Our dependent variable for testing significant reduction is binary; therefore, we chose the binary logit model to test our research questions.

The factors observed in our research are given in

Table 1. Types of observed factors.

Variable	Type	Construction
BIDS	Descriptive	Number of bids
OPEN (open procedure)	Binary	1 if using open procedure, 0 if not
NMV (small-value purchase procedure)	Binary	1 if using procedure, 0 if not
RESTRICTED (restricted procedure)	Binary	1 if using restricted procedure, 0 if not
COMP_D (competitive dialogue procedure)	Binary	1 if using procedure, 0 if not
NEGOTIATIONS_WITH (negotiations with prior publication procedure)	Binary	1 if using negotiation procedure, 0 if not
PI (partnership for innovation procedure)	Binary	1 if using partnership for innovation procedure, 0 if not
CENT (centralized procurement)	Binary	1 if using centralized procurement, 0 if not
FA (framework agreement)	Binary	1 if using framework agreement, 0 if not
AUCTION (electronic auction)	Binary	1 if using e-auction, 0 if not
SME (small and medium enterprises)	Binary	1 if involving SMEs, 0 if not
FOREIGN (non-domestic bidders)	Binary	1 if involving non-domestic bidders, 0 if not
SUB (subcontractors involved)	Binary	1 if involving subcontractors, 0 if not
EPP (electronic procurement procedure)	Binary	1 if using electronic public procurement, 0 if not
SOCIAL (socially responsible procurement)	Binary	1 if using socially responsible procurement, 0 if not
GREEN (green public procurement)	Binary	1 if using green procurement, 0 if not
NOTICE_EU (notification in TED)	Binary	1 if using EU Tender Electronic Daily notification portal, 0 if not
CRITERIA (additional, non-price criteria)	Binary	1 if using additional, non-cost criteria, 0 if not
NEGOTIATIONS_WITHOUT (negotiation without prior publication)	Binary	1 if using negotiation procedure, 0 if not

Source: own elaboration.

.

2.3. Research Questions

We explored the issue of the efficiency of the public procurement system from the cost-effectiveness perspective with different variables, similar to other studies (e.g., [46]). This point of view also reflects the notion of value for money in public procurement and is one of the system’s most prominent goals. To achieve this objective, we set out to investigate the following:

• Whether the usage of transparent procedures and techniques of public procurement are positively associated with efficient government spending.
• Whether the usage of strategic sustainable objectives of procurement is negatively associated with efficient government spending.

The EU directives promote the usage of transparent procedures and limit non-transparent procedures except in exceptional circumstances. This ensures a high level of transparency in public procurement. Several authors [30,42,44,45] have explained that transparency positively influences efficiency in public procurement. Our binary model was used to evaluate the efficiency of various factors from the perspective of significant savings and test our first hypothesis. Additionally, we tested the efficiency of public procurement techniques, such as electronic auctions and the centralized procurement and framework agreement. The usage of these techniques allows public entities to improve their purchases through flexibility, economics of scale, responsiveness and incentives [30,47,48,49,50]. We presume that these attributes contribute to better cost-efficiency outcomes of public contracts.

When governments pursue sustainable objectives through procurement policies, their procurement systems and the tasks within them become increasingly complex [51]. Strategic policies in procurement tend to drift away from the notion of cost efficiency. The complexity of sustainable procurement is often related to the complexity of procurement preparation, in terms of the preparation of tender and other documentation [41]. This complexity can increase prices and, through reduced competition, indirectly increase the government’s operating costs [40,52] and lead to less efficient spending. Socially responsible procurement (SRP) generates social value, and SRP generates social benefits, as demonstrated by rather positive feedback from employees [53]. On the other hand, legislative demands that are obligatory only in public purchasing raise the transactional cost. We presume that these attributes contribute to the lower efficiency and that the usage of sustainable procurement is negatively associated with government spending.

2.4. Methodology and Binary Regression Model

One way to measure cost efficiency is to compare the estimated prices of procurement and the final, winning price. This reduction can be seen as lowering contract values and reducing public spending. We tested the notion of cost-effective reduction, and our dependent variable was binary, taking only one of two forms, 0 or 1. We employed the binary response regression model because of the nature of the problem (a dependent variable that tests the reduction in prices) and the high accuracy of prediction. We tested alternative forms, and no improvements could be achieved. The logit model is considered to be the most effective approach for categorical variable data, and compared to the probit model, it is also mathematically simpler [54]. Additionally, logit models are more suitable for studies with large sample sizes [55], as is the case in this study.

