Should tenured researchers be "encouraged" (forced) to retire at a certain age?

Here in Germany, professors are public servants and automatically retire when reaching retirement age. This has the equally automatic consequence that they are not paid a wage out of the university budget any more but a pension. And pensions of public servants work the way that any money they earn is subtracted from the pension.
Thus retired professors are typically (as long as they don't mess up the working atmosphere of the department) a net gain: any lecture they hold is essentially for free, any exams they take (they keep the ability to act as examiner) save work for someone else in the department. OTOH, they typically have to move into some small office, and can keep lab space only if they have grant/3rd party money to pay for that.

The issue is not the age of these faculty members but that they do not carry their share of the workload and expectations of the department: Their sum of research, teaching, and service work does not match what one would expect from someone with the likely salary of someone who has been on the faculty for a long time.

The problem is then one of setting policies. If you don't publish, then you're clearly not matching the expectations for the position. There are then two options a department and faculty member have: initiate the procedure to revoke tenure; or change the workload assignment. We have done the latter in my former department with some old faculty members: They were simply assigned 4, 5, or 6 classes a year instead of the common 3, since they were not doing much research any more and were not overly active in service either. The policy that instituted this did not mention age but was simply based on research productivity. The net effect was two-fold: (i) a number of faculty decided to retire; (ii) the remainder now had a way in which they could actually contribute significantly to the mission of the department, given that their research productivity had waned a long time ago; this also allowed them to, for the first time in a long time, get good annual evaluations because these are based on the workload distribution assigned to them.

In other words, departments that do not address the problem typically do this because they want to "keep the peace" and not "rock the boat" or simply don't want to deal with the problem. But it's not true that there is no recourse -- it's not terribly difficult to design departmental policies that make sure everyone contributes in some way.

It seems to me that in all the examples you have given, the problematic issue is not age per se, but rather, the low level of productivity of the older academics you mention. It is also notable that the problem stems from the tenure system, where the older tenured academic is effectively immune to labour competition from new applicants. This is a form of 'closed shop' in labour economics, and it has various economic implications relating to productivity incentives and outcomes.

Since age is not the outcome of interest, it makes little sense to impose a policy that discriminates with respect to age. (Setting aside legal requirements, which might also prohibit this.) It seems to me to be more sensible that if a change to policies is made, to try to improve productivity and use of resources, that change ought to be either: (1) abolition of the tenure system, thereby freeing labour competition generally and freeing university resources; or (2) keeping the tenure system but raising the minimum productivity requirements imposed on those with tenure.

In regard to these two choices, there is already a substantial amount of economic literature that analyses labour competition under the tenure system (see e.g., Alchian 1953, Alchian 1968, McPherson and Winston 1983, McKensie 1996, Brown 1997, Winston 1999, McGee and Block 2008, Chen and Lee 2009, Brown 2017), which might be of interest in determining whether tenure is a good idea or not. In the first of these references, Alchian posits that non-profit tax-subsidised enterprises (such as universities) free administrators from the usual profit-loss constraints of the market, and therefore allow greater shirking, and greater arbitrariness in hiring and firing decisions. Alchian theorises that tenure acts as a limiting device to offset arbitrariness in hiring/firing decisions by administrators, and thus, there is a trade-off between reduced labour competition, and increased incentives for applicants to seek tenured positions. Brown (2017) tests this theory empirically using data from the 2006 and 2010 IPEDS surveys. He finds that tenure is more common in non-profit institutions than for-profit institutions, and that in private non-profit institutions tenure is positively related to endowment level and negatively related to reliance on tuition and fees.