The Impact of Research and Development Investment on the Performance of Portuguese Companies

Abstract

This study investigates the impact of Research and Development (R&D) investment on the performance of Portuguese companies, specifically addressing the gap in understanding how R&D influences a company’s value and performance. We employ a dynamic panel data model estimated using the Generalized Method of Moments (GMM) to account for potential endogeneity issues. This approach allows us to analyze the influence of R&D investment on the Return on Operating Assets (ROA) for Portuguese companies with significant R&D investments between 2012 and 2019. The analysis reveals that while R&D investment itself may not have a statistically significant short-term impact on ROA, lagged financial performance, leverage, asset turnover ratio, and accounts payable turnover all demonstrate a statistically significant relationship with the dependent variable.

1. Introduction

The high technological development that has been observed in major world economic powers has triggered in organizations a great interest in valuing themselves through the acquisition or internal generation of intangibles, thus obtaining a competitive advantage over their opponents and some differentiation (Arrighetti et al. 2014; Clausen and Hirth 2016; Leung and Sharma 2021; He and Estébanez 2023; Hutauruk 2024). In this context, different business areas have had the challenging need for constant adaptation and updating by developing rapid mechanisms that respond to the most diverse situations (Osma and Young 2009; Tuna et al. 2015; Ibhagui 2019; Rodrigues et al. 2020; Parveez et al. 2023; Mendes et al. 2022). Thus, Accounting, as a support base for Management, has also assumed a position of extreme importance insofar as these constant updates are reflected in the production of financial information (Anagnostopoulou 2008; Honoré et al. 2015).

Investment in R&D has proven to be a motivating aspect for companies, as it brings benefits not only to the entity that carries it out but also to society in general (Freimane and Bāliņa 2016; Pegkas et al. 2019). However, the increase in R&D costs highlights the demand from managers for evidence of its impact on the company’s performance (Tsai et al. 2016; Rahman and Howlader 2022). Some studies have already shown that investment in R&D made by companies positively affects their market value (see, for example, Ito and Pucik 1993; Chauvin and Hirschey 1994; Bae and Noh 2001; Chung et al. 2019; Coelho et al. 2023). The literature reveals that investment in R&D has become a critical factor for the success and survival of companies, showing that it promotes competitive advantage through differentiation strategies that lead to the production of new and better products and services (Basgoze and Sayin 2013; Muhammad et al. 2024). It also identifies a problem in the production of financial information regarding the valuation of R&D expenses arising from investment in intangible assets, and therefore, the present work addresses the deficiencies of the current accounting model as a general objective of analysis (Driver and Guedes 2012; Akcali and Sismanoglu 2015; Carvalho et al. 2024). It is important to distinguish between capitalized R&D reported on the balance sheet and expensed R&D appearing on the income statement. Capitalized R&D reduces information asymmetry between managers and shareholders, while expensed R&D can increase it.

Knowing that there is no universally accepted model for valuing intangible assets resulting from research and development activities and considering this condition as the motivation for this study, the specific objective of this work was to analyze the impact of R&D investment on the performance and value of companies.

The research questions that this work intends to answer are the following:

(1)

As a foundation for understanding the potential impact of R&D on firm performance, we first examine the following: How is the accounting treatment of intangibles and R&D expenses carried out?

(2)

How does the investment in intangible assets resulting from R&D influence companies’ performance?

Given the relevance of the theme, the present work intends to be a contribution to the analysis of the impact of investment in R&D intangible assets in the Portuguese business environment. After this brief introduction, the theoretical framework is presented, highlighting the regulatory support applicable to R&D and the relationship between R&D and companies’ performance. The description of the methodology used to achieve the proposed objectives follows, and after that, the empirical study and the results obtained are presented, as well as their discussion. Finally, the main conclusions of the study are drawn up, as well as the limitations found and suggestions for future work.

2. Theoretical Framework

Firstly, the Portuguese regulatory framework regarding standards and fundamental concepts for the study will be made. Then, we follow with a reference to the problem of the current accounting difficulty in terms of intangible assets: the treatment of R&D expenses. The relationship between R&D investment and companies’ performance will also be analyzed, and finally, the evolution of investment in intangibles in Portuguese businesses will be presented.

