Is Jobless Growth Valid in Turkiye? A Sectoral Analysis of the Relationship between Unemployment and Economic Growth

This study analyzes the validity of jobless growth in Turkiye on sectoral basis. It analyzes the impacts of agriculture, industry, construction and services sectors on unemployment using annual data for the period 2000-2022. ARDL method is applied within the scope of the analysis. The findings are tested with FMOLS and CCR methods. The results show that growth in all sectors reduces the unemployment. A one-unit increase in the share of agriculture sector in GDP decreases the unemployment rate by 0.471 points, 0.680 points in the industrial sector, 0.899 points in the construction sector and 1.383 points in the services sector in the short-run. The long-run coefficients reveal that the impacts of sectoral growth on unemployment are stronger in the long-run than in the short-run. A one unit increase in the share of the agricultural sector in GDP decreases the unemployment rate by 2.380 points, 4.057 points in the industrial sector, 1.761 points in the construction sector and 3.664 points in the services sector in the long-run. These findings show that jobless growth is not valid in Turkiye in general. On the contrary, economic growth plays an important role in reducing unemployment.

💡 Research Summary

The paper investigates whether “jobless growth” – a situation where economic expansion fails to reduce unemployment – is a valid concept for Turkey, using a sector‑specific approach. Annual data from 2000 to 2022 are employed, with the unemployment rate as the dependent variable and the shares of agriculture, industry, construction and services in gross domestic product (GDP) as independent variables. The authors apply the Autoregressive Distributed Lag (ARDL) methodology to capture both short‑run and long‑run dynamics, acknowledging that the variables are a mixture of I(0) and I(1) processes. Bounds testing confirms the existence of a long‑run equilibrium relationship among the variables.

Short‑run ARDL estimates reveal that a one‑percentage‑point increase in the sectoral share reduces the unemployment rate by 0.471 points for agriculture, 0.680 points for industry, 0.899 points for construction, and 1.383 points for services. Long‑run coefficients are substantially larger in magnitude, indicating stronger persistent effects: the same one‑point increase lowers unemployment by 2.380 points (agriculture), 4.057 points (industry), 1.761 points (construction) and 3.664 points (services). These results demonstrate that growth in each of the four sectors contributes to job creation, with industry and services showing the most pronounced long‑run impact.

To verify robustness, the authors re‑estimate the long‑run relationships using Fully Modified Ordinary Least Squares (FMOLS) and Canonical Cointegrating Regression (CCR). Both alternative estimators produce coefficient signs and significance levels consistent with the ARDL findings, reinforcing confidence in the estimated relationships and mitigating concerns about endogeneity, serial correlation, or heteroskedasticity.

The literature review situates the study within the broader discourse on Okun’s Law, Verdoorn’s Law, and the global phenomenon of jobless growth, highlighting that sectoral composition and technological change critically shape the growth‑employment nexus. Prior Turkish studies are noted for focusing mainly on aggregate growth, with limited sectoral disaggregation. This paper fills that gap by providing a detailed, sector‑by‑sector analysis.

Policy implications are drawn directly from the empirical evidence. Since sectoral expansion clearly lowers unemployment, Turkish policymakers should design growth strategies that explicitly target job‑creating sectors. The services sector, which exhibits the largest short‑run effect, and the industrial sector, with the strongest long‑run impact, merit priority in terms of investment incentives, skill‑development programs, and technology adoption policies that are labor‑intensive rather than purely capital‑intensive. The agriculture sector, while showing a smaller magnitude, remains important for rural employment and should be supported through productivity‑enhancing measures that do not displace labor.

The authors acknowledge several limitations. The use of annual data may smooth over short‑term shocks such as the 2018‑2020 pandemic, and sectoral GDP shares do not directly capture labor intensity or productivity changes. Potential bidirectional causality between growth and unemployment, although partially addressed by the ARDL framework, could be explored further with instrumental variable techniques or dynamic panel models (e.g., System GMM).

In conclusion, the study provides strong empirical evidence that “jobless growth” does not characterize the Turkish economy; rather, economic expansion across all major sectors contributes significantly to reducing unemployment, especially in the long run. The methodological triangulation (ARDL, FMOLS, CCR) enhances the credibility of the results, and the sectoral focus offers actionable insights for designing growth policies that are genuinely job‑creating. Future research could extend the analysis with higher‑frequency data, incorporate sector‑specific labor‑productivity measures, and examine the role of external factors such as foreign direct investment and trade openness in shaping the growth‑employment relationship.