The impact of incentives and subsidies on firms' innovation performance in Turkey

Abstract

Countries' economic growth and global competitiveness are largely dependent on their innovation capabilities. Innovation adds dynamism to economies through the development of new products, processes, and services, playing a critical role in achieving sustainable development goals. Companies can quickly adapt to market conditions and become competitive internationally through innovation. In this process, government incentives encourage investment in R&D and innovation activities, accelerating technological progress and enhancing economic growth potential. Understanding the effects of government support on firm innovation performance allows policymakers to design efficient and effective support mechanisms. Governments use various policy tools to encourage innovation, including direct financial support, tax reductions, grants, and subsidies. These supports are vital, especially for SMEs with limited access to financial resources. There is ongoing debate in academic and political circles about whether government incentives actually enhance firms' innovation capacities. The effectiveness of incentives can vary depending on the industry, company size, and technology intensity. Therefore, a thorough analysis of the real effects of government supports is crucial for implementing these policies more accurately and effectively. The aim of this thesis is to comprehensively examine the impact of government incentives and subsidies on firm innovation performance in Turkey. The research seeks to provide a clear understanding of the effectiveness of these policies by evaluating the effects of various incentive mechanisms on firms' innovative outputs. The lack of a full understanding of the outcomes and performance of incentives and subsidies constitutes the main problem area of this study. In this context, the thesis aims to offer recommendations to policymakers for making incentive policies in Turkey more strategic and targeted. By providing valuable insights that will enable the more effective design and implementation of support mechanisms, this study aims to contribute to both academic research and policy development. In this study, a comprehensive literature review was conducted, covering the impact of innovation on economic progress, various forms of government incentives, and methods for measuring innovation performance. The historical developments related to the research topic were examined in detail, and existing academic studies on the factors influencing innovation performance were analyzed. The literature review revealed that innovation performance is shaped by multiple variables and highlighted the importance of understanding the complex nature of these interactions. This extensive review serves as a foundation for developing the theoretical and methodological framework of the thesis. Based on the literature review, a model suitable for the research topic and dataset was developed. In this model, the presence of innovation in firms was considered as the dependent variable. Frequently mentioned firm characteristics in the literature were selected as control variables, while government incentives and subsidies were identified as the main independent variable. The dataset used in the analysis comes from the World Bank Enterprise Survey (WBES), which covers more than 150 countries. These surveys provide insights into various aspects of the business environment, including access to finance, corruption, infrastructure, competition, and firm performance factors that are not typically found in financial reports. The developed model was evaluated using the Logit regression analysis technique in Python. Model fit and statistical significance were examined using diagnostic measures such as Pseudo R-squared, Log-Likelihood, LL-Null, and the LLR p-value. The results indicate that the independent variables have a statistically significant effect on the dependent variable. According to the findings, the most significant impact on innovation performance comes from the ratio of R&D expenditures to total revenue. The effect of firm size is minimal. Subsidies and export rates have been observed to positively influence the probability of innovation, aligning with the existing literature. The effect of firm age, although very low, is positive and thus considered statistically significant. In the Logit regression model, calculating the actual probability requires considering the simultaneous effects of all variables. Therefore, the survey data of firms were applied to the model using mean and median values to estimate the probability of innovation in Turkey over the years, based on whether firms received subsidies. The calculations indicate an increase in innovation probability between 2002 and 2005, followed by a continuous decline after 2005. While outliers elevated overall performance in all years, the median-based calculation for 2019 shows that the innovation probability remained at only 5% for non-subsidized firms and 19% for subsidized firms. To test the developed model, sample firms were selected from one of Turkey's leading conglomerates, and their annual data for the selected variables were used to calculate innovation probabilities. The obtained probabilities were then compared with the results derived from the mean and median values of firms across Turkey. The test confirmed that the model's results align with actual data. By testing this method, it was assessed that the model could be used to evaluate a company applying for subsidies, estimate its probability of innovation before receiving support, and determine its position relative to the national average. For international comparisons, the 48 countries surveyed in the most recent 2019 survey were ranked based on their Gross Domestic Expenditure on R&D (GERD) as a percentage of GDP. According to this ranking, six countries with a higher GERD/GDP ratio than Turkey and seven countries with a lower ratio were selected. The model was applied to these selected countries, and the effects and consistency of all variables were analyzed. The marginal effects of the variables were calculated and compared. Subsequently, the model was run using the mean and median values of surveyed firms in each country, and the average innovation performance was determined based on whether firms received subsidies. These results were then compared with the findings for Turkey. In this study, the performance results of firms from 13 selected countries were compared. According to these results, Turkey has the lowest probability of innovation in the absence of subsidies. While subsidies have a positive effect in Turkey, the innovation performance increases only to 19% when subsidies are provided, indicating a limited impact of financial support. The country with the strongest effect of subsidies is Slovenia. In all countries, the model is statistically significant, and the effect of subsidies is positive. Similar to Turkey, the most influential variable on innovation performance in the selected countries has been the proportion of R&D expenditures. The effect of the number of employees is close to zero in all countries. The effect of firm age is negative in five countries, indicating that as firms grow older, their innovation performance decreases in these countries. Overall, the impact of firm age on innovation is minimal across all countries. The effect of exports is positive in all countries except for three. Limitations; The World Bank conducts Enterprise Surveys in over 150 countries. However, surveys are not conducted in the same year for all countries. This creates limitations in accessing data for specific years, such as 2019, for countries like the USA, China, OECD averages, Japan, and Korea, which constrains the scope of the research. While the literature highlights regional differences, dividing the dataset by regions results in an insufficient number of incentivized firms for meaningful analysis. Since some questions were not asked every year, they lacked continuity, and certain variables mentioned in the literature could not be included in the model.