2.1. Applicable Regulations: Measurement, Recognition, and Disclosure

When Portugal joined the European Economic Community in 1986, it was necessary to adjust the existing regulations to the Fourth Directive (Directive n° 78/660/EEC). Thus, the Portuguese Official Chart of Accounts, which had proved to be out of step with reality, was subject to several successive changes resulting from the need to adapt the accounting model in force. This dynamic also emerged from the adoption by the European Union itself of International Accounting Standards, aiming at accounting harmonization as much as possible. Years later, with Decree-Law n° 158/2009 of 13 July (Ministério das Finanças e da Administração Pública 2009), the Accounting Standards System (ASS) came into force. The ASS was developed in close alignment with the new EU standards so that the country can achieve alignment with the European Union accounting directives and regulations. Within this modernization of the standards, it became possible to compare national companies and companies in international markets.

Regarding intangible assets, which are the focus of this work, they are accounted for by the Accounting Standard for Financial Reporting 6–Intangible Assets (based on International Accounting Standard 38–Intangible Assets). The Accounting Standard for Financial Reporting 6, §8 (a), p. 99 defines an intangible asset as “a non-monetary identifiable asset without physical substance”, which determines some parameters that an asset must comply with in order to be considered as an intangible asset and specifies how the carrying amount of this type of asset should be measured. This standard mentions three conditions for an asset to be defined as intangible, namely identifiability, control, and future economic benefits. Regarding the criterion of identifiability, §12 states that an intangible asset is identifiable if it is separable from other items in the item category and can be sold, licensed, rented, or exchanged, or if it results from contractual or legal rights, regardless of whether these rights may be separable from the entity. As for control, it is explained in §13, p. 99, which states that “an entity controls an asset if it has the power to obtain future economic benefits flowing to the entity, and can restrict the access of others to those benefits”. Future economic benefits include revenues from the sale of products or services or other benefits arising from the use of the asset by the holding entity, as stated in §17.

At the level of the initial measurement of intangible assets, this should be done at cost (as indicated in §24) that is, the intangible asset should be carried at its cost that is less than any amortization and accumulated impairment losses (§71). After initial recognition, §73 shows that the entity should adopt the revaluation model as a measurement metric, that is, the intangible asset should be carried at a revalued amount that is its fair value at the revaluation date minus the losses of subsequent accumulated impairment and minus any subsequent accumulated amortization. In subsequent measurement, if the entity chooses to account for the intangible asset under a cost model, then all intangible assets in that class must follow the same assumption (§70). On the one hand, recognition is easy to interpret when the asset is purchased, and on the other hand, it is not when the intangible asset is generated internally. When the asset is acquired separately, §27 states that its cost is determined by its purchase price plus any cost that has been allocated to prepare the asset for its use. For internally generated assets, §64 states that the cost includes all costs directly attributed to creating, producing and preparing the asset to be able to work as intended. In the case of intangible assets generated internally, §49 recognizes the difficulty in assessing the recognition of the asset in the cases such as (1) identifying whether and when there is an identifiable asset that generates future economic benefits and (2) determining the reliable cost of the asset. For assets with this particularity, it is important to clarify the concepts “research phase” and “development phase”. Regarding the research phase, §52 clarifies that expenditures obtained in this phase should be recognized as an expense when they occur. Paragraph §53 adds that, at the research stage, the entity cannot demonstrate that an intangible asset that could generate future economic benefits exists. Regarding the development phase, §55 states that an intangible asset arising from this phase should be recognized only if the entity is able to demonstrate very particular conditions such as the feasibility of completion, as well as the intention to complete the intangible asset so that it becomes available for use or sale; this includes its ability to use or sell the intangible asset and to reliably measure expenditure on the intangible asset during the development phase, as well as the way in which the intangible asset can generate future economic benefits and the availability of technical and financial resources to complete the development of the intangible asset. It should also be noted that §51 reinforces that if the entity cannot distinguish the research phase from the development phase of a project generated internally, then the entity should treat the expenditure as if it were incurred only in the research phase.

