1: \begin{abstract} 

2: 		We study the $L^1$-approximation of the log-Heston SDE at the terminal time point by arbitrary methods that use an equidistant discretization of the driving Brownian motion. We show that such methods can achieve at most order $  \min \{ \nu, \tfrac{1}{2} \}$, where $\nu$ is the Feller index of the underlying CIR process. As a consequence Euler-type schemes are optimal for  $\nu \geq 1$, since  they have convergence order $\tfrac{1}{2}-\varepsilon$ for $\varepsilon >0$ arbitrarily small in this regime.

3: 		\medskip

4: 		

5: 		\noindent \textsf{{\bf Key words: } \em

6: 			Heston model, CIR process, $L^1$-error, lower error bounds, optimal approximation, Euler-type methods

7: 			\\

8: 		}

9: 		\medskip

10: 		\medskip

11: 		\noindent{\small\bf 2010 Mathematics Subject Classification: 65C30;  60H35;  91G60 }

12: 		

13: 	\end{abstract}