A binary logit regression model is used to model binary variables, where the regressand is a binary variable, to determine event probabilities of factors where the dependent variable follows the Bernoulli distribution.

We employed this model to test whether different variables in our model raise or lower the odds of contract price reduction. In this case, the logistic distribution is very similar to a normal distribution [56]. Instead of using a linear model, a logistic function should be used for the dependence of the non-normal (binary) variable on the explanatory variables. Therefore, the notation of the equation is as follows:

$$y_i={\displaystyle \frac{1}{1+e^{-({\beta }_0+{\beta }_1{x}_1+{\beta }_2{x}_2+\cdots .+{\beta }_k{x}_k+u_i)}}}$$

(1)

The regression function is non-linear in both variables and parameters, so the ordinary least squares method cannot be used to estimate it [56]. Therefore, the equation is transformed into the following notation:

$${\displaystyle \frac{y_i}{1-y_i}}=e^{({\beta }_0+{\beta }_1{x}_1+{\beta }_2{x}_2+\cdots .+{\beta }_k{x}_k+u_i)}$$

(2)

The ratio ${\displaystyle \frac{y_i}{1-y_i}}$ is called the odds ratio.

We tested whether the observed variable in reduction is statistically significant and then tested the significance of the model itself. After that, we computed the cutoff rates with the estimation of the model. Logistic regression analysis can result in meaningful findings and can lead to accurate predictions with an appropriate choice of cutoff [57]. Lastly, we computed and analyzed statistically significant variables; first, we computed odds ratios, and then the probability of each variable. A table summarizing our procedure and briefly explaining the results is presented at the end of Section 4.

As explained, the observed variable is reduction, which is calculated using the following formula:

$$Reduction={\displaystyle \frac{Final price-Estimated price}{Estimated price}}$$

(3)

In binary response models, the dependent variable takes only two values, 0 or 1, as follows:

$$y=\left\{\begin{array}{c}1 if yes\\ 0 if no\end{array}\right\}$$

We examined how the independent variables (procedures, techniques and strategic policies) in our model affect the dependent variable, cost-effective reduction. Several authors, including the authors of [46], point out that although it is common to find studies that calculated differences between estimated and final values for certain factors using a linear regression model, the effect of ‘imaginary savings’ cannot always be found in the research, as these differences do not necessarily reflect actual market prices. Therefore, in our work, we mainly discuss ‘reductions’ rather than ‘savings’.

In procurement procedures, public entities are predominantly buying off-the-shelf products or premade solutions, with much less innovative aspects in organized, competitive and transparent environments. Rules of procedure that promote transparency and competition tend to reduce procurement prices [27]. Possible reductions in prices (as explained between estimated, average market price and final price) are based on the literature on the average reduction or ‘savings’ ([27] 2–10%, [47] 3,4%, and [46] 6%) and our mean of reduction (0.088). We arbitrarily divided procurements where the considerable reduction was at least 5% (value of 1 > 0.05; 0 < 0.05) to test our research hypotheses. We considered this threshold to represent cost-efficient public procurement.

2.5. Binary Regression Model in Public Procurement

Binary dependent variable models have been used in several public procurement papers as limited dependent variables, for example, ref. [58] on the use of competitive dialogue, ref. [59] on bidder prediction and [60] on start-ups in public procurement. There are also several papers on strategic public procurement, such as [61]; on barriers to innovation in public procurement, [62]; on green award criteria in procurement, [63]; on green procurement and innovation, [64]; on the adoption of green procurement practices, [65]; on financial barriers and green innovation, [66]; and on European funds and green public procurement.

As we have seen, previous authors have studied public procurement mainly from the perspective of different types of procurement goods [15] and from the perspective of projects of varying complexity [21,22]. The authors of [67] studied some of the factors influencing the final price in public procurement procedures, most of which are exogenous in nature; similarly, the efficiency of the public procurement system has been studied mainly from a competitive perspective by [68,69,70].

3. Results

The data for the analysis were extracted from the SQL database of the public procurement portal (enarocanje.si) of the Ministry of Public Administration, which is managed by the public company Uradni list d.o.o. We excluded public procurement contracts where there was an error in the estimated value of the contract or the final value of the tender. We included 120,455 public procurement contracts for goods, services and works during this period.