The dissemination of information about research and development activities has been considered to be an extremely important issue (Ehie and Olibe 2010; Duqi et al. 2011). Kang and Gray (2011) carried out a study to examine factors associated with the practices of the dissemination of information regarding the intangible assets of the 200 largest companies in emerging markets and concluded that the companies in the study voluntarily disclose the accounting information of their intangibles; however, this disclosure is affected by specific factors such as leverage, IFRS/US GAAP3 adoption, industry type, and country-specific indicators such as economic policies and legal systems. It is therefore necessary to create conditions for the dissemination of information about intangible assets resulting from R&D activities (Bandeira and Afonso 2010). This authors analyzed the importance of R&D in the companies’ valuation and the benefits of a correct valuation, which were translated into an increase in the quality of financial reporting. The study was based on a sample of the twenty companies that, worldwide, invested the most in R&D between 1996 and 2006, having concluded that there is a positive relationship between the results and, consequently, on the value of companies and their R&D activities (Bandeira 2010). In the same line of research, Castilla-Polo and Gallardo-Vázquez (2016) carried out a study on the existing literature on the subject and the disclosure of intangible assets; they concluded that the relevance of this information is essential for stakeholders. According to Lev (1992), an adequate strategy of voluntary disclosure of credible information in a frequent and relevant way reduces the information gap between the company and the stakeholders. On the other hand, Nichita (2019, p. 225) analyzed research papers trying to answer the following research question: “How do researchers address the definition, measurement, recognition and potential of intangible assets to generate future economic benefits when a formal framework for reporting them is highly controversial?” In this study, research articles about intangible assets published between 2000 and 2019 were analyzed, and it was concluded that research on intangibles does not have unanimous agreement regarding the definition, measurement, recognition, and disclosure criteria; however, it recognizes the important contribution of these resources to boost the competitiveness, performance, and gains of organizations. Alves and Pascoal (2017) argue that adequate measurement, recognition, and disclosure of information that reflects the company’s economic situation make accounting useful for decision making. Bandeira (2010) proposes that the recognition of R&D expenses can be done through the allocation in full to results or through full or selective capitalization. Regarding the full allocation of expenses to results and considering the principle of prudence, these are considered as expenses of the period in which they occur. This method can be reductive, given that it can become difficult to apply. As a result, when making a prediction of benefits, it will not be possible to quantify them in the best way. Regarding full capitalization, it is based on the balancing principle, that is, the company incurs expenses to collect revenue in the future. This method incurs the risk of recognizing intangible assets, since a substance that does not have a realizable value is being accounted for as an asset. As for selective capitalization, it should treat R&D expenses according to the degree of certainty of future economic benefits.

2.2. Relationship between R&D and Companies’ Performance

Compared to tangible assets, intangibles are associated with higher levels of uncertainty, as suggested by Gu and Wang (2005), since information about intangibles is more complex; see the example of technologies and patent property rights. In their study, Gu and Wang (2005) estimated the relationship between earnings forecasts made by analysts and companies’ intangible assets. The formalization of the hypothesis aimed to test the following: (1) whether the analysts’ forecast errors regarding future earnings are higher for companies with a higher intangible purpose; (2) the existence of a positive association between the analysts’ forecast errors and the diversity of companies’ technological investment portfolios; and (3) whether analysts’ forecast errors are higher for companies with permanent investments in innovation. The results of this study showed the existence of a positive relationship between analysts’ forecast errors and the intensities of intangibles that companies held. The same was true regarding the diversity and innovation of companies’ technology investments. Contrary to what Gu and Wang (2005) expected, the findings showed that industries with higher intensity values of intangibles did not have higher forecast errors, and the regulation related to intangibles of biotechnology, pharmaceutical and medical equipment industries decreased the analysts’ forecast errors and their association with intangibles.

