Maybe better say 'zoo' vs 'forest', or 'very well protected area' vs 'partly protected area'.
If there is only a few habited planets inside grabby alien sphere, they will be very valuable and very well protected so no UFOs will be observed. 

If there are millions of them, they are less valuable and thus less protected and therefore can be used for some practical activity, like turism, hunting or mining unobtanium. Obviously, if UFOs are aliens, local alien authorities let them be visible sometimes, so local aliens laws are not very strict. 

Observation selection effects like SIA favors the hypothesis that there millions habitable planets inside any grabby aliens. 

I think that your model is correct and 'anthropically' supported. 

In some sense it favors 'zoo hypothesis". However, there is an important distinction: is it zoo or natural reserve. In general, on Earth zoos are rare but well kept and natural reserves are more abundant, but less controlled. The same anthropic considerations which favor silent rulers, favor natural reserves vs zoos.

This has bad consequences for us: natural reserves are more likely to be visited by unauthorized visitors and poachers. Or if we will be less anthrophomorphisng, they have less value for cosmic rulers as they are more numerous. UFOs observations and their alleged connections with cattle mutilations and abductions are more favoring the idea that Earth is less protected hunting ground than well protected zoo. 

Speaking about "unobservable" part. Aliens which consist of fields or using something like "5-th" dimension to travel will have much less visible footprint but will have much large sphere which they can grab, as they can travel with almost light speed. The same anthropic considerations favor such aliens as they will have larger sphere of influence. Observations of UFOs also imply that they use some non-typical for us way of propulsion, like instant acceleration, manipulating gravity and moving through objects. 

In other words, if aliens are not interesting in building Dyson spheres and can travel with near-light speed without leaving visible traces, we will see much less signs of their activity. Maybe they are more interested in controlling space than in performing a lot of computations. 

The conclusion is unpleasant: we are typical and neglected planet which sometimes is abused by our mostly invisible rulers. But it is the same as life situation of most people on Earth. 

I think that UFOs are really a wildcard in x-risk research. 
1.Even if UFOs don't have any serious substance behind them, the fact that many serious military people and even presidents believed in them, should update our prior about human irrationality and therefore increase our expectation that nuclear risks and AI risks will be mismanaged.

2.If UFOs have an interesting, but not world-model-shattering explanation, e.g. they are a form of ball lightings, this opens a possibility of creating new weapons after their nature will be learned. 

3. If their explanation is world-model-shattering, all our expectations about x-risks are wrong. 
World-model-shattering explanations are not only classical aliens. 

They can be:
-glitches and viruses in the matrix; 

-space-faring animals consisting from exotic fields; 

-clouds of nanobots from already extinct civilizations. 

-malfunctioning berserks-robots

-elaborated lies from cold war era similar to other memetic mind attacks like "red mercury" or number radio stations.

It also happens in personal life.

Critics: "‘Long Reflection’ Is Crazy Bad Idea" https://www.overcomingbias.com/2021/10/long-reflection-is-crazy-bad-idea.html 

Actually, I am going to write someday a short post "time machine as existential risk".

 Technically, any time travel is possible only if timeline is branching, but it is ok in quantum multiverse. However, some changes in the past will be invariants: they will not change the future in the way that causes ground father paradox. Such invariants will be loopholes and have very high measure. UFO could be such invariants and this explains their strangeness: only strange thing are not changing future ti prevent their own existence. 

Thanks for these details!

The report doesn't mention any likelihood of any of these events happening

May be you discuss a different report than I read. The one I read says: 

In fact there is evidence of eruptions at Kivu in about 1000 year cycles and predictions based on observed accumulation rates (10-14% per year) suggest an eruption in the next 100-200 years, see Report [1].