We ran a qualitative response regression model to study the impact of explanatory variables on reduction. We employed the binary logit regression model, which is used to model binary variables, where the regressand is a binary variable [71], to determine the event probabilities of factors where the dependent variable follows the Bernoulli distribution. In a binary response model, the interest lies primarily in the response probability [72]:

$$P (y=1|x)=P(y=1|x_1, x_2, \dots , x_{\mathrm{k}})$$

(4)

All the factors used in our research, except for the number of bids (offers), are dummy variables. By testing the different factors on the dependent variable, we were thus able to determine the direction and magnitude of their effect. First, we tested the statistical significance of the model.

We found that the following factors were not statistically significant and therefore were not included in the computation: involvement of SMEs, non-domestic bidders, involvement of subcontractors, restricted procedure and partnership for innovation procedure. We chose values lower than p < 0.05 to lower the chance of a Type I error [73]. The chi-square likelihood ratio of 6888.550, with a p-value of 0.0000, tells us that our model is statistically significant. We ran the model and obtained the results shown in

Table 2. Logit analysis.

Variable	Coefficient	Std. Error	z-Statistic	Prob.
C	−0.650443	0.043152	−15.07343	0.0000
BIDS	0.042663	0.002051	20.79769	0.0000
OPEN	0.833665	0.040557	20.55559	0.0000
NMV	0.538723	0.043667	12.33712	0.0000
COMP_D	0.662644	0.188860	3.508652	0.0005
CENT	0.145386	0.034033	4.271919	0.0000
FA	−0.820435	0.014418	−56.90379	0.0000
AUCTION	0.377771	0.070341	5.370587	0.0000
EPP	−0.166179	0.013901	−11.95442	0.0000
SOCIAL	0.294821	0.018330	16.08392	0.0000
GREEN	−0.124506	0.014815	−8.403979	0.0000
NOTICE_EU	−0.442513	0.027942	−15.83682	0.0000
CRITERIA	−0.248120	0.021228	−11.68826	0.0000
NEGOTIATIONS_WITHOUT	−0.236314	0.049532	−4.770958	0.0000

Source: authors’ computation.

.

The descriptive statistics of the included factors are listed in

Table 3. Descriptive statistics.

	BIDS	OPEN	NMV	COMP_D	CENT	FA	AUCTION	EPP	SOCIAL	GREEN	NOTICE_EU	CRITERIA	NEGOTIATIONS_WITHOUT
Mean	2.637	0.516	0.400	0.001	0.034	0.371	0.007	0.700	0.134	0.301	0.522	0.104	0.052
Maximum	155	1	1	1	1	1	1	1	1	1	1	1	1
Minimum	1000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
Std. Dev.	3.504	0.500	0.490	0.032	0.182	0.483	0.084	0.458	0.341	0.459	0.499	0.305	0.222
Skewness	12.686	−0.063	0.408	30.994	5.115	0.532	11.764	−0.871	2.143	0.866	−0.090	2.603	4.043
Kurtosis	274.541	1.004	1.167	961.641	27.159	1283	139.386	1.759	5.593	1.750	1.008	7.774	17.346
Jarque–Bera	3.73 × 108	2.01 × 104	2.02 × 104	4.63 × 109	3.45 × 106	2.05 × 104	9.61 × 107	2.30 × 104	1.26 × 105	2.29 × 104	2.01 × 104	2.50 × 105	3.73 × 108
Probability	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
Sum	317,700	62,127	48,183	125	4136	44,747	852	84,278	16,196	36,296	62,931	12,471	6244
Sum Sq. Dev.	1.479 × 106	3.008 × 104	2.891 × 104	1.249 × 102	3.994 × 103	2.812 × 104	8.460 × 102	2.531 × 104	1.402 × 104	2.536 × 104	3.005 × 104	1.118 × 104	5.920 × 104
Observations	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455	120,455

Source: authors’ computation.

:

Second, the computed value of the coefficient of determination and the adjusted coefficient of determination is not a good measure of the model’s goodness of fit, as actual values of the dependent variable can only be 0 or 1, so the R² and adjusted R² are small. We therefore employed the goodness of fit, which reveals how many observation units in the model prediction were correct, i.e., the proportion of correct predictions [56]. The graph (see Figure 1) below shows the prediction at different cutoff rates.

