How do microkernel-based operating systems enhance security compared to monolithic kernels?

In the realm of operating systems, two primary architectures are often discussed: monolithic kernels and microkernels. Monolithic kernels incorporate various services such as drivers and system calls into a single large block of code running in a single address space. In contrast, microkernels aim to run minimal code in kernel mode, relegating most services to user space. This architectural divergence raises intriguing questions about security implications. Supporters of microkernels argue that isolating services can enhance system stability and security by minimizing the attack surface and allowing for easier updates and maintenance. On the other hand, monolithic kernels are traditionally faster due to reduced context switching overhead. In this context, how do microkernel architectures specifically bolster security, and are there trade-offs in performance or complexity that might offset these benefits?