From the literature point of view, intangible assets represent an important part in determining the value of a company. In this line, Oliveira et al. (2010) tested the relevance of recognizing intangible assets and goodwill. These authors focused their study on the analysis of the financial reports of companies with values not listed on the Portuguese Stock Exchange between 1998 and 2008. They formulated hypotheses to test (1) the relevance of asset recognition on the market value of equity and (2) whether the relevance of the accounting value, gains, and intangible assets recognized based on the IAS differs from the information based on Portuguese accounting principles. The results showed that, with the change to the IAS/IFRS, the increase in the relevance of intangible assets was very small, given the conservative nature of Portuguese regulations and the aforementioned international standards. Ely and Waymire (1999), Barth and Kasznik (1999), and Gelb and Siegel (2000) also developed research aiming to study if intangible assets associated with research and development and information asymmetry are significantly positive in relation to the publication of returns from repurchase shares. The conclusions obtained showed that companies with more intangible assets were more prone to share buybacks and to lower information asymmetry toward investors. Ballester et al. (2003) showed that there are significant differences, since the time series assumes the invariance of the specific parameter of the company combined with time. The transversal approach, on the other hand, assumes that all companies have the same capitalization and amortization rates for their R&D expenses. As stated by Gelb and Siegel (2000), R&D and advertising expenses usually result in patents, technologies, and brand names that, because they are intangible, are difficult to valuate accurately. According to the same authors, an incorrect valuation leads to the creation of financial information that is not very useful or relevant for investors. Thus, they developed research with the purpose of understanding whether companies with significant levels of intangible assets are more likely to highlight the increase in dividends and share buybacks, which is usually seen as a way of recognizing favorable investment opportunities, rather than using ordinary accounting disclosures. Focusing on the theme of the economic value of R&D activities, Ballester et al. (2003) used past information on earnings, accounting values, and R&D expenses to estimate the economic value of the kinds of R&D that investors consider to be an asset. This study adapted the methodology used by Ohlson in 1995 to estimate the existence of unusual incomes, which is the proportion of current expenditure on R&D as a form of future economic benefit for the company and the amortization rate of that same asset. The authors compared a time series of data with estimates of parameters of capitalization, the persistence of earnings and the economic value of R&D intangibles, and a cross-sectional series, and they concluded that, in general, investors consider R&D expenditures as an economic good.

3. Methodology

The empirical study developed aimed to analyze the impact of investment in R&D intangible assets on companies’ value, with a special focus on Portuguese companies that invested the most in R&D in the period from 2012 to 2019.

3.1. Sample

The 25 Portuguese companies that had the most investment in R&D in the period from 2012 to 2019 were selected for the sample, resulting in a total of 200 observations. After selecting the sample, data related to the variables identified in

Table 1. Description of the variables.

Variables
Dependent variable
Financial performance (ROA)	Operating result/Total assets
Independent variables
Financial performance (ROA) (t-1)	Operating result (t-1)/Total assets (t-1)
Size	Log (Total assets)
Liquidity	Current assets/Current liabilities
Leverage	Liabilities/Shareholders’ equity
R&D investment intensity	Investment in R&D/Total assets
R&D investment intensity (t-1)	Investment in R&D (t-1)/Total assets (t-1)
R&D investment intensity (t-2)	Investment in R&D (t-2)/Total Assets (t-2)
Asset turnover rate	Turnover/Total Assets
Inventories turnover rate	COGS/Inventories
Turnover rate of accounts receivable	Turnover/Accounts receivable
Turnover rate of accounts payable	CGSMC/Accounts payable
Dummy variables
Technological intensity	High technological intensity
	Medium–high technological intensity
	Medium–low technological intensity
	Low technological intensity

Source: Adapted from Ayaydin and Karaaslan (2014).

were obtained, and finally, the pml package of R software (version 4.3.3) was used to specify the panel data (R Core Team 2021).

3.2. Variables

The variables used in our model were based on the study of Ayaydin and Karaaslan (2014) adapted to the reality of the Portuguese companies (please see their description in Table 1). The dependent variable of the model is the financial performance of the company measured by the return on assets (ROA), which is calculated by the ratio between the operating result and the total assets, in the current period. This ratio is a measure of profitability and allows us to analyze the return generated by a company’s total assets. As in Lev et al. (2005), the return on assets was used as a measure of a company’s profitability, since it is through it that the company’s position is reflected. The company’s performance was thus measured in terms of profitability and not in terms of innovative results (such as productivity or number of patents). The independent variables used in our model are the firm size, liquidity, leverage, R&D investment intensity, and operating efficiency ratios: turnover of receivables, payables, inventories, and assets, as well as the level of technological intensity, which was coded by dummy variables. The firm size was measured by the natural logarithm of the total assets to avoid any compound effect. The R&D investment intensity was calculated by the ratio between the investment in R&D and the total assets. It became important to add lags of the dependent and independent variables, since the analysis of economic and financial relationships usually observes economic behaviors that can be influenced by past experiences and old patterns.

3.3. Procedure

A common issue when analyzing R&D is the potential endogeneity that can arise between its variables. Existing studies on the relationship between performance and R&D often use static linear regressions, neglecting the temporal dependence of firm R&D investment (e.g., Honoré et al. 2015; Rodrigues et al. 2020). These studies assume no correlation between the past and present values of a firm’s R&D investment, which is an assumption that contradicts intuitive expectations.

To mitigate endogeneity concerns, we employed a dynamic panel data approach with a two-step system generalized method of moments (GMM) and one-year lagged dependent variable as instruments. Additionally, some independent variables have been lagged to further address simultaneity and endogeneity issues.

The generalized method of moments (GMM) estimator, as mentioned earlier, is a statistical technique used to estimate parameters in models with potential endogeneity issues. It is particularly effective for panel data analysis where variables may be correlated with past values and unobserved effects. GMM uses instruments—typically lagged values of the variables—to control for endogeneity, providing consistent and efficient estimates. The two-step GMM approach enhances precision by first obtaining initial estimates and then refining them using a weighting matrix derived from these estimates. This method is robust to various sources of endogeneity, including unobserved heterogeneity and simultaneity. For a more detailed discussion on GMM estimator, see Ullah et al. (2018).

The formulation of the model that allows for determining the impact of R&D on the firm’s financial performance (ROA) is described in Equation (1) and includes the following:

• The financial performance lagged by one period (ROA(t-1));
• The firm size (SIZE);
• The liquidity (LIQ);
• The leverage (LEV);
• The assets turnover rate (ATR);
• The inventories turnover rate (ITR);
• The turnover rate of accounts receivable (TRAR);
• The turnover rate of accounts payable (TRAP);
• The R&D investment intensity (RD) and its lags of one (RD(t-1)) and two periods (RD(t-2));
• The technological intensity of the firm (TI).

$$\begin{gathered} \mathrm{RO}{\mathrm{A}}_{i,t}={\beta }_0+{\beta }_1\mathrm{RO}{\mathrm{A}}_{i,t-1}+{\beta }_2{\mathrm{SIZE}}_{i,t}+{\beta }_3\mathrm{LI}{\mathrm{Q}}_{i,t}+{\beta }_4\mathrm{LE}{\mathrm{V}}_{i,t}+\mathrm{AT}{\mathrm{R}}_{i,t}{\beta }_5+\mathrm{IT}{\mathrm{R}}_{i,t}{\beta }_6+\mathrm{TRA}{\mathrm{R}}_{i,t}{\beta }_7+\mathrm{TRA}{\mathrm{P}}_{i,t}{\beta }_8 \\ +\mathrm{R}{\mathrm{D}}_{i,t}{\beta }_9+\mathrm{R}{\mathrm{D}}_{i,t-1}{\beta }_{10}+\mathrm{R}{\mathrm{D}}_{i,t-2}{\beta }_{11}+\mathrm{T}{\mathrm{I}}_{i,t}{\beta }_{12}+{\epsilon }_{i,t} \end{gathered}$$

(1)

Note that index i represents the company, index t represents time, and ε refers to the error term. As mentioned, several explanatory variables have also been included as lags. Considering that some of the independent variables showed no statistical correlation with the dependent variable, five versions of this model were developed so that it was possible to measure their impact on the target variable.