The optimal cutoff rate, where Dep = 0 and Dep = 1 were the closest, was 0.4 (see

Table 4. Prediction results of cutoff rate.

	Estimated Equation			Constant Probability
	Dep = 0	Dep = 1	Total	Dep = 0	Dep = 1	Total
P(Dep = 1) ≤ C	42,543	15,675	58,218	76,003	44,452	120,455
P(Dep = 1) > C	33,460	28,777	62,237	0	0	0
Total	76,003	44,452	120,455	76,003	44,452	120,455
Correct	42,543	28,777	71,320	76,003	0	76,003
% Correct	55.98	64.74	59.21	100.00	0.00	63.10
% Incorrect	44.02	35.26	40.79	0.00	100.00	36.90
Total Gain *	−44.02	64.74	−3.89
Percent Gain **	NA	64.74	−10.53

* Change in “% Correct” from default (constant probability) specification. ** Percent of incorrect (default) prediction corrected by equation. Source: authors’ computation.

).

The expectation–prediction evaluation for our model at a cutoff rate of 0,4 is correct in 59.21% of cases.

The correlation table is presented in

Table 5. Correlation table.

	BIDS	OPEN	NMV	COMP_D	CENT	FA	AUCTION	EPP	SOCIAL	GREEN	NOTICE_EU	CRITERIA	NEGOTIATIONS_ WITHOUT
BIDS	1000
OPEN	0.071	1000
NMV	−0.023	−0.843	1000
COMP_D	−0.005	−0.033	−0.026	1000
CENT	0.007	0.080	−0.084	−0.003	1000
FA	0.074	0.303	−0.237	−0.018	0.031	1000
AUCTION	0.007	0.001	0.011	−0.003	0.071	−0.026	1000
EPP	−0.048	0.113	0.017	−0.027	0.014	0.021	−0.046	1000
SOCIAL	−0.008	−0.008	0.025	0.000	0.023	0.053	−0.027	0.037	1000
GREEN	0.033	0.151	−0.083	−0.001	−0.008	0.297	−0.017	0.008	0.233	1000
NOTICE_EU	0.026	0.850	−0.854	0.015	0.095	0.302	−0.001	0.085	−0.017	0.119	1000
CRITERIA	−0.041	0.060	−0.047	0.028	0.035	0.037	−0.028	0.016	0.011	0.136	0.073	1000
NEGOTIATIONS_ WITHOUT	−0.098	−0.241	−0.191	−0.008	0.001	−0.124	−0.019	−0.229	−0.029	−0.116	−0.084	−0.044	1000

Source: authors’ computation.

.

Table 5 shows the pairwise correlations. The correlations show the direction of movement among the variables. The test is an important tool for helping us understand the trends and behavior of the relevant variables of our model. We see that correlation is low among procedures that are usually exclusive but not among those that include different techniques. There is a high correlation between publications in the EU portal notices for open procedures and negative NMV (small-value domestic procedure), as open procedures are mostly subject to publication in the EU portal for tenders (TED). In open procedures, we see that there is a positive correlation, while with domestic small-value purchases, it is negative. Both with open procedures and the NMV, there is a weak positive correlation in using electronic procurement. In sustainable procurement, the opposite can be observed. While open procedures are positively associated with using green procurement and rather neutral with regard to social procurement, the NMV is positively associated with social and negatively with green; in both, the correlation is weak. Centralized procurement tends to involve larger values, so it is logical that it is positively correlated with open and negatively with lower values (as in the case of the NMV). Electronic auctions are negatively correlated with sustainable procurement, both green and social, and this is an expected outcome, as auctions are predominantly used for off-shelf products in non-price criteria procurement. We see that e-auctions are negatively correlated with criteria. Non-price criteria are positively associated with sustainable procurement methods, which usually involve additional award criteria, such as lifecycle cost. We see that bids are positively correlated with factors showing a higher probability of reducing prices, meaning that competition plays an important role in cost-efficient procurement.

