The key to understanding the relationship between magnification ratio and coverage percentage is to realize that the other variable is the size of the viewfinder. If one APS-C camera (such as the Canon 30D) gives a 95% view at 0.90x and another APS-C camera (such as the Canon 40D) gives a 95% view at 0.95x, it is because the viewfinder in the 40D is slightly larger than the one in the 30D. If the viewfinders were the same size, increasing the magnification from 0.90x to 0.95x would reduce the coverage from 95% to around 90%. That would be the equivalent to increasing the magnification of a negative during printing without increasing the size of the print: it will crop the image. If you want to print the entire image and you increase the magnification by a factor of 1.056x (0.95/0.90=1.056), then you must also increase the print size by the same factor. A 16x20 print would need to become a 16.88x21.11 to show the same view at the increased magnification.
The main reason more expensive DSLR cameras can have larger viewfinders and still be brighter at the same time is that they tend to be full frame cameras. Within the scope of SLR cameras, the size of the reflex mirror and focusing screen is determined by the size of the sensor. Just as a larger sensor can collect more light (in the same lighting conditions) from the same lens/aperture combination than a smaller sensor can, the larger mirror reflects more light towards the viewfinder than a smaller one does. Remember, the magnification is also affected by sensor/mirror size. Just as the image from a smaller sensor must be enlarged by a higher ratio to yield the same print size as an image from a larger sensor, a FF camera will use a lower viewfinder magnification to give the same percentage of coverage than an APS-C camera will if both cameras have the same sized viewfinder.
Another reason for the difference is that cheaper DSLRs tend to use hollow penta-mirrors (that are cheaper to manufacture) to rectify the upside down and reversed image coming from the lens. More expensive cameras use solid penta-prisms that transmit the light that enters them more efficiently.
Focusing screens also play a part because different focusing screen designs allow different amounts of light to pass through them. View screens that allow accurate DoF at very wide apertures are darker than view screens that cut off the DoF differences at wider apertures. Cameras designed to use interchangeable focusing screens usually have a menu option to tell the camera which screen is being used so that the camera can properly calibrate the results from the exposure meter that measures light after it has passed through the focusing screen. With the rise of auto focus systems in SLRs just over a quarter century ago, the emphasis slowly shifted to viewfinder brightness at the expense of accurate depth of field (DoF) representation when using wider lens apertures. At the dawn of the digital age little more than a decade ago the introduction of sensors smaller than the traditional 36x24mm film frame heavily contributed to the acceleration of that trend. With the smaller sensors/mirrors/focusing screens of 'crop sensor cameras', viewfinder brightness suddenly became a more pressing concern.
With the emergence of consumer level DSLRs, the prism has been forced to share the real estate traditionally given to it with the built-in flash. This is especially a concern with entry level cameras that place compactness and lightness high on the list of desirable attributes. If you want to add a flash without increasing the size of the top of the camera, you have to make the prism or pentamirror smaller. Most pro level cameras can have a larger prism because they don't include a built in flash and their buyers are usually more concerned with performance than with bulk and weight. The lenses they often use have already killed the compact/lightweight possibilities.
So what limits the magnification in the viewfinder?
From a purely theoretical standpoint, the size of the sensor/film and the size of the viewfinder do. The larger the apparent size (as seen by the user) of the viewfinder is, the more magnification can be used and still yield the same percentage of coverage of the Field of View (FoV) captured by the sensor or film. The larger the sensor size is, the less amount of magnification is used to provide the same coverage.
From a manufacturing/marketing standpoint it is also about the cost/benefit ratio. A larger pentaprism costs more to make than a smaller petaprism does. A smaller pentamirror is even cheaper. It is easier to justify the additional expense for a $5K pro model than it is for a $500 entry level model. If the difference in price is (strictly speculative) $100, that would add 20% to the cost of the entry level model making it $600. On the other hand, putting the cheaper viewfinder on the pro model would only reduce the price by 2% to $4,900.