Who Pays for the Dream? Leverage, Dealer Intermediation, and the Distribution of Retail CFD Losses

Retail foreign-exchange and contract-for-difference (CFD) trading is marketed as a route to financial independence. This paper asks a simpler question than whether the route works: when it does not, where does the money go? Starting from the accounting identity of a closed trading pool, I show that the aggregate result of retail clients is bounded above by zero before costs and is strictly negative after spreads, commissions and financing, so that the marketed outcome is arithmetically unavailable to the majority regardless of skill or psychology. Three institutional features then determine who receives what the majority pays. First, leverage: for a fully margined position the adverse move that exhausts equity is 1/L, so a 1:500 account is exhausted by a 0.2% move, a fraction of a normal day's range; the natural experiments of the 2010 U.S. leverage cap (Heimer & Simsek, 2019) and the 2021 Australian product intervention (ASIC, 2021) show that capping leverage cut high-leverage traders' losses by about 40% and aggregate retail net losses by about 91%, while reducing brokerages' operating capital, which identifies the loss as intermediary revenue. Second, dealer intermediation: a broker that internalizes client flow (B-book) earns the client's net loss, so the model is viable only if clients lose on net, and its prevalence is itself evidence about expected client outcomes. Third, introducer compensation: educators paid per lot or per spread share earn from client volume rather than client results, which aligns them with the dealer and, combined with the attribution mechanism of Rakhshani Pourfard (2026b), removes the last party with an incentive to question the method. Jurisdictional arbitrage concentrates the highest leverage and weakest protections on clients excluded from regulated venues. The paper derives testable implications, states policy conclusions consistent with the evidence, and gives a practical rule: size positions by risk, not by available leverage, and treat offered leverage as a signal about the counterparty rather than as a gift.