The findings are significant because unemployment remains one of the most important indicators of regional economic health. Lower unemployment generally reflects stronger economic activity, higher household income, and broader employment opportunities, while rising unemployment often signals economic slowdown, declining investment, and increasing poverty risks.
According to the researchers, Pematangsiantar represents an important case study due to its strategic role as one of North Sumatra’s commercial centers and a gateway to the Lake Toba tourism area. The city's economy is largely driven by trade and services—sectors that are highly responsive to national and global economic changes. Consequently, unemployment levels have fluctuated considerably over the past decade.
Official data from Indonesia’s Central Statistics Agency (BPS) show that the city's Open Unemployment Rate experienced notable changes between 2017 and 2025. The unemployment rate increased sharply from 8.80% in 2017 to 12.14% in 2018, remained relatively high during the COVID-19 pandemic, and then gradually declined after economic recovery began. By 2025, the unemployment rate had fallen to 7.74%, indicating improving labor market conditions.
To analyze future trends, the researchers applied the Markov Chain method, a mathematical model widely used to estimate the probability of future events based on historical transition patterns. Unlike conventional forecasting methods that focus solely on numerical values, the Markov Chain model predicts how likely a system is to move from one condition to another over time.
In this study, annual unemployment data were categorized into two simple states: "Increase" and "Decrease." The researchers then calculated transition probabilities between these two states to estimate how unemployment is likely to evolve during the 2026–2030 period.
The analysis produced encouraging results. The probability that unemployment would continue declining after a previous decline reached 80%. Even more striking, whenever unemployment increased in the historical data, it was always followed by a decline in the subsequent period, resulting in a 100% transition probability from "Increase" to "Decrease."
Based on these transition probabilities, the model predicts that declining unemployment will remain the dominant trend over the next five years.
The projected probabilities are as follows:
- 2026: 80.0% probability of unemployment declining
- 2027: Approximately 84.0%
- 2028: Approximately 83.2%
- 2029: Approximately 83.36%
- 2030: Approximately 83.33%
Meanwhile, the probability of unemployment increasing remains relatively low, ranging between 16% and 20% throughout the forecast period.
The researchers also calculated the model's steady-state distribution, representing the long-term equilibrium if historical transition patterns continue. The results indicate that the labor market will stabilize with an 83.3% probability of remaining in a declining unemployment state, while the probability of experiencing increasing unemployment is only 16.7%.
To evaluate the robustness of the model, the research team conducted sensitivity analysis by modifying transition probabilities under several hypothetical scenarios. Although the probability of declining unemployment decreased when transition probabilities were adjusted, the "Decrease" state remained dominant across all simulations. This suggests that the Markov Chain model is relatively stable and reliable for describing labor market dynamics in Pematangsiantar.
The researchers further validated their calculations using the statistical software R. The software-generated results matched the manual calculations exactly, confirming that the mathematical model was correctly implemented and capable of producing consistent probability estimates.
According to the authors, the study demonstrates that Markov Chain analysis is an effective tool for forecasting unemployment dynamics and identifying long-term employment trends. Rather than predicting exact unemployment figures, the model estimates the likelihood of future changes, providing policymakers with valuable information for planning employment strategies.
The findings also suggest that Pematangsiantar's labor market has shown considerable resilience. Historical increases in unemployment have generally been temporary, with economic conditions returning to a downward unemployment trend in subsequent years. This pattern indicates that the local economy has been able to adapt to external shocks, including economic slowdowns and labor market disruptions.
However, the researchers caution that the projections should not be interpreted as absolute predictions. External factors—including national economic performance, inflation, industrial restructuring, investment levels, and mismatches between workforce skills and labor market demand—could still alter future unemployment patterns.
For local governments, the study provides a practical analytical framework for designing proactive labor market policies. The probability-based forecasts can help authorities anticipate employment challenges, strengthen workforce development programs, expand productive economic sectors, and improve job creation strategies before labor market conditions deteriorate.
The researchers also recommend expanding future studies by using longer historical datasets and incorporating additional variables such as economic growth, inflation, educational attainment, regional investment, and industrial development. Combining Markov Chain analysis with other forecasting techniques could further improve prediction accuracy and provide a more comprehensive understanding of labor market dynamics.
Author Profiles
Nilam Sari is a researcher at HKBP Nommensen University Pematangsiantar, specializing in applied statistics, probability modeling, and quantitative economic analysis.
Gayus Simarmata is a lecturer at HKBP Nommensen University Pematangsiantar whose expertise includes mathematics, statistics, stochastic processes, and mathematical applications in economics.
Yoel Octobe Purba is an academic at HKBP Nommensen University Pematangsiantar with research interests in applied mathematics, statistical modeling, and data-driven decision-making.
Research Source
Article Title: Application of Markov Chains in Predicting the Unemployment Rate in Pematangsiantar for the Years 2026–2030
Authors: Nilam Sari, Gayus Simarmata, Yoel Octobe Purba
Journal: Indonesian Journal of Advanced Research (IJAR)
Volume: 5, Issue 6 (2026)
DOI: https://doi.org/10.55927/ijar.v5i6.16579
0 Komentar