Teknolojinin işsizlik üzerindeki etkisi

One of the most important aims of scientific research is to develop methods and tools for the most beneficial use of human labor. Scientific research put forward for this purpose and transformed into technological outputs over time lead to structural changes in existing jobs and cause some professions to disappear completely or to be replaced by new ones. This situation brings us face to face with a socio-economic problem, unemployment, which has existed since the use of labor as a production factor, and its effect is increasing day by day despite all the social and economic precautions taken by the states. The relationship between technology and employment is a longstanding and highly controversial issue. Although the labor-saving technologies used in the manufacturing sector, in particular, with the Industrial Revolution, kindled the concerns of technological unemployment in the economics literature and focused attention on this issue, the extensive literature in this field has not yet reached a definite conclusion. Based on this conflict, this thesis aimed to examine the relationship between technology and unemployment at the macro level, focusing on both developed and developing countries separately, as well as taking into account different sectors. Using a large data set covering the period 1990-2019, 21 developed and 23 developing countries were discussed separately in the first stage, and in the last stage, three main sectors for 30 developed and developing countries were focused on. In the studies conducted for developed and developing countries, the total unemployment rate was used as the dependent variable in the first place, and the employment rates were defined as the dependent variable in the analyses based on the skill groups. In the sectoral level studies, the dependent variables are the employment rates of the relevant sector. In addition, since the main interest of the study is the effect of technology on employment, per capita gross domestic R&D expenditures in developed and developing country investigations and per capita business enterprise R&D expenditures in sectoral analyzes were defined as the main technology proxy. Alternatively, although the main theme of the thesis is on the technology-unemployment relationship, macro variables such as per capita income, inflation, worker productivity, real effective exchange rate index, real interest rate, trade openness, public unemployment expenditures, and trade union density were also used in empirical models. Therefore, the relationship between employment and many important macro indicators was also questioned. In line with the above-mentioned purposes, the empirical process followed in the study consists of three stages. First, unit root tests of the series in the models that are considered to show the technology-unemployment relationship were performed, and it was shown that the first differences of all series were stationary. In the second stage, the long-term relationship of the series was proven by both Pedroni cointegration tests (Pedroni, 1999, 2004) and the Kao cointegration test (Kao, 1999). In the last stage, the coefficient estimates of the relevant models were made using the Panel Dynamic Ordinary Least Squares (PDOLS) proposed by Pedroni (2001a), Kao and Chiang (2001) and Mark and Sul (2003), and Fully Modified Least Squares (FMOLS) developed by Philips and Moon (1999), Pedroni (2001b), Kao and Chiang (2001). The first empirical analysis on developed economies shows that as R&D expenditures increase, the employment level of high-skilled workers increases, while the opposite is true for low-skilled workers. Thus, in advanced economies, technology has a complementary effect on high skilled workers. However, it can be indicated that the destructive effect of technology outweighs the complementary effect for the low-skilled group. In addition, the estimated coefficient of R&D expenditure in the overall models for the total unemployment dependent variable was positive and significant, providing evidence that higher R&D does indeed have a negative impact on unemployment. The possible reason for this result may be that the density of skilled workers in developed countries is still less than the share of unskilled or low-skilled workers. The impact of technology on overall unemployment could be reversed in the future if the share of low-skilled workers in the workforce declines and the employment structure becomes more skill-intensive and/or low-skilled workers become better equipped and adapt to new technologies. In addition to the findings on the technology-unemployment relationship for developed countries, among other important macro variables, the employment effect of per capita income was found to be positive, while inflation was negative. It was determined that factors such as real interest rate, public unemployment expenditures, and trade union density also increase unemployment. The effects of these macroeconomic indicators on employment by skills are heterogeneous, excluding per capita income. It has been determined that the increase in per capita income has a positive effect on the employment of both skill groups. The second main finding of the thesis is on developing countries. In the literature on the technology-employment relationship, studies that touch on developing countries are extremely limited. Developing countries have been neglected for a long time because the economics literature concerning the relationship between technology and unemployment, especially since the Industrial Revolution, emerged in developed countries, and the machines also first appeared in developed ones. So naturally, employment concerns first started in these geographies. In order to fill the literature gap on developing countries, in this study, the effect of technology on employment was discussed in a wide country group and time period, specific to developing countries. According to the results, the general unemployment effect of technology in developing countries is in the direction of labor-saving, as in developed countries and the effect sizes are close to each other. While both skill groups appear to be negatively impacted by technology in developing countries, the negative impact on high-skilled employment is relatively low. In addition, low-skilled employment in developing countries is less affected by technology than in developed countries. One of the possible reasons for this is that the relative wages of workers in these countries are low and therefore the comparative advantage of labor persists for a longer period. Then again, the employment effect of per capita income is positive in developing countries as well as in developed countries. The increase in per capita income increases the employment level of both skill groups. The effect of inflation, moreover, is heterogeneous by skill groups but negative on overall employment. The relationship between unemployment and interest rate as well as real effective exchange rate index in developing countries is not strongly significant. In addition, it was determined that trade openness has a strong positive effect on high-skilled employment, but its effect on low-skilled employment is also positive but weak. Finally, it should be noted that there are notable differences between the findings of developed and developing countries. Therefore, both the effect of technology on unemployment and the movement of unemployment with other macro indicators are in a heterogeneous structure among country groups. The third important finding of the thesis is on the effects of technology on employment in different sectors. The impact of innovation on employment is expected to differ between sectors, and the findings of the thesis support this hypothesis. The results show that the technology produced in agriculture and industry reduces employment in agriculture and industry. The destructive effect of technology in the agricultural sector cannot be suppressed even through income, and the demand, which is inelastic due to the nature of the sector or less flexible than in the past, cannot create additional employment areas where manpower will be used. On the other hand, economic growth is an important compensation mechanism for the industry, but it does not completely eliminate the labor-saving effect of technology, but partially alleviates it. This is true for developed and some developing countries covered in the study. Finally, the effect of technology on employment in the service sector has been identified as positive. In other words, R&D expenditures in the service sector increase employment within the sector. The results obtained in this chapter also indirectly support the routine/task-oriented technological change and job polarization hypotheses. Keywords: Technology and Unemployment Relationship, Technological Unemployment, Skill Groups, R&D Expenditure, Skill-Biased Technological Change (SBTC), Routine-Biased Technological Change (RBTC), Job Polarization, Panel Cointegration, Pedroni Cointegration, PDOLS, FMOLS.