This is Stulz with some academic foundation. It starts with the theory of the risk management irrelevance proposition which, basically, says that in perfect markets risk management cannot add value (it is has a close relation to, and in some respect is the risk-based sibling of the famous Modigliani-Miller capital structure irrelevance proposition, https://en.wikipedia.org/wiki/Modigliani–Miller_theorem).
But bankruptcy risk and financial distress (distress is almost more relevant as a cost and refers to the threat or possibility/probability of bankruptcy) are real-world frictions (ie, assumptional violations) that render the risk irrelevance proposition incorrect. Stulz puts forth a really important idea: If markets are perfect --> risk management is irrelevant, but markets are not perfect in so many ways.
In the case of bankruptcy, Stulz (Chapter 3, no longer assigned but thematically relevant) uses the example of a gold mining firm called Pure Gold, who is obviously exposed to the price risk of gold. Unhedged, the value of the firm has a future distribution which includes some tail probability that, if gold prices go too low, the firm becomes insolvent (this is actually quite realistic for some commodity producers). Under the perfect markets/irrelevance idea, risk management can't help.
But risk management can hedge the price risk of gold (which may be systematic or diversifiable; under the perfect market argument, it doesn't matter. If the risk is diversifiable, investors don't discount the firm's price. If the risk is systematic, the firm's cost to reduce it ought to be the same as the markets; it ought to reduce the firm's cash flow even as it reduces the firm's beta.). Risk management can, for example, enter a short futures position to hedge the gold price risk. Presumably the cost of this is approximately zero (actually its not zero but that's another story) is the meaning of "the cost to hedge bankruptcy risk is zero" .... and "risk management creates value" because the firm value is not reduced by the present value of (tail probability of bankuptcy * cost of bankruptcy). All other things being equal, a firm who cannot go bankrupt is more valuable than one who can go bankrupt. Re: the market will bear the diversifiable risk: this refers to the CAPM idea that the market won't charge anything for diversifiable risk. If the firm's value = future cash flow/[1 + Rf + β*factor], an increase in diversifiable risk does not need to reduce the firm's value. I hope that helpful!