An economics puzzle

I must not be the only person to have noticed the following peculiarity. Often when shopping for clothes at the end of the season, one finds that there are only a few items remaining in a particular style, and that these items are mostly in sizes XL and XXL. This is not to say that sizes S, M and L can never be found under such circumstances, but just that they tend to be noticeably under-represented. The situation is arguably a little puzzling. Why should certain sizes--in particular, the largest sizes--happen to be over-represented among the items left over at the end of the season?

The obvious though somewhat unsatisfying and answer is that there are many fewer people who take sizes XL and XXL than there are who take size M, and to a lesser extent sizes S and L, meaning that if everyone buys shirts at the same rate, there will tend to be a surplus of the largest sized items at the end of the season. The reason why this answer is unsatisfying is that it begs the question of why sellers don't simply purchase fewer of the largest sized items from their supplies, so as to ensure that their stock accurately reflects the distribution of sizes within the population. In other words, the answer requires one to accept that sellers behave irrationally by consistently purchasing more than the efficient number of the largest sized items. Indeed, the only way it could be reconciled with reasonable assumptions about the behaviour of sellers is if, due to the nature of the manufacturing process, the marginal cost of producing items is so small that it is profitable for sellers to simply purchase the same number of items in each size.

An alternative answer is that the puzzle does not really exist. Sellers do purchase the efficient number of items in each size, and there is no systematic tendency for items in the largest sizes to be left over at the end of the season. Rather, because there are fewer people who take sizes XL and XXL, there is more sampling error in the number of the largest sized items left over. On some occasions there are more of the largest sized items remaining than one's expectation, and on some occasions there are fewer of these items remaining. It is only because I am someone relatively close to the mean of the size distribution, who tends to purchase items in sizes S and M, that I notice when there is a surplus of the largest sized items left over.