Your thought pattern makes sense but a lot of dislocations in these markets are once again due to liquidity. I think the easiest way to summarize this is becaus

Your thought pattern makes sense but a lot of dislocations in these markets are once again due to liquidity. I think the easiest way to summarize this is because of the liquidity transformation ETFs provide, the tail wags the dog until it doesn't. This gets a bit rambling but I am too lazy to edit right now. Liquid hedges are extremely valuable when trading corporate bonds. With almost every sell-off/rally, you will see more liquid instruments be first movers and usually overshoot in either direction. If I wanted to buy/sell 100mm MV of bonds right now, I would have to pay through the nose and there's no dealer on earth (due to post-GFC regulations) that is warehousing 100mm MV of a single bond issue so then I'm probably buying 10-20 different bonds and that doesn't really fit what I'm trying to do. Instead, I can buy/sell 100mm of a bond ETF, pay close to equity-like bid/ask spreads, and do so without picking up the phone or having someone decline to trade with me. I'm a little braindead right now and I know this is a bad example but this makes intuitive sense: if I really need cash immediately for something, I'll go to the pawn shop first before going to the bank. The pawn shop will give me immediate liquidity while the bank might not be open, want me to open an account, fill out some documents, blah blah. The same thing happens with SPY, if I want to hedge my up/downside and need to do it quickly and efficiently, buying/selling single name equities in a lot of cases is not the best way to proceed. Dislocations vs NAV rarely happen with SPY or anything with large liquid underlying assets because the create/redeem process for these is very efficient; I can do it through an API. There have been advances in the corporate bond market but for the most part bonds are traded via chat or phone. So, let's say bonds sell off for whatever reason, everyone immediately reaches for an ETF rather than selling any bonds. The underlying market takes some time to react and there's a further cascade of price discovery that usually happens in order of liquidity. When you have several thousand bonds underlying an ETF and no efficient way to trade them, price discovery takes a while. ETF overshoots, bonds try to catch up, oh but wait now this bond looks cheap/rich vs that other bond, then maybe everyone starts taking their hedges off, and then the process restarts in a miniature fashion. This tug of war can take days to shake out but in the meantime, the create and redeem process still exists so if I feel the ETF is way too rich (simple unreaslitic example but "wow, ETF dividend yield is 5% but the underlying bonds yield 10%") I can buy some shares and redeem them for the underlying bonds and also do the reverse. The create/redeem process was broken during the recent sell-off because there was no consensus on where NAV actually was. Sure, the ETF provider publishes a NAV level but you had several ETFs trading at several percentage point discounts. In normal cases, I can buy the ETF and redeem it for the underlying theoretically pocketing the difference. I'm not going to do that if the underlying price should be much lower or I'm just uncertain as to where it really should be. Long story short, ETFs track NAV closer in more efficient markets. In less efficient markets, ETFs or any liquidity transformation vehicle usually lead the underlying because it's easier/cheaper to transact in them. From there it's price discovery on the underlying which happens at varying speeds and the process goes back and forth.