4. Presentation and Discussion of Results

In a first phase, the results of descriptive statistics were obtained. They consist of measures that allow for a description of the sample, namely by the value of the mean, median, maximum, minimum, and standard deviation. Then, the five versions of the panel data model described by Equation (1) were estimated using the generalized method of moments.

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

	Mean	Median	Maximum	Minimum	SD	No. Obs.
ROA	0.0659	0.0511	0.4939	−0.5389	0.1058	200
SIZE	19.2649	18.8465	24.5094	16.7203	1.8202	200
LIQ	1.4400	1.2138	6.8035	0.3629	1.1092	200
LEV	7.0858	1.5631	560.8876	0.1905	40.9572	200
RD	0.0556	0.0397	0.6220	0.0001	0.0661	200
ATR	1.0591	0.8600	4.4360	0.1022	0.7226	200
ITR	7.1206	4.6212	104.3020	0.4600	9.8259	200
TRAR	12.8931	4.7818	168.2804	1.2494	25.6355	200
TRAP	3.6934	2.5390	15.5081	0.0000	3.0995	200

presents the results of descriptive statistics (mean, median, maximum, minimum, and standard deviation) for each dependent variable, which were obtained with a total of 200 observations. The mean value of the return on assets shows that in the period from 2012 to 2019, the assets had the ability to generate an average positive operating result of 6.59%. The results in this table also show that the intensity of investment in R&D in relation to the asset, and considering average values, was 5.56%. Regarding leverage, it represents the level of indebtedness used to maximize the return on investment. In this case, the average value was 7.09%, which shows that the companies under analysis have heavily resorted to debt to finance their assets. With respect to the effectiveness of the operational cycle, the turnover of assets was around 1.06, which means that, on average, companies were using their assets efficiently to generate sales. On the other hand, the average inventory turnover was 7.12, which means that, on average, stocks were renewed 7.12 times a year, that is, they remained in the companies for about 1 month and 21 days. The average turnover of accounts receivable was 12.89 days, and the average turnover of accounts payable was 3.69.

Regarding technological intensity, as shown in

Table 3. Distribution of the sample’s technological intensity.

	No.	%
High technological intensity	8	32%
Medium–high technological intensity	9	36%
Medium–low technological intensity	1	4%
Low technological intensity	7	28%
Total	25	100%

, the sample is composed of eight companies with a high level of technological intensity (32%), nine companies with a medium–high level of technological intensity (36%), one company with a medium–low level of technological intensity (4%), and seven companies with a low level of technological intensity (28%).

Table 4. Pearson’s correlation matrix.

	ROA	SIZE	LIQ	LEV	RD	ATR	ITR	TRAR	TRAP
ROA	1
SIZE	−0.1485 *	1
LIQ	0.1054	−0.1249 ●	1
LEV	−0.1633 *	0.1371 ●	−0.0403	1
RD	0.1030	−0.4527 ***	−0.0075	−0.0980	1
ATR	0.3049 ***	−0.3862 ***	−0.0718	0.0082	0.2373 ***	1
ITR	0.0178	0.3127 ***	−0.1540 *	−0.0527	0.0582	0.0369	1
TRAR	−0.0476	0.7360 ***	−0.1719 *	−0.0118	−0.2471 ***	−0.2423 ***	0.5085 ***	1
TRAP	0.4078 ***	−0.0209	0.4168 ***	−0.0894	0.1347 ●	0.4754 ***	0.2553 ***	0.1670 *	1

Note: ^●, * and *** correspond to the statistical significance levels of 10%, 5%, and 0.1%, respectively.

presents the Pearson’s correlation matrix for the dependent variables. Pearson’s correlation assumes values in the interval [−1;+1] and measures the magnitude and direction of the linear association between two variables, with −1 meaning the existence of a perfect negative linear relationship (when one variable increases, the other variable decreases) and +1 meaning the existence of a perfect positive linear relationship (when one variable increases, the other variable also increases).

