Author: Jaeyoung Seo(Bank of Korea), Dongweon Lee(Bank of Korea), Yongsung Chang(Bank of Korea), Jay H. Hong(Seoul National University)

<Abstract>

The official measures of employment and average working hours from the Economically Active Population Survey, commonly used as key employment indicators, may have limitations in accurately reflecting the actual labor input in production. To address this, alternative employment indicators are proposed. First, the Economically Active Population Survey classifies “temporary leave due to business downturn or suspension” as employed. Excluding these temporarily inactive workers from the employed category (as done in the United States) increases the correlation coefficient between employment and GDP from 0.7 to 0.83 for the period following the COVID-19 pandemic, when production activities were severely disrupted. This re-classification also increases the elasticity of employment to GDP (the so-called Okun's coefficient) by 63%, from 0.6 to 0.98. Second, the official measure of average monthly working hours are calculated based on the hours worked in the week containing the 15th day of each month. This method introduces significant irregularities when the week coincides with lunar holidays, substitute holidays, elections, etc. Adjusting for these irregularities using interpolation techniques increases the correlation coefficient between total working hours and GDP from 0.53 to 0.69 compared to the official statistics.