> Sam Vimes' boots
This is a silly economic theory. Rich people buy high quality goods that cost several times more money, often thousands of dollars as opposed to tens, for signaling reasons. Those products often are more durable, but not in proportion to how much more they cost. If a $1000 pair of boots lasts five times longer than a $50 pair of boots, you're still paying $750 more in the long run, and paying it as an up front cost, and taking a higher risk because anything that damages the high cost product is a bigger loss.
That isn't to say that you can never find a product which is such garbage that it costs more even though it costs less, but there is no reason to expect that to be the common case, and a big reason to expect the contrary: Sellers will try to charge more to people who can afford it even past the point that higher prices can cost-effectively improve quality, and rich people will pay those prices because they're less price-sensitive or more willing to pay for signaling.
> Poor people relying on a fake review are harmed by a greater proportion than are rich people who can afford to wait to send it back and get a refund.
You're making a proportionality argument here, but that's irrelevant. The question isn't what rich people do. It's if the amount it would cost to prevent fake reviews is more than the cost to ordinary people of them existing.
And the effect you're referring to only applies to a narrow range of products. If you order some paper towels and end up using napkins for a couple of days while they send some different ones, it's not a big deal. Conversely, if they send you an incompatible replacement battery for your device, you're stuck using the one that only lasts 15 minutes until you can get another one even if you make $100,000/year.
> When the buyers can't tell if they're being ripped off, there is a death spiral of quality on the market as the price that consumers will bear has to assume the risk of lemons, but only lemons can be profitability sold at low prices.
The assumption here is that Amazon reviews are the only means for people to evaluate quality. There are retailers that don't even have review systems and still don't become lemon markets because people can rely on third party reviews or brand reputation. And those systems are typically better because they're not tied to a retailer with a dominant market position, which improves competition in retail and routes around the perverse incentive of the retailer to host dishonest reviews.