We performed computation with the relevant factors of our model. The transparent procedures, according to the model, tend to have a higher probability of being used in the procurement with a reduction of 5 percent or more, e.g., if the contracting authority uses an open procedure, ceteris paribus, there are 130 percent odds (and probability of 70%) that the reduction in the contract value or savings will be 5 percent or more. Similarly, we can see a high probability for the small-value purchase procedure (NMV) and competitive dialogue, and there are 93 percent odds (65% probability) of reduction when using this type of procedure, similar to centralized procurement and electronic auctions. If the contracting authority is using an electronic auction, there are 45 percent odds that savings will occur. We can see that these factors tend to be more efficient than the procedures not involving these factors. A negotiated procedure without publication reduces the possibility of reduction (above 5%), as do some other factors, such as green procurement, publication in the EU, additional criteria and, interestingly, the involvement of electronic public procurement. The usage of framework agreements reduces the odds of considerable reduction by 46 percent (the probability is 30%).

We can also see that each additional bid increases the odds of significant reduction by 4%, which means that competition, as measured by the number of bids in the procurement procedure, has an effect on reducing contract values.

As noted above, our model shows that at least one of the independent variables involved is correlated with the dependent variable (lowering), with 59,2% correct.

Regarding our first hypothesis, we cannot confirm that the usage of transparent procedures and techniques of public procurement is positively associated with efficient government spending. We see that all transparent procedures in our model are positively associated with efficient spending, but not all procurement techniques. We also cannot confirm our second hypothesis, as we see that the usage of strategic sustainable objectives of procurement can be associated with efficient government spending.

4. Discussion

The aim of this study was to examine the efficiency of the public procurement system as limited by the rules of procedure, the use of non-binding techniques (in the sense that the contracting authority can choose to use any or none of the techniques, as opposed to one of them) and other factors that influence the outcome of procedures, with a focus on the cost-effectiveness of procurement.

Firstly, we can see that transparent procedures give better results in terms of cost-effectiveness. These methods include open procedures and procedures for small-value purchases, i.e., the national open procedure, which is similar to the open and, as explained, is a more economical procedure. Small-value procedures tend to reduce procurement costs by a smaller margin than open procedures. When an open procedure is employed, odds are raised by 130% compared to reference values of other variables; therefore, the probability is 69% that cost reduction is significant and at least 5% (as demonstrated by the difference in average market prices). Additional bids induce higher cost reduction. We see that transparent procedures raise the odds of significant reductions due to their complete transparency, which not only attracts more competition but also gives an additional opportunity for economic operators to better decide possible business strategies of supply. This is especially true when an open procedure is employed. This procedure is published both in the national (local) and EU public procurement portals. We also see that competition, measured in the number of bids, increases the odds of significant price reduction.

Non-transparent negotiated procedures reduce the possibility of reducing contract values. When a negotiated procedure is used, the odds of savings of more than 5% are reduced by 21%, or there is a 44% chance that saving occurs. On the other hand, the use of competitive dialogue, which includes open notification and negotiation, increases the odds of cost-effective procedures by 93%, pushing the probability to 66%. We see that when a less transparent procedure is employed, the odds of savings are reduced. A negotiated procedure without publication and competitive dialogue involves the use of negotiations to lower prices, but we see that transparency is the key factor of price reduction.

Using different techniques, which can be combined with different procedures (and, unlike procedures, can be used simultaneously), we see that when contracting authorities use centralized procurement, the likelihood of cost reduction increases. This confirms the notion of the cost efficiency of centralized procurement due to market power, economics of scale, economics of scope and the professionalism of the central purchasing bodies with knowledge and expertise in handling both the procurement procedures and acquisition of procurement items. Centralized procurement can be an effective tool but should be chosen wisely, as inappropriate use can hinder competition in the long term, become prone to collusion, and lead to the deterioration of local supplies.

When contracting authorities use electronic auctions, the odds of significant savings are 45% (and there is a 59% probability that this event occurs). This result also confirms the notion of the efficiency of electronic auctions as an appropriate tool for reducing costs in public procurement. However, we see that the probability of reduction is lower than that of some other factors, confirming some recent studies (e.g., [47]) stating that reduction is present but not to the extent that the first studies on the use of electronic auctions in public procurement predicted. Nevertheless, while electronic auctions improve cost efficiency, the impact on the long-term supplier relationship and market competition should be considered, as its deterioration could result in a market for lemons.

On the other hand, our model shows that the use of framework agreements reduces the likelihood of significant cost reductions. Framework agreements tend to be seen as a more flexible technique, giving procuring entities more freedom. However, this technique can hinder competition as only those bidders who have entered the agreement can supply goods or services to the procuring entity for an extended period of time. This ‘closure’ with limited and known bidders is more prone to collusion, especially when the market is small and there are few bidders, making FAs an environment where bid rigging can much more easily occur.

