1: \begin{abstract}

2:   We establish conditions under which Metropolis-Hastings

3:   (MH) algorithms with a position-dependent proposal covariance matrix

4:   will or will not have the geometric rate of convergence. Some of the

5:   diffusions based MH algorithms like the Metropolis adjusted Langevin

6:   algorithm (MALA) and the pre-conditioned MALA (PCMALA) have a

7:   position-independent proposal variance. Whereas, for other modern

8:   variants of MALA like the manifold MALA (MMALA) that adapt to the

9:   geometry of the target distributions, the proposal covariance matrix

10:   changes in every iteration. Thus, we provide conditions for

11:   geometric ergodicity of different variations of the Langevin

12:   algorithms. These results have important practical implications as

13:   these provide crucial justification for the use of asymptotically

14:   valid Monte Carlo standard errors for Markov chain based

15:   estimates. The general conditions are verified in the context of

16:   conditional simulation from the two most popular generalized linear

17:   mixed models (GLMMs), namely the binomial GLMM with the logit link

18:   and the Poisson GLMM with the log link. Empirical comparison in the

19:   framework of some spatial GLMMs shows that the computationally less

20:   expensive PCMALA with an appropriately chosen pre-conditioning

21:   matrix may outperform the MMALA.

22: \end{abstract}

23: