I am not sure if I agree with you. If it is true, then they should be able to prove its value scientifically. Its in their best interest both commercially and as professionals to pay for and run these tests. My guess is that many of these address the symptoms and provides relief such as ordinary painkillers and sedatives but they seldom are the cure.
李时珍 and his 本草纲目 is one of the best documented records on the medicinal properties of plant parts. The description of the plant and its various parts was to the best of its generation. Tested by folks generations over to work. Even modern day science is looking at the TCM herbs: isolating and identifying its active compounds for novel drug discovery.
If you want me to put a figure on it, I'd say about 95% of those individually documented herbs have the medicinal properties as prescribed, put it to double blind test and I say 90% will clear. In terms of the permutations resulting from the types and quantities of herbs used, my confidence level isn't as high as yours. I only put the blame squarely on those who knew too little and yet trying to pull a fast one.