As a home sewist, I'm surprised to hear the 44% figure on an industrial scale. If I made one T-shirt at home, it might result in up to 50% offcuts? With all sleeves cut out of one bolt in alternating warp direction (~tessellated) and all rectangularish body pieces cut from another and ribbons of binding from another, I wouldn't expect to see more than 10-20% offcuts. And then home home sewists often shred the offcuts and use them to stuff other projects, which should be easier to do and funnel to the right place. I didn't try very hard to access the original paper but count me a bit skeptical and confused.