Using panel data covering 30 urban areas for 16 years, this is the first comprehensive study which investigates the rental adjustment process in the German office market. In doing so, we use the recently developed panel cointegration technique for non-stationary data in conjunction with an error correction model (ECM). The application of this approach enables us to overcome the data limitations, particularly existent for the German real estate market. The primary motivation is to provide useful insights into the long-term relationships and short-run dynamics of real office prime rents. To this end, we estimate a reduced-form equilibrium model, in which office employment is defined as proxy for demand of space, and stock data as well as the vacancy rate are included as supply drivers, respectively. The empirical evidence suggests that a one percent rise in office employment increases real rents on average by 2.23% through higher demand of office space. On the other hand, a one percent increase in the supply of office space measured by office stock decreases real rents by 1.71%. Furthermore, due to a decrease in the occupancy rate and an increase in the stock variable at the same time, real rents decrease by 2.17% much stronger. Finally, the results of the error correction model show that deviations from the long-run equilibrium lead to an adjustment process which restores equilibrium within two years.