How To Fix Bond ETFs

October 22, 2009

You want to fix bond ETFs, Dave? Here are a few simple ideas.

Idea No. 1: Optimize The Creation Basket (At Least Sometimes)

The first idea is drawn from the way Vanguard manages its bond ETFs. You touched on this in your blog, but you really didn’t get to the core of what makes their bond ETFs interesting.

In the Vanguard structure, ETFs are one share class of a broader mutual fund. One advantage of this, as you mentioned, is that an ETF like the Vanguard Total Bond Market ETF (NYSEArca: BND) gets to tap into a much broader portfolio of securities. While a fund like the iShares Barclays Aggregate Bond Fund (NYSEArca: AGG) holds 247 bonds, BND is part of a larger mutual fund that holds more than 3,000.

That’s good in general: The more diversified portfolio, the better. But it’s good in a more interesting way, too.

Vanguard can’t ask Authorized Participants who are creating new shares of BND to go out and buy tiny slivers of all 3,000-plus bonds in the portfolio. Instead, they ask for a small subset of those: about 50 or so. Those 50 are chosen so that their performance will generally match up to the performance of the broader portfolio. Over time, Vanguard can either rotate the creation/redemption basket or use fund flows from other share classes to diversify its portfolio so that it doesn’t become overly concentrated in these 50 names.

By contrast, AGG’s creation basket is nearly the same size as its fund: APs must buy up 200 or so bonds to create new shares. So the creation process is more difficult, even as the end-portfolio is narrower.

I’m not saying all ETF providers should switch to the Vanguard model; it has its own issues. But maybe non-Vanguard ETFs should look at optimizing the creation/redemption basket to streamline the arbitrage process. Maybe they could rotate that basket frequently to make it more difficult for others to front-run the trades. It’s an idea worth considering at least.

Idea No. 2: Allow Cash Creations And Maybe Redemptions

Another thing providers could do to limit the premium/discount issue is to allow cash-based creations and redemptions?if not all the time, then in special situations. We know from the muni bond ETFs that cash-based creations largely eliminate the problem of premiums (although they do nothing for discounts). They may raise the internal costs of the ETF slightly, but that is debatable.

Allowing cash-based redemptions would be difficult, because the sponsor would be stuck with the risk of holding the bonds in an illiquid market. You couldn’t do it on a one-for-one basis. But perhaps you could implement a high fee to execute a cash-based redemption: Charge APs the equivalent of 2 percent or so to execute a cash-based redemption. That would at least limit the size of potential discounts, and the net risk to the sponsor would be limited, and in many cases, could be hedged in the derivatives market.

Idea No. 3: Use Swap-Based Structures

I’m traveling in
London
right now, so here’s an idea from across the pond: What about removing the risk of tracking error and creations/redemptions altogether by moving illiquid bond categories to a swaps-based structure?

I know the answer: This would go absolutely nowhere commercially, as investors are rightfully afraid of anything with “swap” in the name. But the truth is that swaps can be managed with very little true counterparty risk. If you “true-up” the swap on a daily or weekly basis, and if you have multiple swap counterparties, you can get the real credit risk down toward 5 percent or less of total assets. If it were handled transparently and on a rules-based basis, I think it would offer a real alternative to the replication structure that dominates in the
U.S.

I Like Bond Index Funds And ETFs

I received a lot of emails about my last blog, so let me clear up a few things. First, I believe in bond indexing: The data are unequivocal that almost all active bond managers trail their benchmarks. I hold bond index funds in my own portfolio. But I don’t think they’re perfect, and I think there are simple ways they could be improved.

I also like bond ETFs, with a few caveats. The Treasury, international Treasury and TIPS bond ETFs seem to work just about perfectly. As ETFs start to incorporate corporate bonds, things get more challenging, as premiums and discounts start to rise.

I don’t think the problems are intractable. I just think they should be recognized, understood by investors and tackled head-on.


Find your next ETF

Reset All