3.1. Methodology
Testing users’ perception of the influence of ERP systems on return on assets (ROA) and on productivity was carried out with a questionnaire distributed in the online environment (social media and email). The respondents were asked to evaluate the influence of ERP implementation on ROA and productivity, using a (1–5) point Likert scale (1, strong disaccord; 2, disaccord; 3, neutral; 4, accord; and 5, strong accord).
The impact of ERP implementation on companies’ performance expressed by return on assets (ROA) and productivity was estimated by the following regression equation, using the Ordinary Least Squares regression:
The performance indicator of the company (i) in the year (t), the dependent variable, was expressed by return on assets (ROA) and productivity. Return on assets was measured as the ratio of net profit and total assets of the company, and productivity was measured as the ratio of revenue to employees. The ERP variable indicated that a certain company (i) had implemented ERP during the analyzed period. This variable had the value 1 for companies that had implemented ERP, and the value 0 for companies that had not implemented ERP. As control variables, we used the following variables: turnover, fixed assets, equity, revenues, field of activity, listed on BVB, and number of employees.
By the coefficient β1 we indicated the impact of adopting an ERP system. If the value of this coefficient was positive, this meant an increase in performance determined by the ERP implementation. If the value of this coefficient was negative, this meant a decrease in performance for companies that had adopted an ERP system.
In the regression equation, the Time variable had the value 1 in the post-implementation years, in the case of companies that had implemented ERP. The Time variable had the value 0 both for companies that had not implemented ERP, and for the period before implementation in the case of companies that had implemented ERP.
The coefficient β1 measured the overall impact of ERP, by comparing companies that had implemented ERP with companies that had not implemented ERP. The β2 coefficient measured the impact of ERP adoption over time, comparing the period before ERP adoption with the period after ERP adoption, only in the case of companies that had adopted ERP.
Hypothesis 1 (H1). ERP implementation produces an increase in company performance, expressed by ROA.
Hypothesis 2 (H2). The implementation of ERP produces an increase in the performance of companies, expressed by productivity.
3.2. Data
We collected information on the implementation of ERP systems by companies, the type of modules implemented within ERP, the date of ERP implementation, and data on respondents’ perception of the benefits of ERP implementation in companies using a questionnaire distributed in the online environment (LinkedIn, e-mail and other social media networks). The questionnaire was sent to 500 companies from Romania and 406 companies filled out the questionnaires. Furthermore, we merged the dataset on ERP implementation with the financial data from the ORBIS database for the period 1999–2020.
The final database included 397 companies from Romania, out of which 267 companies implemented ERP systems. The minimum number of modules implemented by companies was 1 and the maximum number of modules implemented was 9. Most companies implemented the Accounting module (87.64%), Human Resources module (83.15%), CRM (79.03%), Acquisitions module (59.93%), or Financial module (50.19%). Fewer companies implemented the Production module (26.97%), Project Planning module (24.72%), Registry module (23.60%) and Business Intelligence module (19.85%).
Out of the total number of the companies that implemented an ERP system, 64.79% of them, that is 173 companies, implemented at least the Accounting, Human Resources and CRM modules. This feature could lead us to the conclusion that these companies implemented ERP from an administrative point of view, due to the legislative regulations in Romania that impose the need to operate tax statements in a system that processes them automatically, and were less focused on an eventual increase in performance.
Most of the companies for whom the questionnaire was filled out were privately owned companies (93.45%), and of these, 242 companies had implemented an ERP system and 129 companies had not implemented such a system. As for the publicly owned companies (35 companies), only one company did not implement ERP. The significantly higher percentage of implementation for the public sector can also be explained by the high cost of ERP implementation and by the fact that these publicly owned companies are larger and have easier access to finance.
The answers to the questionnaires were obtained from companies mostly with up to 500 employees (56.68% of the total number of companies analyzed had fewer than 50 employees and 15.37% of the companies fell within the range of 151–500 employees). Most of the companies that did not implement an ERP system had fewer than 50 employees and all the companies with over 150 employees implemented ERP. It is possible that the high cost of ERP implementation was the cause of a lower proportion of implementation by companies with fewer than 150 employees. With regard to the companies that implemented ERP, most of them were companies with revenues greater than two million euros and they recorded an average increase in revenue in the post-implementation period of approximately 20%, compared to the year of ERP implementation, and an increase in revenue of approximately 28% compared to the pre-implementation years.
Regarding the sector of activity of the companies, the highest percentage of companies that implemented ERP were companies belonging to the Production and Trade sectors (over 80%). The situation can be explained by the fact that lately online commerce has developed considerably, and having an ERP (CRM) system is necessary for recording and organizing data. At the same time, the companies belonging to the field of Production were large companies that could bear the costs of an ERP implementation.
To study the influence of ERP implementation on ROA and productivity we used a database containing information for 397 companies for the period 1999–2020. The descriptive statistics of the variables are shown in
Table 1.
Univariate statistical analysis using the
t-test (presented in
Table 2) revealed that companies that had implemented an ERP system registered higher values for both performance indicators and the difference was statistically significant. These results mean that, on average, companies that implemented an ERP system recorded a higher performance compared to companies that did not implement an ERP system for the period analyzed.
Regarding the respondents’ perception of the influence of ERP implementation on the companies’ performance, we obtained the following results:
Hypothesis 1 was not confirmed, according to the perception of the 406 respondents (both those who implemented ERP and those who did not implement ERP) (the weighted average score was 2.85-neutral). The perception of the majority of these respondents, totaling 63.05% of the answers, was neutral regarding the assumed hypothesis (
Table 3).
The same results were confirmed by analyzing only the responses of those who implemented ERP and whose perception should be more relevant. The perception of the majority of these respondents, totaling 49.81% of the responses, was neutral regarding the assumed hypothesis (
Table 4) (the weighted average score was 3.02-neutral). Therefore, from the perception of all respondents, Hypothesis 1-The implementation of ERP systems causes an increase in the company’s return on assets (ROA) was not confirmed.
Hypothesis 2 was not confirmed, according to the perception of the 406 respondents (both those who implemented ERP and those who did not implement ERP). The perception of the majority of these respondents, totaling 51.97% of the responses, was neutral regarding the assumed hypothesis (
Table 5). However, analyzing only the responses of those who implemented ERP and who were aware of the cause, the results were the same. The perception of the majority of these respondents, totaling 43.77% of the answers, was favorable toward the assumed hypothesis (
Table 6), but the weighted average score was 3.24 (neutral).