It is commonly assumed that correlation values between 0 and 0.3 indicate the existence of a weak positive relationship, and values between −0.3 and 0 indicate the existence of a weak negative relationship. Values between 0.3 and 0.7/−0.3 and −0.7 indicate that the relationship is a moderate positive/negative. For the range 0.7 to 1/−0.7 to −1, the relationship is assumed to be a strong positive/negative. As can be seen in Table 4, there was a moderate positive correlation between the financial performance (ROA) and the assets turnover (ATR), as well as with the accounts payable turnover (TRAP) (about 0.30 and 0.41, respectively). There was also a moderate positive correlation between the inventory rotation (ITR) and size (SIZE) (about 0.31). There was also a moderate positive correlation between the turnover of accounts payable (TRAP) and the turnover of assets (ATR) (around 0.48). The turnover of accounts receivable (TRAR) also showed a moderate correlation with the turnover of inventories (ITR) in the order of 0.51. The turnover of accounts receivable (TRAR) showed a strong correlation with the size (SIZE) of approximately 0.74. As expected, there was a positive, albeit weak, correlation between the performance (ROA) and the investment in R&D (ID) of around 0.10, although it was not statistically significant (the null hypothesis that the correlation is zero was accepted, since the p-value was greater than 0.1). It was found that the variables LIQ, RD, ITR, and TRAR showed a correlation with the performance (ROA) that was not statistically significant, so five additional models were also considered by introducing them individually and incrementally in order to test their impact on the dependent variable. As already mentioned, if the correlation between two variables is negative, this means that they are negatively correlated, that is, when one variable increases its value, the other decreases, as was the case of the SIZE variable and ROA variable, whose correlation (of around −0.15) was statistically significant at least at the 5% level, albeit in opposite directions. The observed negative correlation between the size and ROA variables could be due to several factors, including (1) large firms being in a growth phase with high investments, (2) potential bureaucratic inefficiencies, (3) diversification into non-profitable ventures, or (4) industry-specific dynamics.

4.2. Estimation

Table 5. System GMM regression results (all variables are defined in Table 1).

Explanatory Variables	Dependent Variable: ROA
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Lag(ROA, 1)	0.194 (0.191)	0.207 (0.181)	0.250 (0.178)	0.477 ** (0.234)	0.513 ** (0.254)	0.471 ** (0.220)
SIZE	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)	0.0003 (0.001)	−0.011 (0.009)
LIQ		0.004 (0.008)	0.002 (0.009)	0.003 (0.009)	0.004 (0.010)	0.004 (0.013)
LEV	−0.0003 * (0.0002)	−0.0004 ** (0.0002)	−0.0004 ** (0.0002)	−0.0002 (0.002)	−0.0002 (0.0002)	−0.0002 (0.0001)
Lag(RD, 0:2)0			−0.136 (0.169)	0.023 (0.284)	0.116 (0.333)	0.156 (0.309)
Lag(RD, 0:2)1			0.071 (0.181)	0.103 (0.294)	0.028 (0.224)	−0.026 (0.206)
Lag(RD, 0:2)2			−0.001 (0.112)	−0.110 (0.215)	−0.099 (0.221)	−0.191 (0.215)
ATR	0.016 * (0.009)	0.016 * (0.009)	0.015 * (0.009)	0.008 (0.013)	0.011 (0.014)	0.006 (0.022)
ITR				−0.0004 (0.001)	−0.0005 (0.001)	−0.0003 (0.0004)
TRAR	0.006 ** (0.003)	0.005 ** (0.003)	0.006 ** (0.003)	0.003 (0.005)	0.002 (0.006)	0.002 (0.005)
TRAP					0.0002 (0.0004)	0.001 * (0.0003)
IT1						0.236 (0.176)
IT2						0.228 (0.181)
IT3						0.259 (0.165)
IT4						0.217 (0.191)
No. of obs.	200	200	200	200	200	200

Note: Robust standard errors are in parentheses. Significance level: * p < 0.1, ** p < 0.05.