Similarly, when electronic tendering is used, the probability of significant cost reduction decreases. This can be explained by the additional administrative burden that e-procurement systems create, especially for small and medium-sized enterprises. In the ECA’s report [3], we see that cross-border procurement between EU countries is rather rare and that the informatization of procurement in the context of remotely doing business is beneficial for a rather small number of procedures. However, we believe that this is only a short-term consequence, as the system of e-procurement in Slovenia was introduced in 2018. We believe that in the long term, this will change as paper-based procurement becomes obsolete and users of the electronic platforms will need to adapt. Additionally, other public sector services will be fully digitalized, and because of this, users will further adapt (increase their digital skills); therefore, remote tendering will become more beneficial for competition.

We also see that when the procurement entity uses non-price criteria, there is a reduced possibility of significant reduction. On the one hand, non-price criteria can diversify the procurement procedure, but on the other, additional criteria can raise the total cost of ownership. When less competent contracting authorities (in terms of procurement procedures and/or market knowledge) use non-price criteria, they could induce unrealistic consideration of non-price criteria, which could lead to the lower efficiency of procurement procedures.

We also see that procurements with EU notices lower the possibility of price reduction, which can be explained by the fact that those procurements are of higher value and are more complex and demanding both for average contracting authorities and economic operators. This results in a higher probability that the procurement documentation will be ill prepared, therefore reducing the incentives of the bidder for price reduction.

The use of strategic procurement, socially responsible procurement (SRP) and green public procurement produces the opposite results. The use of socially responsible criteria in public procurement increases the odds of more cost-effective procurement by 34% (or 57% probability). Socially responsible procurement increases the likelihood of a significant reduction. First, we see that this reduction can be attributed to the notion that SRP is mostly optional, and contracting authorities can choose to implement it in procurement procedures where this suits their needs. Second, in procurement procedures contracting authorities with additional social criteria, companies with greater employment commitment and higher corporate social responsibility (CSR) can be attracted. Companies with higher CSR receive more procurement contracts [37] and have better employee engagement. This engagement increases productivity, and these companies are more profitable and also tend to have lower absenteeism [74], thus significantly reducing the cost of procurement contracts.

Finally, the opposite is true for green procurement, with the use of such criteria reducing the likelihood of cost-effective procurement. While the use of SRP is optional in the public procurement system of Slovenia, green procurement is mandatory for many procurement items, according to the Decree on Green Public Procurement. This adds to the complexity of the system and leads to uncritical, obligatory use of sustainable criteria regardless of the circumstances of each procurement. This can lead to poor procurement documents, especially from smaller procuring entities without sufficient knowledge of the subject matter. Additional requirements in procurement procedures can reduce competition, leading to less efficient spending. On the other hand, legal requirements that are only mandatory in public procurement increase transaction costs, making public procurement less attractive.

In

Table 6. Results of the study.

	Regression	Odds Ratio	Probability	Results
BIDS	0.043	1.044	0.511	Each additional bid raises the probability of a cost reduction. Competition has a positive effect on the reduction in contract prices.
OPEN	0.834	2.302	0.697	The open procedure substantially raises the probability of a cost reduction. Transparent procedures attract competition and lower prices.
NMV	0.539	1.714	0.632	Small-value purchases raise the probability of a cost reduction. Transparent procedures attract competition and lower prices.
COMP_D	0.663	1.940	0.660	Competitive dialogue raises the probability of a cost reduction. Transparency attracts competition and lowers prices.
CENT	0.145	1.156	0.536	Centralized procurement raises the probability of a cost reduction. Centralized procurement using economics of scale and high-skilled competences of central purchasing bodies lowers contract prices.
FA	−0.820	0.440	0.306	Framework agreements substantially reduce the probability of a cost reduction. FAs hinder competition in a closed environment prone to collusion.
AUCTION	0.378	1.459	0.593	Electronic auctions raise the probability of a cost reduction. EAs are seen as an appropriate tool for reducing prices.
EPP	−0.166	0.847	0.459	Electronic procurement lowers the probability of reduction, as this tool can be, in the short term, be perceived as an additional administrative burden for both the contracting authority and bidders.
SOCIAL	0.295	1.343	0.573	SRP raises the probability of a cost reduction. Non-mandatory SRP can attract companies with higher CSR that have better employment engagement, which reduces prices.
GREEN	−0.125	0.883	0.469	Green procurement lowers the possibility of price reduction, which in an obligatory legislative framework adds to complexity and additional transaction costs.
NOTICE_EU	−0.443	0.642	0.391	Procurements with EU notices lower the possibility of price reduction due to the higher value and complexity of procurement, which can more often lead to ill-prepared documentation, reducing incentives for bidders.
CRITERIA	−0.248	0.780	0.438	Non-price procurement lowers the possibility of price reduction. Non-price criteria can diversify the procurement procedure, but additional criteria raise the total cost of ownership.
NEGOTIATIONS_WITHOUT	−0.236	0.790	0.441	Negotiations without publication procurement lower the possibility of price reduction. This procedure lacks transparency, which directly influences competition and contributes to less efficient spending.

