You could take any font with both "normal" and "light" weights, and then apply the leading and pair-kerning from the normal weight to the light weight, to produce a font that reflows text the same as the normal weight, but uses the ink of the light weight. I do wonder whether the results would cause typographers to gag, though.
This is a huge cost. Consider the U.S. Army, one part of the much larger Department of Defense. The Army has literally thousands of pre-prepared forms [1] and manuals. It's likely that these forms get printed hundreds of thousands of times per day. You could indeed save a lot of ink by switching to Garamond but that means that each form needs to be re-done. Some are simple MS Word forms where it might be easy. Others are PDFs and most are XFL (awful Lotus Forms), all of which will take a major effort to convert. Multiply that time and effort across an entire government's worth of forms and the effort required is mind-boggling. It would take many years for this to pay off.
In the military/government, forms, publications, and manuals are regularly updated, distributed, and printed. Rolling out a new font on each update before distribution wouldn't be a huge task.