shows the results of the system GMM regression considering six variations of the model in Equation (1) according to Pearson’s correlation matrix. The variable of the intensity of investment in R&D on the current period and its two lags—the variable ITR, the variable TRAR, and the variable LIQ—were introduced individually. The results in Table 5 show that leverage had a negative effect on the ROA, as in the study by Ayaydin and Karaaslan (2014), demonstrating that companies prefer internal funds when making capital structure decisions. When the ITR and TRAR variables were both introduced, the LEV variable was no longer statistically significant. In terms of operational effectiveness, the ATR and TRAR variables were statistically significant to explain operational profitability (ROA) at least at the 10% and 5% levels, respectively, when the TRAP and ITR variables were not included. On the other hand, inventory rotation (ITR) had a negative impact on profitability, which contradicts the results obtained by Ayaydin and Karaaslan (2014). However, as previously mentioned, inventory turnover had a ratio of 7.12. On the one hand, this ratio is a reasonable value, but on the other hand, the estimation data show a negative impact on the companies’ performance, which is unusual, since the higher this ratio is, more is produced or sold, and therefore, the impact on profitability should be positive (Clausen and Hirth 2016). Regarding the previous year’s profitability (Lag(ROA, 1)), it was statistically significance at least at the 5% level, thus demonstrating a positive impact on current year profitability for models 4, 5, and 6 (Anagnostopoulou 2008). The size variable (SIZE) had a negative impact on the firm’s performance in model 6 with the inclusion of all the variables under study. This result was not expected, since the sample mostly considered big companies, which normally have advantages in terms of investment in R&D (compared with smaller companies), and it is possible to increase the degree of innovation (Muhammad et al. 2024).

Concerning technological intensity, it proved to have a positive impact on companies’ performance, showing that profitability is influenced by the level of technological intensity (He and Estébanez 2023). The results in Table 5 also show that investment in R&D was not statistically significant, with the lags for the previous two years having a negative impact on the performance of the companies under study. This occurs because investment in R&D implies an initial cost and consequently an associated risk, and therefore, only after overcoming this initial risk it will be possible to observe the positive return on investment (Rodrigues et al. 2020). Additionally, and from a more commercial perspective, it is also necessary that innovative products resulting from this investment be accepted in the market so that this initial recovery is possible (Ibhagui 2019). In conclusion, it is possible to state that until this acceptance occurs, investment in R&D will be nothing more than a cost borne for companies, and therefore, the negative impact of this variable on companies’ performance may be related to the recognition of this cost.

5. Final Remarks

R&D has been gaining importance in the Portuguese business world, and the main objective of this work was to analyze the impact of this investment on the performance of Portuguese companies. Regarding the accounting treatment of R&D expenses, the state-of-the-art approaches indicate that there is still no conceptual framework for the treatment of this type of item in order to make it as realistic as possible in terms of value for the company and consequently for its stakeholders.

In fact, the results obtained were somewhat unexpected, since the intention was to demonstrate that, for this type of company, the return on assets is partly explained by the investment in R&D and that there is a relationship between these two variables. However, the results obtained show that most of the variables were not statistically significant to explain the corporate’s performance. It is possible to conclude that R&D investment does not explain the evolution of corporate performance, at least in the short term, that is, this investment, which is initially seen as a cost, may only later prove to be a contributing factor to financial performance. In general, the variables that may contribute (even some of them minimally) to financial performance are the financial performance one-period lagged, the leverage, the assets turnover rate, and the turnover rate of accounts payable. Finally, it should be mentioned that the contribution of R&D investment to return on assets was not fully perceptible in this study, but it can be with the analysis of other types of parameters, such as the number of patents for example. The main limitation of this study was the lack of data, since not all companies in the sample provided information for all years and for all variables, making the study difficult to carry out and somewhat limited the robustness of the results obtained.

To gain a more comprehensive understanding, future research could explore additional performance metrics beyond the ROA, such as decomposing the ROA into profit margin and asset turnover, employing the return on investment (ROI), and incorporating market-based measures to assess the impact of R&D on firm performance in various settings.

We suggest applying the model studied for the post-COVID-19 pandemic years so that it can be possible to recognize its impacts on the investment of intangible assets in various sectors and more precisely in the Portuguese business environment. On the other hand, for future research, we suggest that the impact of investment in R&D on companies’ performance could be assessed considering the number of patents or the number of innovative products that companies have been developing, since the existence of a weak relationship between the investment in R&D and the improvement of financial performance of the companies under study was clear.