1: \begin{abstract}

2: Using Riemann--Stieltjes methods for integrators of bounded

3: $p$-variation we define a pathwise integral driven by a

4: fractional L\'evy process (FLP). To explicitly solve general fractional

5: stochastic differential equations

6: (SDEs) we introduce an Ornstein--Uhlenbeck model by a stochastic

7: integral representation, where the driving

8: stochastic process is an FLP. To achieve the convergence of improper

9: integrals, the long-time behavior of FLPs

10: is derived. This is sufficient to define the fractional L\'evy--Ornstein--Uhlenbeck process (FLOUP) pathwise as an

11: improper Riemann--Stieltjes integral. We show further that the FLOUP is

12: the unique stationary solution of the

13: corresponding Langevin equation. Furthermore, we calculate the

14: autocovariance function and prove that

15: its increments exhibit long-range dependence.

16: Exploiting the Langevin equation, we consider SDEs driven by FLPs of

17: bounded $p$-variation for $p<2$ and

18: construct solutions using the corresponding FLOUP. Finally, we consider

19: examples of such SDEs, including

20: various state space transforms of the FLOUP and also fractional L\'evy-driven Cox--Ingersoll--Ross (CIR) models.

21: \end{abstract}