Schnurr, AlexanderWoerner, Jeannette H. C.2010-12-012010-12-012010-12-01http://hdl.handle.net/2003/2751910.17877/DE290R-15648In this paper we introduce the well-balanced Lévy driven Ornstein-Uhlenbeck process as a moving average process of the form X_t=integral(exp(-lambda*|t-u|)dL_u). In contrast to Lévy driven Ornstein-Uhlenbeck processes the well-balanced form possesses continuous sample paths and an autocorrelation function which is decreasing more slowly. Furthermore, depending on the size of lambda it allows both for positive and negative correlation of increments. As Ornstein-Uhlenbeck processes X_t is a stationary process starting at X_0=integral(exp(-lambda*u)dL_u). However, by taking a difference kernel we can construct a process with stationary increments starting at zero, which possesses the same correlation structure.enDiscussion Paper / SFB 823 ; 47/2010AutocorrelationFinancial modellingInfinitely divisible distributionLèvy processOrnstein-Uhlenbeck processSemimartingal310330620working paper