These days, parallel processing is widely used and quite significant. In parallel computing, the synchronization problem is a crucial issue. Locks are now utilized in parallel processing to address synchronization issues. However, locks have a number of problems. Another method for resolving synchronization issues that does not have the disadvantages of locks is software transactional memory, or STM. The use of Software Transactional Memory (STM) to forecast GDP and unemployment rates has been investigated in this paper. Time-series data-based predictive models are usually intricate and computationally costly. The use of STM in predictive modelling may increase the effectiveness and precision of GDP and unemployment forecasts. STM is a concurrency control method that offers a means of managing shared memory in multi-threaded systems. The integration of STM into forecasting models, the difficulties it solves, and the possible advantages it presents in the context of economic forecasting have all been covered in this study. STM is superior to locks since its performance is on par with locks and it is not affected by their shortcomings. Locks take a pessimistic approach. However, STM is superior since it employs an optimistic approach. Additionally, codes with locks can be changed to codes with STM simply by substituting STM calls for lock calls. Thus, a major benefit of STM is that it may be utilized with legacy code.