The explanation is in the piece. Here's the logic: If you try to reduce demand for a product, but actively maintain a supply of that product, it sends a confusing signal to the market. If you want consumers to stop buying something, surely you also have to communicate that it will no longer be available. For as long as governments stockpile ivory, consumers are likely to buy ivory whenever they can, believing that the trade will one day resume. That undermines the credibility and therefore the efficacy of demand reduction campaigns. These campaigns try to convince consumers that their purchase of ivory directly destroys elephants. Why would they believe that line if they see African governments keeping ivory in the hope of being able to sell it one day?
Thanks Bernhard for your comment. The problem is that 100 tonnes would not be released onto the market at once. It would be sold to legitimate traders (chosen by the government), who would then sell the value-added product to consumers as per a normal business model. There are two problems with your argument. One is that using the proceeds from stockpile sales to fund demand reduction campaigns is contradictory - you'd be selling to the consumers you're trying to convince should not buy ivory. Second, for as long as governments can make money from selling stockpiles, what incentive do they have to actually prevent poaching? They can simply wait until the elephant is poached and then confiscate the ivory to sell it themselves. While the ivory is not a dangerous product, as you point out, its sale sends a dangerous signal to the market that the trade is legitimate. Destroying the exchange value of ivory, and increasing the use value of elephants, means that you have to destroy exchange value at both the supply and demand ends of the spectrum.