I don't know how to say this in any way that does not sound unkind, the avoidance of which is 100% my intention... [and I saw they pulled your other question, slightly unfairly in my more tolerant opinion, so I'll postscript answer that as well ;-) ].
Your gold example is a bad one... precisely because the issuing bank can, and will, hedge its gold exposures via any/all of the bullion, gold futures or options markets.
I can easily issue "gold notes" to my clients that pay 100% of what gold did; and those notes will never need hedge themselves. Or just launch a Gold ETF, which is a derivative (albeit an unlevered one) of the gold price.
And I could this for eg Rhodium, Carbon Prices, or perpetual preferred equity if the mutual urge to consummate was there between us. There's nothing special about gold here.
The problems start to arise when the bank wants to start to create products that are not simply 100% related to the underlying. Then the need to hedge becomes essential. This hedging requires the existence of a futures/options market in that underlying.
[Black-Scholes versus reality. You won't like this answer; but it's an answer nevertheless, maybe not of the EXACT question you intended. Look at XVV... you're long of VIX. VIX being settled against VIX futures. Which are hedged against a wide range S&P options. Which embed expectations of implied vol, that the hedging thereof by all banks etc. cause realised losses when realised vols are (usually) lower than implied. Put simply, the vast majority of equity returns has come from the put-selling versus the call-buying component of investors' buying equities There's a "compensation for insurance" within the "equity risk premium", when one compares to derivatives prices that, by definition of construction, have to assume zero returns on the forward (aka the risk-neutral measure, to prevent arbitrage.)]