You got it backwards. If a stock is priced fairly, a buyback is value-neutral. Reduction of the cash on hand is offset by the increase of future cash flow per share. It's dividends that reduce the market cap.
Perhaps you are thinking of share price instead: there it is true that dividends reduce share price, while buybacks are share-price neutral at least theoretically, though it is commonly believed by many people that they do affect price.
(Hence, under that same theoretical model, market cap must decrease if share price remains the same because market cap is total number of outstanding shares * share price. So if the former decreases while the latter stays the same, market cap must have decreased.)
Uhh...what?
Buybacks are a return of value. Share prices adjust accordingly to the updated proportion of outstanding stock.
Might want to select a new username, friend.
That said, this is all under a theoretical model (as in Miller-Modigliani theorem). In practice/empirically, there is reason to plausibly believe that e.g. the decision to announce a buyback has a signalling effect and so can increase share prices.
Nowadays, as you say, there is no difference in tax rates for qualified dividends. However, one big remaining benefit of buybacks is that a dividend forces one to incur a taxable event when the dividend is issued, even if one chooses to immediately re-invest the dividend in the company (as many people still in the accumulation phase of investing do). On the other hand, a buyback does not force those people to sell.
On the other hand, a buyback should lower the market cap of a stock, so cap-weighted index funds ought to sell and re-balance when a buyback is issued, so most index investors would seemingly end up selling. However, they end up being able to avoid most of the capital gains taxes by re-balancing through redemptions and heartbeat trades.