Lately I have focused on product recalls involving children's clothing - specifically drawstrings or lead. In this case, it is a blanket with a manufacturing defect. The satin border can come loose and pose a strangulation hazard.
This problem is likely a stitching problem. I would have to see a picture of the actual defect. Possibly it is loose tension on the stitches or an improper stitches per inch (SPI). I have seen some manufacturers try to produce goods with stitches that are too large. I have been known to go down the production line and turn the SPI up because you can *technically* sew faster with a larger SPI (say 5-7 SPI versus 8-10 SPI). For an operator that is paid by the piece, this is a huge temptation. An in-line inspector or quality auditor should pay attention to this problem if it shows up. An incorrect SPI is ugly and will interfere with the structural integrity/strength of the seam. The actual SPI needed should be specified in a quality manual or other technical specifications. You can determine what is appropriate during product testing. (Topstitching usually has larger stitches and heavier thread. Not usually a problem unless the topstitching is also a major functional seam. Some operations are sewn and topstitched in one operation and thus must have the proper structural integrity).
Another possibility is that not enough of the blanket fabric was pushed into the binding. Depending on the type of fabric used for the blanket, the binding could just pull off the edge. Some fabrics are just not appropriate for this kind of finish with the expected kind of use. I have seen bindings pull off of all kinds of fabric, say an organza versus a cotton broadcloth. Or a triple velvet versus a fleece.
Again, I don't know the exact mechanism that caused the binding to become loose. Just some things to consider if making a similar product. BTW, any type of trimmery needs to be pull tested. Buttons, snaps, ribbons, flowers, doll eyes, the list could go on. All these things are potential choking hazards.