Source: authors’ computation.

, we present the results of our computation in the binary logistic model and briefly comment on the results.

5. Conclusions

The prevailing contention in the realm of public procurement appears to be an inherent dichotomy between the pursuit of efficiency and the pursuit of social and environmental objectives. Restrictive conditions have the potential to impede the competitive landscape and result in the exclusion of the bidders [75]. Conversely, the European Court of Auditors [3] has advocated for the optimization of public procurement goals in cases where it seems that the efficiency and strategic goals of procurement are incompatible.

The year 2024 was a wake-up call for public procurement in the European Union. In December 2023, the ECA published its first in-depth report on the competitiveness of EU public procurement markets [3], stating that the 2014 reform of EU directives has shown no signs of reversing this trend. Overall, there is a lack of awareness of competition in public procurement. In 2024, Enrico Letta, with the approval of the EU Council, published a report on the EU’s single market, arguing for significant reforms. Similar to the ECA, Letta [4] proposed reforms to public procurement, focusing in particular on the key objectives of procurement and better regulation. At the end of 2024, the European Commission launched an evaluation of whether public procurement rules in the EU are working as intended [5].

The results of our analysis show the importance of cost-effective public procurement procedures in a transparent environment. While open procedures are among the most efficient, national small-value purchases, which are similarly transparent to open procedures but more flexible, with more discretion and cost-effectiveness, lead to similar results in terms of cost-effectiveness. Furthermore, it was confirmed that less transparent procedures are also less efficient in terms of the probability of reducing contract values. The number of tenders received also has a direct impact on the probability of achieving a reduction in contract values.

On the other hand, the use of centralized procurement and electronic auctions leads to the best results, while techniques such as framework agreements, which operate in a controllable, semi-closed environment and are more prone to collusion, are less likely to be cost-effective. With regard to other factors, we observed that their use by the contracting authority (green procurement and electronic procurement) reduces the likelihood of cost-effective procurement outcomes, while the opposite is true for socially responsible procurement. We determined that strategic procurement can have either a negative or positive impact on cost efficiency, so these policies should not be written off as unsuitable.

It can be said that a combination of transparent procedures and an undisclosed market environment at the time of tendering, as well as less binding secondary objectives of procurement, deliver the best cost-efficiency results where savings are considered to be significant. It is recommended that EU legislation promotes a high level of transparency, provides more flexible rules for procurement above the threshold and limits the use of framework agreements. Secondary measures may have a higher policy objective of addressing climate change and rapidly emerging environmental challenges and may result in a fairer and more inclusive and sustainable society. Their cost-effectiveness may remain a challenge.

We observed that socially responsible procurement, which operates in a non-mandatory, demand-driven environment, raises the odds of a significant reduction in procurement contracts. On the other hand, we observe that green procurement lowers the odds and operates in more complex and challenging legal and operational environments.

It seems that we can have both sustainable and efficient procurement. However, in order to achieve the goals of efficient procurement on the one hand and sustainable procurement on the other, strategic goals should be achieved via the professionalization of procuring entities and should be demand-driven rather than uncritically prescribed by the obligation of legislation. The reform of legislation, which promotes public procurement procedures that are transparent and promote competitiveness and that are better tailored to the individual needs of procuring entities, will deliver better results in terms of cost efficiency but will also stimulate the usage of sustainable, strategic goals of procurement. We see this as a system that delivers the best value for society.