Price collars have frequently been advocated to restrict the price of emissions permits. Consequently, collars were incorporated in the three bills languishing in Congress as well as in California's AB-32; Europeans are now considering price collars for EU ETS. In advocating collars, most analysts have assumed (1) collars will be implemented by government purchases and sales from bufferstocks, just like bands on foreign exchange rates or commodity prices; and (2) firms must surrender permits whenever they pollute. In fact, however, no actual emissions trading scheme has conformed to these assumptions. In the current paper, we maintain the second assumption (continual compliance) and show that while a price collar supported by a suffciently large bufferstock can restrict permit prices, a price collar supported instead by auctions with reserve prices cannot. In a companion paper (Hasegawa and Salant, 2012), we show that neither method works once account is taken of delayed compliance.