Your class E would protect against flights in IMC crossing an approach at relatively low level. Since the traffic density in IMC at low level is very much lower than in VMC at low level, the risk of collision is small to start with. To the best of my knowledge, there have been no collisions in IMC in the UK for 30 years, probably much longer. Of those potential collisions with IFR flights, most of the transiting IFR flights would be talking to local ATC, whether there's class E there or not. Thus the class E would be very unlikely to be effective in actually reducing the probability of a collision.
The much greater comparative risk is a collision with a VFR flight crossing the approach track in VMC. The traffic density is much higher, and the high workload of an approach means that the ability to see-and-avoid a random VFR flight, which in a fast aircraft is very limited at the best of times, is low. And VFR flights are less likely to be in contact with ATC. But your class E does not prevent a VFR flight from flying right through the localiser at 3 miles.
If you need controlled airspace, you need at least class D airspace.
Class D is of course the much better option, and much of what you write is very valid. However, I think the larger issue is that presently many airfields operate with basically no genuine form of protection for IFR flights - None at all. There may have been no collisions in IMC in the UK for 30 years. That shouldn't mean we sit on our hands waiting for one. Class C seems to be the preferred choice internationally. In the UK, Class D seems to work well on the whole. But for airfields presently in Class G, I think Class E CTZs are probably as much as we could hope for in today's climate. The goalposts have moved a lot over the years.
There is a reluctance in the UK to give more of the FIR over to the world of civilian regulated airspace. Much more so now than say 30 years ago. I'm quite convinced that the number of aircraft movements and passenger figures at some of today's regional airports in class G would be be on a par if not higher than the figures which led to the introduction of Class D CTZs at many airfields decades ago. It is almost as if the FIR has been 'over fished' and if we give any more airspace over to the formation of new CTZs, then there wont be enough left for PPLs, Military or GA traffic.
I don't want to step on anybody's toes, and I understand the strength of feeling of those who might say, 'why should you have all the airspace?'
That is why Class E might be an answer.
I don't think class E would make a significant difference to the risk, and it would significantly increase the complexity of the airspace.
Class E could not possibly be as complex nor as confusing as the current system of 'controlling' in Class G. Neither could it possibly be as utterly daft as Class F (it only exists if you want to play) airspace. Too many airfields have outgrown the old system of separating those we can - when we can, provided they want separating. It's time to move on....starting with a new look at Class E airspace.
If CTZs were given away cheaply in the past, it is not an excuse for making Airspace changes well nigh impossible today. Any body looked at CAP725 by the way? Hmmmm
CAP 725 - Airspace Change Process Guidance